ORNL-3936.txt

ifi"lW#Mfi%flfififlfifiiififlfiflfififillfi“fl

3 4456 0362LL0 &

ARTIN MMARIETTA RCY SYSTENMS
i 1

i

i

 
 

 

 

 

 

 

 

 

 
ORNL-3936

Contract No. W-7405-eng-26

MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT

For Period Ending February 28, 1966

R. B. Briggs, Program Director

JUNE 1966

OAK RIDGE NATIONAT, LABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the Sfimms
e

W
il
3 4u5kL 0382610 8

£
iii

CONTENTS
SUMMARY 4 v s eeeansesnsosasasssassasscsoesasnsasassasnasssssnnnassanecss VIiL

INTRODUCTION e s v voansssansnsas cenesaen e creeaaeen eeenes ceeenn .. 1

Part 1. MSRE OPERATIONS AND CONSTRUCTION, ENGINEERING
ANALYSTS, AND COMPONENT DEVELOPMENT

1. MSRE OPERATIONS...viesennecanse checrsacscectoateasesrnansceanes 7

Chronological Account...coevesrnanes e nsedmseetcesaneaaasnan cons 7
Analysis of Experiments.. e eacerrerterenoecaronnnnnns s 10
Reachivity BalanCe.iseesreseesssssssnsoeastoansasossassssnass 10
Power Calibration........... eseearesrevesoseesssebaaacas s 12
Flux MeasurelertS s ceeaeeseacosnsssssssossossasnnescannssass 12
MSRE Dynamic Tests.c.cocinennenssnn Ceeeneeean cenaan veeansaenses 13
Description of Dynamic Testt.reer i et eiiierrenonnne 13
Frequency Response TeshbS.iviieeirriesrssssscnssssnans cerans 15
Implementation of Pseudorandom Binary Tests....eceesas.. .o 16
Analysis Procedures...... cenreceaerneonas cCeeraenecans sosnn 17
Results..... et sassastieasat e s et ennenseen e oe eo e 18
Temperature Response Tesh.c.eeeeeiniraittssorsesssssosnassnes 22
Systems Performance..eeecoacasas teoenemassnas ceeaans ternasseanns 23
Off-(a8 SySbelleeeessseeennsnaronssasooranasnanssennsonsonss 23
Salt-Pump Oll SysbemS..cesereessosssnnesssssarssenanssssasana 27
Treated Cooling-Water System......oeee. Creersesesasnassenss 28
Secondary Containment....... . ... tecmesensasaenas cennns veeas 29
Sh1elding.esecasesanseassaosonssssasosansasassoronnaanaenss O
Component Performance.coeeeeeeesss. Ceeeaneseereesesentaanann oo 34
RAALAE O e e e svesvensencuonsonnssoasscnssossansssasnosaasnses
Cther Components........ tasssesesarssiesestnasrasnoannss 36
Inspection of the Fuel Pump e eteeesacenean Creceanaanas .. 38
Heat Treatment of Reactor Vessel...eeeeerescencsocarsssanns 39
Stress Analysis of Reactor Piping and Nozzles...eeeernecann 40
Instrumentation and Conbrols.ee s e nrvacsassaonsas besesa s 41
General.sevesveesanecns feasseseseseesesoaarsannans cesans .o 4L
Safety Instrumentation....... e eresnasseavreeennroennasoes AL
Wide-Range Counting ChannelsS...ceeeseesoveoescanccosononsan 42
Nuclear Instrument PenetratioN......ceceseeses Creneeneeneas A2
BF3 Confidence Instrumentation..veessessoeecanenccaasaseses 44
Personnel Monitoring Systelm...cveieieesteensciececsanns creens 45
Control Instrumentation..s.ceeereessersesscesssosassnansasns 4O
Operating Experience — Process and Nuclear Instruments..... 47
Data Sysbem..seiieensonaearoas e Cerceenanaraoans vesna. A9
MSRE Training SimulatorS..ceseececrsessssosesssscsacnnn casen 50
Documentation..... caseesaansan ceessesaaesessseesotarrneroas 51
iv

2. COMPONENT DEVELOPMENT ... .eeeeeeenanoosas ceeatesaanens e ceen 53
Freeze ValveS.s.eusas Ceeesacens creesecaea teneacannrease ceeteaa 53
Control RodS.eeevasen. thecesreassneres e na e cerieasans e 53
Coutrol Rod Drive Unlts ceranas Ceeseaseas s chearseneas 54
Radiator DoorS.ieesssessensranas ceraees feeet et et e 54
Radiator Heater Electrical Insulation Failure.......... carerree 57
S L e - e e st s v ettt st et ansossostsesesosarensoscnansocnses 58
Coolant Dalt Sampler...coeeeeenss Ceeesereresearsara e ceseenas 59
Fxamination of Components from the MORE Off-Gas System......... 60

Capillary Flow Restrictbor FE 52l.. it eieeinererescranennns 60
Check Valve CV 533............ Ceeeesareerann Ceretsasseseans 60
Charcoal Bed Inlet Valve HV 62l.. ¢ iiiiiinennrencncanrons . 60
Line 522 Pressure Control Valve PCV 522. ...t iiinnennas 62
Line Filber 522. it iiiiuiiteeestnsensessatsenssassanaananas O
Flow Test on the MSRE Filter from Line 522 ........... ceenan 65
Fuel Processing System Sampler...... C i tiieeretees s cee e . 65
Of T ~Ga8 CaND T sttt eeroerossoanosseacnsssasosssnsasssnsssssnss 67
Xenon Migration in the MSRE......cvveineveen.. Ceeteraneceanaas . 69
Remote MaintbenanCe .o reeerisorscsssrssssnanns ceeaceenaas cheean 70
Practice Before Operation..vee e eeieersteceeceraasanscanans 72
Maintenance of Radioactive Systems..iieiiieierescens Pr e 72
Pump Development. oo eseeeeersitotoesosnssassssssscanasnnasssenas T
MSRE PUMPS e e v vevencensns ceeeenne C e easerereiteeetateranans T4
Other Molten-Salt PUmpPS . vesesseetecenssscencsnanaan ceeeaas 75
Instrument Development..oeeeeeeieierveseoassnsasssososas cereans 77
Ultrasonic Single-Point Molten-Salt ITevel Probe..... ceaean .77
High-Temperature NaK-Filled Differential-Pressure
Transmitter.....cieevevvenn. trectravescrsacerassssssanens 77
Float-Type Molten-Salt Level Transmlttef C et et i etn e 78
Conductivity~Llype Single-Point Molten-balt Ievel Probe..... 78
Single-Point Temperature Alarm SwitchesS..viieeereereeeennnn. 78
Helium Control Valve Trim Replacement........... tereeraanaa 79
Thermocouple Development and Testing......... ceenes ceceiens 79
Temperature SCanner....... Gttt eseassecaer et cerenes ceeea 80

3. MSRE REACTOR ANALYSTS..ieeeseaoenaes sesecsnraeenaa seersescensss 82
Least-5Squares Formula for Control Rod Reactivity...ieveeeoensons 82
Spatial Distribution of '3°Xe Poisoning in MSRE Grephite....... 87

Part 2. MATERIALS STUDIES

4o METATILURGY . iveveeennens cirerareas ceeceanaan e itesesieaenar e 95
Dynamic Corrosion Studies......... teeeenrasens b aseecessassenans 95
MSEE Material Surveillance TestsS..ieesreitensoseens ceseaaeraease 96

Reactor Survelllance SpeCimeNnS...eescietsessaessaasearscsnns 96
Survelllance Control SpeCllelS .. oeseessossassssscasasssasss 99

Hot-Cell Metallographic Examination of Hagstelloy N from
Ixperiment MIR-47-6 for Evidence of Nitriding........ cerrraan 100
Ut

Posgtirradiation Metallographic Examination of Capsules 1—4

from Experiment ORNL MIR-47-6.ceirrernnareennnaan e 101
Development of Craphite-to-Metal Joints....... caranaaen vessanae 101
Tests of Graphite-Molybdenum Brazed Joint for Contalning Molten

Salts Under Pressure..cvvrseesenssass feseaenas e censanss 104
New Grades of Graphite . veeiescoerssoctssensacsssassososssnnas . 107
Evaluation of the Effects of Irradiation on Graphite.......... . 108

" EBffects of Irradiation on Hastelloy N.ivesriooneeesesassoneanass 111
Weld Studies on Hastelloy W....... teecnecaes chearesaeans sereasae 115
MY T RY 4 s v sesonoassssonssasnsasanssassassosnssaass heereear e 122
Chemistry of the MORE..veeiereeorensacensesoronsoncssnssssesess L22

Apalyses of Flush, Fuel, and Coolant Selts..... e reas. 122
Fxamination of Materialsg Trom the MSRE Off-Cas System...... 124
Uranium~Bearing Crystals in Frozen Fuel....... teesenaesones 128
Physical Chemistry of Fluoride Melts........eovennns e . 128
Vapor Pressure of Fluoride Melbs...veeiieieeernonneasnan, oo 128
Methods for Predicting Density, Specific Heat, and
Thermal Conductivity in Molten Flucrides.....oevee.. eeees 130
Oxide Solubilities in MSRE Flush Salt, Fuel Salt,
and Their MIxtures........ e N B X
Ceparations in Molten Fluorldes....ccevvearsvcncnse cereanna ceees 136
Fvaporative-Distillation Studies on Molten-Salt Fuel
Components...osne. ceenaen Ceesreenanes Ceresosceeaunenan eas 136
Effective Activity Coefficilents by Evaporative
Disbillatlion.siseseeesrtoccoeesnsresocassconsssnacaans e 139
Fxtraction of Rare Earths from Molten Fluorides into
Molten Mebals. eeeeeeeeacoosncrssnnonsonnsnse eeson cesesaes 141
Removal of Protactinium......... P 75
Radiation Chemistry........... feieesatsesacs e O 57
Tn~-Pile Molten-Salt Irradiation Experiment........c.cocena.. 152
Development and Evaluation of Methods for the Analysis
of the MSRE Fuel..ovvvenvnnannas cersanan feerana cevons cireas. 154
Determination of Oxide in MSRE FUel...ceeerenionnarrccocnen 154

Voltammetric Determination of Ionic ITron and Nickel in
Molten MSRE FUEL.s.uecovrarncossssosasansssscsosnsansnness 1OZ
Development and Evaluation of Equipment and Procedures for

Analyzing Radicactive MSRE Salt Samples..... et ear s eee 165
Sample Analyses....ceveaen Cheeseaarsesresaesaenas cerarnenanseass 167
Quality Control Program........ e e sseanssecannans Ceeeraraaaan 168

MOLTEN-SALT BREEDER REACTOR DESIGN STUDIES. .. .eeeeevensaenaaass 172

MSBR Plant Desigi.vsevecsasscssorascsesessssesoscsarsassason eaee 172
FloWSheet oo venreresannnannnas Ceeeeeseeteeeeeesaneonansanass 172
Reactor Desigheeeseeeeoceonnass cheraraeennas ceteensasaanaes 174
FTuel Processing...ceeeessesesss faareesesaaans serseas eenes  L77
Heat Exchange and Steam SystemsS.....veessescesesesseraeeasas 181

Capital Cost Bstimates......ocevnnnenn.. O < ¥4
Reactor Power Plant..... rteeseastas et eaeanesnnan cessaeeas 182

Fuel Recycle Plant.......... g < 2
vi

Nuclear Performance and Fuel Cycle AnalyseS....vveeeenen. veess. 186
Analysis ProcedireS. . voiersesssanens ceaeans ceeareenes ceeeen 186
Basic Assumpbions..eieeeevieicenean. Gt et rerrtesasesstesens ... 186
Nuclear Design Aralysis...... ceeriaeseanaaaes Cerecesaan ce.. 189

Power Cost and Fuel Utilization Characteristics......ccvcciaens 191

MOLTEN-SATY REACTOR TROCESSING STUDIES. ... iieirevrsesesrosonenns 193

Semicontinuous Distillation..c et ieieneisacssssassosssnansans 194

Fuel Reconstibubion..coieiivr ittt eneoncnearonnnns Crreeaean 199

Continuous Fluorination of a Molten Salt.......coveveennn Ceeaaee 202

Chromium Fluoride Trapping..cvevrieeceesss tettesseaasanenan ceees 202

Degign and Evaluabtion SHuly...ovveeireniiiiniiiiiianoriinanannss 203

Description of Fuel Process.....ceevens Cereeiraeteaes Ceeenaenue 204
Description of Ferbtile Process..ieeeiieeriinenenneens Ceeeaes 208
Waste Treatment........cveinenne.. et eesanenseensnsaneeas 208
OFf-Gas Treatmenl .. e e cnrernsceonserosssnasnsesons Ceraesae 209
Surmary of Capital and Operating Costs............. Ceaaaae 209

Processing CosSt.iernanienrnnncnans e et e cesa e 210
SUMMARY

Part 1. MSRE Operations and Construction,
Fngineering Analysis, and Component Development

 

 

1. MSRE Operations

 

Preparations for power operation were completed, and the MSRE was
operated at nuclear powers up to 1 Mw before the system was snhut down to
replace a space-cooler motor and to relleve plugging problems in the off-
gags system.

The power preparabtions included some system modifications shown to
e required by operating experlence and by continuing development and
analysig work. Remote maintenance techniques were tried and evaluated,
some specilal tests were performed, the operators were trained and guali-
fied for power operation, and the secondary containment was sealed and
shown to have an acceptably low leak rate.

The nuelear performance of the system at 211 powers up Lo 1 Mw was
highly satisfactory. Reproducible reactivity behavior and a lack of
gignificant cross contamination hetween the fuel and flush salis were
demonstrated. Dynamics tests at power showed that the reactor has a
slightly wider margin of stability than had been predicted from calcu~
lations. Preliminary results indicate that xenon poisoning may be lower
than was anticipated. :

The performance of most of the egquipment was satisfactory, but
substantial operational difficulty was caused by plugging of very small
copenings in off-gas system components by organic material. This problem
was extensively investigated after the shutdown from 1 Mw. Other, less
gserious problems ineluded the {reak failure of an electric motor inside
the secondary containment, activation of the corrosion inhibitor in the
treated cooling water, air entrainment in the cooling water, and ex-
cegsive radiation levels in a Tew remote areas. Solutions have been
developed for all the problems except the off-gas plugging, which 1s
still under study.

Formal design of the instrumentation and controls systems for the
MSRE was completed. Additions and modifications are now velng mads as
needed to provide additional protection, improve performance, or provide
more Information for the operators.

The addition of a low-level BF3 counting channel with control func-
tions, the addition of cadmium shielding in the neutron instrument pene-
tration, and changing reactor period Interlock trip points were required
to obtain satisfactory performance of the nuclear instrumentation system.
The remaining changes were mostly of secondary importance. GSome sporadic
viii

difficulties were experienced with individual hardware items such as air
and helium valves and electronic switches. Most of the work done on the
instrument system can be characterized as debugging the original instal-
lation.

The data~logger—computer was put into operation in conjunction with
the reactor. Although the performance has not been up to expectations,
it is proving useful to the operation of the reactor experiment. A power
level simulator was asscmbled, and it operated satisfactorily for the
training of operators.

2. Component Development

 

The "fast thaw" requirement was eliminated in all freeze valves ex-
cept for those which control the emergency drain of the reactor and of
the coolant system. The operation of all the valves which might contain
sufficient radiocactivity in the salt to produce radiolytic fluorine at
low temperature are now operated above 400°F, which is sbove the threshold
Tfor flucrine release.

The braided wire sheath cover for the convoluted hose of control rod
No. 3 was found to be severely torn about 2 f{ below its upper end. The
cause was traced to a Jammed roller in the upper bend of the control rod
thimble. The roller was replaced, and the upper rod sheath was repaired.
There has been no further difficulty after several months of operation.

Control rod drive unit Wo. 3 was replaced because of a shift in the
remote position indication and because of a tendency for the lower Iimit
switch to stick. The shift of the indicated position was eliminated by
removing the excess slack 1in the chain, which had allowed the chain to
slip over the sprocket. The sticking lower limit switch is being cured
by replacing the return spring with a stronger one.

Modifications were made {0 the radiator door guide tracks and lock
mechanisms to allow for thermal distortion, found after the initial oper-
ation of the radiator doors at temperature. Alterations were made o the
limit switch system to prevent a damaging overtravel of the door in the
upper end of the travel.

A "loss-of-tension" device was designed which will stop the radiator
door drive unit should the door support cables show any slack as the door
15 being lowered. This arrangement is intended to prevent damage to the
support cable if the radiator door Jams, as well as to indicate a mal-
function.

Failure of the insulation on the electrical leads to the radiator
heaters was traced to excessive heat leakage into the areca immediately
above the radiator. Changes were made to reduce the temperature in this
area, and electrical insulation with a higher temperature rating was in-
stalled.
ix

Several changes were made to the sampler-enricher to improve the
operation and safety of this system. Among these were the changes made
to the interlock circuit, which require that additional barriers be
present during certain critical operations, thereby assuring double con-
tainment at all times.

Forty samples were taken during runs 4 and 5, nine of which were
large 50-g samples for oxygen analysis. These larger samples caused
some difficulty until the capsule design was altered slightly to make it
hang straighter.

One of the operational valves developed a 20-cc/min helium leak
across one of the two sealing surfaces of the gate. Since this is one
of two valves in the line and the leak 1s clean buifer gas, the valve
was not replaced.

During the same period ten samples were removed from the coolant
system, two of which were the larger 50-g samples. The first sample
taken after an extended shutdown had a black film on it, which was ldenti-
fied as decomposed oil. Although there was oll in the general arca the
exact source was not established. No films were found on subsequent
samples.

The design and installation of the fuel processing system sampler
is proceeding.

A system is being designed to permit analysis of the reactor off-gas
gtream. It will contain:

1. a thermal conductivity cell for on-line indication of the gross con-
taminant level,

2. a chromatograph for quantitative determination of contaminant,

3. & refrigerated molecular sieve trap for isolation of a concentrated
sample for transfer to a hot laboratory for isotoplc asnalysis.

Estimates of the '3%Xe poison fraction for the MSRE were computed
35 a function of several parameters. At 10 Mw the resulls indicate that
the poison fraction is 1.6%. It was found that the mass transfer coef-
ficient from the salt to the graphite is controlling the transfer and
that the properties of the graphite are not important.

The remote maintenance group gained more experience with the reactor
components during the period pricr to power operation. Among these were
removing and replacing the pump rotary element and replacing the graphite
sampler assembly. After a short period of power operation several oper-
ations were performed, using remote maintensnce technigues, on & mildiy
radiocactive gysten.
The MSRE pump test Tacility was modified, and the prototype pump
was operated for periods of 165 and 166 hr at 1200°F to provide shake-
down of the spare fuel pump impeller and the spare coolant pump drive
motor. The spare rotary element for the fuel pump was modified to pro-
vide positive sealing against oil leakage Trom the shaft lower seal
catch bhasin into the system past the outside of the shield plug. The
drive motor containment vessel was redesigned, and the new design will be
used for the fifth drive motor vessel. Modified ejectors were installed
on tne lubrication systems for the MSRE salt pumps, and the lubrication
pump endurance test was continued. The ME=2 fuel pump tank design was
completed and is being reviewed.

The PK-P molten-salt pump continues on endurance operation and has
operated for 22,622 hr. The pump containing the molten-salt bearing was
placed in operation, but the bearing seized after 1 hr of operation.

Efforts to improve the stabllity of the ultrasonic level probe in-
stalled in the MSRE fuel storage tank were continued without success.

Testing of a NaK~filled differential pressure transmitter which
failed in service at the MSRE was continued. Performance of the instru-
ment was improved by refilling with silicone o0il but is stilli not satis-
factory.

Performance of the ball-float-type transmitter iunstalled at the
MSRE continues to be satisfactory. Some difficulties were experienced
with a similar (prototype) transmitter on the MSRE pump test loop; how-
ever, these troubles were anticipated and corrected in the design of the
MSRE model.

Performance of the conductivity~type level probes installed in the
MSRE drain tanks continues to he acceptable.

Observation of the performance of 110 single-point temperature-
alarm switches is continuing. Data obtained Lo date are insufficient to
determine whether set-point drift in these switches is excessive.

Testing of alternate trim meterial combinations for the helium con-
trol valves was terminated. Some additional valve failures have occurred.

Results of final checks indicate that errors in the coclant-salt-
radiator differential temperature signal, produced by thermocouple and
lead-wire mismatch, have been eliminated.

Drift testing of selected MSRE-type thermocouples was concluded.
The Cinal temperature equivalent drift values were between +4.7 and
+6.4°F,

Performance of the MSRE temperature scanning system continues to be
satisfactory. Calibration drift appears to have been eliminated, and
reliability is much better than had been expected.
xi

3. M3HEE Reasctor Analysis

 

For the purpcse of on-line computation of control rod reactivity
with the TRW-340 data logger, a mathematical formula was fitted to the
rod-worth vs position curves obtained from calibration experiments. The
form of the expression used was obtained by applying a periturbation
technique to evaluate the integral expression for the rod reactivity. A
linear least-squares curve-fitting procedure was then used to evaluate
the unknown coefficients in the resulting Tunctional expression. Close
agreement between calculated and experimental curves was cblalned for
those configurations of shim and regulating rods of interest in monitors-
ing the contrel rod reactivity during operation.

Theoretical calculations were made to estimate the inlfluence ol the
overall spatial distribution of 135%e gbsorbed in rores near the grapnite
surfaces in the reactor core. The purpose was to determine spatial cor-
rection factors for use in the on-line calculation of *2%Xe reactivity
with the TRW-340. Basged on an approximate model of the reactor core,
thege calculations indicated that the equilibrium 1353 reactivity at 10
Mw is reduced by a factor of about 0.76 relative to the value obtained
from a "point" calculation. In addition, this correction was found to
depend on the time history of the power level. Results of calculations
are presented for step changes in power level, increasing to and de-
creasing from 10 Mw,

Part 2. Materials Studies

 

4o Metallurgy

Thermal convection loops made of Hastelloy N and type 304 stainless
steel have circulated molten fuel salt for 33,000 and 22,000 hr, re-
spectively, without incident. A Cb—1% Zr loop circulating lead at 1400°F
with a 400°F AT was found to produce columbiwm crystals by mass transfer.

Specimens of Hastelloy N and grade CGB graphite showed no detectable
changes as a result of 1100 hr exposure to molten flucride salts in the
MSRE core during the precritical, initial critical, and assoclated zerc-
power experiments.

The: reactor control specimen rig, which will establish base-line
data by exposing graphite and Hastelloy N survelllance specimens t0 ap-
proximately the operating conditions of the MBRE except for radiation,
has been loaded with salt and is veing calibrated with the computer that
monitors the MSRE.

Metallographic examination of capsules from in-pile experiment MR
47-6 showed no evidence of nitriding of the Hastelloy N. Ho apparent
X111

change in wall thickness or evidence of attack was observed, although an
unexplained change in the etching characteristics of the grain boundaries
at the surface was noted.

Development of methods of Joining graphite to metal has included:
(1) the design of a transition joint to reduce shear stresses arising
from thermal expansion differences and (2) screening tests on potential
brazing alloys.

A small pipe of grade CGB grapnite brazed to molybdenum satisfactorily
contained molten fluoride salts at 700°C under pressures of 50, 100, and
150 psig for periods of 100, 100, and 500 hr respectively. This is the
first of a serles of tests of graphite-to-metal Jjolnts to determine if
such Jjoints are corrosion resistant and mechanically adequate for the re-
quirements of molten-salt breeder reactors.

A few samples of needle~coOke graphite and isotropic graphite have
been obtained and are being evaluated to determine their suitability for
use in molten-salt breeder reactors.

The radiation-damage problems were evaluated for graphite in advanced
molten~-galt reactors, considering growth rate, creep coefficient, flux
gradient, and geometric restraint as important factors. The stress de-
veloped by differential growth in an isotropic graphite should not be
allowed to exceed the fracture strength of the graphite and thus cause
failures. The estimated life of graphite is at least five years before
failure from inability to absorb creep deformation. The major uncertainty
scems to be the ability of grapnite to sustaln dosecs of 2 X 1023 nvt with-
out loss of integrity.

Creep~-rupture life of Hastelloy N was found to be less affected by ir-
radiation as the stress levels are lowered. The effects of irradiation
temperature on the postirradiation creep life of air-melted heats are un-
certain. Vacuun-melted heats show a large dependency on irradiation tem-
perature. Pretest heat treatment can improve the ductility of irradiated
specimens. The creep~rupbure properties of structural material in the
M5RE appear ©o be better than originally predicted on the basis of linear
extrapolation of data for stress vs log of rupture time.

Experimental welds have been made {0 study methods of improving the
weldability of Hastelloy N.

5. Chemistry

Three innovations have been introduced in the chemical analysis of
MSEE salts: a new end point for uranium titrations, a new method for
determining structural-metal ions, and a new method for oxide analyses.
Together they have given increased assurance that fuel conforms to the
inventory composition and that the chemical purity of the salt has been
maintained.
Examination of deposits believed to have been responsible for the
plugging of offwgas lines in the MSRE revealed the presence of oil and
polymer products presumed to have formed from oil. A negligivle amount
of salt was found.

The formula for the uranium-bearing crystals in the frozen fuel has
been found to be I4iF-UF, rather than 7[iF-0UF,; as formerly supposed.

In a study of the physical chemistry of fluoride melts, vapor-pres-
sure measurcments have been made for three compositions in the ILiF-Bels
gystem. Because of vapor-phase association, the apparent volatility of
LiF increases with decreasing concentration of IiF in the melt. Methods
have been developed for predicting density, specific heat, and thermal
conductivity in molten fluorides.

The solubility of oxide in MSRE~-related fluoride mells has been re-
evaluated with improved experiments. When increasing amounts off ZrF,, as
present in the MSRE fuel, are added to flush salts, the capaecity for
oxide first decreases, then Increages.

Interest in reprocessing methods for MSBR fluorides has led (0 coOn-
tinuing studies of distillation and of chemical reduction as a means of
separating rare earths from fuels or protactinium from blankets. The
composition that yields MSRE barren solvent as dilstillate has been found;
this product distills, leaving the rare earths behind. Alloys of bismuth
containing a small amount of lithium have proved very effective for re-
ducing and extracting rare earths inte a liquid-metal phase. Thorium has
been found effective as a reducing agent for removing protactinium from
tlanket melts. Protactinium can also be removed on ZrOs.

The in-pile molten-salt loop experiment and assoclated auwxiliary
equipment are being fabricated and assembled so that modifications to
beam hole HN-1 in the ORR and installation of equipment can begin in
April.

The precision and accuracy of the hydrofluorination method for de-
termining oxide in MSRE salts were established. The method was applied
to the analysls of nonradicactive samples taken during the startup of
the M3R; the results were in reascnable agreement with those obtained by
the KBrr, method. The hydrofluorination apparatus for the determination
of oxide in radiocactive samples was fabricated and tested and is now being
installed in a hot cell.

Tonic iron and nickel were determined voltammetrically on a sample
of molten fuel withdrawn from the MSR. These measurenments indicate that
the major fraction of iron and of nickel in the fuel 1s present in an
un~ionized state, presumably as finely divided metal. Also, a well-
defined voltammetric wave for the reduction U(IV) — U(IIL) was observed.
xXiv

Efforts were continued on the development and evaluation of equip-
ment and procedures for analyzing radicactive MSRE salt samples. The
coulometric uranium procedure was modified to eliminate a negative bias.

Both flush- and fuel-salt samples were analyzed Tor U, Zr, Cr, Be,
¥, Fe, Ni, and Mo. The analyses were routinely performed in the hot
cells of the High-Radlation-Level Analytical Iaboratory.

The quality control program was continued during the past period.

The results obtalned on synthetic solutions established more reallstic
limits of error for the methods employed.

6. Molten-Salt Breeder Reactor Design Studies

 

Design and evaluation studies were made of tThermal molten-salt
breeder reactors (MSBR) in order to assess their economic and nuclear
potential and to identify important design and develcpment problems.

The MSBR reference design concept is a two-region, two-fluid system with
fuel salt separated from the blanket salt by graphite tubes. The energy
produced in the reactor fluid is transferred to a secondary coolant-salt
circuit which couples the reactor to a supercritical steam cycle. On-
site fluoride volatility processing is employed, which leads to low unit
processing costs and economic reactor operation as a thermal breeder.
The resulting power cost is estimated to be 2.7 mills/kwhr for investor-
owned utilities; the associated fuel cycle cost is 0.45 mill/kwhr
(electrical); the specific fissile inventory is 0.8 kg/Mw (electrical);
and the fuel doubling time is 21 years. Development of a protactinium
removal scheme for the blanket region of the MSBR could lead to power
costs of 2.6 mills/kwhr (electrical), a fuel cycle cost of 0.33 mill/kvwhr
(electrical), a specific fissile inventory of 0.7 kg/Mw (electrical),
and a fuel doubling time of 13 years.

7. Molten~Salt Reactor Processing Studies

 

A close-coupled facility for processing the fTuel and Tertile streams
wi.ll be an integral part of a molten~salt breecder reactor system. The
Tuel salt will be processed on a 40-day cycle. The uranium will be re-
moved from the carrier salt and fission products by fluorination, and the
carrier salt will be recovered from the fission products by a semicon-
tinuoug vacuum distillation. Relative volatilities between lithium and
the rare carths have been measured to be 0,01 to 0.04 at 200 to 1050°C.
The reconstitution of the Tuel salt, by combining the purified carrier
salt with the purified UFg, can be done by direct absorption of the Ulg
in fuel salt which already contains some UF, and subsequent reduction of
the intermediate uranium fluoride to UF, with hydrogen. Experimental
tests showed rapid and complete absorption. The primary proolems in con-
tinuous Tluorination of the fuel salt to remove the uranium are corrosion
XV

and getting adequate mass transfer and countercurrent flow fto assure good
recovery. Corrosion can probably be eliminated by the use of a layer of
frozen salt on the wall of the vessel., BExperimental work with a small
countercurrent continuous fluorinator gave recoveries of 290 to 96% of the
uranium. Fluorination during the processing of the fuel from the MSRE
produces volatile chromium flucorides. These can be effectively trapped,
with negligible uranium losses, by use of sodium fluoride beds. A pre-
liminary design study has been made on a conceptual processing plant in-
corporating the above concepts. Among the problems which this study
illuminated were the complications from the handling of high-heat-
generating materials. The fixed capital cost for the conceptual plant was
$5.3 million; the salt inventory cost was $0.196 million, and the direct
operating cost was $787,790.00 per year.
INTRCDUCTION

The Molten-Salt Reactor Program is concerned with research and de-
velopment for nuclear reactors that use moblle fuels, which are solu-
tions of fissile and fertile materials in sultable carrier salts. The
program is an outgrowth of the ANP efforts to make a molten-salt reactor
power plant Tor aircraft and is extending the technology originated there
to the development of reactors for producing low-cost power for civilian
UELS.

The major goal of the program is to develop a thermal breeder re-
actor. Fuel for this type ol reactor would be 233UF4 or 2'35UF4 dissolved
in a szalt of composition near 2LiF-BeFs. The blanket would be Thly dis-
solved in a carrier of similar composition. The technology being devel-
opad Tor the breeder is applicable to, and could be explolited sooner in,
advanced converter reactors or in burners of fissiocnable uranium and plu-
tonium that also use fluoride fuels. Bolutions of uranium, plutoniurm,
and thorium szalts in chloride and fluoride carrler salts offer attractive
possibilities for mcobile fuels for intermediate and fast breeder reactors.
The fast reactors are of interest Loo bubt are not a significant part of
the progran.

Our major effort is being applied to the development, construction,
and operation of a Molten-Salt Reactor Bxperiment. The purpose of this
Experiment is to test the types of fuels and materials that would be used
in the thermal breeder and the converter reactors and to obtaln several
years of experience with the operation and maintenance of a small molten-
salt power reactor. A successful experiment will demonstrate on a small
cscale the attractive teatures and the technical feasibility of these sys-
tems for large civilian power reactors. The MSRE operates at 1200°F and
at atmospheric vrecsure and willl generate 10 Mw of heat. Initlially, the
fuel conbains 0.9 mole % UF., 5 mole % ZrF,;, 29.1 mole % BeF,, and 65
mole % LiF, and the uranium is about 30% °°°U. The melting point is
840°F, In later operation, we expect to use highly enriched uranium in
the lowzr concenbration typilcal of the fuel for the core of a breeder.

In each case, the composition of the solvent can be adjusted to retaln
about the same liguidus temperature.

The fuel cirvculates through a reactor vessel and an external pump
and heat-exchange system. All this egquipment is constructed of Hastelloy
N,* a new nickel-molybdenum-chromium alloy with exceptional resistance
to corrosion by molten fluorides and with high strength at high tempera-
ture. The reactor core conbains an assembly of graphite modzsrator bars
that are in direct contact with the fuel. The graphite is a new material?
of high density and small pore size. The fuel salt does not wet the
sraphite and therefore should not enter the pores, even at pressures well
above the operating pressure.

 

LA so s0ld commercially as Inco No. 806.
“Grade CGB, produced by Carbon Products Division of Union Carbide
Corpe.
Heat produced in the reactor is transferred to a coclant salt in
the heat exchanger, and the coolant salt is pumped through a radiator
to dissipate the heat to the atmosphere. A small facility is installed
in the MS5KEE building for occasionally processing the fuel by treatment
with gaseous HF and Fs.

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

Fabrication of the reactor equipment was begun early in 1962. Some
difficulties were experienced in obltaining materials aand in making and
installing the equipment, but the essential installations 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 June 1, 1965, The zero-
power experiments were completed carly in July. Additiocnal modifica-
tions, maintenance, and sealing and testing of The containment were
required before the reactor began to operate at appreciable power. This
work was completed in December, and the power experiments were begun in
January 1966. The reactor had been operated for a short time at 1 Mw
at the time of this report. lIurther increases in power were delayed by
difficulties with the off-gas system.

Because the MORE is of a new and advanced type, substantial research
and development effort is provided in support of the design and construc-
tion. Included are engineering development and testing of reactor com-
ponents and systems, metallurgical development of materials, and studies
of the chemistry of the salts and their compatibility with graphite and
metals both in-pile and out-of-pile. Work is alsc being done on methods
for purifying the fuel salls and 1in preparing purified mixtures for the
reactor and for the research and development studies. Some studies are
being made of the large power breeder reactors for which this technology
is being developed.

This report is one of a series of periodiec reports in which we de-
scribe briefly the progress of the program. ORNL-3708 is an especially
useful report because it gives a thorough review of the design and con-
struction and supporting development work for the MSKE. It also de-
scribes much of the general technology for molten-salt reactor systems.
Other reports issued in this series are:

ORNT,-2474 Period Ending January 31, 1958
ORNL-2626 Period Ending October 31, 1958
ORNI-2684 Period Ending January 31, 1959
ORNL-2723 Period Ending April 30, 1959
ORNT,~2799 Period Ending July 31, 1959

ORNL,-2890 Period Ending October 31, 1959
ORNL~2973
ORNL-3014
ORNL-3122
ORNL-3215
ORNL-3282
ORNL-3369
ORNL-3419
ORNI-3529
ORNL-3626
ORNL-3708
ORNL-3812
ORNL~3872

Periods Ending January 31 and April 30, 1960

Period
Period
Period
Period
Period
Period
Period
Period
Period
Period

Period

Ending
Ending
Ending
Ending
Ending
Ending
Ending
Inding
Ending
Ending
Ending

July 31, 1960
February 28, 1961
August 31, 1961
February 28, 1962
August 31, 1962
January 31, 1963
July 31, 1963
January 31, 1964
July 31, 1964
February 38, 1965
Auvgust 31, 1965
Part 1. MSRE OFPERATIONS AND CONSTRUCTION, ENGINEERING ANALYSIS,
AND COMPONENT DEVELOPMENT
1. MSRE OFPERATIONS

Chronological Account

 

Preparations for operation at high power were completed, and the ex-
perimental program was resumed. The power ascension was interrupted at
1 Mw, however, by partial or complete plugging at several points in the
fuel off-gas system. The plugging materials were identified as organics,
probably the products of oil decomposition.

Figure 1.1 outlines the major activities in the period covered by
this report. A brief account follows; details are given in later sec-
tions.

Two of the larger modification jobs scheduled before power opera-
tion, the coolant line anchor sleeves and the installation of new ra-
diator doors, were completed in August. Late that month, the assembly
of graphite and Hastelloy N surveillance SpéClmenS,'Whlch had been in
the core from the beginning of salt operation, was removed. While the
reactor vessel was open, inspection revealed that pleces were broken
from the horizontal graphite bar supporting the sample array. The pleces
were recovered for examination, and a new sample assembly, degsigned for
exposure at high power and suspended from above, was installed.

The fuel pump rotary element was removed in a final rehearsal of
remote maintenance and to permit inspection of the pump intermals. It
was reinstalled after inspection showed the pump to be in very good con-
dition.

Tests had shown that heats of Hastelloy N used in the reactor vessel
had poor high-temperature rupture life and ductility in the as-welded
condition. The vessel closure weld had not been heat treated, so it was
heated to 1400°F for 100 hr, using the installed heaters, to improve these
properties.

At the conclusion of the heat treatment, the reactor cell was sealed
for the first time. The drain cell had already been sealed, and some of

ORNL--DWG 66 -3546R

 

 

 

 

 

 

 

 

 

 

FUEL-PUMP } {
INSPECTION CONTAINMENT TESTING FUEL CIRCUIL ATlON
VAT VR T T PR A 22T Ao o A 2 o it -
HEAT TREATING PP FLUSH FOOLANT CIRCU{ ATION
REACTOR VESSEL STREssgs & CETEEZR oo IR0
CORE ] V2T RADIATOR
— 4 \ OFF GAS SYSTEM
%AIVIP FS WIRING T T A
""" DOQFS_'MECHANSMS EZITE I
E’_}?‘;'A&] ................... Q]
OPERATOR TRAININF
'-2 m .........
AAAAAAA -_SEPT OCT fi DEC JAN FEB

 

 

 
the closure devices on containment penetrations had been taested. Now test-
ing of the containment provided by the reactor cell, drain cell, and vapor-
condensing system became the primary effort. After preliminary tests at
pressures up to 5 psig, the program was interrupted on October 21 to in-
stall Masonite sheets between the cell mewbranes and upper blocks to
modify the access opening over the core. Testing at 20 psig disclosed
many small leaks at penetrations, which were repaired (many'while the
reactor cell was cpen for ten days for straln-gage measurements of pip-

ing stresses). After the repairs, leakage rates were measured at 10,

20, and 30 psig. Ixtrapolation to 39 psig (the peak pressure in the
maximum credible accident) gave a leak rate of O.4%/day, compared to
l.O%/day assumed in the safeby analysis. Leakage rates at —2 psig (the
normal operating pressure) were measured to serve as a reference during
subsequent operation. This program was completed on December 5.

Stiresses in the reactor cell piping and vessels were caleulated in
detalil to permit evaluation of the service life of the Hastelloy N parts
under irradiation. Some adjustments of supports were made to minimize
stresses, after which the calculated stresses were acceplably low ex-
cept at the heat exchanger nozzles, where a complicated geometiry made
calculations unreliable. The strain-gage measurements in early Novem-
ber showed tolerable stresses at this point also.

While the contaimment testing and strain-gage measurements were
under way, the operators and supervisors underwent further training
with emphasis on power operation. Classroom lectures were followed by
practice on a simulator which included the actual controls, instrumen-
tation, control rods, and radiator doors.t Examinations and certifica-
tion of qualified operators followed.

Early tests with the radiator hol and the exercises during simula-
tor practice showed that the operaticn of the radiator doors was unre-
liable. TFour weeks in November and December were spent in modifying
and adjusting the door rollers, tracks, seals, and limit devices before
tests showed they would operate reliably hot or cold.

Air leakage from the radiator enclosure when the main blowers were
operated proved to be excessive, both from the standpoint of coolant-
cell ventilation capacity and because of excessive heating outside of
the enclosure. Hoods were installed, into which the radiator doors re-
tracted, and the sheet-metal enclosure was generally tightened and modi-
fied before leakage became acceptable. (Even after the lmprovements it
was necessary to supplement the cell exhaust with ducting to one annulus
blower to attaln a negative pressure in the coolant cell.)

When the main radiator blowers were operated with the radiator hot,
ailr leaking from the top of the enclosure overheated electrical insula-
tion in that area. It was necessary to Install ceramic insulation on
leads on top of the radiator and revoute the leads to cooler locations.
Ducting was also installed to redirect cooling air flow across the top
of the enclosure where tne door hoods had blocked the original flow pat-
terns.

The radiator work lasted from early Novemper to mid-January, delay-
ing the filling of the coolant system and the start of power operation.
As socon as conbaloment testing was Tinished, the insbruments, con-
trols, and equipment were given the checkoubs required prior to startup.
The fuel system was then heated, and flush salt was clrculated for three
days., oSamples of the flush salt taken at this time were analyzed for
oxides by an improved, more reliable method. Results averaged lessg than
100 ppm, well below tolerable levels, (Evidently the measures taken o
avoid oxygen contamination while the reactor vessel and fuel pump were
open were effective.)

Fuel salt was charged into the loop, and the reactor was taken crit-
ical on Decenmber 20. Between then and January 18, when the ccolant loop
was tilled, nuclear experiments were restricted to powers below 25 kw,
Even so, useful measurements were made on {lux distributions in the
thermal shield and beside the reactor vessel, control rod shadowing el-
fects, and zero-power kinetics of the nuclear system. At the same time,
numerous fuel-salt samples were analyzed, showing uranlum in excellent
agreenent with expectations, low oxide concentration, and practically
no corrosion.

With the coolant system in operation, the power was raised to 100,
500, and then 1000 kw as heat was extracted at the radiator. Dynamics
tests and heat balances were conducted at each power. On January 23,
while the power was at 500 kw, the fuel pressure control valve (or its
filter) showed signs of plugging, but the situvation cleared up in a Tew
hours. There was also evidence of an abnormal restriction in the egqual-
izing line between the fuel pump and the drain tanks. The next day the
reactor was taken to 1L Mw, and a few hours later signs of intermittent
plugging in the fuel off-gas line again appeared. Lt was established
then that the equalizing line was completely plugged. When 1t was also
discovered that the auxiliary vent line was almost completely plugged,
the reactor was taken subcritical. Lfforts to blow oub the plugs in the
equalizing line and the auwxiliary vent line failed, and the fuel and
coolant loops were drained to shub down the reactor. (The syslens were
kept hot, with helium circulating.)

The check valve in the vent line to the auwxiliary charcoal bed was
removed, and the line was recomnected. Gas then freely passed through
the lines. The fuel drain cell was opened, and the flow-restricting
capillary in the equalizing line was removed. This line then proved to
be clear. The check valve and capillary were taken to the High-Radiation-
Level BExamination Laboratory (HRLEL), where they were observed to contain
small aocunts of intensely radiocactive material. They were not plugged,
however.

Meanwhile the plugging of the fuel pressure conbrol valve (or filter)
recurred. Blowing backward throusgh the line did not clear the obstruc-
tion, so the assembly of valve and filter was removed. A new valve was
installed, and the assembly was replaced to see 1f it were open. (This
was simpler than flow testing the contaminated parts outside the sys-
tem.) At first the pressure drop was acceptable, but over a few hours,
excegsive pressure drop again developed. After the assembly was removed,
the new valve was proved to be c¢lear, so the filter was taken to the
HRLFEL for examinaticn. The old pressure control valve was alsc cut open
and examined in the HRLEL after flow tests showed its C was a factor of
10

3 below ils original value. There was some solid material in the pores
cof the Tilter, but the pressure drop across the Tilter was not as high
as 1t had apparently been while installed. Inspection showed that the
valve trim was coated with oll and varnish-like material.

The total amount of material in the valves and filter was quite
small. Therefore we decided to resume opevations, using a Tllter with
larger pores (50 p instead of 1 w) and controlling pressure with a hand
valve (CV = 1.0) already in the line instead of the pressure control
valve (CV = 0.07). (The pressure control valve was eliminated.)

During the operation at significant power, activation of potassium
in the cooling water corrosion inhibitor had been more than expected.
While the reactor was shubt down, the original corrosion inhibitor was
replaced with i nitrite and borate, which would give only a small
amount of activity (britiom).

Also while the reacltor was down, pressure drops across the charcoal
beds were measured. 'There had been intermitient indications of excessive
pressure drops during the L-Mw operation and just afterward, bub when
measured before the next startup they were normal. Operators and ex-
tension handles on the charcoal-bed hand valves had been repaired during
this interval. (Ice, rust, and binding had prevented easy operation,
and setscrews in two of the extensions had sheared.)

Operation was resumed, and fuel was circulated for 44 hr before the
power was raised to 1 Mw. Lt was necessary to gradually open the hand
valve being used for pressure contbrol aboub l/4 turn during this period
in order to hold the fuel pressure in the range of 4 to 6 psig. The
implication was that a small amount (perhaps 0.01 in.) of material had
built up on the valve trim. Within 3 hr after going to L Mw, the ap-
parent accumulation of material accelerated. The valve was adjusted to
maintain control, but a short time after makling an unusuvally large ad-
Justment, the differential pressure across the charcoal bed and valves
suddenly increased, indicaling hlockage in those lines.

Jugt at this point, the motor on the space cooler in the fuel drain
cell failed. Rising temperatures in this cell then required that the
Tuel e drained and gecured s¢ the cell could be opened.

The remainder of the period was spent in replacing the space cooler
motor and Investigating the trouble in the off~gas system.

Analysis of lxperiments

 

Reactivity Balance

 

The reactivity balance involves calculation of the effects of changes
in all known variables affecting reactivity. The net effect of changes
between any two times that the reactor is critical should be gzero if
the calculations are acceuwrate., Deviations from zero indicate eifher
error in the measurements and calculations or some unforseen effect.
11

The reference conditions Tor the reactivity balance were established
in run 3 (the zero-power experiments which ended July 4, 1965). Between
then and the startup in December, one variable changed significantly —
the 227U concentration in the fuel. Because of uranium additions to the
circulating fuel, at the end of run 3 the 233 concentration in the loop
was apbout lO% higher than that ian the sgalt remsining in the drain tank.
Thus the drain and mixing at the end of run 3 caused a substantial de-
crease in 227U concentration, but one that could be caleulated. On the
other hand, the fuel loop was flushed al the end of run 3 and the be-
ginning of run 4, resulting in mixing into the fuel an amount of flush
salt which was probably small but was not directly measurable. Before
criticality in run 4 the 2357 conceuntration was calculated, assuming no
dilution with flush salt. The change in concentration from rux 3 was
converted intc reactivity change, using the concentration coefficlent
of reactivity that had been cobserved in the zero~power experiments.2
The resultant reactivity was converted to a predicted control-rod con-
figuration for criticality at 1200°F and zero power, using the control-
rod polsoning representation based on the rod calibrations. (See "MSRE
Reactor Analysis,” this report, for a discussion of the egquation for rod
poisoning.) The predicted critical position of the regulating rod (rod
1) with the two shim rode withdrawn 34 in. was 24.7 in., and the observed
critical position was 23.3 in. This difference, equivalent to only 0.08%
Bk/k, may he due to several factors, lncluding (1) differences in temper-
ature measurement, (2) inventory errors at the end of run 3, and (3)
errors in the rod-poisoning calcwlabion. The most important conclusion
to be drawn from this is that there was no significant dilution of the
fuel salt by the flush-salt operation.

Critical operation of the reactor during runs 4 and 5 provided an
opportunity for checking out some parts of the reactivily-balance cal-
culstion. 'The computer-executed reactivity balance was nct used be-
cause (1) the accurate representation of control~rod poisoning had nob
been incorporated into the program, (2) the calculation of xenon poison-
ing was not fully developed, and (3) there were indications of other
systematic errors in the program that had anot been resolved. However,
several manual calculations were made at very low powers where the fig-
sion product poisoning and fuel~burnup bterms are insignificant. Some
calculations were also made at powers up to 1 Mw put without the xenon
and. samarium terms. :

Tn general, the zero-power reactivity balances showed a variation
of +0.02% 8k/k. This corresponds to a variation of *0.3 in. in the
critical position of the regulating rod or +3°F in system temgerature
and probably represents the limit of accuracy of thig portion of the
calewlation.

Calculations after up to 24 hr of operaticn at 1 Mw showed no mea-
surable change from the zero-power reactivity valances. Neglect of the
fission product terms should have resulted in a change of about -0.2%
&k/k if the anticipated xenon behavior had oceurred. This is much larger
than the scatter in the measurements, so one can conclude that the xenon
poigoning was actually much less than predicted.
12

The difference belween indicated and actual position of control rod
3, which developed after the start of run 4 (see p. 54), was reflected
ag a shift of about 0.05% 6k/k in the zero-power reactivily balance.
This demonstrated the abllity of the reactivity balance to show up small
changes, at leastc under simple conditions, that is, no Tission product
poisoning and low power.

Power Calibration

 

One cbjective of the early power tests was to correlate the thermal
power of the reactor and the oubputs of the various neutron sensing ele-
ments. The principal requirement of such a calibration is an independent
method for measuring the thermal or fission power of the reactor. Vari-
ous methods were used in the apprceach to 1 Mw with reasonably good agree-
mexnt .

The first approximation, used throuvghout run 3, was based on cal-
culatiocns of the inherent (alphamn) source in the fuel and subecritical
multiplication. "The primary purpose of this calibration was to permit
positioning of the safety chambers so that they would provide a control-
rod scrawm at a few kilowatis and still not interferce with the planned
experiments at lower powers. Absolute accuracy was not required since
none of the experiments lnvolved operation at powers where thermal ef-
Tects were gpparent.

A thermal-power calibration was made shortly after the start of run
4 at a nominal 25 kw. At this time, only the fuel loop wasg full and
circulating; the thermal power was evaluated frowm the rate of tenpera-
ture rige and the kncwn neat capacity of the loop. Thisg provided a
basis for subsequent operation at higher powers and, incidentally, showed
that the initial source-power calibration had been in error by only 8%.
A gimilar test was performed al higher power with both the fusl and
coolant loops full and circulating. However, the second test was di-
vided in two parts. First, the nuclear power was set at about 75 kw,
and 75 kw of electrical-heat input to the loop was turned off. These
operations did not lead to a steady temperature, but the change in tem-
perature slope witnh time before and after the increase to 75 kw was used
to correct the thermal power. The nuclear power was then increased by
50 kw, to 125 kw, and the increase in the rate of temperature rise was
measured. The thermal-power calibration from these two steps was within
5% of the 25-kw calibration.

several overall heat balances were calculated with the nominal nu-
clear power at 1 Mw. The results in run 4 varied between 0.94 and 1.18
M, with an average value of 1.0l Mw. In run 5, the steady-state heat
balances at an indicated power of 1 Mw gave 1,19 = 0.02 Myw. The reagons
for this apparent discrepancy have not been established.

Filux Measurements

 

Flux measurements were made in the neutron source tube to determine
1T operation at power would regenerate the external neutron source (Am-
13

Cm-Be) to a significant degree. The re%eneration would take place as a
result of thermal-neutron captures 1n e lAm‘(Tl/g = 462 years), which
would produce the more active 2420 (Tl/g = 162.5 days). The measure-
ments were made using cadmium-covered and bare copper folls to determine
the thermal-neutron flux at the permanent source position, an elevation
of 830 £t & in. The thermal flux at that point was found to be 3.8 X 107
neutrons cm? sec”* with the reactor at 1.02 kw. This extrapolates to
3.7 X 10*1 neutrons cm™® sec™t at a power of 10 Mw. Using conservative
assumptions, it was calculated that the source strength will be kept at
an acceptable level for at least 1-1/2 yvears after power operation has
begun.

Flux measurements were also made in the reactor furnace (™~ in. from
the reactor vessel wall) in the location later occupied by metallurgical
surveillance specimens. The measurements were made using gold and copper
foils. The foils were sealed in a stainless steel envelope from which
alr had been purged to protect the copper from oxidation during the irra-
diation at 1200°F. Fluxes were measured along a vertical line 80 in.
long. The peak flux (at approximately the core midplane), extrapolated
to a power of 10 Mw, was 7 X 1012 neutrons cm™? gec”t. Fast and thermal
contribubions to this flux were 4.8 X 10Y® neutrons cm™? sec™ and 2.3 X
10%2 neutrons cm™? sec™t respectively.

MSRE Dynamic Tests

 

Description of Dynamic Tests

 

A number of dynamic tests were performed during operaltion at zero
power and subseguently at power levels up to 1 Mw. The purpose of these
tests was to investigate the inherent stability of the sgystem, as well
a5 1o evaluate the mathematical models and the paramseters that were used
to predict its dynamic characteristics.

The inherent stability of the system (i.e., without automatic con-
trol) was the subject of an exhaustive theoretical analysisB in which 1T
was concluded that the reactor would be stable at all powers and that
its stability characteristics would improve as the power level increased.
Teste made on the reactor operating without servo conitrol at power levels
of 75 kw, 465 kw, and 1 Mw confirmed these predictions, as shown in Fig.
1.2, which shows the response of the power to arbltrary perturbations.
The improvement in the anatural damping characteristics and the decrease
in the period of oscillation with increasing power level are clearly
shown. Figure 1.3 shows a comparison of the experimental and theoratical
veriods of oscillation.

A subject of considerable interest during the MoRE development was
whether or not the reactor power would "wallow" when operating without
servo control. Wallowing would occur if the system were unstable or il
random perturbations in reactivity or ccoling load induced lightly damped
14

power oscillations. At both 75 and 465 kw some wallowing was observed,
but the power level perturbations were less than ifi%. In voth cases
neither of the main radiator blowers was on, so the bulk of the cooling

load was

due to natural convection and radiation, and random fluctuations

in load were secen. When one radiator main blower was Lturned on for the

ORNL-DWG 66~-47861

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

7O o P —e
- POWER 22 75 kw ‘ \ 1 ‘
£ 60 ff —F | o
x5 | | - i“‘mfi \J
5> , —~T J’»i:fi
5 ¢ T ‘
o = | i
= 1
g e ) e .
= i
,,,,,,,,,,,, | B
P ‘ I
= | 5
o 2 - — . R
Ll oo |
g0 | | - ‘ | e
T = 20 b - e — b e P i . _.l_ .
2 POWER = 465 kw \ i | |
i ! i !
0 ‘ L i : — ;
. 80 — 71— e ey - — T T e e
2 POWER = t Mw | ;
o 5 70 b I . _T‘_ - C e e L -
Y L /‘HMA‘ -------------------------- s ‘ i
D | i
£z 60t ] g b —
& . | i !
— 5 - e — — e a1
o 400 800 1200 1600 2000 2400 2800 3200 3600
TIME (sec)
1..2. TInherent Response of MSRE at Three Low Power Levels.

OQRNL-DWG 66—-4782
100 ooy 4 - e

50
-
SobSNdx e e e
= IS
Z 20
=
a X—EXPERIMENTAL
= 10 X THEORETICAL ..l 1l
(6] - .
o beeed 4y vy ™ Lyt
<o e+ eyt
5
O
<
o=
=S S I 1 (N O I S . I N 0 0
o

2
1 .................
0.03 C.05  OA 0.2 0.5 1 2 5 10 20

Ng, NUCLEAR POWER (Mw)

Fig. L.3. Comparison of Predicted and Meagured Inherent Periods
of Oscillation at Low Powers.
15

1-Mw case, however, the bulk of the cooling was by forced convection, and
the power level fluctuations were reduced considerably. Thus, at higher
powers, negligible inherent low-frequency neutron level fluctuations can
be expected since the load will be steadier and the inherent damping of
the system even greater.

f'requency Response Tests

 

Poth pseudorandom binary tests and pulse tests were carried out in
order Lo determine the frequency response of the system. The pseudorandom
binary tests consisted of specially selected periodic series of positive
and negative reactivity pulses. The advantages of this type of signal
for frequency response testing in the MoRE are:

1. Tne signal may be generated with standard reactor hardware. The
sequences were generated using the MSRE digital computer and a portable
analog computer to control the position of a regulating rod.

2. A large nmumber of frequencies may be analyzed in a single test,
with less time required to carry out the experiments compared with that
required for osecillator tests, which analyze one frequency at a time.

3. 'The loecation and relative amplitudes of the harmonics of the
input signal may be adjusted by varying the basic pulse or "bit time"
and the number of bits in the sequence.

4o The signal power is concentrated in specific harmonic Trequen-
cies. This gives a better effective signal-to-nolse ratlo than that in
nonperiodic signals with a continuum of frequencies.

5. The average value of the perturbation 1s essentially zero, re-
sulting in a small net upset of the system.

Pgeudorandom binary tests were run at power levels of 100 w, 75 kw,
465 kw, and 1 Mw. The particulsr sequenceg and thelr duration were
selected zo that the expected low-freguency resonance 1ln the Treguency
response amplitude? could be identified. Sequences with 19, 63, 127, and
511 bits per seguence were used. As will be shown below, this type of
test worked quite satisfactorily.

Single reactivity pulse {eslts were also run, but the resulting
transients were inadequate for frequency response anslysis. Since the
signal strength of a single pulse is distributed continucusly over a wide
freguency range, pulse tests require a comparatively low-noise signal.

The low-Trequency load perturbations seen at low power reduced the signal-
to-neoise ratio to an unacceptable value in the freguency range of interest
(0.002 to 1 radian/sec). Pulse and step response tests are planned at
higher power levels, where the conditions should be more favorable.
=
o

Implementatlion of Pseudorandom Binary Tests

 

The pseudorandom binary reactivity idnput required a very precisely
controlled series of regulatling rod insertions and withdrawals. Since
the frequency range of interest was from about 0.002 to 1.0 radian/sec,
the rod Jogger was designed so that sequences with a bit time of 3 to 5
sec could be run for as long as 1 hr. The rod Jogger systenm, which re-
quired no special-purpose hardware, consisted of a hybrid computer con-
troller shown schematically in Fig. 1l.4. The portable EAI-TR-10 analog
computer was used to contrel the bit time and the rod drive motor Mon"
times for the insert and withdraw commands. ‘The MSRE digital computer
was programmned to control the sequencing of the pulse train bty means of
a shift-register algorithm.4 The number of hits in the sequence could

q

be varied over a range between 3 and 33,554,431 bits.

The rod Jjogger system performed extremely well, as it was able {o
position the rod with an accuracy of about #0.01 in. (corresponding to
+0.0005% &k/k) out of 1/2 in. peak~to-peak rod travel for over 500

insert-withdraw operations.

The analog computer was also used to amplify and filter the rod-
position and power~level signals prior to digitizing. Throughout the

ORNL—DW(G 66— 4763

PSEUDO-RANDOM

 

BINARY SEQUENCE

 

PORTABLE ANALOG MSRE DIGITAL
COMPUTER COMPUTER

________ q e e e

| EAI-TR-10 I | ar-240

e 1

] BIT-TIME P

[ GENERATOR o

| | ,

| 1L | - S

| = = Ll

|

|

!

 

 

  
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

e INGERT | COMPUTATION
ROD ORIVE [ MOTOR L_l
MOTOR | WITHORAW | ON ™
| ‘ ! TIMER ‘ |
1 , | l
| _____ROD-POSITION | ]
| SIGNAL | L t [
] ’ | || AamPLIFYING — DIGITIZER AND
| AND | MAGNETIC TAPE
ROD [ FILTERING B RECORDER
POWER-LEVEL | |
SIGNAL I |
L S B 1
REACTOR

  

Fig. l.4. Block Diagram of MSRE Rod Jogging and Data Logging Sys-
tem for Pseudorandom Binary Reactivity Input Tests.
17

dynamic tests the MSRE computer was used in the fast scan mode to
digitize and store the data on magnetlic tape at a gampling rate of 4 per
second.

A typical plot of the reactor power level fluctuatlons during a

pseudorendom input test is shown in Fig. 1.5, where the points were taken
from the MSRE computer digitized record of the test.

Anelysis Procedures

 

The test data were analyzed using two independent digltal computer
methods. This duplication was done as a check on the analysis proce-
dures. The "direct method" used a digital filtering technique to obtain
the spectral density of the input and the cross spectral density of the
input and the ocutput. The Trequency response of the system at some fre~
quency ig the ratio of the cross spectral density to the input spectral
density at the frequency. It 1s necegsary that frequencies selected for
analysis be harmonilc frequencies since the input signal is perilodic. The
"indirect method" required a calculation of the autocorrelation function
of the input and the cross correlation function of the input and the
output. Subsequent Fourier analysis of these correlation functicns gave
the input spectral density and the cross spectral density. These were
then used to glve the frequency response as in the direct-method analysis.

 

   

 

 

 

Both methods gave essentially the same resulis.
ORNL-DWG 664764
42 ]
38 |— ]
=
=
0
(€9
© 34
8 |
- : '
* o : I
Lt
=
g 30 +t— —]
nh
m [
<L ] i
W : b
4 .
2
< 26— _
0 417 8.34 12.51 17.09 21.26 25.43 29.60
TIME ( 30-min FULL SCALE )
Fig. 1.5. Nuclear Power vs Time for 511-Bit Pseudorandom Sequence
with Ng = 465 kw.
18
Results

The results of the MSRE Trequency response tests for power levels
up to 1 Mw are shown in Figs. 1.0 to 1.13, along with previocusly published
theoretical predictions.? The experlmental results indicate a consist-
tently lower magnitude ratio for 6N/No)/(bk/ko) than the thecoretical
predictions at all Trequencies and at all power levels. However, the
shapes of The experimental curves clearly are consistent with predictions.
Work is now in progress to determine the reasons for the differences.
Fortunately, the theorctical work in ref. 3 included calculations of the
sensitivity of the frequency response to changes in numercus system
parameters. This information 15 proving to be valuable in establiching
which system parameters to suspect as responsible for the discrepancies.

Farly work suggests that the zero-power discrepancies are due Lo the
effect of fuel-salt plenums above and below the core. These regions in-
crease the cffective volume of the core. An increase of the effective
core volume to include one-third of the upper and lower plenums in Lhe
tnecretical model gilves agreement with experimental resulis.

The at-power theoretical results obtained with the increased ef-
fective core volume agree with experimental resulis at higher frequencies
for the at-power runs, but the discrepancies in magnitude ratio remain
at low frequencles around the peak. The computed sensitivities indicate
that the frequency response in this range is most sensitive to power
level, Np, fuel temperature coefficient of reactivity, Q%, and fuel heat

capacity, (MCp)f. Furthermore, the system equations indicatle that these

4 ORNL-DWG 36- 4765
T 3- bn SEQUENCE — INDIRECT ANALY&S °o v
~ | 63-bit SEQUENCE -~ DIRECT ANALYSIS i

 
  
  

A L%

     

i
1

 
 
   

 

 

 

fifi'}; i;.
- |-
|

 

 

| i
102 {* - LL ey fiA_IA'_L_| L _
0.004 0002 0.005% Q.01 0.02 0.05 0A

FREQUENCY {radians /sec)

Fig. 1.6. TFreguency Response — Magnitude Ratio of (8N/Ng)/(8k/kg)
Tor Ng = 100 w with Fuel Circulating.
12

ORNL-DWG 5€-4T66

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

© | T T T T
63-bit SEQUENCE ~INDIRECT ANALYSIS O

—4Q - &3-bit SEQUENCE —DIRECT ANALYSIS @

_20 ————e o L ——— e e
’é THEGRETICAL T
3 -0 b o "
B 40 |k SR — Pl s

0.5

a 0 g 2_.——07/0?)%3:;1%0'. ¢

-l T e

Q
—60 ] 77fiI- -
-70 //// ®

 

 

Q.01 0.02 Q.05 oA Q.2 0.5 1.0
FREQUENCY (radians /sec)

Fig. 1.7. Frequency Response — Phase of (EN/NO)/(SK/ko) for No =
100 w with Fuel Circulating.

ORNL-DWG 66-4767

CASE SRB-75 kw-92
511-bit SEQUENCE — INDIRECT ANALYSIS o
S41-bit SEQUENCE — DIRECT ANALYSIS 0
Lo - e

______ 1

Wy
B4/

 

 

 

 

2
10 -
0.001  0.002 0005 0.01  0.02 005 04 0.2 05 1.0
FREQUENCY (radians /sec)
Fig. 1.8. Frequency Response - Magnitude Ratio of (SN/NO)/(ak/kQ)
for Ng = 75 kw.
fig.

o

75 kw.

20

ORNL-DWG 66-4768

 

° | i T
60 [ L L | g
50 | — . | CASE SRB-75 kw-9 N

511 ~bit SEQUENCE - INDIRECT ANALYSIS ©

 

 

 

40 — —"lfi : © 511-bit SEQUENCE ~DIRECT ANALYSIS @ |
ol - L By
o A Ll

—~ THEORETlCAL i C

fol- A

;o b

S0 - ‘J(T%firri%*—
ol Tl P

-320 - - — |

 

N
_44

 

 

al e -ij_ff”ffi' -

OOO1 0.002 0.005 O 01 0.02 005 O Q.2
FREQUENCY (radians /sec)

 

1.9. Frequency Response -- Phase of (BN/NO)/(Bk/kO) for Ng =

  

ORNL--DWG 66-4769
I

  
  

CASE SRB-465 kw- 1O

  
 

—_ (SAMPUNG INIEH’VAL 1.0 sec}

~ 1 OH—=bit SEQUENCE~DIRECT ANALYSIS
1 {SAMPLING INTERVAI_ O°5 sec)

L L

IHEORETFCAL ;

 
        
    

 

 

0.Q01 0.002 0.005 0.01 0.02 0.05 (O 0.2 0.5 1.0
FREQUENCY (radians /sec)

..10. Frequency Response - Magnitude Ratio of (SN/NO)/(Bk/kg)
21

ORNL ~DWG 66~ 4770

 

 

 

 

 

 

50 51T T T T T T 11

so Lo LA CASE SRB—465 kw~1{0 jNE
! 51— bit SEQUENCE — INDIRECT ANALYSIS ©

40 | ot b — (SAMPLING INTERVAL =4.0 sec) -+

v ] . 5it- bit SEQUENCE — DIRECT ANALYSIS o | |||

{SAMPLING INTEF?VAL 0.25 sex)

 

 

 

 

 

 

 

 

 

 

 

 

PHASE (deg)
A
S

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_20 ---------- -
"f
30 cé)'
-40 - - |
|
S0 [ .
GO e —— A4
—70 ............ J—-.
—-80
0.001 0.002 0.005 0.01 0.02 0.05 04 0.2 Q.5 1.0

FREQUENCY (radians/sec)

Fig. 1.11. Frequency Response — Phase of (8W/Ng)/(8k/kg) for Ng =

ORN‘ DW(J 66-47T1

e [ T

T T T 1 T T

{CASE SF?B HVlw H) a
‘-311vb|1L SEOUENCE INDIRECT ANALYSIS 1
| 5#-bit SEQUENCE - DIRECT ANALYSIS v

{CASE 5RB-1Mw~12) .
l 127~ bl‘r SFQ‘ rI‘dC'—»\I\JD\REf"'I" ANI:\L.Y IS

 
  
 
  
  
 
 
 
 
 

 

ra
\‘
D’
:..
)

m

G
C
T
2
=
l'"i
9
A
m
o
b
=
>
—
—k
‘:_
n
>0

5”//\/ r 2

8)'(/:(

SAMPLING INTERVALS:
INDIRECT ANALYSIS 1 sec
. DIRECT ANALYSIS -0.25 sec

 

  

 

103}

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Q.004 0.002 Q.005 0.01 0.02 0.05 o 0.2 0.5 1.0
FREQUENCY {rodinns/seq)

Fig. 1.12. Freguency Response - Magnitude Ratio of (bN/No)/ ok/ko)
for Ng = 1.0 My,
27

ORNL-DWG 66-4772

 

 

 

 
    

 

 

TR T fi—'—'r_'“%—*—’“—v—‘ . R o ‘|
- CASE SRB-{Mw—*H .
54 - bit SEQUENCE — INDIRECT ANALYSIS e
~ 5{4—bit SEQUENCE — DIRECT ANALYSIS a
. CASE SRB-4Mw-12
20 : 127-bit SEQUENCE — DIRECT ANALYSIS a
@ ; ; | SAMPLING INTERVALS : INDIRECT ANALYS!
2 30 \— s e ’ l_4seCL S
w L SAMPLING INTERVALS: DIRECT ANALYSIS
w20 b — S —
< : ; —0.25 sec
&40 e — oo - - -
5 ' _—~THEDRETICAL
— L } e R
S0 e e o b —
i !
| o ‘
—20 o — I — e e ——— .< — o TT \‘ e .- -
I . ! Vo ‘ }
-3 e b T
© S firfifiéfif | “?:
A0 e | — a«fififiq; g%& e
e e P Tl M el
C.001  0.002 0.005 0.0t 0.02 0.05 0.4 0.2 0.5 1.0

FREQUENCY (radians/sec)

Fig. 1.13. Frequency Response - Phase of (6N/N0)/(6k/ko) for Ng =
1.0 Mw.

parameters appear as a quantity No@f/(MCp)f. It was found that excellent

agreement helween theoretical and experimental results may be obtained at
all frequencies and all three power levels 1if, in Lhe theoretical calcu-~
Jations, the increased effective core volume 1s used and the guantity

Noaf/(MCp)f is inecreased by a factor of 1.5 to 2. The plausibility of

such changes 15 now belng examined.

Temperature Response Test

 

In order to get a more accurate estimate of the fuel temperature
coefficient of reactivity, the hot-slug transient test as reported
earlier’ was modified and rerun in January 1966. In this test the re-
actor power level was controlled by the flux servo at 100 w with the
fuel loop circulating and the cooclant loop stagnant. After heating the
stagnant coolant salt about 20°F hotter than the fuel sali, the coolant
pump was started, producing a2 hot slug of fuel in the heat exchanger
which was subsequently introduced intc the core. 1In this test, then,
the change in reactivity was due entirely to the change in temperature,
and the rod-induced reactivity required to keep the reactor critical was
approximately equal and opposite to that due to the Tuel and graphite
temperature change. ©Since the graphite responds slowly to changes in
fuel temperature, the immediate effects on reactivity can be attributed
to the fuel temperature coefficient alone. Hurther corrections have yet
23

ORNL-DWG 66-4773

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

9-REGION CORE MODEL bl el +

® EXPERIMENTAL RESULTS FROM A 1
MOT-SLUG TEST USING SAMULON'S \ ‘
02 | METHOD; SAMPLING INTERVAL = A L
0.25 sec ‘
0.4 ’

0.001 0.002 Q.005 0.0 0.02 0.05 A 0.2 0.5 1.0
FREQUENCY ({radians /sec)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 1.14. TFrequency Response — Magnitude Ratio of Reactor Outlet
Temperature Response to Inlet Temperature Perturbations.

to e applied to the results of the test, but a preliminary estimate of
~4.7 % 10° (°F)1 has been obtained. 'This compares Tavorably with the
predicted value of —4.84 x 1077 (°F7)~! used in the theoretical calcu-
lations.

The transfer function of reactor outlet temperature respconse to
changes in reactor inlet lLemperature was also calculated from this test,
and the plot of magnitude ratio vs frequency is shown in Fig. 1l.l4. DNote
that the experimental curve indicates much greater attenuation over most
of the freguency range shown. This is probably due to higher-than-ex-
pected rateg of heat transfer to the graphite, since experimental mixing
studies’® indicated that there is very little attenuation due to mixing
in the frequency range below O.1 radian/sec. Reasons for this discrepancy
are being studied.

oystems Performance

 

Off-Gas System

Some difficulty with plugging and partial restrictions in the MSRE
off-gas systems has been encountered at various times in the operation
of the reactor systems. In the past these obstructions have developed
at in-line Tilters and pressure-control valves with extremely small flow
passages. During early operations, the plugging of filters was attributed
to having filters with inadeduate Tlow areas to accommocdate the material
1o be removed from the gas gLireams. (For example, the original filter
in the coolant off-gas line, 528, was a tubular element only 1 in. long
by 1/2 in. in diameter.) The plugging of valves appeared Lo be a re-
sult of filters which were too coarse to remove all the particles that
were capable of causing plugs.

During shutdown which followed the zero-pewer experiments (zun 3),

larger filters, capable of removing pvarticles greater than 1 p in di-
ameter, were installed ahead of the pressure~control valves in both the
24

fuel and coolant off-gas lines. The fuel-system pressure-~control valve
(Pcv 522), which had exhibited some ervatic behavior in run 3, was tested
and found to have the same pressure-drop characteristics that it had when
first installed. This valve was reinstalled for the next operation of
the reactor. No further difficulties were encountered with the coolant
of f-gas system, bul several problems developed in the fuel system. (See
Fig. 1.15 for a simplified flowsheet of the fuel off-gas system.)

The performance of the fuel off-gas system was satisfaclory during
the suberitical and low-power (up to 25 k) operation of the reactor in
December 1965 and January 1966. The first indication of difficulty in
this run (run 4) occurred on Janvary 23 with the reactor power at 500 kw.
The total integrated power accumulated up to this time was about 3 Mwhr.
At that time the system pressure began Lo rise slowly, indicating either
a restriction in the vieinity of PCV 522 or a malfunction of the valve.

(A routine check of system pressure interlocks on the previous day had
indicated normal behavior of all components.) shortly after the start

of this pressure transient, the response of the drain-tank pressures
indicated an abnecrmal restriction at a capillary flow restrictor in line
521, the fuel-loop-to-drain-tank pressure equalizing line. The excess
pressure in the fuel system was relieved by venting gas through HCV 533

to the auxiliary charcoal bed until the restriction at PCV 522 was cleared
by operating tne valve several Gtimes through its full stroke. These opera-

ORNL-DWG €£6-2422,

FE 524 N
B e

Cv 528

   
  
 

V 524B FROM COOLANT
Eo- b SYSTEM AND
80TH OIL SYSTEMS

   

 

GAS HOLDUP
AMD COOLER

Pd| e
170 liters 556
T
PCY —
3 o2 V5228 VOLUME
(ki ¥ HOLDUP
AN 185 fters

v 5188 % SAMPLE STATION

A .
HOV 533 e D—ba—T VAIEE
v 5188 VAIBCH[ Ty 51ap v 622

CV533 (T
. V51852 ~ vsace
FE 521

 

 

 

 

 

 

 

(T e

V 51883 518C Y
o ¥ V518G

HCY  HCY  HC IS & mem ve20

545 546 544 r :

 

 

 

   
 

V 5384
YV 561

  

AUXILIARY
CHARCOAL BED
IACB)

V5624

 

 

   

FD-1 FROM SAMPLER
ENRICHER

Fig. 1.15. BSimplified Flowsheet of MSRE Fuel Off-Gas Sysbem.
25

tiong slso revealed at least a partial resgtriction in the line Lo the
auxiliary charceoal bed, apparently at HCV 533, The reactor was made sub-
critical after 4-1/2 hr operation at 500 kw.

Satisfactory pressure control was maintalned for sbout 24 hr, but
the plugging at HCV 522 recurred shortly after the reactor power was
raiged to 1 Mw, Again, the operation of this valve fhrougn its full
gtroke was effective in reliceving at least part of the restriction.
However, oun this occasion, evidence of restrictions at the charcoal-bed
inlets began to appear. These restrictions were bypassed by putting
the two spare charccal-bed sections (24 and 2B) in service and cloging
the inlet valves to the two that were restricted (1A and 1B). Within
sbout 6 hr the inlets of beds 28 and 2B also plugged. When the pre-
vicusly plugged beds were checked, at least cne was found to be clear
and off-gas flow was reestablished through bed 1B. The manipwlation of
the charcoal-bed inlet valves revealed substantial mechanical difficul-
tles with the manual operators. These were caused by misalignment of
extension handles and corrosion due to weabher exposure and resulted in
two of the valves becoming inoperative.

The combination of difficulbies in the off-gas system rezulted in
a reactor shutdown to correct the conditions. The restrictions in the
equalizing line (521) and the suxiliary vent line (533) were relieved.
by removing the capillary (FE 521) and a check valve (CV 533) respec-
tively. The valve and filfer assembly, and a second valve that had been
tried vriefly, were removed from line 522. 'The original valve and the
filter were gubsequently subjected to detailed examinations which are
deseribed in "Component Development,” this report. Pressure-drop meas-
urements showed a higher pressure drop by a factor of 3 Lo 4 in the
valve and a factor of 10 in the filter. However, the observed restric~
tions, particularly in the case of the Tilter, were not high enough to
account for the observed pressure behavior in the operabing system.

While the reactor was shut down, several measurements were made of
the pressure drop across individual sections of the charcoal beds.
Measurements made shortly after the shutdown showed essentially normal
pressure drops for three of the four sectionsg. The pressure drop across
the fourth section (1B) was about 60% higher than thal scross the others.
(These meagurements include the pressure drops across the inlet and oub-
let valves of each bed section.) Although nothing was done to correct
this condition, measurements made Jjust before the next startup (run 5)
showed that all four sections had essentially identical, normal pressure
drops. Mechanical repairs had been made Lo the operating mechanisms of
the valves, but these did not affect the flow characteristics. Measure-
ments were also made of the pressure drop across the auxiliiary charcoal
bed; no abnormalities were observed there.

The preparations for the next startup included installation of a
large, relatively coarse (50-u) filter and a large, open hand valve at
the PCV 522 location. This arrangement eliminated the swall, easily
plugged passages associated with PCV 3522, and 1t appeared that the filter
would still remove any particles large enocugh to plug valve 5228, which
was bto be used for system pressure ccentrol. (Satisfactory manual pres-
sure control using valve 522B had been demonstrated during the shutdown. )
26

In the next operation of the fuel loop (run 5), fuel salt was cir-
culated for 44 hr with the reactor suberitical or at very low power
(100 w) vefore the power was raised. The evidence is not conclusive, but
there was some indicatlion ol an dncrease in flow resistance at wvalve
5228 during this time. ©Shortly after the power was raised to 1 Mw,
the increase in flow resistance at 522B accelerated sharply. Freqguent
adjustment of the valve was required to maintaln reasonable pressure
control. The restriction at the charcoal beds which occurred after
one such adjustment developed quite rapidly and resulted in almost com-
plete plugging of the two beds that were in service (1A and 1B). After
the reactor drain, which was reguirved because of the failure of the
space-cooler motor, it was debermined thal the auxiliary charcoal bed
was also restricted. This bed comes into service automaltically during
a drain as a backup to rellieve any excess gas pressure. The helium in
the fuel system was then released through the two standby charcoal beds.

Detailed investigatlions after the shutdown showed that the charcoal-
bed restrictions were probably at the inlet valves. Valve 621, the in-
let valve to bed 1B, was removed for examination in the HRIEL. The re-
sults of these examinatlions are described in the section entitled "Com-
ponent Development.' The pressure drop across bed 1B with valve 621 in
place was very much higher than normal. With the valve oul, the pres-
sure drop was nmucen lower bub still higher than for a normal, unrestricted
ted. An excess of helium was then forced through the bed for a few min-
utes, and the pressure drop suddenly decreased. BSubsequent pressure-drop
measurements under controlled conditions gave values which were normal
for an unrestricted bed.

everal gas samples were taken from the fuel loop during this shut-
down 1n an effort to identify some of the plugging constituents. For
several days after the drain, all helium flow through the fuel system
was stopped, and the hellium in the loop was circulated at ~1100°F. The
helium in the off-gas holdup volume (170 liters) was stagnant at about
122°F. Three successive pairs of samples were then trapped from the
discharge of the holdup volume; each pair of sample traps consisted of
a pipe coil at liquid-nitrogen temperature followed by a U-tube filled
with molecular sieve material, also at liguid-nitrogen temperature. ‘The
first pair of samples was isolated from the gas that had been in the hold-
up volume by displacing it through the traps with helium from the fuel
loop. The helium which entered the holdup volume during this displace-
ment was presumed to be representative of the material circulating in the
loop. A second displacement through anobther pair of traps was used to
isolate material from that gas. For the third pair of samples, the re-
actor cell temperature was allowed Lo rise to heat the off-gas holdup
volume to 173°F before the gas was passed through the traps. The pur-
pose of this was to see 1T any additional material was released from
the holdup volume by the increase in temperature. Analysis of the samples
has not yet been completed because of the high activity associated with
thiem.

bince particulate matter originating al the fuel pump was regarded
as one possible cause of plugging in the off-gas system, the main fuel
off-zas line (522) was opened at the fuel puwp for viswal inspection.
cmall amounts of gray-~green powdery material were found between the faces
of the flanges thal were opened, but no deposits were seen in the part
of the holdup volume that could ve inspected. btamples of the solid ma-
terial and swabs of the holdup volume were removed Tor analysis.

Additional work to be performed before the next startup includes:
1. a blowout of the holdup volume Lo remove any loose material,

2. design, fabrication, and installatlion of a new fillbter-control-valve
assembly at PCV 522, and

3. replacenent of several valves 1in the system with valves having larger
trim,

Salt-Pump 01l Systems

 

Botrained gas in the circulating olil tends to accumulate in the
volute of the standby oll pump. From the beginning of MSRE operation,
it was necessary Lo prime the gtandby pump frequently o keep it ready
Tor operation. (Priming wags done routlinely once a shift.) Observations
in a development loop showed that entrainment and frobthing in the res-
ervoilr were due to the action of a Jjel pump used to scavenge oll from the
bearing housing (uging shield plug oil flow to drive the Jet) and agl-
tation of oll by the shaft and.bearings.6 In Beptember 1965, the jets
on the fuel and coolant pumps were replaced with less powerful jJets to
decrease entralnment. After this modificaticn, it was possgible to re-
duce the frequency of priming the standby pump to once a week in the
fuel o0il system; on the coolant lube oll system it was necessary Lo con-
tinuve as before.

After the Jets were changed 1t was found that changes in Tuel~pump
bearing~-oll flow or reservoir pressure caused changes in reservoir level,
apparently because of holdup of oil in the salt-pump motor cavity. OILL
from the bearing housing can enter the mobor cavity eltner by leakage
past the upper seal or through a passage comecting the motor cavity to
the area beltween the two shaft bearings. Apparently an excess of flow
to the bearing housing causesz oil to be forced up Into the mobor cavity.
Although the breasther line affords a return to the olil reservoir, it has
bveen found that once the oil collection beglins in the motor cavity 1T
msually conbinues until elther the oil zystem is shut down completely or
the bearing Tlow is reduced to below normal. Apn increase in oll {low
to the bearing housing from the normal 4 gpm to 5 gpm causes a level de-
crease in the oll reservoir of over 2 gal. (Much more than this could
collect without reaching the motor, but normally action is taken to stop
collection at less than a gallon.)

Leakage pasl the lower seal drains to the oil catch tank, which is
equipped with a level indicator and which is perlodically drained to the
waste oll recelver. After the installation and calibration of a siphon
tube in the fuel-pump oil catch tank in Novenber, oll was not added to
bring the level indicator on scale. The fuel-pump lower-seal leak rate
during the months of December and January is therefore nobt known pre-
cigely. However, when oll was deliberately added to the oil cabtceh tank
28

in February, it took 900 cc to briang the level instrument on scale. Tae
seal leakage evidently had been very low, because this volume is very near
that calculated to fill the empty pipe from the oil catceh Lank to the drain
valve at the waste oll receiver. After {the oil catch tank level indicator
was brought on scale, the indicasted leak rate was approximately 2 cc/day.
Phe coolant-salt-pump Llower-seal leakasge measured during Januvary and

most of February was zero. Near the end of February a leak rate of ap-
proximately 4 cc/day began.

The levels in tne oll reservoirs are recorded roubinely once a
shift. As was seen above, the amounts of oil held up in the motor
cavities vary with operating conditions, and these cause variations in
the olil reservoirs. DBut there has been no reason why this variable
holdup should show any long-term trend. Therefore, the reservolr levels
recorded over a period of weeks should reveal changes in oil inventories.
Figure 1.16 is a plot of the fuel-pump o0il reservoir levels from December
1965 through bFebruary 1966, when the oil system was in continuous opera-
tion. Lt is apparent that leakage was less than can be measured by this

ethod. The coclant-pump oil reservolr level showed similar behavior
except that the fluctuations were smaller (prdbably reflecting less motor-
cavity holdup).

Treated Cooling-Water System

 

The treated-water system 1s a closed systenm removing heat from the
thermal shield and other items in the reactor and drain cells. Problems

ORNL - DWG 66 -3524

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o Y —— .
| ; |
| 1.
a5 | e { liter ]
U% = 149 liter)
_ : e s B
5 @% 2 o9
; %o sp ©PHEPDE ®
U>J @@@
b 80 e ——— - s 0000 000F 1
i ® e 232 epper o396 ersne | @ s 9 |@ ‘ ® ,
o s ¥ ® @ ose & 23920035950 2L O o
e | ® ’ |S
e | ? [™~DRAIN
e | R , | MOTOR |
j | cavTy
! =
]
sol --------- Ao b ‘ , e |
10 20 30, 10 20 30 10 20 l
| DEC | JAN | FEB |

Fig. 1.l6. Fuel-Pump Oil~Reservoir Ievel in Runs 4 and 5.
29

encountered during this report periocd were neutron activation of the cor-
rosion inhibitor and interruptions of flow through the thermsal shield
slides.

Activation. During power operation in run 4, induced activity in
the treated water rose to an unexpectedly high level. Extrapolation of
observed radiation levels indicated that at 10 Mw, gamma radiation near
the heat exchanger in the diesel room would be about 400 mr/hr; gimilar
levels would exist in the water room. The activity proved to be 12.4-hr
42K produced in the corrosion inhivitor (2000 ppn of a 75% potassium ni-
trite, 25% potagsium tetraborate mixtuce). The level indicated that the
average lux in the system.(about 80% of which is included in the thermal
shield) was equivalent to sbout 7 X 10%9 thermal neutrons cm™? sec™t at
10 M.

A survey of possible cation replacements for potassium led to the
choice of Lithium, highly enriched in the 71i isotope to minimize tritium
production. An adequate amouvnt of lithium nitrite solution was prepared
cormercially by ion exchange from potassium nitrite and lithium hydroxide.
After the 4000-gal treated-water system was diluted with demineralized
water to reduce the potassiwm from 800 to 3 ppm, the desired inhivitor
concentration was attained by adding i nitrite, boric acid, and 14
hydroxide.

When the reactor was next operated at power, in run 5, objecticnable
activation again occurred, this time due to 1.0 ppm of sodium which had
evidently come in with the demineralized water from the ORNL facility.
Condensate wag produced at the MSRE with less than 0.1 ppm of scdium,
and this was used to dilubte the sodium in the treated~water system to
0.3 ppm. The concentrations of sodium and 511 are now low enough so that
shielding and zero-leakage containment of the water system will not be
necessary.

Flow Interruptions. On several cccasions flow through the thermal
shield slides stopped. Each time it was found that when the water supply
line was disconnected and water was allowed to flow backward through the
slides, considerable amounts of alr came oub with the water. The con-
clusion was that air could accumulate in the longest slide until the
nead exceeded the pressure available to cause flow. (This should be im-
posgible if the piping in the slide were installed as designed.) Vent-
ing the inlet line restored the flow, and provislons were made Lo fa-
cilitate venting. The source of the air was evidently a vorbtex in the
surge tank which occurred when the level was low. After the flow through
the surge tank was drastically reduced (it had been about 30 gpm) there
was no recurrence of the flow interruptions.

 

Secondary Contalinment

 

The secondary contalnment, that is, the envelope surrounding the
reactor and drain-tank cells, was tested and put iuto service for the
first time during this report periocd. The testing procedure followed
generally the MSRE Operating Procedures, Sect. 4E.7 This procedure in-
30

cludes testing of individual closure devices on lines penetralbing the

of thne overall leakage rate.

All lines carrying nhelium, air, or water into the cells are eguipped
elther with check valves or block valves actuated by radiation monitors.
All of these valves were tested, in place wherever possible; obherwise
they were removed and bench-tested. "Yest pressures were 20 psig or more,
and leakage rates were reguired Lo meet conservalive criteria. This work
wag done concurrently with maintenance before the cells were closed.,

Leaks which occurred in the cooling-water system were mostly in the
cneck valves. These were the resuwlt of Toreign parcicles, apparently
washed out of the system and trapped between the sliding and moving parts
of the valves. Two hard-seated valves in the treated-waler system had
to be replaced even thiough the scats were lapped in an effort to get them
to seal. One was a check valve with a swinging check, in g wabter line
between the surge tank and condensate tank, and the other was z spring-
loaded hand valve on top of the surge tank, which has to contain air.
They were replaced with soft-seated valves. The leakage rates on water-
contalning valves were determined by collecting the leakage, and a flow-
meter wag uvsed to measure air leakage through valves on tup of the surge
tank.

The majority of the valves in the helium system of bobth primary and
secondary containment had satisfactory leak rates. Those which had ex-
cegssive leakage were check valves and were found to have damaged O-rings
and/or foreign particles, usually metal chips from machining, in them.
AfTter cleaning and lnstalling new O-rings, they were satisfactory. Many
of the check valves had to be removed from the system to pressurize them.
After reinstalling, the lines were pressurized and the fittings were
checked with a helium leak detector. Leakage rates througn the valves
vwere determined by a flowmeter or by displacement of water in a calibrated
tube.

The inshrument-air-line block valves with Tew exceptions were found
to be satigfactory; however, nunerous tube fitlings were found to be
leaking by checking with leak-detector solution when the lines were
pressurized to check the valves. Several quick disconnects on alr lines
inside the reactor cell and drain-tank cecll were found to be leaking when
checked with leak-detector soluticon and were repaired. These leaks do
not constilitute a leak in secondary contalnment, since each line has a
block valve coubside the cell; bull a leak here doss affect the cell leak
rate when air pressure 1s on the line to operate the valve.

The butterfly valves in the 30-in. line used for ventilating the
cells during maintenance operations were {irst checked by pressurizing
between them. The leakage measured by a flowmeter was excesslve, and
the valves had to be removed from the system to determine the cause.
There was considerable dirt on the rubber seats, and one was cub; these
seats were cleaned and repaired. We found that the motor drive units
would slip on their mounting plates by a small amount, thus causing a
slight error in the indicated position of the valve. Dowelg were in-
stalled in the mounting plates to prevenlt this. Small leaks were also
31

found around the pins which fasten the butterfly to the operating shaft;
these leaks were repalired with epoxy resin., The line from the thermal-
shield rupture disks to the vapor-condensing systems has numerous threaded
joints that leaked badly when pressurized with nitrogen. Iach Joint was
broken, the threads were coated with epoxy resin, and the joint was re-
made. All joints were leaktlight when rechecked with leak-detector solu-
tion.

When closing the cells, all membrane welds were dye checked. After
conpletion of the membrane welding, alternate top blocks were installed,
and the cells were pressurized to 1 psig to leak check all membrane welds
with leak-detector solution. DNo leaks were found in the welds. The re-
actor access cover plate, which has a double O-ring seal, was found to
e tight by pressurizing between the O-rings.

With the cells at 1 psig, all penetrations, pipe Joints, tube fit-
tings, and mineral-insulated (ML) electrical cable seals subjected to
this pressure were checked with leak-detector solution. Numerous leaks
were found ia tube £ittings and ML cable seals, and the leak rate was
about 4500 ft3/day, indicating a major leak. Many of the small leaks
were shopped by simply tightening the threaded parts of the seal. How-
ever, the leak rate was still aboub 4500 £t3/day.

The cells were then pressurized to 5 pslg, and leak hunting con-
tinued. Three large leaks were located: one in the sleeve of line 522,
the off-gas line from the pump bowl to the carbon beds, under the benb
house floor, another in an instrument air line to valve HCV 523, and
anobher from the vapor-condensing system to the steam domes and oubt to
the north equipment service arca through a line that was temporarily
open. All penetrations, tube fittings, and ML cable seals were again
checked with leak-detector solution. Many ML cable seals which had not
leaked at 1 psig were found to be leaking, and some of bthose which had
been tightened and sealed at 1 psig now leaked.

The hock gege being used to monitor the leakage rate along with
conventional pressure gages did not work properly, and the trouble was
found to be the resull of extremely swall leaks in the tube fittings on
the reference volume side of the system. This and temperature changes
in the coolant drain-tank cell and special equipment room caused asppre-
ciable difficulty in determining the leakage rate. The piping to the
hock gage was evenbually replaced with virtually all welded tubing.
Temporary closures were put on the special equipment room, and the door
entering the coolant drain-tank cell was kept closed as much a8 possible.
These measures considerably improved the reliabllity of the data.

The MI cable seals as a group accounted for a large percentage of
the remaining leaks. Many were sealed by tightening, but many of them
required tightening more than once. As a result, several of the gland
nubs split and soldering was required. All large leaks were sealed or
greatly reduced, and many of the small leaks were stopped. To stop
these and other small leaks which may not have been located, all ML cable
seals were coated with epoxy resin at the seals oubtside of the cells.
Teflon tape, used extensively on ML cable seals, other threaded pipe,
and tube fittings, did not perform gatisfactorily in providing a gas
seal.,
32

After stopping or reducing the leaks in the ML cable seals and in-
strument air lines, testing was begun at 10, 20, and 30 psig. At these
hignher pressures, large leaks were located in bobh component coolant pump
dome flanges by leak-detector solution, and onc of them was audible. To
correct these leaks, the width of the gaskel was reduced to increase the
loading pressure on the gasket. All penetrations, ML cable seals, tube
fittings, and external parts of valves which are part of the secondary
containment were checked wilh leak-detector solution at each of the above
pressures.

In each of these tests, the containment system was pressurized to
a specific pressure. Then all gas additions and controlled exhausts
were stopped, and the leak rate was measured by the change in pressure
corrected for changes in temperature. The calculated leakage rates are
presented in Fig. 1.17.

ORNL-DWG 66-4753

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BOO [ e e —
400 S
300 - — — r— ]
~ /
D
o
2
ud
|_
<I
& ool | . ]
il
9
<1
7
Pt !
L i e
- &
'_
= 100 -
w
>
z
<l
5 ®
<
O
O e (f) ASSUMED LINEAR RELATIONSHIP
o @ ORIFICE FLOW
(@ TURBULENT FLOW
-100 /// """"""""""""""" T (@ LAMINAR FLOW T
e EXPERIMENTAL DATA
200
-10 c 10 20 30 40

CONTAINMENT PRESSURE (psig)

Fig. 1.17. Pressure Dependence of the leakage Rate of the MSRE
cecondary-Containment Vessel for Possible Flow Regimes. Normalized +o
the maxipum allowable leakage rate under MCA conditions.
The leakage rabte ls acceptable if it is l% peyr day of the contained
gases under the conditlons of the maximum credible accident., This rep-
resents a leakage rate of aboub 430 scfd at 39 psig and 250°F. With this
as bhe condition of maximum leskage rate, the curves in Fig. 1.17 relate
the leakage rate to system pressure for various flow reglmes that can
exist (see ref. 1). Also shown is a highly conservative linesr relation-
ship for leakage rate vs pressure. The data taken during the leak-rate
tests at 10, 20, and 30 psig are all well within the bounds of these
curves; the in-leakage rate at —2 pslg 1s acceptable for all but the
linear relationship. Although the data at —2 psig indicate the conbain-
ment is inadequate for the lineary extrapolation, a comprehensive exami-
nation of the data clearly indicated that the leakage rate was acceptably
Low.

Core Access Opening. Convenlent access to the control-rod dirives
and the core accegs flange is had by removing a 40-in.-diam plug in a
stesl-lined hole in a lower shield block. Originally a flange on the
upper rim of the hole liner was welded to fthe cell merbrane, and a bolted,
gasketed cover plate completed the containment. This attachment of the
menbrane to a lower block presented a problem which became apparent dur-
ing testing and which led Lo modification. DBecause of thelr great weight,
the upper blocks must be supported at Their ends, not by the lower blocks.
To ensure this, there is necessarily some clearance vetween the upper
blocks and the membrane, which normally lies on the lower blocks. When
the cell is pressurized, therefore, the membrane 1ifts up until 1t con-
tacts the upper blocks. This it was free to do everywhere except around
the access opening, where it was fasbened to the flange. Analysls showed
that this situvation would produce excessive sbresses when the cell was
pregsurized for leak testing. Redesign of the liner or [lange was im-
practical, so the membrane was cut loose from the flange and an overlap-
ping patch was welded on the membrane. Thig pateh must now be cut off
whenever access is required.

 

Vapor-Condensing System. Installation of imstrumentation and shield-
ing on the vapor~condensing system was completed, and the gystem wag leak-
tested before water wasg added. OChilelding consists of earth having a min-
imum thickness of 4 ft on top of the gas storage tank. This tank is just
inside the area fence along the HFIR access road, and at least 8 €t of
earth wasg provided on the road side to gilve complete probection to passers-
by under the worst conditions. Before the tanks were covered, all pene-
trationsg were helium leak-tested and found to be tight. Then during the
reactor cell leak test, the vapor~condensing system was leak-tested at
30 pslg, and a leakage rate of only 5.0 42 /day was measured from the
6800~ft§ sysbem. After the leak test, the vapor-condensing tank was
filled with water to the proper level, placing the system in readiness
for operation.

 

Ventilation of Coolant Cell. The cell housing the cocolant-~salt
system 1s connected to the bullding ventilation system by a duct which
exhausts 5500 e¢fm of air. This wasg enocugh to maintain a negative pres-
sure in the area until the radiator blowers were put inbo operation. Then
alr leakage from the radiator enclosure exceeded the cell exhaust capacity.
Even after congiderable work on the enclosure, the leakage wag too much,
34

To alleviate the situatlion, the inlet of the south annulus blcower was con-
nected to tine coolant drain-tank cell by ductwork. This blower can now re-
turn a2bout 7400 efm of air to the duct leading to Lhe coolant stack. Now
when the annulus blower is used to supplement the regular exhaust, the
main radiator blowers can e coperated without raising the coolant-cell
pressure above atmospheric.

shielding

Complete radiation surveys of the reactor areca were made at each
higher power level from 1 to 1000 kw. These showed a Tew areas of un-
acceptably high radiation. The treated-water system has already heen
discussed., ocattering of fast neubrons and gamma rays from the reactor
cell into the coolant dyain cell through the 30-in. cell ventilation
line caused excessive dose rates outside the access door. A wall of
16 in. of concrete block and © in. of borated polyethylene was built in
the coolant drain cell to reduce dose rates ab the door to acceptable
values. (The door is locked during nuclear operation to prevent entry
into the cell.) Gamma radiation from Lhe fuel off~gas line in a 3-f1
gap vetween the resctor building and the vent house reguired earth shield-
ing, as had been expected. We also expected that additional shiclding
would be required directly above the reactor, where there are large cracks
between the shield blocks. However,; present indications are that the
shielding there will be adequate.

Component Performance

 

Considerable effort was spent in modifying and adjusting the ra-
diator to meet operational requirements. Other components performed
well, but there were some failures in conventlona]l equipment.

E@@iator

The original radiator doors proved inadequate in that they bowed
and jammed when the radiator was heated. In August 1965, doors of im-
proved degiga were installed. "Yests showed that these doors deformed
far less when hot and suffTered none of the permanent warping of the
first doors. Heat losses due to air leakage around the doors continued
to present a problem, however. Before all parts of the radiator tubes
could be gotten into an acceplable temperature range for £fillipng with
salt, 1t was necessary to work out several minor changes. These in-
cluded shimming the gasket to conform to the complex bowing of the doors
and enclosure and changing the heater wiring to give a nonuniform heat
input over the face of the wradiator.

At Tirst the new doors did aot move freely. On scparate occasions,
both doors hung up and failed to lower, leading to snarling and kinking
of the lifting cables. Reliable, free movement of the doors was attained
only after several vevisions. The seal strips were modified to eliminate
325

digging into the asbestos and metal gasket and were attached more securely
to the doors after the striops warped and broke in several places. The
upper limit switeh arrvangement was changed to prevent hangups due Lo over-
travel. The door gulde rollers were changed to eliwminate binding. More
clearance bebween the moving doors and the enclosure wag provided by mov-
ing the guide tracks and modifying the closing cams, which force the doors
in against the enclosure at the bobttom of their travel.

When the main blowers were operated to best the doors, other dii-
Ticulties hecame apparent. Alr leakage from the sheel metal enclosure
was clearly excesglve (about 10,000 cfm), even though the original de-
algn bad been modified to include hoods iInto which the doors were raised.
The leakage was reduced by installing asbestos cloth bocots around the
cables, applylng sheet metal Lo some gaps, and packing insulation in others.
Even so, the leakage required an increase in the cell exhaust capacity
to prevent pressurization of the cell. ILeakage of het alr inbo the re-
glon Just above the enclosure caused overheabting of electyical insulation
on the numerous thermocouples and power leads in this vicinity. The prob-
lem was compounded because the new door hoods blocked the circulation of
alr from the cell coolers. lLeak shbopping and more thermal insulation
improved the gituation, bul it was still necessary to relocate junction
vazes in cooler locations along the cell wall and insgtall higher-tempera-
ture electrical inswlation on the thermcoccouple and the electrical leads
over the enclosure. Flexible ducts and vanes were uged Lo redirect cool
air flow over the leads and the lifting mechanisms. See "Component De-
velopment” for additional information about the modifications.

After the ilaprovements in the door seals, air leakage through the
radiator was manageavle, bubt it was still considerable. Oback draft
alone pulled enough alir arcund the closed doors Lo make the heat losses
(and reguired neater settings) sensibive to the wind outside and the
position of the radiator bypass damper. It was found that the door~-seal
leakage also depended on whether the doors had been gerammed or were
lowered normally. The first operation at 1 Mw (in run 4) was with oue
main blower on, the bypass open, the oubtlel door ralsed 15 in., and the
inlet door all the way down. lLater, in run 5, more than 1 Mw was ex-
tracted with both doors down {(but not scrammed), one main blower on,
and the bypass open.

During the 1-My operation in run 4, salt froze in the lines con-
nected to FT 201, the venturi flowmeter located in the upper part of
the radistor enclosure. With the blower on, cool air leakage past the
element made it necessary te turn on local heaters to prevent freezing.
When the blower was stopped, the heaters had to be turned down to aveld
overheating. It sppeared likely that in higher-power operation, with
greater air pressure in the radiator, the element could not be kept from
freezing. Therefore, between runs 4 and 5, the top of the enclosure was
cut open, part of the stacked-block insulation wasg removed, and fibrous
insulation, protected by shim stock, was installed around the alement.
Cperation in run 5 indicabed that the insulation was then adeguate.
ther Components

 

Heaters. Very little trouble was encountered with the heaters on

 

A faulty thermocouple led to failure of two elements in H 106-1, a
heater on the fuel drain line to drain tank 2, in October 1965. Investi-
gatlon of an anomalous drop 1n current showed that two elements had failed
after =2 lengthy period in which the hester was operating near the maximum
power (about 50% higher than other heaters in similar locations). The
thermoecouple whose reading was used o set the heater was then checked.
One lead was grounded in the shicld, causing the indicated temperature
to be 400°F low. All three elements in H 106-1 were replaced. X rays
showed that about 1 In. of resistance wirec in bthe two failed elements had
melted. Inspection of the pipe showed no damage from the abnormally high
temperature .

All heaters performed properly during the heatup of the fuel and
coclant loops, with one minor excepbion. A broken lead turned up on a
heater between the radiator and the coolant pump. It was simply recon-
nected.

After the startup, one of threc clements in one of the heat-exchanger
heaters failed. No action was {taken because the other elements were encugh
to produce the desired temperature.

Drain-Cell Space~Cooler Motor. The fans on the reactor- and drain-
cell space coolers were originally driven by 3-hp motors; 10-hp motors
were subsequently lnstalled to provide the additional power needed to
keep the fTans running in the event of an accident which pressurized the
cells.

 

The motor on the draiun-cell cooler failed at the end of run 5, after
about 2900 hr of operation. The cooler was removed, and examination
showed that the rotor had slipped along the shalt until the rotor fan
blades rubbed, causing the stator windings to overheat and to be destroyed.
Normally a force of 2 or 3 tons is required to force the shaft out of a
rotor of this size.

This very unusual failure does not appear to reflect a weakness in
design, but most likely improper manufacture. Therefore, the ruined
moLor was replaced with one practically identical.

Component Cooling Pumps. The two component cooling pumps are large,
positive~displacement blowers whicn supply cooling air to freeze valvesg
and other eguipment in the reactor and drain-tank cells. Whenever +the
fuel system is hot, one pump is kept in operation, with the other on
standvy. By the end of February 1966, CCP 1 had operated 4995 hr; CCP 2
had operated only 1620 hr.

 

The drive belts on CCP 2 broke after 1454 hr of operation and again
after 11 additional hours. The belts had been operated at loadings in
excess of thelr rating since the beginning of operation when the sheaves
were changed to reduce blower speed. After this was recognized, follow-
ing the second fallure, the sheaves and belis were replaced with a poly-
37

V-belt drive adequate for the 75-hp mobor capacity. The criginal belts
on CCP 1 lasted for 4360 hr before failure. They too were replaced with
a poly-V-belt drive.

At the time of the second failure of the CCP 2 belts, it was dis-
covered that the discharge check valve on this blower was inoperative.
This valve is a 6-in. bubterfly-type valve with an elsstomer hinge hold-
ing two flappers in place. The hinge had broken, allowing one flapper
to fall off. The result was that CCP 2 "motored" because of reverse
flow when CCP 1 was operated. The check valve was rebulllb and rein-
stalled.,

Diegel Generators. UThree Diesel-powered generators, each witn a
capacity of 300 kw, supply emergency ac power to motors and heaters in
the MERE. There has been no ocecasion on which this emergency power was
needed. The diesels are started once a week, and once a month they are
operated under partial lcad. They have also been operated during plaanned
power oubages for uvp to 6 hr.

 

In November 1965, coclant leaked into the crankcase of DG 3, and
it was found that an imperfect head made a tight seal velween head and
vlock impossible. The head was replaced and the bolts torgqued to the
manufacturer's specifications. In February a crack was found in the
bhlock at one of the bollt holes. A stop-leak conmpound was used bempo-
rarily until repairs could be plamned. At the manufacturer's recom-
mendation, repair of the crack was attempted by metallic arc welding.
The attempt was only partly successful, and stop-leak compound was
needed to halt slight seepage from a crack adjacent to the weld. Further
attempts at repair are deemed lnadvisable, and the expense of a new block
does not appear justifiable in view of the purpcse of DG 3 — 1T it were
to faill when called on for emergency power, it would at worst cause delay
and inconvenience but no system damage. FPresently DG 3 is tested routinely
with the cother dlesels, but the crankcase oll is monitored egpeclally
carefully for traces of coolant.

Aiy Compressors. Compressed alr at the MSEE is supplied by bthree
reciprocating air compressors. Two of the compressors, AC 1 and AC 2,
provide instrument air. The other, AC 3, supplies alr for service needs
and, in emergencies, can supply alr to the instrument systen.

 

The service alr compressor is an S~in.-bore, 7-in.-stroke, 5l4-rpm
machine with Teflon rings, installed new in 1962. In wmore than three
years of service, it falled once, in 1964, when blockage of the after-
cooler discharge line by a falled check valve led to excesgive pressure
and a head gasket fallure.

The instrument air compressors are also 8-in.-bore, 7-in.-stroke
machines, bubt they were originally equipped with carbon rings and op-
erated at 600 rpm. DBoth had been used elsewhere, and ovoth had been
overhavled and equipped with Teflon rings before belng placed in service
at the MSRE in 1964. Through Januvary 1966, six head gasket failures had
pecurred in these two compressors, three on each compressor.

Tnvestigabion of the excessive failure rate on AC 1 and AC 2 led
to the conclusion that the failures were caused by overheating of the
38

cylinder liner, which occurred whenever the cooling-water temperature

or the lcad was slightly above normal (but still within the cpecified
range). This was attributed to binding of the Teflon piston rings, which
was observed at times after brief operation at rated load. When the rings
began to vind, the cylinder liner apparently overheated, expanding axially
and deforming the head gasket. Water then leaked from the cooling pas-
sages into the cylinder when the compressor cooled and the cylinder liner
shrank. The compressor manufacturer now reccommends a maximum speed of
514 rpm on 7-in.-stroke compressors; apparently piston speeds at 600 rpm
were too nigh after Teflon rings were substituted for carbon. LIt was pos~
sible to reduce the speed on AC 1 and AC 2 to 512 vpm and still meet the
requirenents for Instrument alr flow. This change was made, therefore,

by changing tne drive sheaves and belts, and apparently the binding and
overhealing were eliminated.

Inspection of the Iuel Pump

 

The fuel pump rotary element was removed for inspection after the
flusn salt circulation at the end of run 3. The pump was checked for
dimensional changes, deposits, corrosion or erosion of hydraulic parts,
and for rubbing of close-clearance rumning fits.

The pump was generally in good condition, and nothing was [ound that
would interfere with the satisfactory operation of the pump. There was
nc indication of rubbing of the running fits or of corrosion or erosion
of the hydraulic parts. The only dimensional change was a 0.006-in.
growth of the pump lank bore diameter wnere the upper O-ring mates with
the pump tank.

Figure 1.18 1s a photograph of the lower part of the rotary clement
in the decontamination cell where 1t was examined. The dark stains on
the shield plug arc the results of an o0ll leak through the gasketed joint
alt the cateh basin for the lower oil seal. This oil had ruan down the
surface of the shield plug, where it had become coked by the higher tem-
peratures near the bottom. Some of the oll had reached the upper O-ring
groove al the bottom of the shield plug and had become coked in the groove,
bul none appearcd Lo have leaked past the O-ring during high-temperature
operation. Scme fresh oil was observed below the ring alfter the rotary
element had been moved to the decontamination cell, but we believe this
oil dripped from the open oil lines during the transfer to the cell.

The photograph also shows part of a deposit of flush salt that
failed to drain from the labyrinth flange just above the impeller. Cal-
culations of [ission product activity after high-power operation indi-
cated that this incomplete drainage would not increase the mainbenance
hazard significantly, provided Tlush salt is circulated to flush fuel
from this locatlon pricr to the removal.. There were also deposits of
salt that contained fuel in both The uvper and lower O-ring grooves.
These deposits would not ve effectively diluted during flush salt opera-
tion and would constitute the major radiation source during malintenance
operations.
39

 

PHOTO 81122

Fig. 1.18. MSRE Fuel-Circulating-Pump Rotary Element After Run 3.

The parts of the rotary element that had been in contact with the
salt had clean, bare metal surfaces, while the parts exposed to the gas
had a thin black film. There was an abrupt transition, which may be
seen in Fig. 1.18, between the two types of surface, but this transition
was slightly above the actual level. The salt deposit above the laby-
rinth flange and a smaller salt deposit in the pump volute were also
partially covered with a thin black film.®

The pump was reinstalled using remote maintenance techniques so
that these techniques and procedures could be evaluated. TFour universal
Joints on the flange bolts that had been found broken during the dis-
assembly were repaired prior to the reinstallation of the rotary element.
The fallures resulted from excessive bolting torque that had been used
earlier to obtain an initial seal on the flange.

lHeat Treatment of Reactor Vessel

 

After the reactor vessel was installed, tests of Hastelloy N from
the heats used in the vessel showed that the closure weld between the top
head and the flow distributor ring would be expected to have poor me-
chanical properties 1in the as-welded condition. Further tests showed
that the rupture life and ductility, which were unacceptably low, could
40

be practically restored to those of the base metal by stress relieving
for 50 to 100 hr at 1400°F.° Although this is well above the normal op-
erating temperature of the MSRE, it was attainable with the installed
heaters, and calculatiocns showed no harmful thermal stresses would be
involved. Therefore the vessel closure weld was heat treated in place.

Temperatures were monitored by six thermocouples evenly spaced
around the vessel Just below the weld. Because of gaps in the shroud of
vertical heater tubes around the vessel, a perfectly uniform temperature
could not be attained. A day was spent at about 1300°F, adjusting heaters
to minimize the temperature spread. Then for 90 hr the lowest tempers-
ture was held at 1400 * 20°F, while the hottest thermocouple was at
1460 = 20°F. After the vessel was cooled at the end of the treatment,
inspection showed that the furnace and insulation were undamaged.

otress Analysis of Reactor Piping and Nozzles

 

The MSRE fuel system was designed for the fuel pump and the heat
exchanger to move horizontally and vertically. This was done to keep
the stresses low in the piping and eguipment nozzles while accommodating
the expansion and contraction of the closely coupled system as it is
cycled between 150 and 1300°F. During the prenuclear and critical test-
ing we found that the pump mount could not be depended on to move ver-
tically. At the same time we learned that the creep properties of the
metal in the reactor vessel and the piping inside the thermal shield
would deteriorate with neutron irradiation. The former condition could
cause the stresses in the piping and nozzles at the reactor vessel to be
high with the system hot or cold. The latter requires that the stresses
be kept low when the reactor is at high temperature, although normal de-
sign stresses are permissible below about 800°F.

According to our calculations, acceptable stresses at the reactor
vessel nozzles could be obtained by raising the pump 1/2 in. and the
heat exchanger 1/4 in. from their existing cold positions and fixing
them against further vertical movement. This "cold springing" would
result in the stresses being high at low temperature, and expansion of
the piping on heating would lower the stress as the temperature is raised.
Because of the relative locations of the supports, fixing the pump and
both ends of the heat exchanger was found by calculation to result in
excessive stresses in the heat-exchanger nozzle during the thermal cy-
cling. However, acceptable stresses could be obtained there by mount-
ing the heat exchanger on spring supports at the end closest to the pump.

The changes were made as indicated above, and the fuel system was
heated to 1200°F. There was no appreciable vertical movement of the
heat exchanger on the spring supports where a downward movement of 0.1
in. had been expected. This indicated that some of the assumptions used
in the calculations were not accurate and the calculated forces at the
heat exchanger were too high, or that the junction of the nozzle to the
shell of the heat exchanger yielded under small forces.

After the system was ccooled, strain gages were installed on the heat-
exchanger nozzle and on the piping to the pump. The heat exchanger was
47

raised and lowered about 1/8 in. while the forces on the spring supports
and the strains were measured. The measurements indicated that the maxi-
mum stress was in the heab-exchanger nozzle, bub that high stresses in
the nozzle were accompanlied by large forces on the springs. We concluded
that the spring mounting should act Lo relieve any large vertical forces
on the nozzle and that the revised installation was satisfactory.

ITnstrumentation and Controls

 

General

Formal design of the MSRE instrumentation and controls system is
now complete. The need for further additions and modificaticns to the
instrument and controls system has become apparent as initial overations
progress. bome provide additional system protection, but in the majority
of cases the purpose is to improve performance and provide more informa-
tion for the operator. A few minor design errors were corrected, and
some instrumentation was added to siuplify maintenance procedures. Ex-
cept for some specification sheet revisions and preparation of a design
report, documentation is complete.

Safety Instrumentation

 

No major changes to the Tlux and tewmperature safety system were re-
guired. Minor modifications were made ag follows:

1. The model Q-2623 relay safety element was altered by replacing
the 115-v ac relays with 32~v dc relays. This eliminated ac pickup on
the other modules through which relay coll current is roubted and reduced
the likelihood of spurious trips, particularly when the Q-2602 fiux am-
plifier trip points are reduced by a fTactor of 1000 from 15 Mv to 15 kw
of reactor power.

2. The voltage adjusting resistor in the rod drive clutch circuit
was increased from 500 to 750 ohms to facilitate setting the cluteh cur-
rent to the necessary and sufficlient minimum value.

3. Current meters were placed in the rod drive motor circuits.
Gross changes in motor Josd, caused, for example, by a sticking rod,
may be inferred from an increase in meter reading.

4, Dynamic braking circuitry for rod drive 1 (servo-controlled
rod) has been designed, built, and tested, with installation scheduled
for March 1966. Braking action is obtained by discharging direct cur-
rent from a capacitor through the motor field windings when the drive
motor is turned off. This break will provide a substantial reduction
in coasting of the rod drive after the power is switched off. As a re-
sult it will take less back-and-Torth Jogging to get small incrementbs of
rod motion when manually shimming. Serve performance is expected to im-
prove for the same reason.
D~
NS

To asgure a drain, additicns and revisions to existing safety clr-
cults were designed to open the fuel drain~tank vent valves on signal
from high neutron flux seram) and keep them open until manual resel 1s
used to permit reclosing.

Wide-Range Counting Channels

 

Reactor power and period information Irom these two flission counter
channelstt provides interlocks which govern permissible reactor power
levels (control system modes)12 and reactor f£ill, which inhibit rod
withdrawal during start, and which provide control-rod reverse actlon.

Very generally and typilcally, reactor period signals Ifrom counting
channels operating at low input levels are characterized by slow response.
This inherent delay produced a problem witi servo-controlled rod with-
drawal during start. In a servo-controlled start, the demand signal
causes the regulating rod to withdraw until the period-controlled "with-
draw inhibit" interlock operates. Tf the period continues to decreasc,
the "reverse' interlock acts to insert the rods in direct opposition Lo
the servo demand. These "withdraw inhibit" and "reverse' trip points
were originally established at periods of +20 and +10 sec respectively.
The delayed low-level response of the "inhibit" interlock allowed suf-
ficient inecremental rod withdrawal Lo produce a 10-sec period and thus
cause a reverse. The situation was aggravated by coasting of the ghim-
locating motor in the regulating rod limit switch assembly.’?

As remedial measures, bhe "withdraw inhibit"” and "reverse' period

trip points were changed to +5 and +25 sec, respectively, and an electro-
mechanical clutch-brake was inserted in the shim-locating molor-drive
train. The dynamic brake described above is also expected to reduce this
period overshoot problem.

Provision for testing the period and log power interlock Trips in
the wide-range counting channels was provided. Testing is done routinely
by Instrumentation and Controls Division maintenance personnel and is ac-
complished by push buttons iunside the idnstrument chassis.

Nuclear Instrument Penetration

 

We agssumed that the neubron flux in the instrument penctrations would
tend to decrease exponentially with increasing distance from the core ves-
sel.t% This implies that most of the Tlux within the penebration originates
with neutrons entering the lower end face of the penetration. Figure 1.19
shows that concrete shielding was added in an effort to snield thne pensc-
tration from othner sources of neutron Iflux. The vernistat in the wide-
range counting chamnel's model Q-2616 function generatort® will compensate
for any reasonable departure from an exact exponential attenuation of
flux; however, it became apparent during the first series of critical ex-
periments that flux attennation within the instrument penctration was
deviating, grossly, from the assumed exponential curve. This unseemly
behavior is illustrated by curve A on Fig. 1.20, which shows fission counter
response (normalized count rate) vs withdrawal from the lower end of the
43

penetration in guide tube 6. It was reasonable to conclude from this,
and from similar curves, that the excess neutrons responsible for the dis-

torted part of the curve were entering the venetration along its lengtn.

The count rabes in other guide tubes nearer the upper half of the guide
tube were even more distorted than curve A,

A flux field with attenvation per curve A precludes successiul op-

eration of the wide-range counting instrumentation. We decided to shield
-he fission counbers from stray, side-entry neutrons with shields of sheet
cadmitm which were inserted into guide tubes 6 and .

As a counter travels
up a guide tube during withdrawal, it first enters a partially shielded
section of the guide tube.

The partial shielding is obtalined by using
wedge-shaped strips of cadmium laid around the guide

tube, which provide

ORNL- DWG G4 -828R
r e _eant 2

¢ FLANGE FACE

 
 
  
  
  
      
   
  
  
  
  
 
 
 
  
  

NEUTRCN INSTRUMENT ‘ A

JUNCTION BOX o -— s
‘ i4iGH BAY AREA - e /// 410 50’
- - - Y
[ - WATER-SAND ANNULUS -ANGE ‘ P

o s
l / :——ligflz/“ |
b T R T TR TR L T FINISHED FLOCR - ; |

FACE OF INST. |
|17 SHELL FLANGE i
; EL. 83 ft8in

 

FL.8521 0, \.,\

f‘ o REMOVABLE CONCRETE SHIZLDING

 

A A s “WATER
L A / I LEVEL
—— R e ; 849 ft ain.

 

 

 

L. INSTRUME
L L VEL INSTRUMENT
| T CHAMBER GUIDE
TUBE (TYPICAL)
REACTOR Ao o ¢ END OF 43-n 0D SLEEVE
/ THERMAL S
{ SHIELD , -'
|

CONCRETE e,

EL. a4 ft 12 in.
SHIELDING - CELEV, -

3

     
 
  
 
  

 

\*48“ DD DUTER SLEEVE
ASSEMBLY (WITH EXBANSION JOINT)

 

-

EXR JOINT

 

 

 

SECONDARY
| & CONTAINMENT
PENETRATION EL. 834 532 in
2IDPLANE
OF CORE

83011 8in

. EL83CH 3in

 

 

 

 

 

 

 

 

/ T OUTER STEEL
o SHELL

T~ REACTOR CONTAINMENT
CELL

 

MSRE Nuclear Iastrument Penetration, Elevation.
Z‘J'_ Z:—

ORNL-DWG §6-2741

    
 

   
   

 

   

 

 

 

107
. MSRE GUICE TUBE NO 9 e
N COUNT RATES IN GUIDE TUBE ‘ &
P NO.& BEFORZ AND &FTER CONCRETE FILL » N
ot ADDING CADMIUM SHIELDING >
A _ SECTIGN THROUGH
S REACTOR SHIZLD ™. NUCLEAR INSTRUMENT
| . 7. PENETRATION
. A
\ C by \,\
10—2 \ Iy ’\1 // >
- ; !; A
w ; \ ‘
= ' . i
= | \ i . GUIDE TUBE
= ; .\l " NO. 6
2 5
= Y oo
O i T\ - e m e
o ‘ - Y i REACTOR
i | \‘ . A CONTAINMENT
Tt . | \ Y MIDPLANE |+ Bia /i CELL
3 i c OF CORE 4 L
& ! ‘ ) \\ i 158 |’
= ; ; \‘ ‘ . e
oS CURVE "B" COUNT \ : } :‘I‘ o
RATE WITH CADYIUM ‘. ‘ }eH !
SHIELD \ : L
, . { ¢
e | .) REACTOR o
101 | CURVE "A"COUNT RATE - \ VESSEL ; {10}
) L OWITHOUT CADMIUM ot et e
SHIELD | !
i | : !) 2
1 P | Lo CADMIUM SH:ELDING ~ {5} (4)
ol . o SUBASSEMBLY — TYPICAL N e s
o 20 40 60 80 IN GUIDE TLUBES ~THIS LENGTH 100% .
x, DISTANCE WITHORAWN (in.) 8 AND 9 CADMIUM WRAP ~— T
. 10f1-Qin 1011-0in SECTIOM A-A
o Y LOCATION OF
T L GUIDE TUBES
Eh 1B R A 1O

 

“GUTER SHELL REMOVED
FOR CLARITY--SHADED WEDGES
ARE CADMIUM —UNSHADED WEDGES
ARE ALUMINUM

Fig. 1.20. Guide Tube Shield in the MSRE Instrument Penetration.

ever-increasing shielding from side-cntry neutrons as the counber is with-
drawn. ‘The bases of these long narrow wedges occupy the full periphery

of the shield tube at its upper end and thereby provide a smooth tran-
sition to the 100% wraparound cadmium sleeve which Tollows. This shield-
ing, Fig. 1.20, was an eminently successful answer to the problem; curve
B, normalized count rate vs distance, wag oblained after this shield was
installed in guide tube 6.

Bl Confidence Instrumentation

 

Because of very unfavorable geometry the strongest practicable neu-
tron source would not produce 2 COuntb/SPC Trom the fission counters in
the wide-range counting channels until the core vessel was approximately
nalf full of fuel salt; neither would it produce 2 counts/sec with flush
salt in the core at any level. This i1s the minimum count rate required
to obtain the permissive "confidence" interlock which allows filling the
core vessel and withdrawing the rods. Therefore a counting channel using
a sensitive BF3 counter was added to establish "confidence" when the core
vessel 1s less than half filled with fuel salt. With the revised system,
a Tuel-sall f1ll may begin when the reactor vessel 1s empty if BFs count
45

ORNL-DOWG 862446

ELECTRONIC RESEARCH
ASSOCIATES
2.5/10UC

   
      
  

CHAMBER | pECRDE
POWER SCALER
SUPPLY Q-1743-C1

 

 

 

BF; COUNTER
REUTERS-STOKES
RSN-28A

 

 

PULSE |
AMPLIFIER 3t ] RCI3-10-59

RC13-10-57 LOG COUNT |_ = * = L
RATE METER
e e e RCA3440-58 7

 

 

   

 

[ TRIP COMP] TRIP COMP
LCR>10cps LCR=10% cps L

 

     
 

 

 

CAL USE|+ conso e e
o - PRI e e
RCI3-10-56 |

 

 

 

 

 

IO BF
CONFIDENCE CIRCUIT

 

Fig. 1.21. Block Diagram of BF3 Counting Channel.

rate "confidence" is established but cannot be continued after the re-
actor vessel is half full unless "confidence’ is established with either of
the two wide-range counting channels., When f1lling the reactor vessel with
flush salt, rods may be withdrawn and the £ill allowed to proceed at any
level if BF; count rate "confidence” is established and if the drain-

tank selector switch is in the fuel flush tank (FFT) position. Selection
of the ¥FY position bypasses the half-Tull weight interlock and requires
administrative approval. Figure 1.21 is a block dlagram of this BFa count-
ing channel.

Personnel Monitoring oSystem

 

The reactor building radiation and contamination warning system
was revised to correct some deficlencies and to Improve its effective-
ness as indicated by tests and experlence during reachtor power operation.
An alarm relay was installed in the personnel and stack monibtor alarm
system circultry to provide an ammunciation upon loss of the 24-v de
power supply. Monitron RE-701Z2, located in the south end of the high
bay, was moved east about 20 £t to monitor Tor possible radiation escaping
from the nuclear instrument penetration. CAM RE-T700L was moved closer
to the sampler-enricher to monitor its operation. Several of the ¢-2091
beta-gamma monitors were relocated to provide move protection, One unit
was installed in the instrument shop, one wiit was removed from the health
physics office and installed in the hall between Buildings 7503 and 7509,
and. the unit in the vent house was relocated to reduce its background.
response from lines in The vent house. Two sdditional alr horns were
added to increase the area covered by the horn evacuation signal. One
norn was installed on the southeast corner of Building 7509 and the other
in the diesel houge. Four additicnal beacon alarm lights were installed
in areas where the horn might not be heard. A light was installed in the
vent house, coolling~water equipment room, bthe switch house, and the Plant
46

and Eguipment shop building. A Q-2277 rate meter and Q-2101 alpha probe
will be installed in the hot change house.

Control Instrumentation

 

Containment. To maintain the integrity of the reactor secondary
containment, solenoid block valves with safety-grade wiring were In-
stalled in the fuel-dralin-tank steam dome drain plping. These valves
are interlocked to close when reactor cell alr activity or pressure 1is
above limits. 'The interlocks override a manuval switch used for normal
operation of the steam dome drain system. Actuating signals were ob-
tained from existing circuits.

Safety-grade control circuits were installed to operate four weld-
sealed solenoid valves serving the fuel off-gas sample system. The
valves arec speeially designed and constructed Tor use on MSRE contain-
ment systems and are identical to those previously purchased for use in
the fuel-pump-bowl bubbler level system. The procurement of four addi-
tlonal valves from a commercial source is nearing couwpleblion.

Fuel Sampler-Enricher. The fuel sampler-enricher safety-grade con-
trol circuits were revised to increasce the relisbilitiy of the two con-
talmnent barriers between the primary system~(fue1 pump bowl) and the
operator. Originally, the circults were designed so that the sample
access port could be opened if either the operational valve or the
maintenance valve was closed. With Uhis arrangement a condition existed
wherein a single failure could permit the access port to be opened when
both The operatiocnal valve and the maintenance valve were open. 1f this
had cccurred when the manipulator cover was removed, the manipulstor boot
could have become the only containment barrier belween the operator and
the fuel pump bowl. ©Since the boot could he ruptured by system pressures
in excess of 10 psig, this condition was considered to be hazardous.

The circuits are now designed so that both the maintenance valve and

the operational valve must be closed before the access port is opened.
Also, the removal valve and access port must be closed and the manipulator
cover must be in place before either the coperational or the maintenance
valve can be opened. The positlon of the cover is detected by a newly
installed vacuum pressure switch. Additional protection for the manipu-
lator oot has also been provided by a new circuit that prevents develop-
ment of excessive differential pressure across the hoob during sampler
evacuation operation. This protection is accomplished by closing a valve
in an exhaust lioe to the vacuum pump when the differential pressure across
the boot exceeds 30 in. (water colimm).

 

Fuel and Coolant Pumps. To satisfy established operating criteris,
the fuel and coolant salt circulating purp control. elircuits were revised
as fTollows:

 

1. In both the fuel and coolant salt civculabing pump circuits, the
existliong pump bowl level swilech actuvation valve was changed to re-
duce the level at which the pump is stopped, and one new switch was
installed to prevent pump startup until normal Ti11 level is reached.
Since the salt level drops 8 To 12% after pump startup, the single
47

level switch system previously used did not leave enough operating
margin to prevent normal level [luctuations from stopping the pump
and shutting down the reactor.

2. To prevent unnecessary shutdowns, the control which indicates that
#V 103 is frozen is now sealed oubt afier the fuel pump starts. The
only purpose of this interlock is to prevent pump startup until the
freeze valve is closed.

3. Jumpers were added avround the coolant salt system helium off-gas ra-
diation contacts in the fuel pump circuit to prevent the pump from
stopping each time a’ cireult test is conducted. The radiation monitor
ig a safety-grade device, and its primary function is to initiate an
emergency drain. Both channels must be tested routinely.

Fuel Processing System. A mass flow-rate meter was installed to
indicate the flow of HF to the fuel storage tank during fuel processing
operations. The purpose of this second flow measurement is to provide
an independent check on the existing orifice meter because the HY flow
rate is important in calculating the amount of oxide removed from the
fuel salt.

 

To prevent possible diffusion of Hp into the HF gas suoply cylinder,
interlocks were installed to close both the Hp and HF gas supply station
valves when the main gas supply valve to the fuel storage tank is closed,.

Operating Experience — Process and Nuclear Instruments

Control System — Relays. Little or no trouble has been experienced.
Virtually all design changes to the relay control gear have been made to
meet new requirements developed by operating experience, not to correct
malfunctions of this equipment. ‘

 

Valves. The difficulty experienced with the pressure control valve,
PCV 522A, in the pump bowl off-gas system ls a part of the larger general
problem of off-gas contamination:® by carryover of solids and vapors which
are deposited in the off-gas lines. Valve-selection criteria for the off-
gas system did not include considering nongaseous forelgn matter in the
off~gas stream.

Two pressure control valves, PCV 500J and PCV 510A1, in the main in-
let helium line gave difficulty from galling. These valves are being re-
worked.

Pressure Transducers. One straln gage unit in the sampler-enricher
suddenly shifted calibration, bubt during operation returned to its original
state. The most likely explanation is moisture which, after geliing into
the device, was baked out during service.

 

Thermocouples. Thermocouple performance has been excellent. Only
one in-cell thermocouple was lost during the six-month period covered by
this report. The plastic insulation on the radiator thermocouple lead
wires suffered from local temperaturesl7 which were substantlially in ox-
cecs of those anticipated. Remedial measures such as insulating indi-
vidual wires with ceramic beads, directed flows of cooling air, aand in-
48

sulating the high-temperature regions from heat sources have reduced, hut
not eliminated, the problen.

ILigquid-Level Bubblers. The helium bhubbler instrumentation used for
fuel salt level instrumentation experienced a mechanical faillure in two
of the differential pressure-sensing instruments. The Tailure was a
leaky connection caused by weak welduments fabricated by the vendor and
used to attach autoclave Titltings Lo the pressure inlet tubing. With
the assistance of the Metals and Ceramics Division, the weldment was re-
designed and rewelded. No further difficulties have been expericnced.

 

Weight Tostrumentation on Salt Tanks. The system has not been en-
tirely satisfactory. The basic input Instrumentation (weigh cells, mount-
ing, etc.) has functioned satisfactorily, but the readout has given trouble.
Manometer readout is accomplished by seclecting a particular weigh cell
channel wiblh pneumatic selector valves. The valves are composed of a
stacked array of individual valves operated by cams on the operating handle
shaft. The valves leak; proposals to eliminate the problen are being con-
sidered.

 

Nuclear Safety Instrumentation. The electronic instruments have
given little trouble. The solid-state modular instruments, hitherto un-
tried In an operating installation, have needed very little service, and
no ma jor problems have developed with use. Channel 3 of the safety system
produced an abnormal numver of spurious trips. A large number of {aese
are believed to have originated In faullty, vibratilon-sensitive relays in
a commercial electronic swibceh which provides the high-temperature trip
signal in the channel. Vibration isolation and substitution of a dif-
ferent relay are being considered as possible antidotes. Another source
of occasional false trips is believed to have originated with a chabter-
ing of the relays which change the sensitivity of the flux amplifiers in
the safety circuits. This has been corrected. Very generally the source
of spuriocus trips has been difficult to trace, since they are random and
usvally appear Lo be unrelated to events clsewhere in the reactor sys-
tem. As electrical noise-producing components elsewhere in the system
are eliminated, the number of false {rips is expected Lo be reduced.

 

Wide-Range Counting Channels. Moisbure penetration has been ex-~
perienced with the wide-range counting channel "snake"'® assenbly. An
improved waterproof Jacket is expected Lo cure tnis affliction. A faulty
vernistat and an overloaded geary reducer in the drive unit of the wide-~
range counting channel required replacement.

 

Linear Power Channels and Servo Controller. 'The compensated ion
chambers use a small electric mobor to change compeunsation. One motor
drive has given some trouble and has been responsible for the maintenance
required by these chambers. The servo rod controller has been used for
automatic control in both the Tlux and tewmperature modes. kExcepting the
provlem associated with the wide~range counting channel, the servo's per-
formance was very satisfactory.

 

Flectrical Power System. Substitution of the new 50-kva solid-~state
converter for the existing 25~kva motor generator is expected to reduce
or eliminate mwany of the problems slemming from instability, noise, aod
poor voltage regulation.

 
o~
O

False alarms from the monitron in the esst service tunnel were trace
to electrical anclise from the sampler~enricher vacuum pumps. An electrical
nolse Tilter is being designed to remedy this. A faulty solid-state switeh
in the sampler-enricher was responsible for nolse 1n the cubput of the
Sorensen regulator.

Data bystem

The installation of extensive modifications®® to the data-~-logger—
computer was completed on August 31, 1965, and testing began immediately.
ceveral Talled compubter components were found and replaced, and two loose
connections in the analog input system were found and repaired. Additional
deglgn changes and adjustments were required as a2 result of the modifica-
tlons, and, with the exception of the digital filter-integrator, which re-
guired complete redesign, these changes and testing were completed on Sep-
tember 16, The seven-day acceptance test was restarted on Sepbember 16,
1965, and completed September 23, 1965, The remainder of Septepber was
spent correcting miscellaneous hardware problems, principally locse cir-
cult card connections. An alr-conditioning failure caused the computer
room ambient temperature to rise to 85°F, and, simultaneously, the ac supply
voltage Tell to 103 v. These two conditions, although simultaneous, were
not coupled. I[o these cilrvcumstances it was impossible 4o keep the systenm
on line for more than a few hours at a tiwme. It was concluded that suc-
cessful operation of the camputer regquires that ambient Temperature be
held below 85°F and that ac supply voltage be maintained between 105 and
120 v.

catisfactory operation was restored, and the system was accephed on
October 1, 1965, provided that Bunker-Ramo Corporation (L) supplies and
installs a digital filter~integrator for the analog input signals, and
(2) provides and installs, at OBNL's option, cirecuitry which prevents
damage to the core memory by restart transients subsequent to a power
logs and which provides a contbrolled sequence automatic restart when
power is restored.

October was spent checking and calibrating inout instrumentation
and in troubleshooting hardware fallures, principally in the analog in-
put area of the system. Instrument calibrations were completed in No-
vember, as were substantlal modifications to lmprove reliability sched-
uled by Bunker-Ramo. The digital filter-integrator and restart circuitry
were alsce installed during this shutdown.

In the period of December 1965 through February 1966, the logger-
computer achieved operating status. In addition to the roubtine and
periodic collection of operating data, 1t was used to obhain transient
and Trequency responses and to determine fusl temperature coefficients
of reactivity. It was programmed to operate the control rod for the
pseudorandom bilnary tests reporbed in "MSRE Dynamic Tests," this report.
The log of operating data, which did not seem significant during opera-
ticn, became extremely useful during analysis of the off-gas problem.
During this period the various compubing and logging programs were being
modified in accordance with the reguirements developed during reactor op-
eration,
50

84

Figure 1.22 charts, on a weekly basis, the "in service” or "on" time
as a percent of total time. The shaded arcas on the figure are scheduled
shutdowns and cannot be charged against the system as malfunctions. Duriung
the period immediately following sysbem acceptance, performance, as noted
above, was disappointingly low and did not approach the specified require-
ments. 'The system has shown a slow but steady overall improvement 1n op-
erating reliability as debugging progressed. IFor example, during February
the logger-computer "in service” time was 99.7%.

MSRE Training Simulators

 

The power level tralining similatort? was operated successfully in

October 1965 as part of the operator training program. It was set up on
two general-purpose, portable BAT-TR-10 analog computers and tied in with
the reactor instrumentation system. No special hardware was required.
The control of the simulator was effected cntirely from the operator's
console, where actual rod moticn, radiator door positioning, and blower

CRNL-DWG 66-2622

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'INTEGRATED
,/ "ON" TIME
100 - / ) .
| L—j . -
_T\"!‘ - *_&H-a.iw}-o—w— -m-».,a,__g..-;—‘l‘w *- ”“’\
] \,,—m" i e
; 95 98
20 ‘
S |
801 E 3 "ON" TIME
4 \ SCHEDULED SHUTDOWNS
§ FOR MAINTENANCE AND
70 \ MODIFICATIONS
4 j {] SCHEDULED SHUTDOWNS
i Y FOR MAINTENANCE OF
€0 ‘ PERIPHERAL EQUIPMENT
;\; -
w504 |
= 4
£ ] N
40
30
20
10
OCT A NOV. 5. DEC.3. JAN.7 FEB.4. MAR.4.

Pig. 1.22. Availability Record of MSRE Data System, October 1965 to
February 1966, Inclusive.
51

menipulation were used as inpubs Lo the simulator. Readouts of the siwmu-
lated linear and log power, key system temperatures, and healt power were
provided by the reactor instrumentation. The simdator was used Lo check
out the reactor flux and lcad conbrol systems and the procedures used to
gwiteh from flux to temperature servocontrol.

Documentation

Except for some revisions and additions to instrument speciflication
sheet and preparation of a design report, documentation of the MSRE in-
strumentation and controls system design is complete. During the past
report period, instrumect application and switch tabulations were com-
pleted and issued, and design drawings were revised to incorporate as-
built revisions and recent additions to and revisions of the system.

References

1. S. J. Ball, Simulators for Tralning Molten-sSalt Reactor Experiment
Operators, ORNL-TM-1445 (in preparation).

 

2. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, ». 19.

3. 8. J. Ball and T. W. Kerlin, Stability Analysis of the Molten Halt
Reactor Experiment, ORNI~-TM-1070 (December 1965).

 

 

4 R. L. T. Hampton, Simulation 4(3), 179-90 (1965).

5. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNI-3872, p. 22.

 

6. MSR Program Semimnn. Progr. Repb. Aug. 31, 1965, ORNL-3872, o0p.
61—62.

 

7. R. H. Guymon, MSRE Design and Operations Repord, Part VIIL, Operating
Procedures, vol. 1, ORNL-TM~-908 (December 1965).

8. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL~3872, pp. 118-19.

 

2. MSR Program Semiana. Progr. Bept. Aug. 31, 1965, CRNL-3872, vp. 94—948,

 

10, MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 43.

11. S. E. Beall et al., MSRE Design and Operations Report, Part V, Re-
actor Safety Analysis Report, ORNL-TM-732, Fig. 2.27, p. 98, pp.
100-10L1.

 

 

12. Ipid., p. 104.
13.

1.

16.

17.
18.

19.

52

lpid., pp. 108-12.

Ibid., p. 100.

 

Thid., p. 98, Fig. 2.27.

R. B. Briggs, Status of the Problem of Plugging in the Off-Cas Lincs
in the MSRE, MSR-66-3 (Mar. 1, 1966) (internal use only).

 

See Chap. 2 of this report.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 38.

 

Ibid., p. 40,
53
2. COMPONENT DEVELOPMENT

The efforts of the development group were devoted Lo assisting in
the prepower operation and testing of the reactor. Several changes in
the equipment and procedures were mades, and these are described velow.

Freeze Valves
The specifications Tor all but three of the Treeze valves were sim-

plified by eliminating the "rapid" thaw requirvement. This requirement is
needed in FV103, which controls the emergency drain of the reactor, and
in FV204 and FV206, which control the emergency drain of the coolant
gystem, but had been included in the remaining valves only as an opera-
tional convenience. Since the maintenance of the proper temperature dis-
tribution needed to ensure a rapid thaw was more difficult than was con-
sistent with good operaling practice, the rapid thaw requircment was
eliminated, and these valves are now operated elther in the thawed or
desp~-frozen condition. Thne valves which might contain suffiecient radio-
activity in the salt to produce radiolytic fluorine at low temperatures
are maintained above 400°F at all times, Experience with the in~pile ex-
periments and other tests have shown That essentlally wo fluorine is re-
leased avbove this temperature.

The addition of the modulating air-flow controllers on FVI03, FV204,
and FV206 and the separation of the heater control circuits on FV204 and
206 have resulted in greatly improved operation of the freeze valves.

Control Rods

Control rod units 1 and 2 have operated without difficulty since the
initial installation at the reactor. Examination of rod 3 at the end of
the criticality tests revealed that the braided wire gheath had besn torn
2l a point 28 in. below the tow block. The inner convoluted hose was
worn bub not completely through the wall. Cause of the damage appeared
to be the upper roller in the control rod thimble, which had jammed and
would not rotate. The roller had a galled flat area on one side. The
thimble roller and upper control rod hose were replaced. Recentl examina~-
Lion indicated no further dAifficulty after zeveral months of operation
at, temperature.

Rod drop times for 51 in. fall remain at 0,72 sec to 0.8 gec for all
three rods. Overall rod lengths remained within 0.10 in. of the installed
lengthg.
54

Control Rod Drive Units

 

Thermal switches were installed in the lower end of the drive housing
and are gcet to alarm when the temperature rises above 200°H, 'The air flow
through the drive wunlt cases is abhout 1.4 scfm, which appears to be ade-
quate to maintain the temperature within the housings at less than 200°F,

1t was necessary to replace the No. 3 drive unit because the position
indicators revealed an 0.8-in. deviation from the original zero set point.
The deviation was caused by a partially restricted spring actlion of the
preload spring operator of the drlve chain. This allowed enough slack in
the drive chain to permit the chain links to slip over the sprocket gear
teeth. The unit continued to operate without difficully with the 0.8-1n.
crror until replaced.

The lower limlt switch of unit No. 3 showed an intermittent tendency
to stick when the control rod was dropped from zbove 24 in. The switch
could be released by fully withdrawing the coutrol rod. We found that
the shock absorber stroke was 25% greater than normal when the control rod
was dropped from 51 in. There is evidence that the inertia of the switch
actuating arm was sulficient, when the control rod was dropped from 51
in., to overcome the force of the actuator recovery spring to a point
where the lower end of the actuating rod could strike the lower flange of
the drive unit, The actualing rod had been slightly bent duve to striking
the flange, cauging it to bind in the guide bushing. A strooger recovery
spring will be installed to prevent the overtravel.

figure 2.1 is a photograph of the control rod drive units in position
on the reactor. The shielding and access hatch have been removed, showing
the air and electrical disconnects in the drive unit cover. When a unit
is removed,the gmall hatch on the drive unit cover is removed (note unit
No. 2 in Fig. 2.1), which permits access to the cortrol rod tow block Tor
releasing the control rod from the drive unit, and also access to the
lower drive unit flange Tor releasing the drive unit from the thimble
flange. This method of removal has been utilized for control rod mainte-
nance since the initial installation at the reactor.

Radiator Doors

Thermal warping of the radiation doors and door cgeasls permitted air
to leak through the radiator enclosure at an excessive rabe, and modifi-
cations had to be made to reduce the leakage.

The following modifications were made (Fig. 2.2 shows the face of the
radiator with the inlet door raised after repairs and modifications):

1. The door trip locks. The Jocks were modified and adjusted so that,
when the door wag Tully closed, the trip locks Torced the metal door,
by a rCorward camming actiocon, against the soft seal on the face of the
radiator enclosure. The forward force 1s exerted by movement of the
door in the downward direction into the rolling cam locks located on
the radiator frame.
23

   
 
 

CONTROL-ROD ACCESS |
H REMOVED

dREMOTE MAINTENANCE
PICKUP BAIL

    
 
  

GRAPHITE SAMPLING

ACCESS HATCH FOR 'N;TAL*’-‘«TION ,E

DRIVE—-UNIT AND
CONTROL-ROD REMOVAL

   

    

 

 
 
 
   
  

 
 

< CONTROL-ROD~COOLING
AIR-SUPPLY DISCONNECT

ELECTRICAL
DISCONNECT

 

DRIVE - UNIT HOUSING
.~ COOLING-AIR-SUPPLY
DISCONNECT

 
 

L

 
 

% \-\\
" DRIVE UNIT NO.

    

 

 

Fig. 2.1. Control Rod Drive Units in Operation Position at MSRE.

: =
PHOTO 52447
, RADIATOR DOOR (N
RAISED POSITION

CAM PULLOVER BRACKET

A

 

LOWER -LIMIT

HOCK-ABSORBER ARM

i

 

Fig. 2.2. Inlet Side of Radiator Door in Raised Position.
56

Door guide tracks. The door moved too close to the radiator seals
when it was raised or lowered. Warpage of the door caused some of
the seal elements to hang on the door, and they were torn lcose. The
position of the door track was moved to provide 1 in. of clearance
between the seals and the face of the door when the door was released
from the fully closed position.

Cam follower bracket. The doors tended to drag and jam in the guide
tracks. Cams had been installed in the tracks to force the doors away
from the seals when they were not fully closed. The cam follower
brackets are mounted on the doors, and contain rollers which ride on
the track cams. These rollers were damaged by the wedging action of
the trip locks against the door rollers and tracks when the doors were
dropped into the closed position. The cams were changed to provide
clearances of 1/8 in. between the rollers and cams when the doors were
closed. A short movement of a door in the upward direction brings

the roller into contact with the cam, forcing the door away from the
soft seal.

End rollers were installed on the cam follower brackets to prevent
side motion from jamming the doors against the tracks.

Reliable operation of the limit switches is important in reducing the
number of ways in which the doors can malfunction, and the original
switches were not very reliable. New heavy-duty switches and actu-
ators were installed to obtain more positive action. An additional
upper limit safety switch was added plus a mechanical "hard stop"
above the added upper limit switch. In the event of complete switch
failure the door strikes the "hard" stop and an overload switch stops
the drive motor before the door seals can be damaged.

The door position indicators are synchro-driven devices located on
the drive shafts. The doors are lifted by steel cables which are
wound and unwound on chain-driven sheaves on the drive shafts. If
the doors jam while being lowered, the sheaves continue to turn and
give incorrect indications of the true door positions. Continued ac-
tion of the drive unit, after the doors have stopped moving, causes
the 1ifting cables to unwind from the sheaves and attempt to rewind
in the reverse direction. The cables usually Jjump out of the grooves
in the sheaves and become snarled and kinked.

A "loss-of-tension" device has been designed but not installed which

will stop any action of the drive unit should the cables become slack.
It is simply a switch which will be actuated by movement of the cable from
i1ts normal tightened position.

7.

When the clearance between the door and door seals was changed to 1
in. with the door open (see No. 2 above) it was necessary to provide
additional clearance between the door and the radiator enclosure hood.
The hoods on both inlet and outlet doors were shifted 1-1/2 in. to
provide this clearance.
B
Radiator Heater Electrical Insulation Failure

The electrical leads from the radiator heaters extended vertically
up through the radiator enclosure and terminated in Jjunction boxes on
the floor of the area between the radiator door hoods. These leads are
insulated with ceramic beads. The electrical leads extending from the
Junction boxes to the supply and control circuits were plastic-insulated
wires which were rated for 140°F. Heat leaks up through the radiator
enclosure along the wire bundles overheated the junction boxes, and the
plastic insulation melted.

The heat leakage into this area was reduced by welding the original
sheet-metal enclosure wherever possible, by patching some openings, and
by reinsulating. The Jjunction boxes were removed from the high-temperature
area between the hoods to a cooler position on the east wall of the pent-
house. Electrical wire extensions between the relocated Junction boxes
and the beaded heater lead wires were insulated with 194°F thermoplastic
and asbestos composition. Figure 2.3 shows the present arrangement of
the heater wiring between the hoods over the radiator doors. Increased

-+ . PHOTO 82264

   
  

‘.

CROSSOVER DUCT i

OUTLET DOOR SHROUD ;i

7

 
 

TOP OF RADIATOR
(8-in. INSULATION)

e

ey

e
o=

  

Fig. 2.3. Area Between Radiator Door Shrouds and Top of Radiator
Enclosure.
58

circulation of cooling air in that area was provided by rerouting some

of the existing exhaust ducts. Cool air is directed downward through
the drive mechanism into the heater lead area; exhaust ducts remove the
heated air from the flcor between the hoods. The south exhaust duct can
be seen in Fig. 2.3. There are two additional exhaust ducts on the north
end of the ares; these are 10-in.-diam flexible tubes. Also a duct was
installed between the door hoods so the hoods will be cooled when the
main blowers are running.

The average ambilent temperature in the area between the doors after
the above modifications was approx 100°F with the blowers off. Tempera-
ture within the beaded electrical lead bundle penetrations at the top of
the radiator was 600°F with the blowers off. These temperatures are sat-
isfactory, and the temperatures also were satisfactory with the blowers
omn.

Sampler-knricher

 

During the shutdown period following run 3, the low-power experiment,
several modifications were made to the sampler-enricher to increase the
efficiency of operation. Most of these were discussed in the previous
semignnual report.l The major modifications made at that time included:
(1) replacing the removal valve with a modified version, (2) adding a
knob to one linkage pin of each access port operator to aid in opening
it with the manipulator if it should fail to function properly, (3) put-
ting steel rings in the convolutions of the manipulator inner boot to
hold it free of the manipulator arm, and (4) adding a pressure switch to
prevent an excessive pressure gradient across the manipulator boot.

The manipulator arm which was bent during run 3 was replaced. The
hand was straightened, and a 1/4 X 1/4 in. projection was welded to the
bottom of each finger to aid in handling the capsule cable. Clearances
in the Castle Jjoint were increased to reduce the force needed to operate
the manipulator.

Additional changes have been made to improve the reliability of op-
eration. The brass keys used to attach the capsules to the latch were
replaced with nickel-plated mild steel keys. In case a capsule is dyopped
it can now be retrieved with a magnet.

Two changes were made in the interlock circuit. A pressure switch
installed on the manipulator cover requires a negative pressure in the
cover before the operational or the maintenance valve can be opened.

Also, both the operational and the maintenance valves must be closed be-
fore either the access port or the removal valve can be opened. These
changes were made to ensure that there is double containment at all times.

Installaticn of lead shielding around the sampler-enricher has been
completed. Ten inches of lead or the equivalent was installed around
the main components and 4 in. around the off-gas system. While the re-
actor was operating at 1 Mw, the radiation level around the equipment
was < 1 mr/hr.
During runs 4 and 5, 40 samples were taken, 9 of which were 50-g
samples for oxygen analysis In place of the usual 10-g samples. Two
of the large capsules failed to trap a sample. The first assembly was
not long enough to submerge the £ill opening of the capsule in salt in
the pump bowl. All other assemblies were made longer. The second fallure
to trap a sample probably resulted from the capsule catching on the latch
stop at the top of the pump and not entering the pump bowl. The capsule
has been modified slightly to make it hang straighter and is therefore
less apt to hang.

While withdrawing one of the 10-g capsules, the position indicator
stopped, indicating that the capsule had hung or the cable had jammed.
After repeated attempts, the cable was withdrawn completely. The cable
was examined to determine the cause of trouble. Apparently the capsule
or latech had hung on the gate of the maintenance or coperational valve
causing the cable to coll up. The cable also backed up into the drive
unit box and caught ian the motor gears. The cable was straightened, the
open limit switches on the operational and maintenance valves were reset
to open the valves wider, and the cutside diameter of the latch was re-
duced. No similar trouble has been experienced.

The equilibrium buffer pressure Lo the operational valve dropped
from 55 to 35 psia during one sampling operation. This represents a
leak of aboul 20 cc He/min through the seal on the wpper gate of the
valve. A particle of salt or metal apparently dropped on the gate while
the capsule was being removed or attached to the latch and lodged be-
tween the gate and the seat. Repeated efforts to blow the particle from
the sealing surface have falled. The leak rate has not increased with
continued use. The leak of heliun from the buffer system is malnly an
inconvenlence, as 1t slowly pressurizes the volume above the valve, and
the gas must be vented to the auxiliary charcoal ved about once a day.
Besides this seal, three more sealing surfaces between the pump bowl and
the sample access area provide adequate safety; therefore the valve will
not be replaced at this time.

The sarple capsule used while the reactor was operating at 1 My left
conslderable contamination in the transport contalner. The open end of
the bottom cup read 40 r/hr near ccontact after the capsule was removed.
Replaceable liners have been Tabricated to reduce the contamination left
in the transport container.

Coolant Salf Sampler

 

During runs 4 and 5, ten samples were isclated from the cooclant
punp bowl, two of which were 50-g samples. During one sampling the
latch failed to slide freely through the transfer tube. No reason for
the difficulty could be found, but the outside diameter of the latch
was reduced to lessen the possibility of future trouble.

During the shutdown period following run 3 a capsule was left hang-
ing on the latch. A dry helium atmosphere was maintained in the sampler.
o0

When the capsule was retrieved from the pump bowl in taking the first
sample of run 4 (CPH4—1), a black film covered the salt and adhered to the
outside of the capsule. The fl1lm was identified as decomposed oil. An-
other sample taken 4 hr later was clean with no evidence of black £ilm.
The sampler was opened and examined for oil contamination. A small

amount was found on the top flange of the glove box near the elastomer
gasket. No evidence of oil was found around the drive unit or capsule.
he source of the oil was not located, bub subscqueni samples have been
free of the film.

Ixamination of Components from the MSRE Off-Gas System

 

The difficulty with plugging and partial restriction in the MSRE
of f~gas system 1s described in the section on systems performance. bSev-
eral of the components which had given trouble during the early operation
at 1/2 and 1 Mw were removed to the HRLEL for examination. Photographs
were taken through periscopes, and many samples of the foreign material
found in the components were removed and transferred to the analytical
chemistry hot cells for analysis. The results of these analyses are re-
ported in the section on chemistry. A summary of the visual examinations
made in HDRLEL is given below.

Capillary IFlow Restrictor FE 521

 

The flow restrictor consists of a short length of 0.08-in.-ID tubing
welded into the line at ong end and coiled to fit inside a 3-~in.-ID con-
tainer. The other end of the line 1s left free. The container was cut
open, and the entrance and exit ends of the restrictor tublng were ex-
amined. The entrance region wag clean, and only =a small deposit was
found on the container wall near the end of the discharge region. This
deposit is shown in Fig. 2.4. The restrictor was not completely plugged
when examined in the hot cell.

Cheeck Valve CV 533

 

The check valve is a small spring-loaded poppet-type valve installed
in the gas line to prevent a backflow into the fuel pump gas space during
the venting operation on the drain tanks. Some foreign material was found
on examination of the poppet, butl this was nol sufficient {c stop the
gas flow. The valve poppet is shown in Fig. 2.5, and for the purpose
of the examination is shown in the open position. The soft O-ring which
makes the seal can also be seen In its normal position ou the poppet.

The other material on the cone of the poppet is an amber varnish-like
material, which was identified as a hydrocarbon.

Charcoal Bed Inlet Valve HV 627

 

The valve examined was one of four which control the gas fiow into
each of four parallel charcoal sections which make up the main charcoal
Bl

bed. The valve had given indications of almost complete plugging before
removal from the system. The valve was cut apart with a saw in the hot
cell, and the valve stem is shown in Fig. 2.6. The small metallic chips
shown near the large end of the cone are from the sawing operation. The
entire cone was shiny as though wet with an oillike material. There was
some white amorphous powder on the tapered section of the valve trim.

 

Fig. 2.4. Deposit in Flow Restrictor FE 521.

 

Fig. 2.5. Check Valve CV 533.
62

 

Fig. 2.6. Charcoal Bed Inlet Valve Stem HV 621.

The powder seemed to be adhering fairly strongly to the metal surface,

although there were some discontinuities. A similar material was found
in the valve body. The general appearance of this area indicated that

the powder had moved up and down with the motion of the valve stem and

was therefore not dislodged by the small stroke of the valve. Samples

of the white material were removed for analysis.

Iine 522 Pressure Control Valve PCV 522

PCV 522 controls the gas overpressure on the reactor system by vary-
ing the gas letdown (off-gas flow). The flow passage at the valve seat
was quite small (about 0.010 in.) in order to control the relatively low
gas flow (4.2 liters/min). The valve was mounted with the seat above
the stem. Flow was down through the seal (reverse of normal). The
valve stem was covered with an amber oillike coating, and there was an
accumulation of fluid in the recess formed by the bellows-to-stem adapter
piece as shown in Fig. 2.7. The tapered flat (flow area) on the stem
had small accumulations of solid material. The valve body was coated
with similar material and had a semisolid mass of material on the sur-
face near the seat port as shown in Fig. 2.8. Samples of these materials
were removed for analyses.
63

[ R27812

 

Fig. 2.7. PCV 522 Valve Stem.

R27813

 

Fig. 2.B. POV 522 Valve Body.
64

Line Filter 522

 

Filter 522 is a l-in.-diam by 15-in.-long cylindrical type 316 stain-
less steel sintered metal element in a l-l/2—in. (iron pipe size) pipe
housing. Element thickness 1is l/l6 in. Flow 1is from the outside in.

The filter area is 0.34 ftz, and the removal rating is 98% particles

> 0.7 w. The pressure drop (clean) at 4.2 liters/min helium (normal MSRE
flow) is 2 in. Hz0. The filter was installed immediately upstream of the
reactor gas pressure control valve (PCV 522) to protect the very fine
valve trim.

The filter assembly was removed from line 522 on February 8, 1966.
Initial visual examination was made at HRLEL on February 9, 1966. The
upper one-third of the element was largely steel gray in appearance.

As shown in Fig. 2.9, the steel gray also appeared in the lower portion
at the seam weld and in a few isolated spots. The major portion of the
lower two-thirds of the element had a frosty white appearance. At no
place was there any evidence of a measurable buildup, or cake, on the
filter surface.

Visual examination was continued on February 10, 1966, at which
time it was noted that the frosty white areas had become darker in ap-
pearance, tending toward the steel gray. Inspection of the interior of
the filter housing showed an accumulation of foreign material in the
bottom as shown in Fig. 2.10. ©Samples were scraped from the side of the
element. The filter was then reassembled, and a flow pressure drop test
(see below) was run on February 11, 1966. The filter was once again dis-
assembled, and the element was sectioned into short pieces. The foreign
material in the bottom of the housing was removed readily by rinsing with
carbon disulfide.

R27776

 

Fig. 2.9. Frosty Deposits on the Surface of Filter 522.
65

R27814

 

Fig. 2.10. Deposit in the Bottom of Filter 522 Housing.

Flow Test on the MSRE Filter from Line 522

 

The filter and filter housing which had been removed from line 522
on February 8, 1966, was tested in the hot cell of the HRLEL as indicated
in the flow diagram given in Fig. 2.11l, and flow tests were conducted on
February 11, 1966. The gas was discharged into the gas disposal filter
system of the hot cell. By taking readings of the filter pressure gage
through the hot cell window for various helium flow rates the relative
plugging of the filter was determined.

The data obtained are shown in Fig. 2.12. The graph also shows the
results of a helium flow test conducted on a duplicate but clean filter.
The data indicate that for a given pressure difference the 522 filter
was passing only 5% of the corresponding flow of the clean filter. How-
ever, extrapolation of the data to the normal MSRE off-gas flow indicate
that, even though 95% plugged, the contribution of the filter (0.075 psi)
to the total system pressure drop (5 psi) was negligible.

Fuel Processing System Sampler

 

The mockup that was used to develop the sampler-enricher for the
MSRE is being modified and will be installed as the sampler for the fuel
66

ORNL-DWG 86-4754

DISCHARGE
UPSTREAM —
PRESSURE :
GAGE T

HOT~CELL
WALL =

 

gl
il
iy
b

 

 

=== FILTER ASSEMBLY ?ggfifi&gRE
P NO. 522 LINE 522
THROTTLE
/ VALVE

 

 

SINTERED METAL
PREFILTER

HOT CELL

 

NN

tFLOWMETER

HELIUM SUPPLY

Fig. 2.11. Test Arrangement at Building 3525 Hot Cell. ZFlow vs
pressure-drop test. Filter No. 522 removed from MSRE on Feb. 8, 1966,

ORNL-DWG 66-4755

HOT-CELL TEST 2-11-66 ON FILTER 522
REMOVED FROM MSRE LINE 522 ON 2-8-66

 

.’§

= HELIUM

o

G

0

w

z NORMAL MSRE

9 OFF-GAS FLOW

9 R

o

& od
BENCH TEST BY ORNL ON
CLEAN TYPE-522 FILTER

0.0t

100 10' 10° 10° 10

HELIUM FLOW RATE (liters/min/ft%)

Fig. 2.12. Flow vs Pressure Drop. MSRE filter No. 522. Filter
element: Pall Trinity type G. 316 ss sintered metal. Area: 0.34 f£t2.
N
~J

processing system., Design of the modifications to the mechanical equip-
ment is finighed except for the ghielding. Modifications to the equip-
nent to upgrade it for 15-psig service are nearly complete.

Redesign of the instrumentation is complete. This work involved
some revisions Lo panel and field instrumentation, some changes in control
circuitry, preparation of installation and interconnection drawings re-
quired to install the system at the MSRE site, and revisions of nomen-
clature and numbering of instruments and circuits to conform with the
MSRE practices.

Functionally, the ingtrumentation and control of thie chemical proc-
ess sampler is similar to that provided on the fuel sampler-enricher
system; however, the detail design of the chemical prccess system sampler
ig simpler. This reductlion in complexity resulted from reduced require-
ments Tor containment, which, in turn, resulted from the lower radiation
levels present in the chemical processing system. The reduced contain-
ment requirements permitted the use of conventlional (sifigle tracked)
control circuitry instead of the redundant (dual tracked) "safety grade
circultry used in the fuel sampler-enricher. Also, it was possible, in
scme cases, Lo use commercial-grade components instead of special weld-
sealed components. In other cases the reduction in requirements for con-
tainment and/or redundancy elimlinated the need for components.

1

O0ff-Gas Sampler

 

A system is being designed to permit analysis of the reactor off-
gas stream. The arrangement is shown schemstically on Fig. 2.13. A
side stream of 100 cc/min is taken from the reactor off-gas stream at
a point downstream of the in-cell volume holdup. The sample stream
passes to the sample unit inlet manifold, whence it may be routed to one
of three alternate paths:

1. a thermal conductivity cell for on-line determination of gross con-
taminant level,

2. a chromatograph cell for batch-wise and gquantitative determination
of contaminants,

3. a refrigerated molecular sieve trap for isocolation of a concentrated
sample, which will be transferred to a hot cell for analysis.

The sample equipment will be housed in a containment box located in
the pipe trench south of the vent house. A recirculating air system
coupled with a radiation detector is provided to permit rapid detection
of system leaks.

In addition to the off-gas sampler, an abtempt will be made to study
the nature and character of off-gas contaminants by inserting various
gample coupons into the pump-bowl vapor space through the sampler-enricher
line.

Instrumentation and controls design for the off-gas sampler is in
progress and nearing completion. Instrumentation is being provided for
68

ORNL-DWGC 56-4756

CONTAINMENT {5¢f, O psig
30 cfm - T T T T T T T T T T

‘]_ . VENT

"TO STACK

JTHERMAL—CONDUCTIVITY
o CELL

AIR SYSTEM

|
[
| |
| _ - !
CHROMATOGRAPH
t) : —— {2 CELL B 1
| T |
| i TO AUXILIARY
!
i

e b —
CHARCOAL BED
NOTE: DELAY TIME ‘——ee ——i
BETWEEN PUMP BOWL AND @ wfl

     
 

RECIRCULATING

 

 

SAMPLE UNIT APPROX |

|
|
|
|
|
)

 

 

 

 

 

 

 

 

 

Fomin L — A X REFRIGERATED SAMPLE TRANSFER
| MOLECUILAR-SIEVE TRAP SOTTLE
ENRICHER SAMPLE-ISOLATION |
SAMPLER BOMBS e Ll
\ 3 .
L 100 em™/ min
3psig
. 4200 cmd/min
volume | L. | o | | voume i — CAD:(AJ woon BE“E;‘
HOLDUP HOLDUP "L_i_Lm 1?5
---------- o ' STACK
FUEL PUMP LINE 522

BOWL REACTOR OFF-GAS

Fig. 2.13. Schematic Diagram — MSRE Off~-Gas Sampler.

on-line chromatographic and conductivity analysis; for measurements of
flows, pressures, and temperatures required for proper operablon of and
interpretation of data from the chromatograph and conductivity analyzers;
for control of temperature of a molecular sieve trap and of the level of
a liquid-nitrogen bath in which the molecular sieve is immersed; for de-
tection and annunciation of undesirable operating conditions; and to pre-
vent the occurrence of harzardous conditiocns.

Since the sampler will be an integral part of the primary contain-
ment during sampling cperations, and since some components of the sampler
would not meet the requirements for primary containment system components,
solenoid block valves will be installed in the inlet and outlet lines
which connect the sampler to the reactor system. Two valves will be in-
stalled in series in each line. These valves will automatically close
and isolate the sawmpler from the reactor system in the event of high pres-
sure in the reactor containment cell, high pressure in the fuel-pump bowl,
or high air activity in the sampler enclosure. High reactor cell pressure
is indicative of a rupture of the primary contalinment and the occurrence
of the maximum credible accident. High fuel-pump-bowl pressure indicates
that conditions exist that could result in a rupture of The sampler pri-
mary containment. High sampler-air activity indicates that a rupture of
the sampler primary containment has occurred. Closure of the block valves
resulting from high sampler-air activity (and tne aCCcompanying alarm) will
also provide protection to the sampler operator against the occurrence of
high background radiation resulting from small leaks in the sampler.
Sampler-air activity will be detected by two GM-tube-type (ORNL model Q-
1916) gamma monitors, which will monitor two separate and independent
69

alr samples collected from and rebturned to the sampler enclosure. The
isolation block valves and assoclabed detecting instruments and control
cireuitry were designed in accordance with the recommendations of the
ORNL "Proposed Standard for the Design of Reliable Reactor Protective
Systems.” At least two independent devices were used to supply input
signals to the system and to effect the corrective blocking action.
Cne-out-of-twe logic was used in the control circuitry, and separation
and identification were maintained in the detail design of the wiring.

Panel~-mounted instrumentation will occupy 5 lin £t of panel (6 £t
nigh X 2 ft deep). These panels will be located in the vent house, south
of the reactor building. The sampler and associated instrumentation will
be located in a trench outside the vent house. ©OGince all major sampling
operations will be carried out at the sampler, all readout of ianformation
will be presented at the sampler panels; however, occurrence of an alarm
condition at the sampler will actuate an annunciator in the main control
room, and provisions are being made to permit some information to be trans-
mitted to the Computer—-Data Logger. Also, a sample permissive switch will
be located in the main control room. This switch, which is connected in
the vlock valve circuits, will be used to prevent operation of the sampler
without knowledge of the reactor operators.

Most of the instrument components used in the off-gas sampler were
salvaged from the ORNL MIR-47-6 experiment located at Idaho Falls or were
on hand in ORNL Reactor Division stores. Reconditioning and callbration
of this equlipment and procurement of additional components are nearing
completion. Preparaticn of instrument applicatlion drawing is essentially
complete. Panel design is approximately 75% complete. Interconnection
wiring and piping design is under way, and fabrication of panels is also
under way.

Xenon Migration in the MoRE

 

Based on results of the ©7Kr expérimenta and xenon stripper effi-
ciency reported by the University of Tennessee, the steady-state 135%e
poison fraction for the MSRE was computed. The following major assump-
tions were made:

1. Wo bubbles circulating with the fuel salt.

2. Todine remains in solubtion, that is, it 1s not adsorbed on any sur-
faces and is not volatilized in any form.

3. Xenon is not adsorbed on any surfaces.

4. The mass transfer of xenon across the salt-gas interface in the graph-
ite pores will not be lnterfered with, for example, by the accumula-
tion of sludge or fission products at this interface.
70

To compute the 135%e poison fraction, thne 135%c concentration in the salt
must first be calculated. This 1s accomplished by the following rate bal-
ance:

135%c generation rate = 12°Xe decay rate in salt
g S
+ 135%e burnup rate in salt + 135%e stripping rate (1.)
+ 135%e migration rate Lo graphite,
where
135%e migration rate to graphite = 135%e burnup rate in graphite

+ 127%e decay rate in graphite.(2)

At a given power level the left gide of Eq. (1) is a constant, and
the right side is a fuaction of the '?7Xe concentration in the salt.
After the concentration is computed the 135%e poison fraction may then
be computed.

The results of This calculatlion at eguilibrium are shown in ['igs.
2.14a, b, and ¢, as a function of the variable indicated when all other
variables remain constant. The circle on the curve indicates the ex-
pected or design value. For this calculation the core was divided into
72 reglons, and average [luxes and adjoint fluxes were used for each
region. Figure 2.14c shows that the poison fraction is considerablg
lower at low power levels. This is because a higher fraction of 135%e
decays alt these power levels. One notable result of the calculation is
that the xenon poison fraction did not change in value when the diffu-
sivity of xenon in graphite ranged from 1077 to 5 X 1077 ftz/hr. This
is because the mass transfer coefficient from salt to grapnite is con-
trolling the transfer process. Therefore, 1f future reactors have low
mass transfer coefficients as in the MSRE, it may not be necessary to
purchase the more expensive low-void-fraction and low-diffusion-coefficient
graphite, ifT r35%e poison fraction is the only consideration. The final
measured *37Xe poison fraction may be different Irom these predicted values
because of the agsumptions involved. As more information is gained about
the reactor system and its chemistry, the equations will be adjusted to
provide a better model for calculation of the migration in future systems.

Remobe Maintenance

 

The activities of the remote maintenance group were dictated largely
by the condition of the reactor. During the last stages of constructlion
and for as long as the cell was open, attention was given to trying the
remote maintenance techniques on the installed eguipment and in check-
ing to ensure and improve maintainablility. Once the cells were closed
the effort was turned to recording cxperiences and planning for later
work. Tlater, actual maintenance was performed in several areas. A de-
scription of this work is given below.
71

ORNL -DWG 66 -4757

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3

2 \fi\

\“\

1
S
=
o (@)
5 0
L
2 0 5 10 15 20 25
1w STRIPPING EFFICIENCY (%)
=
Q
o
g
-3
€
>
w
0

2

_.-——-‘—"—"_-—-_-—-
.‘-_-'-,‘—--"
/.’

q
8
= (£)
S 0
5 0 0.02 0.04 0.06 0.08 0.10
g
o MASS-TRANSFER GOEFFICIENT TO BULK GRAPHITE (ft/hr)
=
Q
0
o 3
D
e
w0y
M

 

 

 

 

 

 

 

 

(e}

 

1 10 100
REACTOR POWER (Mw) LOG SCALE

Fig. 2.14. (2) Effect of Stripping Efficiency on 135%e Poison
Fraction; (b) Effect of Masgs Transier Cogfficient on 13°Xe Poison Frac-
tions; (c) Effect of Reactor Power on 135%e Poison Fraction.
=

Practice Before Operation

 

Practice with reactor components was had in handling the pump ro-
tary element and in changing the existing graphite sample assembly for
the final uwnit. The sampling procedure included the handling and/or
operation of (1) an inert atmogphere in the standpipe, (2) a special
heater tool to thaw out a frozen salt annulus around the outside of the
sample holddown assembly, (3) disconnects, valves, and flanges inside
the standpipe, (4) stowing the holddown assembly, and (5) the sample it-
self. In addition to these a limited visual inspection of the core was
conducted using a 7/8—in.~diamASCOpe, A special fitting had to be in-
stalled in the strainer basket above the core to acconmodate the revised
sample assembly.

The pump rotary element was lifted out of the cell for inspection
using the lifting yoke and crane with in-cell direct guiding. During
the 1ift, clearances of all the auxiliary eguipment were observed and
points where damaging interferences could occur were noted. The unit
was reinstalled by remote means using the view afforded through the port-
able maintenance shield. The process of torgueing up the twenty-two 1-
1/2-in.-diam bolts in the pump's lower flange was started in the portable
shield bul was finished with direct means to save time in the reassembly
process. While set up in this area, representative electrical and ther-
mocouple disconnects, pump-bowl heaters, and auxiliary Tlanges were han-
dled with the long-handled tools.

Procedures were revised, and in cases like the pump, graphite sam-
pling, and the control rods were written up in minute detail. Tabula-~
tilons were prepared relating neaters, Lhermocouples, spare disconnects,
and other eguipment with shield-block locations and elevations. Additional
tools were designed, and some existing tools were revised.

Maintenance of Radiocactive Systems

 

In January, after a short period of power operation, it was necessary
to use remote maintenance on a number of tasks. [Table 2.1 shows the op-
erations which were done and the actual time and manpower used. Radiation
levels were significantly lower than those which are anticipated after
prolonged power operation. While this required only miniwal shielding,
enough of the shielding was used for the operations with the control rod
and. PCV 522 to provide experience in setting up and in using the tools
through the portable shielding. Although PCV 522 and its filter assembly
was quite a large radiocactive source, 100 r/hr at the cuter surface of
the filter, no significant personnel exposure was encountered. 1In general
the tools, technigues, and previously made preparabions proved more than
adequate. In some cases it was necessary to fabricate special tools or
revise existing ones, but these cases did nob hold up progress to any
great extent. Good cooperation belween management, reactor operations,
maintenance planning, health physics, and the craft foreces contributed
toward efflicient and smoothly run operations.
Table 2.1. Summary of Remote Mainternance Work at MSRE — January 27 Through February

 

Perscanel Involved

 

 

Llapsed
Description . Tim Dat
: Rlgger . : s Remote (hr? ’
and Millwright  Pipe Fitter Maintenance
OUperator enanc
1. Remove check valve from line 533 G 1 2 2 3-1/2 Jan. 27
(use Cam-loc on extensicn)
2. Remove FE 521 and repisce with a 0 o 4 2 2 Jan. 28
new unit
3. Control rod drive No. 3: remove 2 1 0 1 6 Jan., 29
shielding, disccunects, and cover
plate
Qontrol rod drive No. 3¢ remove 2 2 O 2 g8 Jan. 31
drive ard install new drive
Control rod drive No. 3: hook up 2 2 0 2 2 Feb. 1
air and electrical disconnects
Control rod drive No. 3! replace 2 2 0 1 2 Feb. 2
cover and shield
4. POV 522, unsuccessful abtempi in-cell 2 0 2 2 3 Feb. 3
replacemen
5. PCV 522, remove and put intce deconteami- 3 1 2 2 6 Feb, 5
nation cell
PCV 522, revuiid and replace in systenm 3 1 2 6 Feb.
PCV 522, remove from system 2 1 i 2-1/2 Feb. &
PCVY 322, cut filter loose from valve, 2 2 1 2 Feb.
put in can and carrier
9, PCV 522, attach flanges for fiow l{ests 0 0 i 1 2 Feb, @
and 10
10, PCV 322, install new assembly 2 G 1 1 1 Feb. 10
Total {hr) g9 48 4y 76.5 46

 

€L
T4

Pump Develobment

 

MSEE Punps

Molten-Salt Pump Operation in the Prototype Pump Test Facility. The

 

modifications® to the test loop were completed. The venturi flowmeter
was relocated upstream of the orifice flow restriction, and a new flow-
straightening section was installed near the pump discharge. 'The new
flow straightener was modified o provide additional weld attachment of
the blades inside tae pipe. The new arrangement of the system is shown
in Fig. 2.15.

"wo test runs were made with the prototype pump circulating the salt
TiF-BeFa-Zr¥ ,-Tht 4-UF 4 (68.4-24.6-5.0-1.1-0.9 mole %) at 1200°F. The

ORNL-DWG 66-2048

  

AlR

§
4

" ] WELL -~

_____ {\\ _fl/ ”

SALT
PUMP

AR COOLER
VENTURI
METER

  
     
  
  

,,,,, :Ih L . ' ST nql--m

) . e :::-"t-:-~Tf.i;f.‘.f,i_‘1j:_j_'_';:_:-_—‘k— o J i

DISCHARGE = “SALT FLOW < FH;- .n.%fin_iFT\\\\\

PRESSURE STRAIGHTENER — THROAT
INLET PRESSURF —~*j[ PRESSURE

).

41

_8-in. PIPE ) i

i} _&-in. PIPE [ FREETE FLANGE J
- e T

\ R s L gifiifif mm J

     
  

e SALT FLOW . T ,
T T T \ QRIFICE FLOW
L:J RESTRICTER
_____ L FREEZE VALVE \\
’ MSRE FREEZE
FLANGE

 

Fig. 2.15. MSRE Salt Pump Hot Test Facility, Modified.
75

first run covered a period of 165 hr to provide hobt shakedown of the im-
peliler (13-1in.-0D) for the fuel pump spare rotary element. The other
run covered a period of 166 hr to provide shakedown of the drive motor
for the spare coolant pump.

The prototype pump ie being prepared for test with an 11-1/2-in. im-
peller. Measurements will be made with the radiation densitometer to
determine the undissolved. gas content of the circulating salt. OCther
tests will be made in the pump tank off-gas circuit in connection with
the plugging incidents experienced at the MORE.

Pump Rotary Element Modificaticon. The spare rotary element for the
fuel salt pump was modified to provide a positive seal against leakage
of oil from the catch basin past the outslide of the shield plug and into
the pump tank. Previcusly, a solid copper O-ring compressed between the
bearing housing and shield plug was used to provide the seal. Incidents
have occurred wherein oil leaked past the O-ring. Cross-sectional views
of the part of the pump which ineludes the modifications are shown In IMig.
2.16. The larger section shows the relationships between the pump shaflt,
the shaft lower seal, bearing housing, cabch basin, shield plug, and tThe
pump tank. The exploded views indicate the nature of the modificatlon.
Positive containment of the oil is obtained by providing a seal weld be-
tween the bearing housing and the shield plug. The rotary element is
being assembled for cold shakedown, and after satisfactory operation
will be prepared for MERE use and stored. The spare rotary element for
the coolant pump will receive the same modification.

 

Tubrication System. Modified Jet ;pum.ps4 were ingtalled in the re-
turn oil lines of the MSRE salt pump lubrication systems.

 

The lubrication pump endurance test’ was continued, and the pump
has now operated continuously for 22,622 hr circulating oil at 160°F
and. 7C gpm.

MK-2 Fuel Pump. The pump tank design was completed, and it is being
reviewed.

Drive Motors, A new design was conpleted for the containment vessel
for the drive motors of the MSEE salt pumps. It provides single contain-
ment for the electrical power penetrations, in contrast to the double con-
tairmment provided in the original design. The resulting simplification
should eliminate a serious fabrication problem previously experienced with
weld-induced laminations 1In the vessel wall.

Other Molten-Salt Pumps

 

PK-P Fuel Punp High-Temperature Imdurance Test. Endurance wperation6

of this pump was continued, and it has now operated continuously for 5160
nr circulating the salt IiF-BeF,-Thi,4-UF, (65-30-4-1 mole %) at 1200°F,
800 gpm, and 1650 rpm.

 
70

Pump Containing a Molten-Salt-lubricated Journal Bearing. This

pumpV'B’g was placed in operation circulating salt at 1200°F, but it ex-

perienced a seizure after 1 hr of operatlion. IExamination revealed the
seizure to be in the molten-~salt-lubricated bearing. As was the case
with the test previous To this one,s two of the snap rings for the bear-
ing sleeve gimbals support were lost and the other two were only loosely
retained. Whether the bearing seizure or loss of the snap rings occurrcd
first is a mool guestion. The design of the gimbals support is being

ORNL-DWG 66-2049

SOLID COPPER O-RING BUNA N O-RING

Ve
LOWER SEAL ////
CATCH BASIN 7 ///

   
  
 

7 7 - LOWER SEAL
7 — //// \  CATCH BASIN
[N,
77 A d =

 

 

 

 

IR

 

 

 

 

 

 

SRR
BEFORE MODIFICATION AFTER MODIFICATION

 

   
 
    
   
   
   
   
    

SHAFT LOWER SEAL--- / -
BEARING HOUSING

SEAL OIL LEAKAGE OQUT

 
   

T

,‘ g

   

OIL
LEAKAGE

 

s

PUMP TANK

 

SHIELD PI.UG SHAFT § -7

Fig. 2.16. MSRE Salt Pump Rotary Llement, Modification.
7'

modified to eliminate the use of the snap rings, which are intended for
retaining the four fulcrum pins in the support. Instead, the pins will
be retained by plugs tack-welded at thelir outer ends.

Instrument Development

 

Ultrasonic Single-Point Molten-8alt Level Probe

 

As no fuel has been processed during the last report period, the
ultrasonic level preobe installed in the MoRE fuel storage tank'® has not
been used.

Efforts to correct this instrument's deficiencies have been unsuc-
cessful. The excitation osecillator which had proved so unstable was
modified considerably (using information supplied by the manufacturer)
in an attempt to eliminate its excessive frequency drift. These modi-
fications made no improvement in the stability of the oscillator. As
the design and success of other intended modifications depended upon the
stability of this conmponent, none of them have been attempted. To cor-
rect this situation, the ORNL Instrumentation and Controls Division Elec-
tronic Design Group has been requested to investigate the electronic equip-
ment asscciated with the ianstrument and make recommendations as to the
most practical action to take.

High~Temperature NaK-Filled Differential-Pressure Transmitter

 

The coolant-gsalt system flow transmitter that failed in service at
the MSREYL, 12,13 yag refilled with silicone oil and tested with the seals
at room temperature. Prior to refilling the transmitter body, a vacuum
punp wag connected to the instrument in such a manner that the pressure
could be reduced on the process side of both seals and both sides of the
silicone~filled body at the same time. Liguid braps were installed in
the body evacuation lines to catech any oil that might be Torced out of
the transmitter body by the expansion of trapped gas during evacuation.
When the pressure wag reduced to 28 in. Hg vacuum (the lowest pressure
attainable with the system at that time), oil was forced from the capil-
laries: 8.2 ml from the low-pressure side of the transmitter, and 15.8 ml
from the high-pressure side. This indicates that there was gas trapped
in both sides of the transmitter body, and that the amounts trapped were
unequal. The exact amount of gas trapped cannot be calculated from the
amounts of oil caught in the contalners, because the capillaries are not
connected to the two sides of the transmitter body in the same horizontal
plane. After it was refilled, the temperature sensitivity of the instru-
ment was 1 in. (water column) change in indicated outpul for each degree
Fahrenheit change in ambient temperature. This sensitivity is greatly
reduced from that observed prior to refilling but is still excessive.

A1l the above testing was done with the seals and transmititer body

at room temperature. The seals are now being heated to 1200°F, at which
temperature the test will be continued.
78

The new NaK-~filled differential-pressure transmitter ordered as a
sparc for the MSRE is scheduled for delivery in March.

Float-TLype Molten-8alt Level Transmitter

 

Performance of the ball-float-type transmitter installed on the MSRE
coolant-galt pum_pl4 continues to be satisfactory. The necessary actions
have been completed to correct the previously reported errors in calibra-
tlon.t%Y% A these were calculated corrections, some additional future
adjustment may be reguired.

The ball-float level transmitter on the MsSRE pump test loop con-
tinues to operate satisfactorily, although on one fill the float would
not rise untlil the temperature of the heater on the bottom of the float
chamber was increascd. bGither uanmelted salt or curvabure of the bottom
inside the float chamber may now be assumed to be the cause of this stick-
ing. A previous inspection of the core tube showed no deposits that would
cause it to stick in the core chamber. The bottom inside curvature of the
float chamber on the MSRE pump test loop fits The curved bottom of the
float very well and may block the entrance of the molten salt into the
chawper. This design error was noted prior to the fabrication of the
iloat chamber for the MSRE and corrected on the assumption that this
valving action might occur. The bottom of the float chamber was flat-
tened so that the round-bottom float could not block the flow of the
molten salt into the chamber.

Bxcept for detailing of the differential transformer assembly, de-
sign of the ball-float transmitter installation in the MK-2 MSRE fuel
circulating pump has been completed.

Cne of the prototype ball-float level transuitterst® installed on

the level test loop completed four years of operation at temperature
this month. Tt is still operating satisfactorily. The other was re-
moved last year so the ultrascnic single-point level indicator could be
installed. It was operating satisfactorily when removed.

Conductivity-Type Single-Point Molten-5Salt Level Probe

 

Performance of the conductivity-type level probes installed in the
MSRE fuel, flush, and coolant drain tanks continues to be satisfactory.
There have been no further failures of excitation and signal cables on
these probes.

Single-Point Temperature Alarm Switches

 

Observation of the performance of 110 single-point temperature alarm
switches installed at the MSRE has continued. As reported previously17
the switch modules were resel prior to power operation of the reactor.
Data obtained from subseguenlt spot checks of module set points indicated
79

that a few modules had shifted excesgsively; however, these data are con-
sidered to be inconclusive because records indicated that the modules may
have been readjusted. Some additional cases of duvual set points were dis-
covered and corrected. Routine checks and observations of module perfor-
mance will be continued until sufficient data are obtained to permit an
accurate evaluation of the reliability of thege devices.

Helium Control Valve Trim Replacement

 

Testing of alternate material combinations for helium control valve
trim has been terminated. BSince tests had shown that all the alternate
material combinations under consideration would gall when operated in
dry helium without lubrication and that none would gall if a minute amount
of lubricant was present,l7 the valves which had failed in MSRE service
were refitted with spare trim using the original (L7-4 PH plug and Stel-
lite No. 6) material combination. Close attention was given to alignment
during reassembly of the valves, and all trim was given a light coat of
machine oil before installation. All the repalred valves operated satis-
factorily in shop tests. One of the repaired valves was installed in the
MSRE main helium supply line and operated for several months before stick-
ing. One additional heliuvm control valve failure (also due to sticking)
has occurred. Both valves are being disassembled and will be refitted
with spare (17-4 PH to Stellite No. 6) trim. Because the failures may
be due to evaporation of the lubricant, an attempt will be made to find
a less volatile oll or grease to use for trim lubricant. Previous at-
tempts to use grapnite-base lubricants were not successful.

Thermocouple Development and Testing

 

Coolant Salt Radiator Differential Temperature Thermocouples. Noise
in the coolant salt radiastor differential temperature thermocouple circult
which was under imvestigationla was finally reduced to an acceptable level.
A final check on the effects of thermocouple and lead-wire material mis-
match was made by heating thelr disconnects both individually and simul-~-
taneously to 150°F. No readable change in the output voltage was nobted.

A geven-point calibration was run on the thermocouple pair between 1040
and 1250°F, which resulted in a constant error with the ocutlet thermo-
couple reading 0.220 mv high with respect to the inlet thermocouple. The
resistance of the thermocouple loop was determined to be high encugh to
reguire recalibration of the receiving instrument to maintain reguired
accuracy. Also, the zero of this instrument was offset to correct for the
thermocouple loop error.

 

Thermocouple Drift Tests. The checking of elght metal-sheathed
mineral-insulated Chromel-~Alumel thermocouples fabricated from MSRE ma-
terial for calibration drift at 1250°F was concluded.!® A1l thermocouples
continued to show some drift to the end of a 26-month test period. The
final temperature equivalent drift values were between +4.7 and +6.4°TF.

 
80

Temperature Scanner

 

Performance of the temperature scanning system developed for use
at the MSRE?? has continued to be satisfactory. The modifilcation previ-
ously reporte&jzl together with improved calibration and maintenance pro-
cedures, appears to have eliminated the calibration drift problems. Per-
formance of the mercury switches has been excellent. We had expected that
the swiltches would need freguent attention and that the mean 1life between
routine cleaning or repair would be about 1000 hr. The switches have given
very little trouble, and the mean life of the switches hag been much greater
than 1000 hr. Since the start of operation of this system in September
1964, there have been no bearing or other mechanical failures of the five
switches installed. IFour switches developed excessive noise during this
period and required cleaning and replacement of the wercury. In one of
these cases the switch failure was caused by a failure in the nitrogen
purge gas supply system. All five switches were cleaned and reconditioned
before the start of power operations as a routine precaulbionary measure.

References

1. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL~-3872, pp-
56-58.

 

2. Ibid., p. 55

3. MSR Progrem Semiaonn. Progr. Rept. Aug. 31, 1905, ORNL-3872, p. 61.

 

4. Tbid., p. 61.

5, TIbid., p. 62.

 

6. TIpid., p. 65.

 

7. MSR Program Semiann. Progr. Rept. July 31, 1963, ORNL-3529, p. 54,

 

8. MSR Program Semiann. Progr. Rept. Jan. 31, 1964, ORNL-3626, p. 40.

 

9. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 52.

 

10. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 66.

 

11. Tpid., p. 70.

12. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, v. 48.

k2

 

13. MSR Program Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p. 45,

 

14. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 7L1.

 
81

 

 

 

 

MSR Program Semiann. Progr. Rept. Feb. 28, 1965,
MEBR Program Semiann. Progr. Reph. Feb. 28, 1962,
MSR Program Semiann. Progr. Rept. Aug. 31, 1965,
lbid., pp. 73, 74.

Ibid., p. 73.

MSR Program Semiann. Progr. Rept. Aug. 31, 1962,
M3R Program Semiann. Progr. Rept. Feb. 28, 1965,

 

ORNL~3812,
ORNL-~3282,

ORNL-3872,

ORWI-3269,

ORNL-3812,

P

D.

.
62

3. MSRE REACTOR ANATYSIS

least-bquares Formula Tor Control Rod Reactivity

 

One routine fTunction of the MSRE data logger will be the periodic
calculation of the separate reactivity effects assoclated with reactor
operation above some base line, or zero-reactivity condition. For this
purpose, one of the guantities the computer must calculate is the nega-
tive reactlvily corresponding to each configuration of the shim and
regulating rods. By use of a least-squares curve-tfititing procedurc, we
have obtained an analytical formula which closely approximates the in+ew
gral curves determined from rod calibration experiments at zero power
The Tunctional expression used for fitting the experimental curves was
determined by anplying a perturbation technigue to the integral ex-
pression for the rod reactivity, as outlined below.

If a reference state is chosen to corregpond to zero rod ingertion,
an expression for the static reactivity change when 2 single rod (or rod
group) is inserted to a position Z in a core with effective height H and
radius R 1is

fflR ffiZ ¢§ BA @r dr dz
Apg = 0 - 0 —— . o)
fO fO Pof0r dr dz

 

In this Tormula, @ is the column vector of group fluxes corresponding to
the condition with the rods inserted, ®O is the row vector of adjoint
fluxes for the reference state, 8A 1s the local perturbation in the neu-
tron removal operator due to absorptions in the rods, and P is the neu-
tron production operator. This expressiocon can also be applied if the

shim and regulating rods are inserted to different positions. To simplify
the analysis, we have assumed that the tips of the shim rods are at egual
insertions, Z, which is always equal to or above the position of the regu-
lating rod, Z;. We Turther specialilze the formula to two-~group diffusion
theory and one dimension, corresponding to the directlon of rod Insertion.

Equation (1) becomes

Z1 % (1) . Lo ¥ (2)
fO dpo d)zap ¢'2 Az +le ¢02 o5

a0 ¢2 dz

 

0
j; o1 (vEg, o + Vi, 0,) de

where subscripts 1 and 2 designate the fast and thermal groups respec~

(1)
a?

gsorption cross section in the region O = 2 = 75, polsorned by the combined

tively. The quantity 875 1s the effectlve Increment in thermal ab-
83

(2)
az
in Zp & 2 £ 7o, polceoned by the regulating rod alone.

shim and regulating rod group, and BJ; is the corresponding increment
[ &) 8) -.L’ p &

In the usual approximation of perturbation theory, the flux distri~-
butions ¢1(z) and ¢5(z) are replaced by the unperturbed fluxes ¢pi1(z) and
¢02(Z), corresponding to the reference state. IF the approximations

nZ

92(z) ® ¢pp(z) = sin T (3)
* * . 7

dp1r ~ Pgp ~ sin 3 (4)

are introduced into (2) and the integrations performed, the usual formula
for the reactivity-worth curve for a partially inserted control rod is
obtained.® This result was tested by adjusting the parameters 8282 and
ég) in order to fit the formula to the rod-worth curves obtalined from
experiment. While reasonably good agreement was obtained in fitting the
curve for the single (regulating) rod, poorer results were obtained when
combinations of insertions of shim and regulating rods were considered.
The functional expression obtained from standard Tirst-order perturbation
theory was, therefore, Jjudged inadequate for representing the rod-re-
activity curves.

02

We have found that a significant improvement in the fit can be made
by assuming that the axial distribution of thermal flux, ¢, 1s perturbed
according to the position of the shim-regulating rod bank (Z;) but is un-
affected by the position of the regnulating rod (Z5). "The approximate
formula for ¢, used for this analysis was:

 

0=z = Zy:
0s(z) = _—-—A----O; sinh oz + —— sin 2, (5)
2 2
D,BZ + D07 DEg, + DaZ
7hn = 2 = H:
95 (2) = B sinh Q(H — z) + — T sin =, (6)
D. B2 + D.OF DB2 + Dor?

2 T 2
B

 

wihere
. (0)
e - ...:a..:?m
Do 7’
62(;)
O = O 4 i
p Dy
- (3)°
T H ’
and D, and Z;g) are, respectively, the thermal diffusion coefficient and

the macroscopic thermal cross section of the core in the absence of the

control rods. The coefficients A and B depend on the strength of the
1)

shim-regulating rod bank absorption, 52;2’,

tors. They can ve determined by requiring continuity of Llux and current

at the interface, Z;, between the "rodded” and "unrodded" regions. The

final result of applying this analysis in Eq. (2) was

together with gecmetric fac-

_Co + C1 F1i(Zy) + Co Fo(Zy, Zp) + C3 Fa(2q) + Cy Fu(Zy, 72)
Ap_ = 2 (7)

 

 

 

” Cs + Cg F1(2y)
where
217 . 217
1 (Zl) = ""iLé‘;"* — 310 -——«Hl ) (8)
I I ) ‘_;
FolZa, Z2) = gflififififfiiéi# + sin 2151 m-sin42?;2 , (9)

FB(ZT) = (COSh a7y sin %3:, — aj:i[-“j: sinh 071 cos Tlgl >

T T ginh ol Zy) cos ZZA J ,  (10)

x

 

o .
[ cosh O(H —~ Z;) sin T = -

 

. | . . N [ 107
AVS = h & — e QL ol
b4( 1, Zo) (\0051 71 sin 5 i sinh 73 cos T >

X { cosh G — Z,) sin Moz, T sinn C(H -~ Zo) cos ) ] . (11)

 

OH
The parameters Co, Ci, ... , Cg were found to depend on the rod absorp-
: (1) (2)

tion strengths, 82a2 and 62&2

ters characterizing the reactor core in the absence of the rods. In

addition, it was found that

» together with other macroscopic parame-

<< 1, 0273 sH. (12)

ng 7y (Zy)
Cs

 

This latter result is useful in modifying expression (7) Turther, in
order that linear leasl-squares analysis could be used to determine the
numerical values of the parameters. Thus,

I
[Cs +Cg Fr(Z2)]™t = 6%- (1.-£€?L ) . (13)

By introducing this approximation in Eq. (7) and redefining the constants,
it was found that

Ao, (Za, Z2) =ag + 81F1 + apF, + asFy + agf, + asFf

+ 2gF1Fs + agfiFa + agfF, . (14)

Bxpression (14) was fitted to the experimental rod calibration
curves, using standard linear least-squares analysls to determine the
coefficients ag, «.. , 8g. The position of the rods relative to the
extrapolated zero of the unperturbed axial flux distribution (Z = 0) is

Zi,2 = %o + X1,2 , (15)

where Zg is the "zero point" insertion of the rods when withdrawn to their
upper limit, relative to the extrapolated zero of the flux distribution,
and X1,z are the measured rod insertions. Adequate approximations for

the parameters characterizing the unperturbed flux distribution, Zo, &,
and I, were obtained from earlier core physics studies. The results of
this analysis are swmarized in Table 3.1 and Fig. 3.1. 1In Fig. 3.1 the
magnitude of the rod reactivity is plotted as a function of rod position
for various configurations of shim and regulating rods. The ordinate
scale in thnis figure i1s arbitrarily normalized to zero measured reactivity
when the regulating rod is fully inserted and the two shim rods are with-
drawvn to 51 in. The leftmost curve represents the reactivity change when
the regulating rod is moved with the shim rods fully withdrawn. The
rightmost curve represents the case when the three rods are moved in a
banked position with the tips of the rods at equal elevations. The re-
maining curves represent the reactivity change when the regulating rod

is moved with the shim rods held fixed at variocus intermediate positions.
86

The so0lid curves are the reactivity magnitude calculated from the least-
The points shown as solid dots are sample points deter-
mined from the experimental rod calibration curves. Over most of the
range of rod movement the calculated and measured reactivity are in

squares formula.

agreement within about 0.02% 3k/k.

Table 3.1.

Near the extreme positions of the

Numerical Values of Parameters in Ieast-Squares
Formula for Control Rod Reactivity (Eq. 14)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
  
 
 
  

 

 

 

 

 

 

 

 

Parameter Value Parameter Value
o, cmt 0.082 as ~1.944 x 1077
7o, Cil 24 .6 8, 5.891 x 1078
ag 2.125 g 2.865 x 1073
aj —0.3935 an —1.747 % 210™7
a, -0.3585 ag —4.358 x 108
55 ORNL-DWG 66-4758
, o
20 1~ e POINTS OBTAINED FROM ROD-CALIBRATION
EXPERIMENTS
.8 = = CALCULATED FROM LEAST-SQUARES
: FORMULA
T 16 - {ROD REACTIVITY NORMALIZED TO |
< 7171 kg 235U IN LOOP)
82
* g ol
T |
-
5
S2
>—
—
5 10
5
3
& 08 |- Lot
i | |
% oe-———L—‘fi -
Lo |
=
04 |— .
. ®
d/ /
0.2 ')fl//%
L ]
0 4 8 12 16 20 24 28 32 36 40 44 48 52
POSITION OF ROD MOVED (in. WITHDRAWN )
Fig. 3.1. Comparison of Control Rod Reactivity from kxperimental

Curves and from IlLeast-Squares Formula.
g7

rods the error is somewhat larger. However, the least-squares formula
should be adequate for automatic moniteoring of the control rod reactivity
under most normal operating circumstances.

Spatial Distribution of '2°Xe Poisoning in MSRE Graphite

 

To supplement the experimental studies of the behavior of noble gas
injected into the MSRE fuel,? theoretical calculations were made to de-
termine the influence of the spatial distribution of 135%e sbsorbed within
the graphite core. Tt is expected that the concentration of xenon In the
salt will be relatively uniform throughout the volume of circulation.
However, within the graphite pores, the 127Xe would tend to assume an
overall spatial distribution governed by the burnout rate in the neutron
flux. This distribution would be concave, with minimum concentration
occurring near the position of maximum thermal Flux and maximum concen-
tration near the boundaries of the reactor core.

It is intended that the resactivity due to 135%e poisoning will be
periodically calculated during operation by use of the TRW-340 data
logger. TFrom a practical standpoint we are limited to the use of a
relatively simple "point" kinetics model for on-line calculations.
Therefore, any corrections for the spatial distribution of the 135%e
poisoning must be predetermined from theoretical studies with a more
elaborate model of the reactor core.

If we consider a step change [rom one ?Qwer level to another, the
correction for the spatial distrivution of 2%¥e within the graphlite
region can be determined from

X ___g . .
) Srapaze ¢ ) Fln G, 6] o) av,

= AN
NXE((I):L) t) f

 

W(Po, P1, % , (16)

o™ (r) o(r) av_

graphite
where
Pgp, P1 = initial and final power levels,
t = time after power level is changed,
@l(r) = thermal Tiux at position r relative Lo the center
of the core after the power level 1s changed,
@1 = thermal flux after the power level is changed,
averaged over the graphite volums,
'fig [(P (,.) .-] = ] ...-,_:1_135- - trati s sraphl be
% 1lr), tl = locally averaged Xe concentration in graphlite

at position r and time €,
88

fig (61, t) = locally averaged 135%e concentration in graphite
at time t, corresponding to a fictitious "flat”
neutron flux distributiocn equal tTo the spatially
averaged flux,

= normalized distribution of thermal group importance
(adjeint flux).

) = normalized distribution of thermal flux,
)

As defined by Bq. (16), W is the factor by which the *3°Xe reactivity
calenlated according to a "point" kinetics model must be multiplied to
account for the spatlal distrivution of the poisoning. In this equation

the "local average" '?°Xe concentration, N%e, is the radial average over

the graphite volume associated with a single fuel channel. It is useful
to compute this quantity before performing the integrations over the
entire core graphite volume indicated in Eq. (16). Although nearly all
the xenon in the graphite would be expected to be in the pores nearest
the graphlite-salt interface, this local averaging procedure can be used
because the radial variation of neutron flux across a single graphite
stringer is negligible.

We have used a one-dimensional model for the fuel channels in order
to simplify the caleculations. The fuel and graphite widths corresponding
to a single channel were chosen 50 that the ratios of mass transfer sur-
face area to volumes were eqgual to those of the actual channel. 'The
eguations governing the production and mass transfer for 135%e between
the salt and graphite were:

(17)

.
g (fifl - l ) , (18)
yo o

 

VL Xe Xe
NG _ . SN o
St T DXe ”axg - [GXe (=) + 7\Xe]NXe ? (19)

4 ¥1
By V1 J; Nie(x, 1) ax . (20)
g9

In these egquations

N%e(x, )

=0
Nxe(t)

=0
NI,Xe

Z
NXe

yo

Xe

Xe

Yo
Ji

VC/VL

il

local concentration of *3%%e at position x within the
graphite stringer, measured from the graphite-salt
interface, atoms per cublic centimeter of graphite,

local concentration of 135Xe, averaged over the graphite
velume assoclated with & single fuel channel,

average concentration of 1357 ang *35%e in the circu-
lating fuel salt, atoms per cublc ceatimeter of liquid,

concentration of '3°Xe in salt nearest the graphite
interface, directly exposed to the graphite pores, atoms
per cubic centimeter of liguid,

fission density in core salt, fissions per cubic centi-
meter of liguid per second,

thermal neutron flux at position r, measured from the
center of the graphite core, neutrons cm™? sec'l,

fission yield of Y?°T and 137xe,

radioactive decay constants for *3°I and 135Xe, sec"l,

effective 1?7Xe removal rate due to external strippiog,

sec” ,

135Xe 2

thermal neutron sbsorption cross section of , cm®,

diffusivity of xenon in graphite, square centimeters
of graphite per second,

coefficient for liquid film at the
interface, cm/sec,

mass transier
graphite-salt
half-width of single fuel channel, cm,

half~width of graphite assocliated with a single fuel
channel, cm,

ratio of volume of salt within the graphite-moderated
region to the total volume of circulating salt.

Boundary conditions are required to solve the differential equations (18)

and (19).
fuel channel,

At the outer boundary of the graphite associated with a single

g -
BNXG
~= | -

X=y1
90

At the graphite-salt interface,

 

2 g
N = BNXe(x = 0) , (22)
J0
>
B Xe _ =L a8 _
0o\ Se = h { Moo BNXe(x =0)| , (23)
x=0
where
B = RT/HXee,
R = universal gas constant, 82.07 cn® atm mole™ (°K)™%,
T = temperature, °K,
HX° = Henry's law coefficient, cm® atm mole“l,
€ = graphite porosity, cubic centimeters of void per cubic centi-

meter of graphlite.

AL time © = 0, equilibrium conditions corresponding to the initial
power level (PO) were assumed. The initial concentrations are the solu-
tions of Fgs. (17) through (23), with the time derivatives cgual to zero.
For times following the change in power level, the differential equations
were solved by means of laplace transforms. A sufficient approximation
to the exact solutlion was obtained by use of the condition

a2

Yo~ o 4 (24)
B
- Xe

Physically, this condition implies that most of the resistance to mass
transfer between salt and graphite 18 due to the fluid film. Results of
krypton injection experiments appear to support this conclusion.?»?

some typical numerical calculations based on the preceding model

are given in Figs. 3.2 and 3.3. Numerical values Tor the effective mass
transfer coefficient, stripping rate, and graphite porosity characteristics
were obtained from ref. 3. Figure 3.2 shows the time variation of the
correction factor, W, as the power level ascends to 10 Mw. In the limit-
ing case of a step change from O to 10 Mw, W decreases monotonically to

an equilibrium value of approximately 0.76. Curves are also glven for

the case when the power level 1s increased in successive steps, each time
1.00

91

CRNL-OWG 66-4759

 

 

0.20

0.80

 

///fll
|

 

1

w Y3Bye POISONING SPATIAL CORRECTION FACTOR

 

(

 

 

 

 

 

5.0-=10.0 Mw

 

 

Q=10 Mw

 

 

2550 Mw

 

 

 

 

 

 

 

 

 

 

070

O
o
@

Fig. 3.2. Time Dependence of the Spatial Correction for the
Poisoning in MSRE Graphite:

16 20 24 28
TIME (hr)

32 36 40

Ascending Power Level.

ORNL-DWG 66-4750Q

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 - 1 ,
{O—=0 Mw
a e e g
| et
5 ] 10,00 Mw
‘&) |
L // ‘_'____.-—M
5 /,,ff”"“ 2.5-m10 Mw
Eé 090 / e 1
© 5.0 =25 Mw
I prommn
% ]
o
)
2 100~ 50 Mw
z I . M
£ 0.80 | e LT = R
v
o
o
Q
=
@£
0
¥
0.70
0 4 8 12 16 20 24 78 32 36 40
TIME (hr)

¥ig, 3.3. Time Dependence of the Spatial Correction for the
Poisoning in MSRE CGraphite:

Descending Fower Level.

135Xe

135Xe
92

allowing equilibrium poisoning conditions to prevail before further in-
creasing the power. The initial "dip" in the curve representing Lthe 5-

to 10-Mw step is a consequence of the relative time lag between the in-
crease 1n the burnocut rate in the graphite and the ilncrease in the pro-
duction rate of *2°Xe in the salt. Figure 3.3 gives the analogous results
for power levels descending from 10 Mw.

Further studies will be made with this theoretical model in order
to determine the sensitivity of the calculated corrections to the ¢f-
Tective mass Lransfer coefficlent and stripoing rate, and to determine

.......

 

1. The Reactor Handbook, vol. IIT, Part A (Physics), ed. by H. Soodak,
pp. 204—6, Interscience, New York, 1962.

 

2. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNI~3812, pp. 12-14.

3. R. J. Kedl, personal communication, dJanuary 19606.
Part 2. MATERIALS STUDIES
95

4o METALIURGY

Dynamic Corrosion Studies

 

A test program is in progress to study the compatibility of struc-~
tural materials with fuels and coclants of interest to the Molten-Salt
Reactor Program. Thermal convection loops described previouslyl’z are
used as the standard test in this program.

Circulation of lead in a Cb—1% Zr alloy thermal convection loop?
was terminated after 5280 hr. This loop operated with a hot-leg tem-
perature of 1400°F and a 400°F AD. Metallographic examination of the
hot leg, shown in Fig. 4.1, revealed no evidence of attack; however, a
small guantity of dendritic crystals was observed in the cold leg.
Electron probe analysie indicated these crystals to be columbium, as
shown in Fig. 4.2. Mass transfer of columbium in lead has not been re-
ported previously.

Two 2-1/4% Cr—1% Mo steel thermal convection loops were started
with lead coolant to study the effect of magnesium additions on mass

 

OO INCHES

 

 

 

Pig. 4.1. Section Through the llot Leg of a Cb—1% Zr Loop Which Op-
erated with Lead for Over 5000 hr at 1400°F,
96

Y-67208

 

LIGHT OPTICS COLUMBIUM L, X-RAY IMAGE

Fig. 4.2. Results of Electron Probe Analysis Showing Presence of
Columbium Crystals on the Cold Leg of a Cb~1% Zr Loop Which Operated for
Over 5000 hr with a Hot-Leg Temperature of 1400°F and a 400° AT. 250X.
Reduced 35%.

transfer of primary elemenis. The loops will operate with a hot-leg
temperature of 1100°F and a 200° AT. Magnesium was added to act as a
deoxidizer and an inhibitor. An original plan to include titanium in
the lcad was deferred because of difficulty in making a Ti~Pb alloy.

Two loops containing molten fluorides continuecd to operate without
incident. A type 304 stainless steel loop with removable specimens has
operated for 22,000 hr, and a Hastelloy N loop containing specimens of
Hastelloy N modified with 2% Cb has operated for 33,000 hr.

MSRE Material Surveillance Tests

 

Reactor Surveillance Specimens

 

Specimens of Hastelloy N and grade CGB graphite were exposed for
approximately 1100 hr to molten fluoride salts in the core of the MSRE
during the precritical coperation and the 1nitlial critical and assoclated
zero-power experiments. The purposes for these were: (1) to monitor
materials during these preliminsry operations and (2) to be the mass
eguivalent for the similar specimens that replace them for surveillance
of the power experiments.3 These specimens showed no changes as a re-
sult of exposure during these Tirst, mild experiments.
97

The graphite specimens were nominally 0.8-in.-diam by 10-in.-long
rods. They were machined from MSKEE graphite, grade CGB, having a rela-
tively high concentration of cracks, with the idea that this would magnify
adverse changes that might occur in the relatively short, mild exposure.
There was essentlally no salt on the graphite, and the machining marks
appeared unaltered. As in the laboratory tests, radiographs showed that
surface-connected cracks tended to be filled with salt, with no salt
penetrating into the graphite. The volume of salt in the graphite aver-~
aged 0.06% of the bulk volume of the graphite, which is approximately
one-tenth of the maximum permitted in the MORE design specifications.
The relatively small dimensions of the specimens and the presgence of a
more-than-typical quantity of cracks would tend to give apnormally high
salt pickunp.

The Hastelloy N specimens, in the form of tensile specimen rods,
also drained free of salt. They had lost their bright, shiny, machined
surface and had a bright, gray-white matte surface similar to that ob-
tained in hydrogen firing of the metal.

 

 

 

 

Fig. 4.3. Microstructure of Hastelloy N Tensile Specimen Exposed
in MSRE During Zero-Power Testing of Reactor. Microstructure is ex-
tremely fine grained. Black band approximately 0.0005 in. below surface
is 0.0005 in. wide and is composed of grains whose boundaries have been
enriched or depleted in some constituent. Etchant: aqua regia.
28

Two of the specimens were sectioned and were examined metallographi-
cally. For comparative purposes, two control specimens of the same
composition were also examined. Longitudinal and transverse sections of
the shoulder and gage-length regions of the lensile specimens were ex-
amined.

Both the tested and control specimens exhibited extremely fine-
grained microstructure (see Figs. 4.3 and 4.4). The grain poundaries in
a 0.0005-in. band approximately 0.0005 in. below the surface of the
tested specimens were darkened. 1t appears that the boundaries were
perhaps either depleted or enriched in some unknown constituent. A
section of a tested survelllance specimen has been submitted for analysis
with the electron microprchbe analyzer; however, the analysis has not been
completed yet.

a4

A longitudinal view of the control specimen is shown in Fig. 4.5.
The microstructure of the control specimen indicates that the surveil-
lance specimens were in a severely worked conditicon prior to testing.
During exposure in the reactor at approximately 1200°F, the alloy re-
crystallized to the smaller grain size shown in Fig. 4.3.

R-27463

 

T

—3.007 IHCHES
= 500X

™

 

 

 

Fig. 4.4. Microstructure of Control Hastelloy N Tensile Specimen.
Although the microstructure is extremely fine grained, the specimen is
coarser grained than the one removed from the MSRE.
99

 

 

 

 

 

 

 

Pig., 4.5, Longitudinal View of Control Hastelloy N Teusile Specimen.
Microstructure indicates that specimen was in severely worked condition.
Etchant: agua regia.

Lurvelllance Control Specime

 

The reactor core control speclimen rig4 wag charged with fluoride

c
salts in the latter part of December 1962, and 1t has beein operating
satisfactorily. This unit containsg three sets of vahlt@ and Hastelloy
N specimens that mabch the sets of reactor core specimens” in the MSRE.
The function of this unit 1s to subject its zpecimens to approximately
the temperature profile and the major temperature and pressure fluctua-
tions of the reactor. These unirradiated specimens will be used to db-
tain base-line data for those thaet are irradisted in the reactor.

The reactor control specimen test unit will copy changes of the re-
actor operation, as specified above, through directions relayed to it by
the computer that monitors the MERE. Calibration of the unit with the
computer is in progress.
100

Hot-Cell Metallographic Examination of Hastelloy N
from Experiment MIR-47-6 for Evidence
of Witriding

 

When the MSRE 1s operating, the atmosphere in the reactor cell and
the drain tank cell will be nitrogen containing about 3% oxygen. There
has been some concern that the nitrogen, when ionized by radiation, will
nitride and embrittie the Hastelloy N of the reactor components. Since
the capsules in experiment ORNL MI'R-47-6 were ccooled with air and with
alr and nitrogen mixtures during irradiation, we examined the bottoms of
capsules 1 and 2 for evidence of nitriding of Hastelloy under irradiation.

The capsules were contalned in a copper heater block so that only
the tops and bottoms were exposed to nitrogen. Construction of the cap-
sules and the operating conditions during lrradiation have been re-
ported.6 Unfortunately, the outer surfaces of the boltoms were damaged
during initial disassecmbly of the capsules at the MIR hot cells. Ma-~
chined grooves on the capsule bottoms were still visible, and the bottoms

 

 

 

 

R-27480
o - z - &
- e A . i
o L i
"’ £ / o ¥ p— S " ’ — - 2
{ sR f > : ./'.’ : E ' # B
Y e <
- i ; (\ . - 0y A “
& ¢ # ©
.« . S - v -
i . * *
3 7z~ % - .
. I . . -~ . Lfié
». - L : : F P
# = * . ' . I
¥ . ? e ! ¥ C f( € - e O
. 5\ - . ] . &
L . % « - o e - <O
* “ { { 3 i ‘ i ~ %
] Vemk Sy 5 o - o
, s o O
7 & \v‘: . e £ s » 4 ‘\ ) Foe Ofl
e ‘: w [ - L P
Vo . - ‘ s
Pt
o .7 / s 7
1 5 . ~ .
- e . -, et
T & S _ ! "; < v . 5
- oy - 5 4 : |~
“ e ¢ N P
£ 7 el AP ey B - iy P x %
s £ 3. " e
* Sy ) *
" < Jfl’ . :
F - o . ~ x
i r H 2y, o ©
Fo ; : T - *
£ 5 i ’ - - i -
s ; C {‘“ e
% L . " Ty i <
. 3 (] “~
! ¥ i j 3 s i ) fj .
- * - . et i f e
-~ . * s o / -
- - - / 5 s - . Pl e LE®
st . N - R .
= ; ' -
i o §~J S ¥ 3 i : % e =
/ i $ 5 - S o
o, ; g v -

Fig. 4.6. Bobttom of the Machined Groove on the Bottom of Capsule 2,
MIR-47-6. A thin oxide scale is visible on the exposed surface; there
is no evidence of nitriding. Etchant: agua regia.
101

of the grooves, which were also exposed to nitrogen, were undamaged.
Tonglitudinal sections of the bottoms from capsules 1 and 2 were examined
matallographi cally. Attention was centered on the bottoms of the un-
demaged grooves. A thin oxide layer was present on the exposed surface,
as shown in EHD. 4.6; however, we saw no evidence of nitriding on either
of the specimens we examined.

Postirradiation Metallographlic Bxamination of
Japsules 14 from Dxperiment ORNL MIR-47-6

 

 

Four Hastelloy N capsules were irradiated in experiment ORNL MIR-47-
6. Irradiation conditlons and a description of the experiment have been
reported.® Metallographic examination of the graphite cores and mo-
lybdenum coupons, contained in capsules 1 and 4, has also been ?eported.7

A lougitudinal sectlon of each capsule vody, at the vapor-liguild
interface, was examined metallographically. A transverse section of
each capsule hody was also examined, and the wall thickness was measured.
Two unirradiated contyrol capsules dfld a section Tfrom an as-fabricated
capsule were also examined. One of the control capsules had been tested
oub-of-pile under condltions similar to those for fhe irradiated capsules
and the other capsule was a spare [or the irrsdiation capsules.

There was no apparent change in wall thickness of tThe capsules
during irradiation. A comparison of the wall thickness of the control
O
capsules and the irradiated capsules 1s shown in Table 4.1.

The interior surfaces of all Tfour irradiated capsules were smooth,
and we saw no evidence of scale or illm formation. However, the gral
boundaries at the interior surfaces of the capsule wallg appeared to be
slightly enlarged and/or Jerkened. The effect was spotbty In all Tour
capsules, and the depth to which grain-boundary darkening occurred was
ereater in the vapor phase. Darkening occurred to a depth of approxi-
mately 5 mils in capsules 2 and 3, and slightly less in capsules 1L and 4.
We saw no evidences of graln darkening on either of the control speci-
mens.  The imner surfaces of the walls from capsule 2 and {rom the tesied
control capsule are compared in figs. 4.7 and 4.8.

We canuct explain what caused the grsin-boundary effect. Because
of their high activity level, the specimens cannot be analyzed with the
electron microprobe. An unirrediated corrosion specimen that showed a
similar effect was examined with the electron microprobe, and no signifi-
cant difference between the affected edge and the center of the specimen
was ohgerved,

Development of Graphite-to-Metal Joints

 

The Jolning of graphite to structural metals such as Hastelloy N is
of prime interest in advanced molben~salt reactor concepts Iin particular,
102

Table 4.1. Wall Thickness of Control Capsules and Capsules
Irradiated in Bxperiment CORNI MIR-47-6

 

Wall Thickness

 

Capsule No. (in.)

Control, tested under same conditions 0.0525
as irradiated capsules

Control, spare capsule 0.0215
Capsule 1 (ORNL MTR-47-6) 0.0525
Capsule 2 0.0530
Capsule 3 0.0515
Capsule 4 0.0525

 

i , R-23971

    

 

 

 

 

o
L
N - . . - “
- o N
i
-
- -~ - .
[
i
U
]
o = o
s - -1 2
= e
e Q
a : o7
9]
e
™ " -
- 2 v :
R
X .
e o Y
i
-
i
— - - '
o
w» B
? [ = .
- et -
-
=

Fig. 4.7. Microstructure of Imner Surface of Hastelloy N Capsule
Wall, Capsule 2, Experiment MIR-47-6, Longitudinal View. Grain bound-
aries at surface are darkened to a depth of 0.005 in. Also note fine-
grained microstructure of Hastelloy N. FEtchant: aqua regia.
103

R-27717 °

 

ES

CH

OO X

Q.035 1IN
Tra

 

 

 

Fig. 4.8, Microstructure of Inmer Surface of Hastelloy N Capsule
Wall from Control Capsule That Was Tested Out of Pile Under Same Condi-
tions as Irradiated Capsules, Longitudinal View. Also note fine-grained
microstructure. Etchant: aqua regia.

1t is highly desirable to Join graphite pipes in the reactor core to
Hastelloy N headers. The basic obstacle encountered in attempting such
a Jolnt is the very large difference between the thermal expansion coef-
ficients of the graphite and the metal. Due to this difference, a Jjoint
oI graphite directly brazed to Hastelloy N cracks upon cooling from the
brazing temperature. One possible method of circumventing this problem
is to make the transition with one or more materials having expansion
coefficients intermedlate between those of the graphite and Hastelloy N.
This reduces the stress gradlents. Thus, this program 1s concerned with
the development of transition pieces, ag well as acceptable brazing
alloys.

Molybdenuwn and tungsten are two metals which have expansion coefl-
ticients intermediate between those of graphite and Hastelloy N, and were
selected for further study. However, due to cost and availabllity con-
siderations, nearly all work was performed on molybdenum. This decision
was further Justified by the fact that nearly all alloys which braze
molybdenum will also braze tungsten.
104

Studies are being conducted, using the newly designed transition
joint presented in Fig. 4.9. The design incorvorates an 11° tapered
edge to reduce shear stresses arising from the thermal expansion differ-
ences. This technigue may enable the joint between the molybdenum
Lransition piece and the Hastelloy N to be made directly, that is, with-
out another transition. Several assemblies are being prepared for evalu-
ation of brazability and effects of thermal cycling.

The ORNIL-developed brazing alloy, 35 Au—35 Ni-~30 Mo (wt %), is very
useful for brazing graphite; however, transmutation of the gold may limit
its appiication in a high neutron flux. Consequently, a study was under-
taken 1o develop a gold-free brazing alloy which is compatible with moiten
salts, contains a carbide former, and has a reasonably low melting point.

Tests of Graphite-Molybdenum Brazed Joint for Containing
Molten Salts Under Pressure

 

 

A first test was conducted in which a small pipe of grade CGB graph-
ite, brazed to molybdenum, contained molten fluoride salts at 700°C under
pressures of 50, 100, and 150 psig for periocds of 100, 100, and 500 hr
respectively. Tais test, which operated satisfactorily desplite a primi-
tive Jjoint design and a relatively thin-walled graphite pipe, suggests
that larger graphite-metal Joints may be fecasible for molten-salt breeder
reactors. Current MSBR designs have a maximum pressure difference of 30
psi across pipe walls, less than one-fifth the maximum pressure used in
this test.

The test configuration is shown in Fig. 4.10. The short graphite
pipe, 1.25 in. OD x 0.75 in. ID x 1.00 in. long, was machined from a bar
of MSRE graphite with its axis parallel with the extrusion direction of
the bar. Pure molybdenum caps werc brazed to the ends of the pipe by
the Welding and Brazing Group cof the Metals and Ceramics Division with
35 Au-35 Ni—30 Mo (ANM-16) (in wt %) braze, which is one of their earlier
developments in brazing ailoys.

ORMNL-DWG 66-4774

 

 

 

 

GRAPHITE 4o 45" MOLYBDENUM HASTELLOY N
7; ’
G e T T
L --------- 134 in, - L— ------------ 2V in e ned 13 in e

 

 

ANGLE CONCENTRICITY: =+ 0.001 in.
ALL ANGLES SAME AS THE ONE DETAILED ABOVE.
SPECIMENS ARE 1Y4—in. OD WiTH A 3/g-in. WALL.

Fig. 4.9, Schematic Drawing of Proposed Graphite—to—Hastelloy N
Joint Using a Molybdenum Transition Piece.
105

METAL - SHEATHED THERMOCOUPLE ——=7 |

ORNL-DWG 66-4775
_ . COMBINATION GAGE
TEE
}‘Lfl )

TO VACUUM —=—

  
 
    
   
 
  
   

INCONEL ——— o |
Y67886

 

 

 

 

 

 

——— PROTECTIVE CONTAINER

(@)

——— Mo

H— CGB GRAPHITE

SHEATH ——

OPEN-END THERMOCOUPLE
MEASURES TEMPERATURE

J
AND DETECTS SALT LEAKAGE

3
INCHES
Fig. 4.10.

Test for Small Graphite Pipe and Graphite-Metal dJoints.
106

The molten salt was fed to the inside of the graphite pipe through
an Inconel tube brazed into the top molybdenum end cap (see Fig. 4.11).
The salt was pressurized by a cover gas of pure argon. A vacuum pump
maintained a pressure of <5 pu Hg in the annulus between the graphite pipe
and the protectlive container,

The appearance of the outside of the pipe was essentially unaltered
by the test (see Fig. 4.11p). These photographs also show the good
wetting of the brazes to the metal and graphite. A small circumferential
separation of the braze and the graphite can be seen at the bottom joint
next to the graphite. Microscopic examination indicated that the ef-
fective bonding and sealing was in the crevice between the pieces. Minor
cracking tended to be confined to the fillets. The microscopic examina-
tions, also, showed that the 35 Au—35 Ni—-30 Mo (wt %) braze wet the graph-
ite but did not penetrate its structure. The braze was brittle. It did
not show any signs of attack by the molten fluoride salt ILiF-BeF,-ZrF,-
ThF¥ ,-UF,, (70-23-5-1-1 mole %).

The 50 Au—50 Ni (wt %) braze used to join the molybdenum to the
Inconel salt-supply tube bonded strongly to the Inconel. In the di-
rection transverse to the rolling direction of the molybdenum, a few

PHOTO 83567

 

(D)

Fig. 4.11. Grade CGB Graphite Pipe Brazed to Molybdenum End Caps
(a) Before Test and (b) After Containing Molten Fluoride Salts at 700°C
Under Pressures of 50, 100, and 150 psig for Periods of 100, 100, and
500 hr Respectively.
107

molybdenum grains were pulled out at the molybdenum-braze interface.
This was not apparent in the as-brazed material and will be studied more
thoroughly as the work progresses.

The radiograph in Fig. 4.12 shows that none of the salt penetrated
the matrix of the graphite. This is consistent with past experience
with this small-pore-size, high-density grade CGB graphite used in the
MSRE.

New Grades of Graphite

 

Procurement and studies of grades of graphite potentially promising
for a molten-salt breeder reactor are in the initial stages. Attempts
are being made to obtain graphite suitable for irradiation studies in
order to secure the required irradiation exposures (102 nvt, E > 0.18
Mev) reasonably soon. Samples of needle-coke graphite and isotropic
graphite are being tested. A needle-coke graphite was used in the MGSRE
to minimize irradiation contraction and associated stresses. Isotropic

Y-69255

 

Fig. 4.12. Radiograph of Thin Sections Machined (a) Transversely
and (b) Longitudinally from the Graphite-Molybdenum Brazed Joint Test
Showing That No Salt Penetrated into the Graphite Pipe Walls. (salt
would appear here as a white phase.)
108

graphite has been included in these studies because it may be superior
in mechanical properties and more resistant to damage under the high ex-
posures required in the MSBR. 8» 9

A part of this effort is to obtain graphite with the configuration
required for current designs of molten-salt breeder reactors. This is
being done because the shape and size can significantly affect the final
properties that can be bullt into the graphite. Current designs require
graphite pipe nominally 4 in. 0D X 3 in. ID X 8 to 12 ft long and 2-1/2
in. OD X 1—1/2 in. ID X & to 12 ft long. Secondary attention is given
to the small-scale laboratory-produced materials.

The prime requirements of the graphite for the initial procurement
are that it have pores small enough to prevent the entry of molten salts
and that it have a low gas permeability. The pore entrance diameters
should be less than 0.5 p, and permeability to helium at 1 atm of pres-
sure should approach 10™7 cmz/sec. The suppliers currently propose ma-
terial with a permeability in the range of 107 to 1072 cm?/sec, with
the higher values being favored.

To date we have obtained graphite samples from the Carbon Products
Division of the Union Carbide Corporation, Great Iakes Carbon Corporation,
Poco Graphite, Inc., Speer Carbon Company, Stackpole Carbon Company, and
the Y-12 Chemical Engineering Group of the Development Division.

The material obtained from the Carbon Products Divislon is a needle-
coke graphite, and the materials obtained from the others are isotropic
graphite. Short pieces of pipe 4.7 in. OD X 3.5 in. ID and 3.6 in.
0D X 2.5 in. ID have been supplied by the Great Lakes Carbon Corporation
and the Carbon Products Division respectively. The materials from the
others were rods or blocks.

To determine if these grades of graphite (and future grades) are

potentially useful for an MSBR, we are routinely examining them for the
entrance diameter spectrum of the accessible pores, permeability to
helium gas, permeation by molten fluoride salts, microstructure, specific
registance, flexural strength, and coefficients of thermal expansion.
The current samples are in this stage of examination. Those that appear
to have promise will be given additional tests for purity and crystallo-
graphic development and will be included in the irradiation studies and
graphite-metal Jjoint development mentioned previcusly.

Evaluation of the Effects of Irradiation on Graphite

 

Recent progress in the development of graphite for reactor use may
be applied directly in estimating properties of a graphite designed for
the MSBR. There are, of course, several unknowns which 1limit the ability
to state categorically that any graphite will withstand the MSBR environ-
ment. The major limitation is the lack of evidence to demonstrate that
any graphite can sustain massive doses of 1023 nvt or greater and still
109

retain its integrity. The tubular thin-wall design in the MSBR is one of
the better configurations for reducing the differential-growth problem

to within the capavilities of the graphite. The tubular shape is also
easy Lo fabricate and to test reliably and nondestructlvely in order to
ensure maximum integrity.

The properties of the most promising grade of graphite can be pro-
Jected Trom available grades with a Tair degree of certalnty (see Table
4.2). "These estimations are based primarily on properties obtained from
isotropic grades with densities required for the MSBR grade.

The magnitude of the stress generated by differential growth can he
fairly well approximated. The main uncertainty is in the flux gradient
across the tube wall. Using the conservative estimation of 2.4 X 1024
in. in.™?* nvt™t as the growth rate and a 10% change in flux across the
wall, a differential growth rate of 2.4 X 1072% in. in.™t nvt™t is ob-
tained. The restraint is internal; therefore, only about half of the
differential growth is restrained, so the effective strain rate is 1.2 X
10725 in. in.7t netT.

The creep rate coefficient is 4 X 10727 in. in.”* psi“l nvt"l, and
the stress 1is simply and directly calculated to be

D5
= Loz X 10777 = 30 psi

g o= <

=z —
4 % 10727

Thig stiress could hardly be a cause for failure, and to obtain stresses

in excess of 100 psi would require the use of grossly unreslistic ma-

terial properties.

Failure, however, could result from the inability of the grapnite
to absort creep deformatlion even though the stress level is much less
than the fracture stress. For lifetimes of 1.5 x 10?7, 3 x 10?3, and
6 x 10?2 nvt, the strain to be absorbed would be about 1.8, 3.6, and
7.2% respectively. This corresponds to 5-, 10-, and 20-year lifetimes
with a dose rate of 9 x 1014 nv, the maximun fast flux in the present
design of the MSBR. The consideration of a strain limit for failure is
realistic; however, the strain limit for fracture has not been estab-
lished. It has been demonstrated that graphite can absorb stralns in
excess of 2% in 1072 nvt without loss of mechanical integrity. There 1s
also some evidence that the growth rate will diminish after a 1022 nvt
dose; thus, the graphite might not be forced to absorb the total quantity
of strain calculated. Therefore it appears that failure by reaching a
strain limit will require at least five years of service.

The main uncertainty, as mentioned earlier, is simply the ability of
the graphite to sustain the massive dose without loss of integrily. There
18 no experimental evidence beyond 2 X 1022 nvt on which to extrapolate
the irradiation damage of graphite, so extrapolation of data to 1023 yould
be pure conjecture. There are, on the other hand, several factors which
force one to be optimistic about the ability of graphite to sustain the
Tablie 4.2. Properties of Advanced Reactor Grades of CGraphite

 

 

 

Grade
Property
1-207-85% H-315-A% E-319% MSBR

Density, g/em’ 1.80 .85 1.80 ~1.83
Bend strength, psi >6000 4500 >£000 >5000
Modulus of elasticity, psi 1.35-1.65 x 10° ~2 % 10°
Thermal conductivity, 1622 21=2"7 23 24

Btu hr™t £t (°F)TH
Thermal expansion, ¢ 5.4=6.,0 4.8-5.8 3.-4.4 0 R5.0

10°%/°¢
Electrical resistivity, 8.9 9.9 =9

omm-cm X 10%
Isotropy factor, CTE}}/CTEL 2.11 1.20 1.18 N1.15
Creep coefficient at 700°C, N4 x 10727

in. in.”t psi"‘~L vt
Permeability to helium, cu®/sec 8 x 1072 2 x 1072 <10-2

Dimensional instabiiity? at 700°C  1/2 to 1/3 of AGOT

<i/2 of AGOT

 

8G. B. Eagle, The Effect of Fast Neutron Irradiation from

495°C to 1035°C on Reactor Graphite,

 

GA-6888 (Feb. 11, 1966).

bBased or. very preliminery results on H-207-85 andé type B

ilsotropic cata to 2 X

104%,

OTT
111

damage. The Tirst factor is the recognition that in the 700°C temperature
range, the dimensional changes for the Tirst 10%? nvt do nothing more
than repair the damage caused by cooling of the graphlte from its
graphitization temperature. The additional dimensional changes of the
crystals after 1022 nvt produce stress in the graphite in the opposite
directicn from that produced by the thermal coolling. In effect, the
stress produced on the basal planes will be compressive instead of
tensile. This, of course, is very much preferred in that the crystal can
sustain this type of loading more readlly without cracking. The ability
of graphite to sustain very large crystal shears in this manner has heen
demonstrated by irradiating pyrolytic carbons and observing shear strains
in excess of 100% without any observable cracks in the structure. This
internal deformation was so great that lrregularities were evident on the
surface. Although these factors do create optimism about the ability of
graphite to sustain the irradiation dose, experimental evidence 1s needed
to demonstrate this ability.

Effects of Irradiation on Hastelloy N

 

The objective of this program is twofold: (1) to determine the be-
havior of the MSRE structural materials under neutron irradiation, and
(2) to investigate various ways of improving the resistance Lo radiation
damage by making small changes in chemlistry or by specified mechanical
treatments. Hastelloy N is a compllcated alloy, and we do not have
encugh data for a statistical analysis; so some of our present interpre-
tations may seem contradictory. Our findings are summarized below.

1. 'The dependence of the creep-rupture life of Hastelloy N at
650°C on the integrated neutron flux is shown by the data in Fig. 4.13.
These data are Tor an air-melted heat (5065) which was irradiated cold.
The quantity £ is the ratio of the rupture life of the irradiated ma-
terial to that of the unirradiated material. There seems to be a general
trend for the rate of reduction in rupture 1life with flux to decrease ag
the stress level is decreased. Three slopes are indicated which are used
in subsequent calculations.

2. The results of the postirradliation creep tests of cold-irradiated
specimens were used to predict the creep behavior of the alloy under
simultaneous stress and irradiation. The basic assumption made in this
treatment is that the creep-rupture curve is not altered by irradiation
except that the rupture life is reduced. Hence, one gives up a specific
fraction of 1life Tor an increment, of flux. The lines drawn in Fig. 4.14
have equations 0i the form

f= A/dB ) (1)

where

rupture life of irradiated material  ©

~ Tupture life of unirradiated material  tg 7

 
112
A, B = constants,
¢ = integrated thermal neutron flux.

To predict what will happen when a specimen is simultaneously being
stressed and irradiated, Hq. (1) is written as

 

 

 

 

THERMAL NEUTRON FLUX (nu/)

Fig. 4.13. Reduction in Rupture Life of Hastelloy N (Heat 5065) by
Irradiation.

ORNL-DWG 66-4776

  
  
 

TN AR
o P H IR 1 vemm -
| N "L ’ 650 °C |

  

|
|
|
||
LN e L]
i ‘ } !

STRESS {1000 psi]

1 10 100 1000 10,000
TIME (hr]

Fig. 4.14. Creep Rupture Properties of Hastelloy N (Heat 5065) at
650°C.
113

The constants A and B are evaluated from the slopes of the lines in Fig.
4.,13. With a value assigned to the fiux and a stress level chosen so
that tg is known, it remalns to determine the value of t that satisfies

(2).

The procedure Just descrived was used to calculate the lines shown
in Fig. 4.14. The constants obtained from the lower slope appear to more
closely predict the actual test data. Points are shown on Fig. 4.14 for
two in-plle creep experiments. The agreement is reasonably good at
stresses above about 20,000 psi, but there is a distinct talling off at
lower stresses. This observation 1s important in that it indicates a
longer rupture life for structural materials at the low stresses in the
MSRE than was previously vredicted on the basgis of a linear extrapolation.

3. The results of in-pile creep tests and postirradiation creep
tests are compared in Figs. 4.15 and 4.16. 'The data from test ORR 140
for heat 5085 (Fig. 4.15) seem to show that the rupture life in poste
irradiation tests 1s conslderably longer than the 1ife in in-pile tests.
The effect increases with decreasing stress. Data for two specimens at
32,350 psi indlcate that holding for several hundred hours at 650°C and
low stress after irradiation result in a longer rupture life at high
stress. The data from test ORR 138 of specimens from heat 5065, shown
in Fig. 4.16, follow the same pattern.

The specimens for the postirradiation tests in ORR 138 and 140 were
held at 650°C for about 1200 hr under little or no stress during the ir-
radiation bvefore testing. The in-pile test specimens were heated only
during the test. TIU seems possible that the difference in thermal history

ORNL-DWG 66-347%

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

60 T T T
h ]
\\\\\ ’
ummmomw
5O [t el el
\\;E
LT
| T
~ 40 e A : \ ——————
a ‘ ™
O \‘\
Q ; & =
g 30 —J el L --T---,u(;u- e ‘»\'fs. AAAAAAAAAA - S
o IN-PILE TEST o A \
W o-, =6x10% v N
= o A1 N
» 20 [~ POSTIRRADIATION TEST t—tite oot e G
A-IRRADIATED AT 850°C TO 2»(1020nw ~
*INDICATES SPECIMEN THAT WAS HELD AT 850°C
10 |- AND 15,000 psi FOR 667 hr BEFORE BEGINNING CREEP TEST -1t
ALL SPECIMENS ARE FROM TFST’ORR140
ol Ll J L] HHH
y 10 100 1000 10,000

TIME TO RUPTURE (hr)

Fig. 4.15. Comparison of In-Pile and Postirradiation Creep FProper-
ties of Hastelloy N (Heat 5085) at 650°C.
114

ORNL-DWG 66-4778

[T ST

 

 

 

]
50—+

 

 

 

 

40 A

 

 

30

 

 

 

 

 

i
‘ |
i

 

+
+ —.
e ede

 

 

 

 

STRESS {1000 psi)

 

 

 

 

 

 

 

 

 

 

 

 

 

IN-PILE TEST || |
20 ro =g, = 6 x40 ny — ] ’T—T—" L R
POSTIRRADIATION TESTS | | | ‘ ’T o | ]
IRRADIATED TEMP DOSE l &= f
10 -4 - 650 3mo§°4— = *INDICATES SPamMFN THAT WAS |4
0 - 150°C 5x10°° HELD AT 650°C AND 15,000 psi FOR
b - 150%C 5 101 657 hr BEFORL BEGHHN!NG CREEP
o L ] TEST | bl
1 10 100 1ooo 10,000

TIME TO RUPTURE (hr)

Fig. 4.16. Comparison of In-Pile and Postirradiation Creep Proper-
ties of Hastelloy N (Heat 5065) at 650°C.

Table 4.3. Effect of Irradiation Temperature on the Creep~Rupture
Properties of Hastelloy N

Thermal dose = 2.5 X 1029 navi

 

 

Specimen Irradiation Stress Rupture Strain Elongation
fié&geé Temperature (psiby Life Rate (%)
(°c) (hr) (%/1r)
Heat 5065
301 650 39,800 14.8 0.117 3.4
302 650 32,350 300.1 0.0034% 1.37
303 150 32,800 42 .0 0.08 5.5
304 150 32,350 114.0 0.0046 1.25
Heat 2477
N50 650 32,350 13.6 0.0137 0.28
WRM 43 32,350 515 3.4
Heat 7304
3ty 650 39,800 2.0 0.073 0.30
342 650 32,350 8.55 0.024 0.72
345 150 39,800 42.3 0.017 0.87
346 150 32,350 85.4 0.0076 1.0

 
115

might account for at least part of the difference in rupture life. This
suggestion is partly supported by results of other postirradiation creep
tests that are included in Fig. 4.16. Except at 40,000 psi stress, the
rupture lives of specimens irradiated cold sre generally lesg than those
of specimens irradiated hot. Although there are not many data for com-
parison, the rupture lives obtained by postirradiation tests of the cold-
irradisted specimens are shorter than those of the in-pile specimens.

The thermal history appears to be an lmportant factor in determining the
creep-rupture l1ife of irradiated Hastelloy N.

4. In both creep~-rupture and tensile tests, the properties of
vacuum-melted heats have shown a large dependence on the irradiation tem-
perature, whereas alr-melted heats were less sensitive to drradiation
temperature. This is 1llustrated by the data in Table 4.3. Heat 5065
wag alr melted, and 1ts creep properties show no consistent dependence
on irradiation temperature. Heats 2477 and 7304 were both vacuum melled,
and the rupture 1ife and ductility were much worse when the material was
irradiatad at 650°C than when irradiated at 150°C. Thig effect is
presently uvnexplained.

5. 'The effect of pretest heat treatment on the tensile properties
is shown in Table 4.4 for two heats of Hastelloy N. The range of heat
treatments used on heat 5065 had no detectable effect on the ductility.
A wider range of heat treatments was used for heat 65-552. 1t 1is apparent
that the duetility improved as the annealing tempersture was increasad.
The grain slze increased considerably, and this would be expected to in-
crease the ilrradiation effect. However, some other effect, possibly re-~
moval of boron [rom the graln toundaries by the high-temperature anneal,
geems Lo have been more important. '

6. The effects of postirradiation ammealing on the tensile proper-
ties of Hastelloy N are shown by the data in Table 4.5. Amnealing re-
duces the ductility of the air-melted heat 5065 by sboul a factor of 2.
The ductility of the vacuum melt, heat 65-552, 1s virtually unchanged oy
annealing up to above 871°C. Ann=aling at 1200°C increases the ductility
by about a factor of 3. '

Weld Studies on Hastelloy N

 

The weld studies on Hastelloy N have two general objectives: (1)
to improve the weldability so that welds have grealter strength and duc-
tility, and (2) to study the effects of irradiation on the properties of
welds and investigate ways of improving thelr resistance to irradiation
damage. It was reported previouslylo that welds involving standard air-
melted heats of Hastelloy N exhibited lower rupture life and ductility at
650°C, as compared with the base metals. However, the properties could
bve recovered by postweld annealing. . It was also shown that welds in-
volving vacuum-melted naterials exhibited better properties. Hlectron
microprobe studies have since shown that significant silicon segregation
occurs in the weld metal. These observatlions have lead to the proposal
Table 4.4, Effect of Pretest Annealing on the Tensile Properties of Hastelloy N

(650°C, 0.C02 in. in.”

 

 

 

Average Yield Total Reduction
Experinment Armesal Grain Size Strength Elorngstion Eilongation
(inr2) (psi) (%)
Feat 5065
a None g.05 40,000 A 11.56 !
O None .05 40,800 2 13.1 1.9
b 8 hr at 871°C C.05 38,1C0 .7 14.3 7.3
Heat 65-552
& None G.025 47 ,80C 5.1 5.3 .2
b None 0.025 47,500 4.7 4.9 .1
b g hr at 871°C .05 41,200 6.6 0.7 O
D I ohr st 1177°C 0.10 37,600 14.4 14.5 ol
a 1 hr gt 1177°C 0.1C 48,400 15.5 5.8 H 17
a I hr at 1316°C 0.25 46,200 21.8 23.3 .0
&RTR, 65°C, 5 x 10%° nvt thermal.
b

CRR, 43°C, 8.5 x 10%% nvt thermal.

o1l
117

Table 4.5. Effect of Postirradiation® Annealing on the Tensile
Properties of Hastelloy N

(650°C, 0.002 in. in.~" min™t)

 

 

 

Postirradiation Yield Uniform Total Reduction
DbAfiheal Strength Flongation Blongation in Area
' (psi) (%) (%) (%)
Heat 5065
Control 46,300 22.8 24.0 28.1
None 40,800 12.2 13.1 21.9
400°C for 1 hr 41,900 11.0 11.5 16.3
650°C for 1 hr 42,100 10.9 12.4 15.5
871°C for 1 hr 43,900 6.6 6.7 10.7
1200°C for 1 hr 37,200 7.0 7.5 12.5
Heat 65-552
Control 41,600 22.0 22.4 23.5
NOHE 4!‘/)500 ‘j"r 7 4‘-9 '?Dlo
400°C for 1 hr 51,100 4.1 4t 13.6
650°C for 1 hr 46,100 5.6 5.8 10.2
871°C for 1 br 43,000 3.3 3.7 8.82
1200°C for 1 hr 38,800 12.7 12.8 18.0
%Irradiation temperature = 43°C; @, = 2.5 x 10°° nvt.

th

that silicon may be responsible for the poor weldability of Hastelloy N,
and studies have been directed toward establishing this point more firmly.
irradiation damage studieg have involved determining how irradiation
changes the tensile properties. Based on the premise that a large grain
size in the weld metal may cause degradation of the properties under ir-
radiation, attempts have been made To reduce the grain size by adding
grain-refining impurities to the filler metal. The work in both of these
areas is presented briefly.

Several experimental welds were made to study the effect of silicon
on the properties of the weld metal. A single J-groove joint configu-
ration with an included angle of 50° was selected.

A second group of experimental welds was made in an effort to refine
the weld-metal gralin size and/or provide innocuous precipitates upon which
the helium (from the n,@'reaction) molecules can assimilate. Experi-
mentally, this has been done by spray coating an alir-melted Hastelloy N
filler metal (heat 35055) with A1503. The weld made from this filler
metal was designated No. 7. For comparative purposes, weld No. 4 was
made with unsprayed, air-melted filler metal from heat 5101; weld No. 6
118

was made with a vacuum-melted filler metal (heat 65-552). The compositions
of the jndividual Tiller metals are given in Table 4.6. The compositions
of the base metal and the deposited weld metal for welds 6 and 7 are also
given in Table 4.6. As can be noted in Table 4.6, most of the Al,0; (weld
metal No. 7) was lost as a slag during welding.

The creep-rupture properties of welds 4, 6, and 7 are compared in
Fig. 4.17 with those of the basc metal and weld No. 1 (duplicate of weld
No. 4). 'The properties of weld No. 4 were quite close to those previously
observed for weld No. 1. Using weld No. 4 as a base line, weld No. & was
slightly weaker, and weld No. 7 was slightly stronger. The rupture
elongations at a stress of 40,000 psi were 3, 6, and 3% for weld Nos. 4,
6, and 7 respectively. All failures were observed to occur in elther the
weld metal or at the fusion line. Metallographic studies are in progress
to determine the exact location of the failures.

Several of the specimens were heat treated after welding for & hr
at 871°C. As chown in Fig. 4.17, this treatment produced a significant
improvement in the creep strength; the ductility was also improved. A
comparison is made in Table 4.7 of the degree of improvement caused in
the different welds by postweld heat treating. The heat treatment made
the rupture lives of the welds more nearly the same and equal to or better
than that of the original base metal.

ORNL-DWG 66-4779

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

70 : — e P
| N | T T ] ‘
N AS STRESS
WELDED RELIEVED
50 1 oAl e . WELD 4
‘\\\ a A WELD &
N w ; [l u WELD 7
\\ ! \\ 550°C
50 i ™ R ™ | L |
\\\ \\ ! i ‘
2 ‘ ™
A0 e 4 A Lo
o ! i ? \
o N
e N
= \\
& N
w30 - e o
= N
w
20 ol i I 4o
N
M)
10
| o
O L J ) . ! e 1 |
1 10 100 1000 10,000
TIME (hr)

Fig. 4.17. Comparison of the Creep-Rupture Properties of Welds and
Bage Metal.
Table 4.6,

Chemical Analysis of Welds

(in wt %)

 

 

_ Al,05-
Heat 5065 fleat 5101 Heat 65-552  Heat 095 oated  Weld Metal  Weld Mebal
A - Weld Wire Weld Wire Weld Wire . .
Element Base=-Metal or - - T - Wire for from from
. on Weld on Weld on Weld tr - a . , ey e
Welds 4, 6, 7 o Weld Weld No. 6  Weld No. 7
' No. 4 To. 6 No. 7 .
No. 7
Chromium 7.26 £.92 6.89 7 .36 .93 767
Tungsten 9.04 Q.05 C.006 0.03 0.01 0.08
Iron 3.91 3.91 4 .06 3.76 4u3 o7
Carbon 0.06 0.05 0.04 0.0% G.038 0.025
Silicon 0.60 0.63 0.14 0.61 0.16 0.52
Manganese 0.55 C.44 0.45 .69 0.48 0.53
Molybdenum 16 .47 16.39 1€.15 16.20 16.21 16.39
Vanadium 0.21 0.34 =5 x 10™% C.21 0.1 0.1
Phosphorus 0.004 0.001 0.002 0.006
Sulfur 0.007 0.00% 0.006 0.008
Alumd o 0.01 } 0.05 0.25 0.06 0.84 0.28 0.08
Titanium 0.01 v 0.02
Boron 0.0030-0.0040 0.00035 0.0C5
Nitrogen 0.000s 0.0081
Oxygen 0.0013 C.11 0.0042

 

6TL
Table 4.7.

120

Propertles of Hastelloy N

(650°¢, 40,000 psi)

Effect of Stress-Relieving on the Creep-Rupture

 

 

 

 

 

 
   

 

 

 

 

 

 

 

 

 

Weld As~Welded Stress-Relieved Ratio,
Nou Rupture Tife Rupture Tife Stress-Relleved
’ (hr) (hr) to As-Welded
25 300 to 550 12 to 22
& 15 420 29
"/ 48 370 7.8
ORNL-DWG 66-1877
80 - S — ; : e e
| | | IRRADIATED CONTROL
' | = . o BASE METAL
201 ! n_%w, - s & WELD SPECIMENS .
i ‘ | HEAT 5085
( ! } ¢ = 0.05 in./in.- min
60 | M;ifi e Ly, = 8.5x10%0 avt .
T ; IRRADIATION TEMPERATURE: 43 °C
3 | f ’
ol — | [m__m_{ -
g | |
2 |
g 40 — - —n ‘ Cmees o .
C.Zfi: ./Jrflw—g%m T
pr |
- i
230 -
Q ;
b
20 &= _;E_mmfmfim T
0 ?fim — -
ol |

Fig. 4.18.

 

 

 

400 500
TEST TEMPERATURE (°C)

300 700

Properties of Welds and Base Metal.

Comparison of the Effects of Irradiation on the Tensile
121

Since only a few of the irradiated specimens from welds 4, 6, and 7
have been tested, it is necessary to speak in generalities concerning the
properties of these welds after irradiation. Speclmens of all three welds
were irradiated in the ORR at 43°C to a thermal neutron dose of 8.5 x 1029
nvt. Figure 4.18 compares the tensile elongation of the base metal and
weld No. 4 specimen before and after irradiation. In general, the base
metal is affected the most. The welds initially have lower ductility,
but the ductility is not reduced much by irradiation. At a test tempera-
ture of 871°C, the weld and the base metal have the same ductility. At a
strain rate of 0.002 in. in.”™  min™' and a test temperature of 650°C, the
total elongations of welds 4, 6, and 7 were 5.2, 7.0, and 6.4% respec-
tively. Thus, it seems that the doped weld wire had little, if any, el-
fect on the rupture ductility. Although postweld heat treatment improved
the unirradiated ductility, it was found that the ductilities of the as-
welded and postweld heat-treated materials were comparable after irradi-
ation. All the irradisted specimens Tailed in the weld metal. These
same welds will be evaluated further through postirradiation creep test-
ing.

References

1. G. M. Adamson, Jr., et al., Interim Report on Corrosion by Zirconlum
Base Fluorides, ORNL-2338 (Jan. 3, 1961).

 

 

2. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp. 81—
87.

 

3. TIbid., pp. 87%-88.

4.  M3R Program Semiann. Progr. Rept. Feb. 28, 1965, ORNI-3312, p. 85.

 

5. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNI~3872, p. 90.

 

6. MSR Program Semiann. Progr. Rept. July 31, 1964, OENL-3708, pp.
281-86.

 

7. E. J. Manthos, Metals and Ceramics Div. Ann. Progr. Rept. June 30,
1965, ORNL-3870, pp. 251-55.

 

. J. H. W. Simmons and A. J. Perks, Nature 206, 610 (1965).

9. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNIL-3872, p. 93.

 

10. Ibid., P Q4.
122

5. CHEMISTRY

Chemistry of the MSRE

 

Analyses of the Flush, Fuel, and Coolant Salts

 

It is important in molten-salt reactor operations to be able o
agcertain quickly and accurately that the uraniwn concentration in the
molten fuel solution conforms to design specifications and that chemical
purity of the salt is maintained.

Significant advances in analytical methods have been incorporated
into routine practice since the zero-power experiment was completed in
July 1265. The increased accuracy which is apparent in chemical analyses
shows more clearly than ever that composition and purity of the reactor
salts can be ascertalned accurately and economically on a routine basis
and that current methods now afford excellent monitoring practices to
molten-salt technology.

Application of three innovations to the analysis of MSRE salts
has been made within the current report period: a new end point for
uraniuvm titrations, a new method for structural-metal ions, and a new
method for oxide analyses. Together, they have provided increased
assurance that the compogition ¢f the fuel conforms o the inventory
values and reactivity balances and thal the chemical purity of Lhe
salt is even grealer than could be affirmed from previous results.

Uranium Assay. Controlled-potential coulometry is used as the
primary method for determination of uranium in MSRE fuel because of
the high order of precision inherent in the method.® From 50 to 100
Hg of uranium can be titrated with a precision of the order of less
than 1%.

Evidence was noted during the precritical and zerc-power experi-
ments that a distinct bias between nominal and analytical values Tfor
uranium existed and that the analytical values for uranium were ap-
proximately 1% lower than nominal. We inferred this phenomenon to be
a conseguence of dilution of the fuel salt by flush salt, although the
amount of dilution required to rationalize the bias was of the order of
140 1b, somewhat higher than was compatible with the weigh-cell data.

A recent refinement in the method of determining the uranium end
point -in the controlled-potential coulometric analysis has been made.
The end point had previously been estimated as that point at which the
electrode current is reduced to 5 pa. The end-point values arec now
computed by extrapolation of the reduction-potential curves to zero from
a 50-ua cutoff point. 1In recent control analyses in which synthetic
standards were used, the 5-Ha end point was demonstrated to have a
negative bias of 0.8%, whereas the bias for the 50-pa extrapolation end
point was zZero.

2

Book wvalue for the uranium concentration of the fuel salt which
ig currently circulated in the MSRKE 1s 4.646 wt %. The mean analytical
value ig 4.642 * 0.028 wt %. Thus the improvement in the method of
123

computing the uranium end point has reduced the bhiag between inventory
and analytical values to much less than one gtandard deviation of the
analytical method.

Uranium Lsotopic Analysis of the MSRE Fuel. The composition of
the circulated and static fuel salts in the MSRE became compositionally
homogeneous for the first time after completion of the zero-power ex-
seriment.  The 22°U enrichment fraction of the =alt cirewdating at zero
power was found experimentally to be 33.7L wt %, as compared with the
book value of 33.5 wh %. After circulation, the salt was drained into
the storage tanks, blending there with other fuel of lower 235UF4 COm~
centration. No significant burnup has, as yet, occurred in the Tuel.
We have had, therefore, an opportunity at the beginning of the full-
power experiment to establish an analytical base line of enrichment.
The experimental results show that the fuel now contains 2220 at 33,241
wt %, in good agreement with a calculated value of 33.2 wt %,

 

otructural -Metal Tmpurities. We have inferred that nickel is
present in the MESRE fuel salt only as a metallic phase and that iron
is also present predominantly, if not entirely, in metallic form.?
In an attempt to gain evidence to support this inference, samples of
the static fuel sgalt were ovtalned from the Tuel storage tanks during
the two-weck period following the zero-power experiment. The fact
that the concentrations of the structural-metal impurities were not
found to change during this storage period led us to conclude that
thermal convection in the storage tank was adequate to prevent settling
of Fine metallic particles of iron and nickel in the drain tank.

 

In recent experiments a sample of the MSRE fuel was removed from
the pump bowl before significant power was generated by the reactor;
this sample was analyzed by a newly developed method using controlled-
potential voltammetry.4 The salt was found to contain approximately
g to 10 ppm of iron in the divalent state. The concentration of
nickelous ion was found to be lower than is debectable by this method,
that is, less than 1 ppm. The net concentrations of iron and nickel
in the salt specimens were 131 and 40 ppm respecltively. The salt
also conbains 48 ppm of chromium in ionic form. These results are
particularly useful in that they indicate that the total concentration
of lonic structural-metal conltaminants in the fuel is not greater than
58 ppm.

Oxide Analysis. Most serious of the adverse consequences of
chemical impurities in MSERE are those which would arise from contamina-
tion by oxides. Of great significance is that oxide content of the
MSRE salts can now be determined routinely and accurately al concen-
trations lower than 100 ppm by purging samples of the molten sgalt with
Hy and HF, Oxide concentration is then calculated from the amount of
water recovered from the effluent gas stream.” This hydrofluocrination
method for determining the concentration of oxides in molten fluorides
was applied to analysis of the MSRE flush, fuel, and coolant salts for
the first time at the beglioning of the full-power experiment. AL that
point the radioactivity of the salts was negligibly low, and analyses
could be performed in the development laboratory. As activity wag
generated in the fuel salt, it became imperative thal subsequent chemical
124

analyses be conducted in the High-Radiation-Level Analytical Facility
(HRLATF). Currently, apparatus which is to be used for routine analysis
of the oxide concentration of MSRE salts is being installed and tested
in this facility.

The resultes of the analyses performed to date in the development
laboratory reveal that the oxide concentration in the MORE flush, fuel,
and coolant salts is 75, 95, and 40 ppm respectively. On the basis of
the results obtained by Baes and co-workers,® saturation of the fuel
by oxide is considered to occur at approximately 500 ppm al the reactor
operating temperature of 1200°F, The current analytical data indicate,
therefore, that the MSRE salts, which have remained in the molten state
for some ten months, have been well protected from moisture contamina-
tion during this period.

Salt Composition Analyses. The sampling schedule which wag followed
at the beginning of the MSRE full-power experiment afforded the first
cpportunity to obtein statistically meaningful analytical data. The
resulte of all chemical snalyses of the MSRE salts which have been per-
formed in connection with the Tull-power experiments are given in
Table 5.1. ‘The composition of the fuel salt, as indicated by these
data, is compared in Table 5.2 with the results of previous analyses.
It is apparent from these data that, but for the still unexplained
minor bias between nominal and analytical values for lithium and
beryllium, the agreement between inventory records and analytical re-
sults 1s excellent.

 

Examination of Materials from the MOSRE Off-Gas System

 

The program for bringing the MSEE to full-power operation has
been interrupted on two recent occasions by flow restrictions in the
off-gas system. Symptoms of plugging became pronounced in eacn case
after about 12 hr of operation aft 1 Mw., Small specimens of the materials
which were suspected to have causced plugging were removed from the
affected parts of the off-gas system and subjected to various tests
in an attempt to determine the reasons for flow restrictions. Tt was
assumed that chemical ijdentification would indicate the cause ol the
restricitions.,

Restrictions were first noted in the capillary restrictor in line
521; at the same time the 533 check valve became inoperative. Restric-
tion was also noted in the sintered stainless steel line filter and
522 wvalve assemblage. In order {0 resume operatlon of the reactor, the
capillary restrictor and 533 check valve were replaced by nonrestric-
tive secltions, and the filter-valve assembly was replaced. The filter
element which was removed was capable of blocking passage of more than
90% of the particulate matter of 0.7 p in diameter. It was replaced
by a filter designed to block particulates of >50 p in diameter. On
resuming operation at 1 Mw, restrictlions appeared to develop in three
locations: at wvalve 522B, at the entries to charcocal beds 1A and 1B,
and in the lines ahead of the auxiliary charcoal bed. Currently, &
specimen has been obtained only from valve 621, which controls flow
into charcoal bed 1B. Examinations and tests are still being performed
Table 5.1, Summary of MSRE Flush and Fuel Salt Analyses, Full-Power Experiment

 

Concentration {wt %)

 

Concentration (ppm)

 

 

Sample Circulation

Number Period (ar) Li Be Zr v F > Fe Cr Ni Mo o” 0°
Fiush Salt

FP4-1 13 35

FP4s2 19 13.65 9.83 <0.0025 0.0210 80.52 104,02 110 <10 33 <15 56

FP4e3 25 46

FP4ed 27 13.55 9,35 <0.0025 0,0207 79,34 102.26 212 62 30 <15 74

FP4e5 29 72

P46 a2 13.50 9.96 <0.0025 0.0200 80.07 103.55 125 <10 <20 <15 180

FP4.7 40 13.65 9,46 <0.0025 0.0241 77.85 100.98 180 54 <20 <15 150

FP4-8 48 106

FP4.9 50 13.35 9.98 <0,0025 0.0221 75,80 99.15 210 " 57 <20 <15 142

FP4-10 64 13.55 9.49 <0,0025 0.0230 75,05 98,11 128 60 <20 <15 1300¢

Fue! Suf'r

FP4-11 3 120

FP4-12 19 10,45 6.33 10,52 4.673 66.45 98.42 96 56 41 <15 144

FP4.13 24 105

FP4-14 27 10.20 6.41 10.77 4,634 64.67 96.68 79 41 69 <15 92

FP4-15 31 80

FP4.16° 48

FP4-17 53 65

FP4-18 64 10.25 6.68 11.24 4.651 67.44 100,26 144 a3 44 <15 85

FP4.19! 68

FP4.20° 75

FP4.21 99 10,47 6.40 10,82 4,671 66.79 99,15 121 60 52 <15

FP4.22 124 10,54 6.54 10.95 4,664 64,68 97,37 116 44 84 <15

FP4.23 148 10.27 5.74 10.85 4,642 65.06 97.57 99 46 35 <15

FP4-24 172 10.65 6,55 10.96 4,655 67,66 100,58 116 48 45 <15

FP4-25 159 10.60 6.37 11.41 4.646 65.44 98.47 26 35 42 <15

FP4-26 217 10.60 5,63 11.20 4,642 66,90 99,07 89 48 7 <15

FP4.27 245 10.55 6.53 11.07 4.618 67.68 100,45 227 50 41 <15

FP4-28 268 10.60 6.42 11.10 4.663 66,19 98,97 211 49 34 <15

FP4.29 314 10,70 6,71 11.54 4,654 69,75 103,35 111 39 31 <15

FP4-30 362 10,63 6,81 11.19 4,661 57.32 100.58 83 49 27

FP4.31 435 10,30 6.63 11.26 4,632 68,51 101,33 150 37 41

GCt
Table 3.1

{continued)

 

Concentration {(wi %)

Concentration {(ppm)

 

 

Sample Circulation
Number Period (hr) i Be Zr ye r 2, Fe Cr Ni Mo o° 0°
FP4.32 457 10,55 6.71 11,80 4,625 67.66 101,34 173 43 33
FP4-33 529 11,208 6.75 11.07 4,596 56.25 99,97 55 30 39
£P4-34 601 11.35% 6.4% 11,13 4,601 68.20 101.82 164 58 16
FP4-35 659 11.36% 6.68 10.86 4,721 69,35 103,93 74 54 <5
FP4-36 i1.25% 6.32 11,20 4.632 56,76 100.16 125 47 <5
FP4-37 11,302 5.33 11,08 4,622 69.35 102,68 189 53 80
FP4-38 10.65 6.54 11,46 4,608 67.25 100.56 311 51 25
FP4-39 10.60 6.33 11.54 4,519 68.33 101.42 78 51 40
Coolant Salt
CP4=1 7
CP4=2 il 83 25 <5 13¢
CP4-3 59 13.78 8,91 75.70 99,39 21 41 37 1i0 185
CP4-4 67 46 53 24 60
CP4-5 180 14,20 8.87 <0.002 76.80 99.87 50 50 20 60
CP4-6 194 33
CP5-1 236 13.86 B.B5 <0.002 76.7 99,41 50 35 <10 150
CP52 i3.82 B.55 <0.002 70.6 99,67 59 35 <10 110

 

8values corrected to 33.241 wt %

bHF-purge method.
CKBrF4 method.

dSample exposed to dry-box atmosphere for 48 hr,

e gL
No sample obiained.

f . .
For amperometric analysis,

35,
235y

gErrc;m~.=ously nigh; attributed to failing batteries in automatic pipette.

act
Table 5.2.

Summary of MSRE Fuel Composition Anslyses

 

 

Component Nominal Book Zevo-Power, s FP4-11 to FP4-39
Experiment Tank
(mole %)
TiF 65.00 64,88 62.31 64,40 63.36 = 0.567
BeF, 29.17 29,26 31.68 29,68 30.65 * 0,533
ZrF, 5,00 5.04 18 5.11 5.15 £ 0.116
237.003yp, © 0.83 0.82 825 0.803 C.825 * 0.011
(wt %)
Ti 10.95 10.93 10.25 10.94 10.51 £ 0.137
Be 6.32 6.34 .71 6,49 6.55 £ 0,161
77 10.97 11.06 11.11 11.32 11.14 £ 0.295
237,003 473 4 646 4,602 4611 4 642 + 0.028

 

“Based on four samples.

Based on four samples obtained on completion

"237.003y - 33,241 w4 235y,

of the zero-power experiment.

L1
128

with the available specimens, esults are sumarized in Table 5.3.

On the basis of the results shown 1in Table 5.3, we would conclude
tentatively that the restrictions in the off-gas line may be attributed
to varnish-like organic material. Gulfspin-35 is used as the lubrical-
ing oil for the rotary element. It is composed primarily of a mixture
of long-chain aliphatic linear and branched hydrocarbons. Recent
measurements show that its refractive index is 1.473, The refractive
index of the heaviest 10% volume fraction obtained on vacuum distillation
of the o0il was found to be approximately 1.50. The high refractive
index of the varnish-like materials removed Trom the off-gas system
suggests that radiation polymerization has produced the plugging phases.
This inference appears to be strenglthened further by the observation
that the varnish-like materials appeared to have limited or no solubility
at room temperature in xylene, petroleum ether, carbon tetrachloride,
or accltone.

Uranium-Bearing Crystals in lFrozen Fuel

 

X-ray diffraction studies of uranium-bearing salts from solidified
MSRE fuel and/or concentrate have yielded information of chemical
interest. A complete crystal structure analysis of "7IiF*6UF," has
shown this tetragonal substance actually to be LiUFs and Lo have a
basically different structure from the rhombohedral compounds which do
have the 7:6 ratio. Compounds of the latier stoichirmetry do not
appear in the present fuel mixtures.

In LiUFs5, each U*" ion is coordinated by nine F ionms in the shape
off a trigonal prism with each rectangular face bearing a pyramid. This
polyhedron is similar to those in UpFs5, but different from those of
g-coordinated U* in Ut¥ze It is not presently known whether this dif-
ference is significant or Just a coincidence of packing of F~ ions.

A determination of the structure of Li,UFg is partially completed:
the orthorhombic crystals have unit-cell dimensions a = 2.96 A, b = 9.88
A, ¢ = 5,99 A, and the space group is Pama or InaZy. Four formula
weights of Ti,UFg in a unilt cell correspond to a calculated density of
4.71 g/em?.  The positions of U%' ions have been determined, but the
other ions are yel to be located.

Physical Chemistry of Ituoride Melto

 

Vapor Pressure of Fluoride Melts

 

Apparatus wag constructed to obtain vapor pressures by the carrier-
gas method in order to determine (1) vapor composition in the LiF-BeFs
system (to complement the manomelric pressure data already obtained for
this system’), and (2) rare-carth vapor concentrations in equilibrium
with liquid mixtures of importance to the molien-salt reactor distilla-
tion procec (from these concentrations more accurate decontamination
factors for the rare-~earth fission products will be obtained).

The apparatus, shown schematically in Fig. 5.1, closely resembles
that used by Sense ¢t g&.,g for studies of flucride nmelts. The reliability
129

Table 5.3. Results of Examinations of Specimens Removed from the MSRE O#f-Gas Lines

 

 

Predominant
Specimen Description Refractive Activity at Isctopes Spectrochemical
and Qrigin Morphology Index Contact (from gamma Data
(t/hr) scan)
1, Deposit from exit orifice of Isotropic particles, 1.520 Li, 99 ug; Be,
capillary restrictor appearing as 124 lg; Zr,
partly coalesced 100 jig
amber globules
2. Scrapings from spool] piece  Same as 1 1.540
adjacent to capillaty
3. Deposit from check Isctropic, amber, Be, 0.95 ug; Li,
valve 533 varnishelike 2 pgy Zr, <0.5
particle, ™~ 50 x be
100 u
4, Scrapings from valve 522 Isotropic, amber, 1,540 2.5 No Zr, Nb,
poppet varnish-like Ce
malrix contain~
ing embedded
isotropic(?}
crystalline
material of
lower refractive
index
5. Scrapings from valve 522 Same as 4 1.544 to 1.5 8981’, 140}33,
seat 1.550 1407 4
6. Qil drops Ffrom 522 valve 1,500 No Zr, Nb,
body Ce
7. Bcrapings from stainless Isotropic, faintly 1.524 to 14DBa, 14°La,
steel filter element colored material, 1,526 msRu,
more nearly 13703;. no
scale-like than Ce, Zr, Nb
glassy in ape~
pearance
8. Metallic scrappings from Granular opaque
stainless steel filter particles; low
element index, transparent,
birefringent crystal-
line material
spalled off metal
on microscope
slide
9, Deposit from HV 621 valve Isotropic, faintly ~1.,526 132.p¢

stem

colored material,
varnish-like in ap-
pearance with peb-

bly surface

 
120

ORNL-DWG 6€5-13115

/—MOLECULAR SIEVE DRYING COLUMN

/ HEATED COPPER TURNINGS
—

PRESSURE GAGE (O—10in. Hz0)

 

P A THERMOCOUPLE NO 3-_
N|CK$L STAINLESS STEEL SLEEVE ‘
A

BO (5Yerin LONG), THERMOGOUPLE
(5in LONGH =~~~ J/ NO. 2i
, E,—f e g = e L*_%

ARGON TO DRYING

 

 

     
  
 

THERM(OCOUPLE NO1
L

  

 

U e el )
SALT TEMPERATURE) - MOLTEN 3 J , COLUMN; THEN
/ SALT - ’ ; ON TO WET
/ LAVITE )
, ‘ P , TEST METER
HIGH TEMPERATURE ALUMINA TUBE -~ NSULATOR /

(36-in. LONG) ——T ,,,,, /
MARSHALL FURNACE (16-in LONG)- !

i,

"~ CONDENSER TUBE
{CONDENSER HAS 0.03%-in, CPENING IN
END ABOVE CENTER OF SAMPLE BOAT)

Fig. 5.1. Transpiration Apparatus.

of the apparatus and the efficacy of the washing procedures for remov-
ing condensed vapor were tested with pure LiF. Satisfactory agreement
with the reliable transpiration data obtained by Sense’ was attained.

Three compositions of the Lib-BeF, system have thus far been
investigated. The vapor-pressure data are summarized in Table 5.4,
based on the assumption that the wvapor consists only of monomeric LiF
and BeFp. It should be noted that the apparent partial pressure of
LiF increesecs with decreasing concentration of LiF in the melt. This
behavior is consistent with the expectation that the vapor species is
predominantly a compound of LiF and Bels, such as TLisBeF, or LiBeF,,

On the basis of the observed vapor composition, it appears that a
liquid compogition of 88-12 mole 9% LiF-Bel's will provide a vapor composi-
tion of 67-33 mole % TiF-BeF,. Hence 88-12 should be the correct composi-
tion in the still pot for the MSR distillation process., This melt com-
position is currently being measured to confirm that the vapor has the
composition 67-33 mole % LiF-BelF,. The latter composition would serve
as MSBR fuel solvent (this solvent will probably not contain ZrF.).

Methods for Predicting Density, Specific Heat, and Thermal Conductivity
in Molten Fluorides

 

 

Densitx.lo Several years ago,t’ after the published data on density

of molten fluoride had been examined, it was oroposed that the simple
Tule of additivity of molar volumes might be very useful Tor estimating
densities of fluoride melts., Since that time, the results of all
additlional experimental investigations have been studied. 'The rule of
additivity of molar volumes appeared to hold quite well except for one
system; this was the NalF-UF, system™? where positive deviabions as
great as 6% were obgerved. Thus it appears that, although there may be
131

Table 5.4. Vapor-Pressure Constants, Assuning that the Vapor

Phase is Composed of LiF and BeF,

 

 

 

 

 

log plmm) = A — —2
T(°K)
Composition
(mole %) Temperature Range p LiF p BeFs

e (°c)

LiF  Bef, A B A B
20 10 8971052 10,370 13,330 10.431 13,890
85 15 8391036 10.43%7 13,330 9.483 12,270
75 25 895-1055 9.720 12,350 8.611 10,710

 

exceptions, the rule of additive molar volumes describes the experimental
data on molten fluorides quite well and remalns the simplest, most
accurate method for predicting densities of Tfluoride melts.

To improve the method of estimation, a revised set of empirical
molar volumes is given in Table 5.5. The origin of these values is
dizcussed in ref., 10. For estimating a densilty expression of the form

o =a +bt, (1)

first solve for densities at two temperatures by using the squation

Il
- (.M. )
iZ—;:‘L s (2)
pt T e e

1

\
v ()]
i=1

where Ni and. Mi are the mole fraction and gram-formula weight of com-
ponent 1, and Vi(t) is the molar volume of compornent i at temperature t.

Substitute molar volumes from Table 5.5 at the twd different temperatures
in order to obtain the two values of pt; since density is linear with
temperature, substitution of the two pairs of values of p, and t in Eq.
(1) provides the solution for the constants a and b.
cpecific Heat. Examination®® of the heat-capacity wmeasurements of

molten fluorides indicated that the heat capacity per gram-atom of melt
is approximately & cal/°C. Hence an expression similar to that of Dulong

t
132

Tahle 5.5. Empiricasl Molar Volumes of Fluorides

 

 

Molar Volume (cc/mole)

 

 

 

At 600°C AL 800°C
LiF 13.46 14.19
NoF 19.08 20.20
KF 28.1 30.0
RbF 33.9 36.1
Cal 40.2 43,1
BeFs 23,6 XA
Mg T 22,4 273, 3
CaFs, 27.5 28.3
ST, 30. 4 31.6
Bal, 35.8 37.3
A1T5 26.9 30,7
YT, 34.6 35.5
TaF, 37.7 38.7
CeT, 36. 3 37.6
Prl, 36.6 37.6
Smf'a 39,0 39.
7rF,, 47 50
ThI', L6.6 477
UF,, 45.5 46"

 

and. Petit may be used to estimate specific heats of fluoride melts. 'The
expression 1s

n
\ 1
g L (Nipi)
c = ral e 3 (3)
1
, (NiMi)
i=1

where ¢ is the specific heat in cal (°K) * g *, p, is the number of

. . i .
atoms in a molecule of component i, and Ni and M, are the mole fraction
and the gram-formula weight, respectively, of cofiponent 1.
133

Sample calculation:

For MSEE coolant, 66-34 mole % LiF-Def,,

) ()
), (W)

i

0.66(2) + 0.34(3) = 2.34

i

0.66(26) + 0.34(47) = 33.1 ,

L B(2.34) 4 s o)1 7L
¢ = —55== = 0.57 cal (°K) + g * .

Thermal Conductivity. A new semltheoreticsal expression based on

 

Bridgman's theoryl3 of energy transport in liguids has been developed;
the derivation of the method is given in ref. 10. The expression is
13
k = -, (4)
V2/3

 

where kX is the thermal conductivity, V i1s the molar volume, and f is
the velocity of sound in the melt; all three variables should be in

cgs units. To use Eq. (4), the molar volume and velocity of sound are
necessary. Molar volume 1s easlly estimated by the rule of additivity
as outlined above. The velocity of sound may be estimated from thermo-
dynamic quantities, using the expression

[(c_/c.) —1lc
W2 = v P (5)
aZrmd.I

 

where Cp and CV are the molar heat capacities at constant pressure and

volume, respectively, @ is the expansivity, T is the absolute ftempera-
ture, CP/M is the specific heat [and should be expressed in ergs (°K) 1

g 1], The ratio Cp/CV varies between 1.2 and 1.35 for most fuzed salts

(1.2 works quite well). The expanslvity is the negative of temperature
coefTicient of density, divided by the density; that is,

. ?.9)
a = .
,(aT p

By using the additivity rule, & is easily estimated.

o -

Oxide Sclubilities in MSRE Flush Salt, Fuel Salt, and Their Mixtures

Bgtimates of the oxide tolerance of MSRE flush-salt—fuel-salt
mixtures reported previouslyl4 wvere based upon transpiration meagsure-
ments of the following equilibria:

HaoO(g) + 2F (d) =207 (a) + 2ur(g) ,
. 2 271 . .
134

H,0(g) + BeFy(d) +=80(s) + 2HF(g) ,
= P2 [i .
Uy = PHF/PHzO 5 (7)

PH,0(g) + ZrF,(d) =%r0,(s) + 4EF(g)

Q, = PfiF/(P§20[2r4+]} : (8)

By combining the results for reasctions (6) and (7),

BeO(s) + 2F(4) = Bely{d) + 0% (a) ,
L L (9)

estimates were obtained for the dissolved-oxide concentration al BeO
saturation of 2LiF-BeF, (approximately the composition of MSRE flush
salt). For 2LiF-Bels melts containing more than 0.1 mole/kg of Zri, —
vherein ZrO, becomes the least-soluble oxide — reactions (6) and (8)
were compined in a similar way to obtalin estimates of the concentration
oif dissolved oxide at Zr0, saturation:

$ir02 (5) + 26 (a) w=2ur¥,(a) + 02 (a) ,
/2 1/2 _ 4t/ 2rn2
U0, = Qn/Qz' " = [zx ¥ 2[0% ] . (10)

However, these estimates were of limited accuracy, owing to difficulties
in the determination of QO by the transpiration method. In particular,
this quotient was fTound too large ©o be measured with useful accuracy in
the MSRE fuel composition, and hence it was necessary to estimate the
oxide tolerance of the fuel by extrapolation from results for fuel salt
diluted by flush salwv.

During the past year, a more direct method for determining these
oxide solubilities has been adopted. A measured volume of salt, which
previously had been saturated by equilibration with excess solid Zr0O;
or BeO, was passed upward through a sintered nickel filter into a heated
reaction vessel (Fig. 5.2). There it was sparged with an H,-HF mixture
to remove the dissclved oxide as H,0. The effluent HF-U,0-H, mixture was
passed through a sodium fluoride column maintained at 90°C to remove the
unreacted HF, and then 1t was passed through a Karl Fischer titrstion
assembly where the water content was determined.

Typlcally, 125-g samples of salt were fillered into the upper
vessel and sparged with a gas flow of 150 cmB/min at a temperature of
500°C. The HF partial pressure was 0.07 atm or less. The time requirecd
to remove egsentially all of the oxide increased as the ZrF, content
of the mixtures was increased, hut generally did not exceed 4 hr. The
blank was usually equivalent to less than 0.002 mole per kg of oxide in
the sample,
135

The results obtained by this method for simulated flush-salt—fuel-
salt mixtures (Fig. 5.3) are in reasonable agreement with previous
estimates; % that is, as ZrF, (present in MSRE fuel) is added o 2LiF-
BeF, (flush salt), the oxide tolerance at first falls, reaches a mini-
mum, and then rises. This behavior reflects a monotonic increase in
the solubility product QZrOg with ZrF,; concentration (the dashed lines

shown in Fig. 5.3 Indicate the behavior which would be found if QZr02

remained constant). The temperature dependence of the oxide tolerance
is indicated in Fig. 5.4 for three compositions: (1) the flush salt,
(2) the approximate composition of minimum oxide tolerance, and (3)
the Tuel salt.

When the present directly measured values of QZr02 are comblned
with previocusly determined* values of QZ in the expression

QO - (QZrOQQZ)l/2 ?

improved estimstes are obtalned for the equilibrium quotient Qg (Table

5.6), corresponding to reaction (6). This quotient should be useful in
predicting the efficiency of oxide removal from MSRE fuel salt by HF
sparging during fuel reprocessingl” and during oxide analysis,d®

The solubility results obtained thus far for BeO in 2LiF-BeF, (Fig.
.4) are somevhat above the previous values obtained by combination of
B and Q0,14 but these new results showed considerable scatter, This

5
Q

ORNL-0OWG €6-4780

PROBE FOR
DEPTH MEASUREMENTS .

   
 
 

Hp FROM
CONSTANT-FLOow e Hy —
CONTROL

 

MaF TRAP

   
 
 
 
 

 

 

   

‘ ~WET-TEST

HE-Hp
‘ METER

MIXING

   

~~REACTION
VESSEL

  
 
 
 

 

 

o

   
 
  

—~ 00015 -in, NICKEL
FILTER BUBBLE-O-

METER

FURNAGE -—»

 

NaOH
SOLUTION

 

KARL FISCHER
ANHYDROUS HF SOLUTION
IN CONSTANT —

TEMPERATURE BATH

 

 

 

EQUILIBRATION
VESSEL

Fig. 5.2. Apparatus for Determination of Solubilities by Hydro-
Tluorination.
136

ORNL-DWG 66-4781

wt % FUEL SALT IN FLUSH SALT
e > 19 20 50 100
P T TR T

  

oo BeO SATURATION « + — & o

700 °C

  

002 Lo

0.04

0.005

 

OXIDE CONCENTRATION (moies /kg)
OXI{DE CONCENTRATION {ppm)

0002 |~ -

 

0.001 ‘ - ; _
0005 Q.01 002 005 Q4 0.2 0.5 1 2

ZrF, CONCENTRATION (moles/kg)

Fig. 5.3. Oxide Concentration Required to Produce Precipitation
of BeC or ZrO, from Simulated MSRE Flush-Sali—~Fuel-Salt Mixtures.

is thought to be caused by the presence of finely divided BeO which was
not Tiltered completely from all the samples. These measurements are
being repeated with the use of longer equilibrium times, in The hope

that this will improve the filterability of the BeC. The oxide tolerance
of 2LiF-BeF, (flush salt), indicated by the lines in Figs. 5.3 and 5.4,
represents our best estimate at present.

Separations in Molten Fluorides
vaporative-Distillation Studics on Molten-Salt Fuel Components

Vacuum distillation separation of molten-calt fuel or fuel components
from the rarc-ecarth fission products is an attractive method of decreas-
ing neutron losses by capture. To design process equipment for this
task, both the mass rate of distillation and the relative volatility of
the rare earths must be known for the particular salt system used.
Completed experiments concern the process demonstration planned on MSKEE
fuel, but are directly applicable to any proposed thermal MSBR fuel.

For MSRE fuel, removal of the uranium by fluorination is proposed.
The fuel solvent and remaining fission products would then be fed at
the distillation rate to a2 vacuum still charged with LiF-Bel,-ZrF, at
that composition which will yield the fuel solvent as distilled product;
the rare-carth fission products would concentrate in the still. The
regidue would be discarded or procegsed, when necessitated by heat from
the Tission products or concentration of rare earths in the product.
137

ORNL~DWG 66-47872
TEMPERATURE (°C)

800 530 500

    

 

0.05

 

 

 

0.02 t-

 

  

FLUSH SALT

 

 

 

 

 

 

0.005 S— - - B —
15wt%FU;?Z;}\\ """

IN FLUSH SALT

100

 

 

 

 

OXIDE CONCENTRATION {moles /kg)
OXIDE CONCENTRATION {ppm)

 

0.002 |

 

 

 

 

 

 

 

0.001 -
1.0 14 1.2 1.3

1OOO/T )

Fig. 5.4. Temperature Dependence of Oxide Tolerance: (1) in Flush
Salt (BeO-Saturated), (2) in Flush-Salt—Fuel-Salt Mixture of Minimum Oxide
Tolerance, and (3) in Fuel Sals.

Table 5.6. Summary of Equilibrium Quotients for
Reactions (6), (&), and (10) in MSRE Fuel Salt

 

 

. 2=
Tenperature [0<7] QZTOQ Q, QO
(°c) (moles;kg)  (moles3/kg?) (atm® kg mole™t) (atm mole kg—+)
500 0.012 1.8 % 10 % 7.1 x 10 ? 3.6 x 10 2
600 0.024 7.3 x 10 % 3.95 x 10 * 1.7 % 10 2

700 0.041 2.1 % 10 3 150 5.6 x 10 2

 
138

A 10-ml graphite cylinder, containing ~17 g of salt with a free
surface (when molten) of 1 om®, was used for the still pot. This
cylinder fitted into a "[I' shaped Hastelloy N tube heated electrically
to a given temperature as measured by four thermocouples in the graphite
cylinder. When the desired temperature was reached, distillation was
initiated by evacuating the assembly. Vaporized sall then passed up
and over the top of the "O" with a small portion (attributed to thermal
reflux) collecting at the base of the graphite cylinder. The product
salt condensed in a cooler (~450°C) collecting cup in the opposite leg.
Distillation was stopped by helium addition after preselected time
periods.

Mass-rate data for ‘LiF weve determined first, and then MSEE solvent
was added to the 'LiF and distilled in aliquot portions from it. Bach
repetition brought the BeFp and ZrF, concentrations in the pot closer
to those which would yield fuel solvent (LiF-BeFs-ZrF,, 65-30-5 mole %)
as product. After the equilibrium concentration was approached, 2200
ppm of neodymivm (as NAFs) was added to the still bottom, and several
distillate samples were taken. Then the neodymium concentration was
raised to 22,000 ppm, and the sequence was repeated. For both "TiF and
the nominal eguilibrium composition, the effect of temperature on mass
rate was determined.

The data of Fig. 5.5 show the effect of temperature and mass rate;
single experiments show considerable scatter, so inclusive bands are
chown. The slope of the bands is consistent with the heat of vaporiza-
tion of the components and strongly suggests that ebullition does not
occur. Solvent surface heat flux ig <1/100 of the available heat to
the graphite cylinder; it is doubtful that the distillation is heat
limited. The fact that the observed rate for ‘LiF is only ~10% of

-2 ORNL-DWG GG-966

(g/cmzv sec)

._,_
-
=

DISTILLATION RAT

s

100°C 1000=C

 

72 73 74 715 76 7.7 7.8 7.9

fig. 5.5. Observed Distillation Rate vs lTemperature.
139

Theoretical is unexplained, but it has been observed in many distilla-
tions using several experimental configurations.

The solvent for any proposed MSBR should distill at rates above
that shown for ‘LiF with the posslble exception of Th¥F,-bearing systems.,
The effect of ThF, is being studied.

Bffectivenesse of separation from rare earths was determined by
activation danalysis for neodymlwia in the product, the still bottom, and
the refluxed salt deposited around the base of the graphite.

The data are shown in Table 5.7.

The equilibrium composition in the still pol undoubtedly changes
with temperature due to the change in activity coefficients of the
components. The compobjtiom found at nominal equilibrium after several
solvent additions of 5 to 7% by weight and afier distillations at 1030°C
wag LiF-BeF,-ZrF, (85,4-10.7 3.9) in the presence of 22,000 ppm of
neodymium as {luoride.

mifective Activity Coefficients by Evaporative Distillation

 

Evaporation into a vacuum from a quiescent wmolten-salt mixture of
low vapor pressure should not permit vapor-liquid equilibrium at the
surtace of the melt; transport from the surface should be controlled
by the evaporation rates of the individual components, The amount
vaporized is a function of The equilibrium vapor pressure, molecular

mass, temperature, and surface area; according to Langmuir,l7
_ grams of 7 T
W, = 5—-—-~—-—-2 - KP, \fMZ/'I‘ , (11)
cme -gec

where P, 15 vapor pressure of component 4, M is molecular weight of
component Z, T is absolute temperature, and K is a constant dependent
on units.

In a multicomponent system at equilibrium,

P, = 7ZNZP7 s (12)
W, o= w = Ky _N_ Q/ /T (13)
Z cme -sec Z 7

IT we use a fixed mechanical configuration and, for simplicity,
define the activity coefficient of the major component (in this cas
LiF) as unity, the activity coefficient For anmy other constituent mdy
be determined by using the ratio

W Ky, M, ) NM, /T

1
gy KON, P7 MLlr*/ T

 

 
140

 

 

 

 

Table 5.7. Councentration of Neodymium in Fractions from
Vacuum Distillations
- e
Nd Concentration
Fraction (ppm)
2200 ppm Added 22,000 ppm Added
Still bottom 2570 22,000
Product 21 6OOb
ReTlux 133 34

 

aMeaa values from several determinations; data show little scatter
exceplt for reflux specimens.

bThese product samples alsc showed Ce and La, which undoubtedly
represent contamination during grinding and handling of specimens; Nd
analysis, therefore, may well be toc high,

altered to

 

N_o.. W e —
y _ L1k . 7 . LiF . JI\JLLD /M . (15)
2y W pC 1z
7 LiF z

Proceeding with this technique, the data shown in Fig. 5.6 have
been taken from LiF-BePFs-ZrF, melts. TFor comparison purposes, selected
data from other sources are included,t820

Although internal consistency exists, the values are dependent on
vapor-pressure data for the pure components. Calculations were made
using the values (PO) for the pure components shown in Table 5.8, The
leagt-precise measurements are from MSRE-colvent distillations; thesc
are included to substantiate the surmise that when ZrF, is included in
the melt, it effectively removes LiF from the solvent. The 1000°C
LiF-Bel’; line has been extended to 100 mole % IiF, since the initial
LiF-BeFo-ZrF, experiment contained <0.35 mole % ZrF,, a guantity so
small that the system can be assumed to be Li¥-BeFs. On the other hand,
as 7ZrF, builds up in the mixed system (to approximately 3 mole %), en-
hancement of The BeF,; activity dis noted; this is consistent withh re-
sults from MSERE solvent (LiF-BeFs-ZrF,, 65-30-5 mole % initially).

The temperatures reported are those of the bulk melt, and 1t is rec-
cgnized that the surface temperature may be significantly lower,

causing an indeterminate {(for the present) error. Both surface-tempersa-
ture effects and the pure LiF-ZrF, system are being studicd. It is
interesting to note that the assumption of unit activity coefficient
for TiF does not lead to incompatibility with the data of others.
141

Txtraction of Rare Farths from Molten Fluorides into Mcolten Metals

 

Studies of the removel of rare earths from a molten-fluoride
solvent and their dissolution in molten metals have been oriented
toward the development of a liquid-liguid extraction process for re-
moving rare-earth fission products from molten-salt reactor fuels.
Fluoride mixtures obtained by dissolving a selected rare earth iato
LiF-BelFy (66-34 mole %) have been usged to simulate the fuel solvent
of the reference-degign MSBR. In fuel reprocessing schemes proposed
for the reactor, uranium will be removed prior to the removal of non-

O ORNL-DWG 66-267

Y COEFFICIENT

i

ACTIVIT

1000°C
e LiF - BeF,, BAES, 700°C 200%¢
Lik- ZrF,, SENSE ef af.
—rr—2 Lif - Bel, + Zr¥, , BAES
— = =~DERIVED FROM MSRE SOLVENT
DISTILLATIONS 900 ~1050°C
LiF - Bef, , BUCHLER, 600°C
LLiF - BeF,, 1000°C
LiF«8eF,, 900°C
LiF - BeF, » ZrFy , 1000°C

 

5 70 75 a0 g5 30 95 100
LiF IN MELT (mole Ys)

Fig. 5.6. Effective Activity Coefficients Caleculated from BEvapo-
rative-Distillation Data for MSBER Scolvent Compositions.
142
a,b,c

Table 5.8. Vapor Pressures for FPure Fluorides

Vapor Pressure

 

 

Compound ()
1000°¢ 200°C
LiF 0. 47 0.072
Bels 65 12.0
LrE, 2700 780

 

“Handbook of Chemistry and Physics, 44th ed., p. 2438, Chemical
Rubber Publishing Co., Cleveland, Ohlo, 19562.
b

 

B. Porter and E. A. Browa, J. Am. Ceram. Soc. 43, 49 (1962).

C : .
8. Cantor, personsal communication.

volatile [ission products. When this simualated fuel mixture is contacted
with a molten bismuth-lithium mixture, rare earths are reduced to their
elemental valence states and are dissolved in the molten-metal phase.
Ixperimental studies conducted thus Tar have examined the distribution
of rare earths between the two liquid phases as functions of the
lithium concentration in the metal phase. BSubsequent experiments will
elaborate on these findings and will examine the back extraction of
rare earths from the liquid metal into a second salt mixture. The
removal of rare earths from the reactor fuel, followed by thelr concen-
tration in a solid, disposable salt mixture will constitubte the re-
procegsing method.

Fluoride starting materials were prepared in nickel equipment by
adding sufficient rare-earth {luoride to approximately 2 kg of the gol-
vent, LiF-BeF, (66-34 mole %), to attain a rare-carth concentration of
about 10 % m.f. in the salt phase. When possible, a selected radio-
isotope of the rare carth was added for analytical purpcses. The mixture
wag treated with an HF-H; mixture at 600°C to remove oxide impurities
and at 700°C with H, alone to reduce concentrations of structural-metal
difluorides in the fluoride melt. The bismuth was purified by sparging
the 2.35-kg batch with Hp at 600°C to remove oxide impurities. The
bismuth purification was carried out in the experimental extraction
vessel of type 304L stainless steel lined with low-carbon steel. Follow-
ing the materials-preparation phase of the experiment, the fluoride mix-
ture was transferred as a liquid to the extraction vesgsel. ILithium
metal was added directly to the metal phase without prior contact with
the salt phase; this was done through a loading port that extended to
near the bottom of the extraction vessel, TLithium, Tor incremental
additions to the experiment, was freshly cut and weighed under mineral
0il, affixed to a small-diameter steel rod, rinsed in benzene, and
dried in the flowing inert atmosphere of the loading port prior to its
143

insertion and dissolution in the molten bismuth. Filtered samples of
each phase were taken under assumed equilibrium conditions (approxi-
mately 2-hr periods) after each addition of lithiwn. Radiochemical
analyses of each phase for rare-ecarth gamma sctivity and spectrographic
analyses of the metal phase for rare-earth and lithium concentrations
provided data for calculating the distribution of the rare earth in the
system and its dependence on the lithium concentration of the metal
phase.

In experiments conducted thus far, the distributions of lanthanum,
cerium, neodymium, samarium, and europium in the extraction system have
been examined separately, but under essentially comparative conditions.

A summary of thege results, illustrated in Fig. 5.7, shows that a
mixture of bismuth containing 0.02 m.f. of lithium metal sufficed for
removing egsentially all cerium, lanthanum, and neodymium and substantial
guantities of samarium and curopium from the barren fuel solvent. 1In
all of the experiments, rare earths that were reduced from solution in
the salt phase were found as dissolved components of the metal phase.

Bach incremental addition of lithium to the system resulted in a
near-linear increase in the concentration of lithium found in the metal
phase, However, this dissolved quantity accounted for only 25 to 50%
of the amount added to the system. The results of a blank extraction
experiment in which rare earths were omitted from the system could not
be distinguished from those obtajined when rare earths were present.
Fach incremental addition of lithium to an extraction system was an
approximate threefold excess over the amount required for the stoichio-
metric reduction of all of the rare earth contained in the experiment.

 

 

  

 

 

 

 

 

 

 

 

 

Wi SAMARIUM
29 CERIUM
FITa00—— o LANTHANUM —|
N MEQDYMIUM
= FUROPIUM
L5
z\z
|z 300 v — : 4 L S
g
|
< [ =T
wilfur
Ll
i
<t | &
el 200 Lo - e ] ]
L L
Qo
=z
cio
i
z|g9 B
El& 100 —
-l b
-
o | 8 g S Y A A

0 1 2 3 4 5
LITHIUM FOUND IN METAL PHASE (mole fraction x 10%)

Fig., 5.7. Extraction of Rare Earths from LiF-BeF, (66-34 Mole %)
into Bismuth by the Addition of Lithium Metal at 600°C.
144

In earlier experiments, beryllium metal was successfully used in
place of lithium metal for the reduction of rare earths. Spectrographic
analyses of the metal phase also showed that the lithium concentration
in bismth increased as the extraction of rare earths proceeded. These
results implied that the reaction

Be® 4 2IiF = Bel, + 214°

was of significance to the reduction of rare earths. Accordingly,
lithium was used as a soluble component in the metal phase in place
of beryllium, which is not soluble in either ligquid phase.

The reduction of rare-carth fluorides by lithium is expected Lo
proceed by the reaction

mi% + (8™ =REO + mrat

2

where m 1s the effective valence of the rare-earth cation. If unit
activities prevail for all metal species in the salt phase and for all
ionic species in the metal phase, then the activity of lithium dissolved
in the metal phasc can be expressed as a function of other activities

in the system as

(aREo)metal (aLiF)m salt

%130 7 Ké(aRE) salt :

 

By assuming thal the activity of LiF and the activity coefficients of
Lio, REOJ and RE remain constant, the dependence of rare-earth distri-
bution on the lithium concentration can be expressed as

where D is the ratio of the mole fraction of rare earth iIn the metal
phase to the mole fraction of rare earth in the salt phase, and

 

m
Ym0 v do .
KQ . RE etal ik
m e
K = v_ .9 * Y
a L1 metal REsalt

A plot of the experimental data according to the logarithmic form of
this equation is shown as Fig. 5.8. Values for m and K. calculated

Q

from the slopes and intercepts of this piolt are as follows:

Rare Earth m KQ

Lanthanum 2.7 2.5 % 107
Cerium 2.3 3.8 x 106
Neodymium 245 2.5 x 108
Samarium 1.6 1.8 x 10%
Europium 1.9 5.9 x 10°
 

 

  
 

-

 

 

   

 

 

YT T F— I
- L NEQDYMIL
o |

 

 

  
 
 

b b

 

  

 

 

 

  

..............

 

LITHIUM FOUND 1IN METAL PHASE
{mole fraction X 32)

 

 

 

 

 

 

 

 

I A

MOLE FRACYION OF RARE EARTH IN METAL
MOLE FRACTION OF RARE EARTH IN SALT

 

Fig. 5.8. Effect of Tithium Concentration in Metal Phase on the
Distribution of Rare Earths Between LiF-BeFs (66-34 Mole %) and Bismuth
at 600°C.,

Although the apparent fractional exponents for the reduction re-
actiong are as yet unexplained, the results are in rough agreement with
the occurrence of lanthanum, cerium, and neodymium as trivalent ions in
the salt mixture; samarium and europium are probably reduced to their
divalent states prior to their extraction into the metal phase.

Removal of Protactinium

Removal of Protactinium from Molten Fluorides by Oxide Precipita-
tion. Since a single-region molten-galt breeder reactor would in-
corporate the fertile material in the reactor fuel mixture, chemical
reprocessing schemes for recovering 2337 could be made more effective
if 233Pa, its precursor, could be removed without alteration of the
relatively large uranium concentration in the fuel mixture. The pre-
cipitation of an oxide of protactinium by the deliberate addition of
oxide ion may provide the basis for such a reprocessing method if the
simultaneous precipitation of U0y can be avoided. Previous studies
have demonsbrated the chemical feasibility of oxide precipitation for
removing protactinium from a fluoride mixture, LiF-BeF,-ThF, (67-18-15
mole %), proposed as the blanket of a two-region molten-salt breeder
reactor.

 

In the fluoride fuel mixture of the MSRE, sufficient ZrF,; has been
added to accommodate gross oxide contamination without loss of uraniwm
rfrom solution as UOp. Farlier studies had demonstrated that U0, would
not precipitate at 700°C from the solvent, LiF-Bel, (66-34 mole %) with
added UF, and ZrF,, until the concentration ratio of Zrf, to UF,
dropped below about 1.5.%? fTherefore a preliminary study was made of
the precipitation of Pal; from a fluoride mixture known to have ZrOs
as the stable oxide phase.23

The fluoride mixture consisted of TiF-ReF, (66-34 mole %) with

added ZrF, equivalent to 0.5 mole of zirconium per kilogram of salt
mixture. About 1 me of 232Pa was included in the salt preparation as
146

irradiated ThO,. The mixture was pretreated with anhydrous HI' and Hp
to convert oxides to [luorides.

The deliberate introduction of solid-phase oxide to the melt was
made by adding ZrOs, in small increments. Filtered samples of the salt
mixture were taken at assumed equilibriuvm conditions after each oxide
addition and were analyzed for “2°Pa by gamma spectrometry. The results
of these analyses showed that approximately 80% of the 233pg activity
was removed after the addition of about 67.5 g of ZrQs; (equivalent to
0.125 mole per kilogram of salt) to the mixture.

1T protactinium either formed a labile solid solution with Zr0O:
or was removed ifrom solution on the salt mixbture by surface adsorption
on Zv0s, then its distribution coefficient should have remained constant.
The fraction of protactinium remaining in the liguid phase could then
be expressed as a linear function of added ZrO, by the equation

1/8 (16)

by = L F (D/W

salt) ’ WZrOZ ?

where D = [Pa] /[Pa]salt’ F a0 = fraction of Pa in the salt, and W

oxide P
is the weight of the designated phase. An interpretation of the eXperi-
mental data according to this linear function, shown in Fig. 5.9,

ORNL-DWG 66-969

 

 

 

 

 

 

   
 

 

 

 

 

 

 

 

 

L |
L _ i (WZroz)
0 =—=1+0 [—5 ‘ '
< Fpa WsaLt ' ‘
o .
o WHERE £, = FRACTION OF #33pa IN LIQUID
= | o
2 Wzro,
~5 = WEIGHT RATIO OF PHASES
=z L SALT
g ° B _ CONC. OF Pa IN SOLID PHASE
= CONC. OF Pa IN LIQUID PHASE 4
<_£ [ ]
Z s
v o
a
o
L. 4
o
=
o
6 3 -
<1
o
IS
|
3 2|
o Wsair = 3-3 kg
o D= 237
moo “——W——‘*T 1
o
o L. i _
0 40 50 60 70

WEIGHT OF Zr0O, ADDED AS SOLID PHASE (g)

Fig. 5.9. Removal of ?3?Pa from Solution in LiF-BeF, (66-34 Mole %)
with Added ZrF, (0.5 Mole per Kilogram of Salt) by Addition of Zr0Op at
600°C.
147

illustrates the constancy of the disgtribution coefficient at a cal-
culated value of about 237. The data further suggest that about 7 g

of the initial ZrO, addition partially dissolved in the salt phase or

was otherwise lost from the reaction wixture. Further experiments will
include studies of the elfect of Zr0, surface area on protactinium re-
moval from salt mixtures proposed for a single-region moliten-salt breeder
reactor.

Removal of Protactinium from Molten Fluorides by Reduction Processes.

 

The effective recovery of 223Pa from a molten-salt breeder reactor will
provide more economic production of fissionable 232U by substantially
reducing blanket inventory and equipment costs and by improving neutron
ubilization. Accordingly, chemical development efforts supporting the
reference~-design MSBR are concerned with the removal of protactinium
from the blanket mixbure, LiF-BeFp~ThF, (73-2-25 mole %}, by methods
which can be feasibly adapted as chemical processes. An experimental
program has been initiated to study the reduction of protactinium fluo-
rides from this salt mixture by molten lead or bismuth saburated with
thorium metal at about 400°C. It was hoped that protactinium, as Pal,
in the salt phage, would be reduced to its metallic state by thorium
metal and could be recovered in the molten lead or bismuth.

The primary objective of initial experiments with this program has
been the study of protactinium removal from the sgalt phase of the ex-
traction system. For these experiments sufficient ¢°?Pa was obtained
for radiochemical analysis by neutron lrradiation of a small gquantity
of ThO,. The simulated blanket mixture was prepared from its comporents,
together with the irradiated ThOp, in nickel equipment. This mixture
was treated at 600°C with an HF-H, mixture (1:10 volume ratio) to re-
move oxide ion and at 700°C with Hs alone to reduce structural-metal
impurities., The metal-phase extractant, lead or bismuth with added
thorium metal, was prepared in the extraction vessel (type 304L stain-
less steel with a low-carbon steel liner) by treatment with Hy, at 600°C.
The extraction experiments were started by transferring a known quantity
of the prepared salt mixture into the extraction vessel containing the
metal-phase extractant. Filtered samples of the salt phase were taken
periodically for radiochemical analysis of dissolved protactinium. In
each experiment, °32Pa was rapidly removed from the salt phase and
remained absent from the solution during the approximately 100 hr at
600°C while tests were made. Subsequent hydrofluorination of the ex-
traction system with an HF-H; mixture (1:20 volume ratio) showed that
233Pa could be rapidly and almost quantitatively returned to solution
in the salt phase.

Typlcal results of these experiments are shown in Fig. 5.10. In
this experiment, thorium metal was added after the salt mixture was
introduced into the extraction vessel to demonstrate the necessity of
the reduction reaction.

The objective of experiments now in progress 1g to examine methods
for recovering 233pg from the extraction system, The proposed use of
macro quantities of 23+ps may be required to circumvent the anticipated
adsorption of the micro quantities of 233Pa, currently used, on the
walls of the container, or on other insoluble species in the systen.
148

Additional studiecg of the deliberate precipliation and adsorption of
233Pa on solid, stationary beds such as steel wool will be made for
comparative evaluation.

Solubility of Thorium in Molten Lead. Current studies of tiae re-
moval of protactinium from a molten fluoride mixbture, which simulates
the blanket of the reference-design MSBR, have been directed towsrd
the development of a liquid-liquid extraction process. The metho.l
proposes that protactinium, as PaF, Iin a salt phase, can be reduced to
its metalliec stabte and extracted into & molten metal phase. The poossible
use of a molten mixture of thorium in lead would provide & convenient
method for combining the reducing agent with the metal-pnase extractant
and for replenishing thorium to the fluoride blanket mixture. The
objective of this study has been to establish the solubility of Thorium
in lead over the temperature range of inlerest to this program and to
provide a lead-thoriuvm solution of known composition for subseguent
protactinium extraction experiments.

 

The exphrlmenbal mixture, contalned in low-carbon steel, consisted
of approximately 3 kg of lead and 100 g of thoriwn-metal chips. Values
for the solubility of thorium were obtalned by analyses of filtered
samples withdrawn from the melt at selected temperatures over the
interval 400 to 600°C under assumed equilibrium conditions. Samples
were withdrawn during two healing and cooling cycles and submitted for
sctivation and spectrographic analyses, These results, plotted as
the logarithm of the solubility vs the reciprocal of the absolute
temperature in ¥Fig, 5,11, indicate that the heat of solution of thnorium
in lead is ag prOXJmatOl* 19 kcal/mole and that its solubility at 600°C
is about 1.85 x 107 m,f.

Protactionium Studies in the High-Alpha Molten-Sall Laboratory.
A new glove- ~box facility, called the High-Alpha Mollen-Sait Laboratory,
for bbUdlCo of protactinium removal from molten Tluoride mixtures was
degcribed in the previous progress report.24 Experiments on extraction
of protactinium,from a breeder-blanket mixture, LiF-ThF, (73-27 mole %)
by ligquid-metal extraction at the tracer level (Irdnulofl of a part per
billion parcs of 233pa) are described in the section "Removal of Prot-
actinium from Molten Fluoride by Reduction Processes' of this report.
A similar experiment with an initial concentration of about 25 ppm of
©3lps 4in the molten fluoride mixture was conducbed in the High-Alph:
Molten-5Salt Laboralory to determine whether a 10° times higher protac-
tinium concentration would result in significant diflerences in protac-
tinium behavior during extraction and recovery experiments Resultrw ro-
ported here jndicate that a large fractlion of the reduced pr Lactinium
disappears Lrom molten lead in a tantalum contalner, jfl,agreement with
the tracer-level experiments, but one sample indicated that 17% of the
protactlnlum remained after about 6 hr at 625°C.

 

The completed experiment, run 1-12, was conducted in several
svages, with the furnace cooled to room temperature betlween stages.
In the first operation a 17 M HF solution containing 9 ng of purified
231Pa was mixed with 4.0 g of irradiated ThF,, containing about 1 mc
of ?2?Pa. The sall was heated in & platinum dish to evaporate water
and HF, and the dried mixture was added to a nickel pot contalning 330 g
Fig. 5.10.

Fig. 5.11.
Molten Iead.

233p4 1N SALT PHASE (%)

Effect of Thorium Metal on the Extraction of #°2Pa from
LiF-BeFp-ThF, (73-2-25 Mole %) in Salt-Lead Systen at 600°C.

CONCENTRATION OF Th IN LEAD (mole fraction X10%)

Temperature Dependence of the Solubility of Thorium in

149

ORNL-DWG 86-870

 

 

100 L’/ P -0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

80
WEIGHT OF SALT: 600¢
WEIGHT OF LEAD: 900g
|
60 e ]
40 b
WITHOUT ADDED WITH O wt %
THORIUM o Th IN Pb ADDED
20 o B
<J
Q
Lt
}._
0 A o
0 20 40 0 20 40

EXTRACTION TIME (hr)

ORNL- DWG 66 - 9271

 

 

TEMPERATURE (°C}
800 550 500 450 400
30
l [ ] l |
. |
2.0 e BY ACTIVATION ANALYSIS

 

0 BY SPECTROGRAPHIC ANALYSIS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12

i3
1o,ooo/flaK)

14
150

of pretreated LiF-ThF, (73-27 mole %). A copper filter unit was placed
in the pot which was sealed, evacuated to 450 p pressure three times,
and filled with helium each time., The furnace was then heated with
heliuwn flowing through the pot under a slight positive pressure. After
reaching 600°C, a mixbure of HF and Hp was passed through the melt Tor
50 min to convert any oxide impurities to fluorides, and then the melt
wag treated with Ho, for 4 hr to reduce NiFp, which resulted from the
hydroflucorination, to metallic nickel., Apparenltly because of a too
high I, Tlow rate, part of the melt splashed up to the top of the pot,
where it solidified. On attempting to remove the first filleved melt
sample, the copper filter was torn loose fram the l/Smin, nickel tubing
to which it was attached, and the copper filter fell back into the
melt. After cooling, the cake at the top of the melt was broken up,

a 1.0~z sample was taken, the remainder of the material was returned

to the pot, and the HI-Hp; treatment of the melt was repeated., No
difficulty due to the copper filter in tThe melt was observed.

In the second stage, the pot containing the treated Lil-Thi',
mixture was connected to a similar pot lined with tantalum and con-
taining 692 g of Pb-Th alloy that had been saturated with thorium at
600°C. The molten salt was transferred to the tantalum-lined pot at
655°C through a transfer line at 600°C by application of 10 1b of helium
presosure to the salt pot. Previous efforts to effect salt transfer had
been frustrated by low temperatures in the transfer line resulting from
the high vertical temperature gradient above the furnace block, which
wags metnioned in the previous report.24 A coiled heater above the
hlock and better insulation at the floor lewvel of the glove box over-
came this problem. After the transfer was completed, helium sparging
wag usged to causc mixing of the lead and salt phases. The sparging
was interrupted at iantervals to perniit phase separation., Samples of
the salt phase were removed with copper [ilter units, and samples of
the metal phase were normally removed with sintered stainless steel
filter units. Measurcments of the total gamma activity of 1-g portions
of the salt samples after recovery from the filter units showed that
the protactinium content of the salt phase was decreasing slowly after
about 2 hr contact time between salt and metal phases. At this Time,
5.0 g of thorium-metal turnings was added to the mixture; sampling of
the two phases 75 min later showed a sharp decrease in the gross gamms
activity of the sall phase without a consequent increase in the activity
of the stainless steel sampler and their contalned lead-phasc sample.

A further 3-hr contaclt time failed to produce & significani change 1in
the activity of the samples with one exception: a sample of the metal
phase inadvertently obtained in a copper sampler showed a significantly
higher activity level, later confirmed by 231pg analyses, than the
stainless steel filtered samples taken before and after the copper
samples. The furnace was then allowed to cool to room temperature.

Finally, the lead phase was transierred to an unlined nickel pot
containing LiF-1hF, of the same composition used previously but with-
out protactinium. A wmixture of hydrogen and anhydrous HE was bubbled
through the lead and salt phases in an effort to convert metalllc
protactinium in the lead to PaF, and transfer it to the salt phase.
Time limitatiocans prevented a thorocugh test of this procedure.
Data obtained in the above-described experiment, run 1-12, are
shown in Table 5.9. The gross gamma values were obtalned by weighing a
sample, usually 1.000 £ 0.002 g, placing it in a small glass vial, seal-
ing the vial in a plastic envelope as it was removed from the glove box,
and then placing the envelope over a sodium iodide crystal that was
connected o an RIDL scaler. These samples were then sent for solution
and analysis by the alpha-pulse-height method to determine their 231pg
content. The whole metal samples were counted for gross gamma activity
in a similar manner. The stainless steel samplers and their lead con-
tente were dissolved separately, and 231pg, analyses were made on the
solutions; but the copper sampler wag dissolved along with the lead
sample that it contained. The amount of lead in each sampler was also
determined.

The data in Table 5.9 confirm the conclusion, based on earlier
experiments with tracer concentrations of #2?Pa (see "Removsl of Prot-
actinium from Molten Fluorides by Oxide Precipitation,"” this report),
that protactinium fluoride is reduced to the metallic state by thorium
dissolved in lead. It further appears thalt most of the protactinium
does not stay dissolved in the lead phase long enough to permit transfer
of the lead to another container for recovery of the protactinium by
hydrofluorination, bub a small fraction of the protactinium was ap-
parently recovered in this Tashion.

The variable concentration of protactinium in the lead, small in
all caseg as compared with the amount in the stainless steel samplers,

Table 5.9. Run 1-12: FEguilibratioan of LiF-ThF,
(73-27 Mole %) with Pb-Th Alloy at 600°C

 

Total “31ps Content

 

Sample of Phaze
(mg)
Initial salt 8.5
Lead, 34-min contact (ss) 0.4
Salt, 57-min contact (ss) oo ds
Lead, 88-min contact (ss) 0.09
Salt, 110-min contact (ss) b2
Salt, 216-min contact (Th added) 0.09
Lead, 251l-min contact (ss) 0.07
Tead, 34l-min contzcet (copper) 1.46
Lead, 359-min conbact {ss) 0.08
Salt, 404-min contact (ss) 0.08

 
152

may indicate that the tendency of protactinium to absorb on, or alloy
with, iron causes the protactinium to be removed from the lead sample
during its passage through the sintered stainless steel filter material.
An effort was made to remove any protactinium-containing material adher-
ing Lo the outside wall of the samplers after they cooled to room temper-
ature prior to analysis. They were scraped and also given an acid
treatment before they were dissolved. It is not clear, at present,

why the lead-containing copper sampler showed a much higher protactinium
content than the samples obtained with stainless steel samplers, but

it seems significant that this sample indicated that 17% of the pro-
tactinium remained in the lead after almost © hr in contaect with tantalum
ab 625°C,

The transfer of a small fraction of the protactinium [rom the lead
phase to a molten fluoride mixture by HF treatment, as indicated by
preliminary results, is encouraging, but means of preventing loss of
the major part of the protactinium by adsorption on or alloying with
container walls or other materials obviously will be sought in continu-
ing experiments.

Radiation Chemistry

 

In-Pile Molten-Salt ITrradiatlion Mxperiment

Design, development, and fabrication of the in-pile molten-sall
experiment continued. Modifications to beam hole HN-1 in the ORR and
installation of experimental equipment are scheduled to begin early in
April, and in-pile operation of the first ilrradiation experiment will
begin in Juns 1966.

fuxiliary equipment needed to modify beam hole HN-1 for the molten=-
salt experiments has been fabricated. This equipment consists of (1) a
new aluminum beam-hole liner, (2) a beam-hole exbension sleeve, and (3)
an equipment chamber, located at the face of the reactor shielding,
which will contain tanks and valves required to remove salt samples
and add makeup salt to the autoclave, during in-pile operation.

Component parts of the experiment package which have been fab-
ricated are the shield plug, the loop container can, and the connector,
ag shown in Fig. 5.12. All component parts of the thermal loop have
been fabricated and are beling assembled. A photograph of the partially
asserbled thermal-loop experiment is shown in Fig. 5.13. The 12-Tt-long
sample line with electric heating elements, which are bonded to the
sample line by means of nickel metal spray, has also been completed
along with the main loop-heater-cooler unit. The heater-cooler-unit
design, previously described,25 has been changed by substituting cool-
ing tubes of Zirecaloy-2 instead of stainless steel and Zircaloy-2 for
the heagter-cooler jJacket in lieu of graphite. This change was made to
reduce the attenuation of neutron flux and to provide for improve-
ments In the efficiency and operational {lexibility of the heater-cooler
unit by using a Zircaloy-2 metal-sprayling technique to attach the heaters
and cooling colls,
152

PHOTO 73060A

CONTAINER CAN

  
   
    

CONNECTOR \
J/,flr

SHIELD PLUG ——

Fig. 5.12. ©Shield Plug, Container Can, and Connector for In-Pile
Molten-Salt Loop.

PHOTO 73059A

CORE BQODY WITHOUT
HEATER —COOLER

RESERVOIR
TANK

COLD LEG WITH
HEATER- COOLER

Fig. 5.13. Partially Assembled In-Pile Molten-Salt-Loop Experiment.
154

The iInstrument and control panels are being fabricated in the
Instrumentation and Controls Division shops and are approximately 60%
complete. The instrument panels are to be set in place in front of beam
hole HN-1 in April.

Development and Evaluation of Methods for the
Analysis of the MSRE Fuel

 

 

Determination of Oxide in MSRE Fuel

 

The study?® of hydrofluorination as a way to determine oxide in MSEE
salts was continued. This method?”? is based on the reaction

02" + 2HF(g) =H,0(g) + 2F ,

which occurs when a molten-salt sample 1s purged with an Ho-HF gas
mixture. The amount of water evolved is taken as a measure of the
guantity of oxide in the molten-salt sample, Data have been reported
previously27 Tor the Karl Fischer titration of the water evolved from
standard additions of ZrO, and U0, to a 50-g fuel melt.

Figure 5.14 is a schematic flow diagram of the hydrofluorination
apparatus now being used. Since it would be difficult to perform the
Karl Fischer titration in the hot cell, a Pp,0,; water electrolysis cell
(Beckmann moisture-monitor cell with rhodium electrodes) was installed
in a component test facility to monitor the effluent-gas stream from
the hydrofluorinator for water.

To establish the conditions necessary to use the electrolysis cell
as a water monitor, ZrO, samples were added to 50 g of MSRE fuel in the
hydrofluorinator. Because there 1is a maximum rate of water electrolysis
beyond which the cell may be damaged, it was necessary to split the
effluent-gas stream and to pass only a portion of it through the cell.
ATter the optimum gas pressures and flow rates were established, the
results given in Table 5.10 were obtained.

Figure 5.15 is a recording of the water evolved from two standard
additions of Zr0Os, to a fuel sample. The first two peaks of each record-
ing result from atmospheric contaminants introduced during loading of
the ZrO, intc the fuel sample and removal of contamination from the
internal metal surfaces of the apparatus. After the fuel-Zr0; sample
was melted, it was purged with Hp-HF at 625°C to evolve the oxide.

The plateau of the curve for the 139-mg Zr0, addition apparently re-
flects the 1limit of solubility of ZrO, in the melt.

Also, analyses were made on ~50-g portions sampled from 6 kg of
simulated molten MSRE fuel of unknown oxide concentration. These samples
were taken in copper ladles, cooled to room temperature, and then trans-
ferred to the hydrofluorinator under an inert atmosphere. Three of the
samples were exposed to the atmosphere, since the samples from the re-
actor will be exposed while being transferred to the hydrofluorinator.
The sample ladle was sealed in the hydrofluorinator with a delivery
tube spring-loaded against the surface of the salt (Fig. 5.14). An
ORNL—-DWG 65--8355A

  
 
    
  

MOISTURE

/ MONITCR
r

115 oo sHiELD

L He BACK FLUSH

5N —NoF — HF TRAP

 

 

 

  

PURIFIER

   
   
 

7

 

 

T

T

CELL WALL

 

 

Fig. 5.14. Schematic Flow Diagram of the Apparatus for the Determi-
nation of Oxide in MSRE fuel by Hydrofluorination.

Table 5.10. Recovery of Oxide from Standard Additions of
7Zr0, to MSRE Fuel

 

Effiuvent Gas

 

 

Passed Through ZrQ, (mg) Recovery (%)
Cell
( Taken Found
%)
10 18.6 18.3 93.6
25.6 26.0 10L.6
25.8 27.1 104.9
5 2.2 2.6 100.5
131.0 134.8 102.9
135.0 136.5 101.1

139.0 140,0 100.7

 
757

0

45

H,0 N EFFLUENT (pg/min)

  
 

HF, 0.2 atmn
SALT MELT , 50 ¢

 

Hy, FLOW, 200 cc/min

156

 

 

 

 

 

 

F'ig. 5.15.
lated MORE Fuel by Eydrofluorination.

Table 5.11.

 

 

 

 

    

ORNL-DWG 66- 4038

-t

139 mg Zr0»
(100.7 % REZCOVERED)

 

 

 

 

| i
1 !
| |
— 18.6 mg Zr0,
(98.6 % RECOVERED)
7N !
HF INTRODUCED - | / N
: i { \ i
SALT MELTED | 1 | ~~
L 1 N 1
{ \
l \
\ |
I \ i
f \
f |
!/
//
s ey
50
TIME (min)

Determination of Standard Additions of Oxide in Simu-

Precision of Determination of Oxide Iin

Molten Simulated MSRE Fuel

 

sample Number

Oxide (ppm)

 

350
345
340
340
375°
360%

4057

 

aEXposed to atmosphere for at least 24 hr.

bExposed to atmosphere for two weeks.
15%

initial purge with Ho-HF was made at a temperature below the melting
point to remove oxide Irom the metal surfaces and oxide contamination
from the surface of the salt. With further temperature increase, the
delivery tube was driven below the surface of the salt as it melted to
sparge the melt with Hp-~-HF, Table 5.11 lists the results.

We must hers consider whether these results represent & quantita-
tive removal of oxdde or simply a return to an original equilibrium base
line. It may be seen by rearranging the equilibrium expression for the
reaction of oxide with HF that 1f the proportion of HF reagent in the
purge gas is increased, the residual-oxide concentration will be
reduceds:

271 _ 2 Y v const
[0% ] (RHQO/PHF) w constant ,

Because the HF reagent contains traces of water, the HF concentration in
the influent cannot be increased independently. However, if the concen-
trations of HF and H,0 in the purge gas are doubled, the residual oxide
in the melt should be reduced by a factor of 2. When the pressure of

HF in the sparge gas through terminal (oxide-depleted) melts was doubled,
no additional water was obtained, Also, squilibrium constants are tem-
perature dependent, and no additional peaks were observed with tempera-
ture variations of 200°C on depleted melts.

Semples of Tlush and fuel salt taken during the December startup
of the reactor were analyzed for oxide in a component test facility.
Table 5.12 summarizes the results. Figure 5.16 is a recording of the
water evolved from the first fuel-sslt sample taken after the fuel was
loaded into the reactor.

The results of the analyses by the hydrofluorination method were
in good agreement with those by the KBrF, procedure?® (Fig. 5.17).
The KBrF, values paralleled the trends shown by the hydrofluorination

Table 5.12. Oxide Concentrations of Flush and IFuel Salt from the MSR

 

 

Time of Salt Clrculation Oxide Concentration
Sample
(hr) (ppm)
Flush salt 2.7 46
29.1 72
4'7.6 106
Fuel salt 3.4 120
23.8 105
32.2 80

52.5 65

 
158

method, but averaged slightly higher. This bias was
since the pulverized samples required for the KBrF, method are easily
contaminated by atmospherlic moisture.

ported above.

not unexpected,

The apparatus (Fig 5.18) designed for the determination, in the
hot cell, of the oxide in highly radioactive fuel samples is functionally
identical with the components test facility used for the analyses re-

furnace, which is external to the valve compartment.

HpO IN EFFLUENT { ug/ min)

Figure 5.18 shows the hydrofluorinator positicned in the

The hydrofluori-

ORNL~-DWG 66-4028A

 

 

 

 

 

SALT MELTED

 

 

 

 

My, FLOW, 200 cc/min |7 |
HF, 0.2 otm ]
FUEL SALT, 48.7g

OXIDE FOUND, ‘20 ppm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

| |
100
TIME {min)
Fig. 5.16. Determination of Oxide in MSR Fuel Sample FP4-11 by
Hydrofluorination.
ORNL-DWG 66-2026A
I | \
‘ A HYDROFLUORINATION
| ® KBrf,
250 f—— - e — — | |
200 ——F—F——— —
e
£ :
o I A R - 3 _ - o
o 190 — ® a |
o |
o o |
100 L1 AL e A ]
® o
A A
o 8
50 1 A - — e - ] T_- .
® ' | f
;—-FLUSH SALT CHANGED TO FUEL
) | {
0 10 20 30 40 50 10 20 30 40 50 60
TIME OF SALT CIRCULATION IN REACTOR {hr)
Fig. 5.17. Results of Hydrofluorination and KBrI'y Analyses for

Oxide in Flush and Fuel BSalts During December Startup of the MER.
159

PHOTO 82177

 

Fig. 5.18. Hot-Cell Apparatus.
160

nator is put on stream through O-ring ball Jjoints that are closed or
opened by a remotely operated argon-actuated air piston. The valve
compartment contains the sodium fluoride trap, the capillary water
splitter, the P>05 water electrolysis cell, and the soda-lime trep for
cleaning the effluent gas stream. These components are shown in Fig.
5.19, where the cover plate has been removed. The modular construction
was chosen in hopes that any maintenance necessary can be performed
remotely. If not, any component can be removed manually in 2 to 5 min
if the hot cell is sufficiently decontaminated.

Figure 5.20 shows the disassembled hydrofluorinator. The copper
sample ladle is contained in the nickel liner, which fits into the
bottom of the hydrofluorinator. The baffles on the spring-loaded sparge
tube fit inside the liner and confine the molten salt to the liner
during the bubbling operation of the analysis. This arrangement allows
the solid sample, after analysis, to be removed from the bottom of the
hydrofluorinator and then to be disconnected at the Swagelock fitting

NaF TRAP
(UNDER VALVE PLATE)
-

CAPILLARY SPLITTER
COMPARTMENT

 

Fig. 5.19. Top View of Valve Compartment of Hot-Cell Apparatus.
161

PHOTO 82187

Wi

AMOLVHOBY T YN

[
)

P

P

|

3

OlLLYN

 

 

 

Disassembled Hydrofluorinator.

Hir, 5.2,
162

and discarded along with the ladle, liner, and bottom of the sparge
tube. This design affords considerable saving in cost per analysis,
because it permits the repeated use of the complete hydrofluorinator.

Figure 5.21 shows the Hp-HF mixer section, which is located in the
access area behind the hot cell. The HF pressure is regulated by con-
trolling the temperature of the HF tank shown at the left. The flow
rates of the Hr and HF are controlled by nickel and platinum capillaries
respectively. This apparatus also contains safety interlocks and limit-
ing capillaries to prevent release of HF and excessive pressures inside
the hot cell.

Figure 5.22 shows the master control panel, which is located in
front of the hot cell. This panel contains the moisture recorder (not
shown), helium and hydrogen pressure controls, HF control switch, hydro-
fluorinator coupler switch, temperature recorder, furnace power controls,
and power control for the coupler-line heater.

After they were fabricated, the apparatus and auxiliary equipment
were assembled in the hot-cell mockup, and all mechanical operations
were performed successfully with Master Slave manipulators. The equip-
ment was then transferred to the laboratory and was assembled on the
bench top for trial analyses of samples. Individual components were
tested exhaustively and, when necessary, were modified or replaced to
obtain maximum dependability and to minimize hot-cell maintenance.

A final check of the splitter and cell was made by injecting a
known quantity of water into the flow system; quantitative recovery of
the water resulted. The entire system was then checked out by analyz-
ing fuel-salt samples. Satisfactory results were obtained.

The equipment is now being transferred to the hot cell and installed
therein.

Voltammetric Determination of Ionic Iron and Nickel in Mclten MSRE Fuel

By controlled-potential voltammetry, Fe2+ and Ni2+ were determined
in a 100-g sample of MSRE fuel. The sample was withdrawn from the re-
actor in enricher ladles and was transferred under an inert atmosphere
to the graphite crucible of an electrochemical cell assembly for remelt-
ing and analysis. The cell assembly and electrodes developed for
electrochemical studies of molten-fluoride salts are described else-
where 29721 The iron and nickel in the molten fuel are suspected to be
in the metallic state, as well as in the form of soluble ionic species.
However, only in the divalent oxidation state are these metals electro-
reducible in the melt and can thus give voltammetric reduction waves.

By voltammetry and by the standard-addition technique, the concentration
of Fe?” was determined to be ~10 ppm. The concentration of Ni2™ was be-
low the limit of detection by voltammetry (<1 ppm).

Average total concentrations of iron and nickel, determined by
conventiocnal methods, are about 125 and 45 ppm respectively. Thus it
appears that most of the iron and nickel in the fuel is present in the
metalliic state, probably as finely divided particles.
163

PHOTO 82182

 

Fig. 5.21. Hp-HF Mixer Section of Hydrofluorinator Apparatus.
164

PHOTO 82189

 

 

Fig. 5.22. Main Control Panel of Hydrofluorination Apparatus.
165

Chromium concentration, determined conventionally, is ~50 ppm.
Interference from uranium prevents the voltammetric determination of
chromium in the molten salt. A well-defined wave was oObserved that
corresponds to the reduction U(Iv) —U(II1). Possibly, this wave can
be used to continuously monitor uranium in the system; more work is
contemplated in this area.

Development and Evaluation of Equipment and Procedures for

Analyzing Radioactive MSRE Salt Samples

 

 

As stated earlier,32 statistical evaluation of the control data

indicated that a negative bias of approximately 0.8% existed in the
coulometric uranium results. The uranium procedure consisted primarily
of six steps.

. Determination of blank.

. Preparation of sample.

Prereduction of sample at +0.125 v vs S.C.E.
Reduction of uranium and copper at —0.325 v vs S.C.E.
Oxidation of copper at +0.125 v ve S.C.E.

. Calculation.

O W

In steps 1, 3, 4, and 5, the titration was allowed to proceed
until the cell current had decreased to 5 Ha.

The calculation was performed using the following formula:

[(A=B) —C — (bt/E)IDE = micrograms of uranium per test solution |,

where
A = readout voltage for uranium and copper reduction in miliivolts,
B = blank of electrolyte in millivolts,
C = readout voltage for copper oxidation in millivolts,
D = 1.233 x 10 ? ug of uranium per microcoulomb,
E = coulometer calibration constant for 10-microequivalent range

in microcoulombs per millivolt,
b = background current in microamperes (5 na),
t = time of uranium and copper reduction in seconds.

In an attempt to eliminate the bias, steps 1, 4, and 6 were
modified. The determination of the blank was eliminated. The cutoff
point for the uranium and copper reduction was changed from 5 pa to
50 pa. An X-Y recorder (Fig. 5.23) was attached to the coulometer to
monitor the titrations. When the readout voltage and cell current were
plotted on the X and Y scales, respectively, a straight line with a
negative slope was obtained after the recorder pen had reached a
maximum Y value. By extrapolation of the curve (Fig. 5.24), it was
possible to obtain the readout wvoltage corresponding to a cell cur-
rent of O amp. Due to the error introduced by this technique, a poten-
tiometer was connected to the integrator circuit of the coulometer in
order to obtain more precise readout voltages. This necessitated
terminating the titrations at a specific end point. A cell current of
166

PHOTO 82047

Fig. 5.23. High-Sensitivity Coulometric Uranium Apparatus.

 

50 pa was found empirically to be the most practical point of termina-
tion. The coulometer was automatically turned off when the ammeter
needle indicated a cell current of 50 pa. At this point, the titration
was approximately 99% complete. This portion of the readout voltage
was taken from the potentiometer; the remaining readout voltage was
obtained by extrapolating from 50 to O pa an expanded cell current vs
readout voltage curve (Fig. 5.25). The two readings were combined to
obtain the total readout voltage for the uranium and copper reduction,
Based on more than 200 titration curves, the readout voltage represented
by the extrapolated portion of the curve was approximately 5 mv; there-
fore, the readout voltages used for the uranium and copper reductions
were oObtained by adding 5 mv to the potentiometer readings. Due to the
above changes, the formula for calculating the uranium present in the
test solution then became

[(A — B) + CIDE = micrograms of uranium per test solution ,

where

= readout voltage for uranium and copper reduction in millivolts,

readout voltage for copper oxidation in millivolts,

= 5 mv (value obtained by extrapolation of the cell current vs
readout voltage curve from 50 to O ua},

= 1.233 x 10 ° pg of uranium per microcoulomb,

= coulometer calibration constant for 10-microequivalent range
in microcoulombs per millivolt.

Qe
I

=
|
167

The uranium present in the test solution could be calculated,
since the voltages read out on the potentiometer were proportional to
the coulombs required in the reduction of uranium and copper and the
oxidation of copper.

Sample Analyses

 

From December 16, 1965, through January 26, 1966, six flush-salt
and twenty-two fuel-salt samples were received in the High-Radiation-
Level Analytical Laboratory. All 28 samples were analyzed for U, Zr,

ORNL~-DWG €66-4784

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S 5
g '\
'_
=
Lt
o
S
S 4
]
-
Lt
Q
3 A\
2
, \
0
0 01 0.2 0.3 0.4 05 0.6
VOLTAGE

Fig. 5.24. Cell Current vs Readout Voltage Curve.
168

ORNL-DWG 66-4785

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.6
©
£ 12
|_
=
w
[1 g
o
3
o 08
|
W
(0]

OA; \

EXTRAPQLATED
VOLTAGE
POTENTIOMETER VOLTAGE
o |
o 0.1 0.2 0.3 0.4 0.5

VOLTAGE

Fig. 5.25. A Portion of the Cell Current vs Readout Voltage Curve.

Cr, Be, F, Fe, and Ni. All six flush-salt samples and the first
twelve fuel-salt samples received were analyzed for molybdenum.

Quality Control Program

 

The quality control program initiated prior to precritical sampling
was continued during the past period. Synthetic soluticns similar to
dissolved nonradioactive fuel-salt samples were analyzed along with
each flush-salt and fuel-salt sample. Due to the relatively smali
number of samples analyzed to date, the accumulated control data is
insufficient tc calculate the true percent standard deviations of the
methods. The values shown in Teble 5.13 were obtained during the
fourth quarter of 1965. The 2S5 and average values shown were obtained
by four different groups of shift personnel. Molybdenum values are not
shown in the table since it was not added to the synthetic solution.
The nickel values indicate that a positive bias exists in the method.
The values given for the coulometric uranium procedure indicate that
the negative bias was eliminated.
169

 

 

Table 5.13. Control Program, Fourth Quarter of 1965
Average
Determination Shift Num?er ?f Added Value F%und 25 Percentage
Determinations (ug/ml) :
(hg/ml)
Coulometric,
uranium A 19 656.6 658 1.11
B 40 656.6 656 1.55
C 30 656.6 659 0.57
D 34 656.6 658 0.29
Amperometric,
zirconium A 26 1117 1122 7,27
B O 1117 1077 13.53
C 14 1117 1119 5.17
D 0
Amperometric,
chromium A 22 11.8 12.202 9.15
B 16 11.8 12.278 11..06
C 8 11.8 12.520 16.84%
D 5 11.8 12.270 14.38
Colorimetric,
iron A 18 6.36 6.286 2.30
B 10 6.36 6.355 10.91
C 16 6.30 6.302 5.51
D 3 6.36 6.253 2.40
Colorimetric,
nickel A 16 6.08 6.713 3.92
B 10 6.08 6.748 6.24
C 18 6.08 6.669 14.32
D 0
Neutron acti-
vation,
beryllium A 3 847 853 1.08
B 0
C 2 847 849 1.83
D 4 847 856 1.48

 
18.

19.
20.

21.
22.
23.
2.

170

References

MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 321.

 

See "Development and Evaluation of Equipment and Procedures for
Analyzing Radiocactive MSRE Salt Samples,” this report.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 117.

See "Voltammetric Determination of ITonic Iron and Nickel in Molten
MSRE Fuel," this report.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 140.

See "Oxide Solubilities in MSRE Flush Salt, Fuel Salt, and Their
Mixtures,” this report.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp.
123-25.

K. A. Sense et al., J. Phys. Chem. 58, 223 (1954).
K. A. Sense and R. W. Stone, J. Phys. Chem. 62, 1411 (1958).

S. Cantor, Reactor Chem. Div. Ann. Progr. Rept. Dec. 31, 1965,
ORNL-3913, pp. 2729,

5. Cantor, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1962,
ORNL-3262, pp. 38-41.

E. A, Brown and B. Porter, U.S. Bur. Mines Rept. Invest. 6500
(1964).

P. W. Bridgman, Proc. Am. Acad. Arts Sci. 59, 162 (1923).

MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 129—
38,

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 152.
Ibid., pp. 140-43,

T. Langmuir, Phys. Rev. 2 (Ser. 2), 329 (1913) (and subsequent
papers).

Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp.
59-62.

K. A. Sense and R. W. Stone, J. Phys. Chem. 62, 96 (1958).

 

 

 

 

 

 

 

 

A. Buchler and J. L. Stauffer, Symposium on Thermodynamice with
Emphasis on Nuclear Materials and Atomic Transport in Solids,
Vienna, July 22-27, 1965, paper SM-66-26, p. 15.

J. H. Shaffer et al., Nuel. Sci. Eng. 18, 177 (1964).
Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1961, ORNL-3127, p. 8.
Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, p. 56.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp. 137
41 .

 

 

 

 

 
25.

27
28'

29,
30.
31.
32.

171

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 109,
Fig. 5.2.

MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 140.
MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 162.

G. Goldberg, A. 8. Meyer, Jr., and J. C. White, Anal. Chem. 32, 314
(1960).

D. L. Manning, J. Electroanal. Chem. 6, 227 (1963).
D. L. Manning and G. Mamantov, J. Electrosnal. Chem. 7, 102 (1964).
D. L. Manning, J. Electroanal. Chem. 7, 302 (1964).
MSR Program Semlann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 148.

 
172
6. MOLTEN-SALT BREEDER REACTOR DESIGN STUDIES

Design and evaluation studies have been made of thermal molten-salt
breeder reactors (MSBR) in order to assess their economic and nuclear po-
tential and to identify the ilmportant design and development problems.
The reference reachor design presented here contains design provlems re-
lated to molten-sallt reactors in general.

The MSBR reference design concept i1s a two-region, two-fluld system,
with fuel sall separated from the blanket salt by graphite tubes. The
fuel salt consists of uranium fluoride dissolved in a mixture of lithium-
beryllium fluorides, while the blanket salt is a thorium-lithium fluoride
containing 27 mole % thorium fluoride. The energy generated in the re-
actor fluid is transferred to a secondary coolant-salt circuit, which
couples the reactor Lo a supercritical steam cycle. On-site fluoride
volatllity processing 1ls employed, leading to low unit processing costs
and economic operation as a thermal breeder reactor.

MSBR Plant Design

 

Flowsheet

Figure 6.1 gives the flowsheet of the 1000-Mw (electrical) MSPR
power plant. Fuel flows through the reactor al a rate of about 44,000
gpm (velocity of about 15 fps), entering the core at 1000°F and leaving
at 1300°F. The primary fuel circuit has four loops, each loop having a
pump and a primary heat exchanger. Racn of these pumps has & capacity
of sbout 11,000 gpm. The four blanket-salt pumps and heat exchangers,
although smaller, are similar to corresponding components in the fuel
system. The blanket salt enters the reactor vessel at 1150°F and leaves
at 1250°F. The blanket-salt pumps have a capacity of about 2000 gpm.

Four 14,000-gpm coolant pumps circulate the sodium fluoroborate
coolant salt, which enters the shell side of the primary heat exchanger
at 850°F and leaves at 1112°F. After leaving the primary heat exchanger,
the coolant salt is further heated to 1125°F on the shell side of the
blanket~salt heat exchangers. The coolant then circulates througn the
shell side of 16 once-through superheaters (4 superheaters per pump).

In addition, four 2000-gpm pumps circulate a portion of the coolant
through eight reheaters.

The steam system flowsheet is essentially that of the new TVA Bull
Run plant, with modifications to increase the rating to 1000 Mw (elecm
trical) and to preheat the working fluid to 700°F prior to entering the
heat-exchanger—superheater unit. A supercritical power conversion sys-
tem 1is used, which is appropriate for molten-sall application and takes
advantage of the high~strength structural alloy employed. Use of a super-
critical fluid system results in an overall plant thermal efficiency of
about 45%.
   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

    

ORNL-DWG 66-4786

    
   

    
 

   

 

  
   
  

 

     
   
   

 

 

 

FUEL _ NET OUTRUT 1000 Mwe .
BLANKET ~—-— GROSS GENERATION  1034.9 Mwe A ——s ———
COOLANT —--— BF BOOSTER PUMPS 9.2 Mwe \ |
STEAM  ————- STATION AUXILIARIES 25.7 Mwe |‘ (58,5 10.067* 7152 % ;
H,0 o REACTOR HEAT iNPUT 2225 Mwt jlocoer T T |
* 07 ib/hr ; - 550 | S .
, NET HEAT RATE 7601 Btu/ kwhr | 3309 I [ | | 1518.5h -540p
P psia ) e 570p~650°F 551.7°F
NET EFFICIENCY 49% | o e _ 220 ! o
h Btu/b REACTOR VESSEL (1) I T - : - g '_P i | 1000%F
—3— FREEZE 2225 Mw: : 3600p,  REHEAT ' —ry I | |
VALVE ! ; 00051 STEAM ‘ | | |
: | razan | PREHEATERS(S)‘ | 00p_ |
o ; i | 100.5 Mwt | 2587 |
\ | | 350Cp I
BLANKET SALT FUEL SALT | | | B65°F I
PUMPS (4) PUMPS (4} ! ) | 1307.8h !
2000 gpm |, L, 14,000 gom ++_ J . a d___ o F e
(NOM) EACK (Now) EAcH | \ | ) ! -4 1P TURBINE 1-527.2 NMwe}—
t q72e” i 57.352"% | | ; GROSS
: 17.3 fts/sec\ \ | I ; ;
: . PREHEATER COOLANT | : BOiLER-SUPERHEATER i !
1180° F P SALT PUMPS (4) : ; ‘ : —
| 0°F L = o o i , CODLANT SALT PUMPS (4) T Em
—————- e ¥ § ; 14,000 gpm (NOM) EACH N LP e L
fi Y 1250°F o | {NOM)EACH i ; : ! TURBINE | 7‘52;'(7)5?&
‘ IS [T e - 850°F ; i ' S
K LT . Vi | ™ j 7 850°F
1 Ig SLANKET SALT HEAT 1125°F | ‘ L E CONDENSER AND
i1 1| EXCHANGERS (4) : L - 1125°F | 1769.2h -3800p FEEDWATER SYSTEMS
’: T 1 Mt (oL ) P > —— - | 200° 7 | SEE STEAM SYSTEM
o ‘. o ; : : FLOWSHEET)
L ) 1 S BCILER- :
) FUEL SALT HEAT ; REHEATERS | — 850°F SUPEREREATERS ! i ]
| EXCHANSERS (4} ) N (8) - : i (i6! : U 3500 p - 548.6°F
: 2114 Mwi (TOTAL)- T . 293.5 Mwi LA 1934.5 Mwt N 546.3h
I 7 —
. —_— . ——— _— +
T 1/ 3475p-695°F - 766.4 h

BLANKET SALT
i DRAIN TANKS (2]

 

 

! BOILER FEEDWATER

 

i PRESSURE BOOSTER

 

FUSL SALT

COOLANT SALT

PLUWPS (2}

 

{ DRAIN TANKS {4)

Fig. 6.1.

 

 

DRAIN TANKS i4)

 

Flowsheet of MSBR Power Plant.

20,000 gpm (NOM) EACH
4.6 Mwe EACH

el
174
Reactor Design

Figure 6.2 shows a plan view of the MBBR cell arrangement. The
reactor cell is surrounded by four shielded cells containing the super-
heaters and reheater units; these cells can be individuvally isolated for
maintenance, The processing cell, located adjacent to the reactor, is
divided into a high-level and a low-level activity area.

Figure 6.3 shows an elevation view of the reactor and indicates the
position of egquipment in the various cells. Figure 6.4, a plan view of
the reactor cell, shows the location of the reactor, punps, and fuecl and
blanket heat exchangers. Figure 6.5 is an elevation of the reactor cell.
The Hastelloy N reactor vessel has a side wall thickness of aboutb 1-1/4
in. and a head thickness of about 2-1/4 in.; 1t is designed to operate
at 1200°F and 150 psi. The plenum chambers, with 1./4-in.-thick walls,
communicate with the external heat cxchangers by concenbric inlet-outlet
piping. The inner pipe has slip joints to accomumodate thermal expansion.
Bypass flow through these slip Jjoints is about 1% of the total flow. As
indicated in Fig. 6.5, the heat exchangers arc suspended from the top of
the cell and are located below tThe reactor. kach fuel pump has a free
fluid surface and a storage volume which permit rapid drainage of fuel
fluid from the core upon loss of flow. In addition, the fuel salt can
be drained to the dump tanks when the reactor is shut down for an ex-
tended time. The entire reactor cell is kept at high temperature, while
cold "fingers" and thermal insulation surround structural supporl members
and all special equipment which must he kept at relatively low tempera-
tures. The control-rod drives are located above the core, and the con-
trol rods are inserted into the central region of the core.

ORNL-DWG 66-785

  
     
 
 
       
       

| REHEAT STEAM *W N
FEED H,0- T T
30 H.P. STEAM T
He sTEAM--—1 |
FEED Hp0- - i T
I.P. STEAM -~

WASTE GAS CELL

 

~FUEL HEAT

¢

: EXCHANGER

!
4

— =z
=D

COOLANT SALT PUMPS-
{ \

      
 
 
     
    
   
 
   
 

iy

20 —==\ ) ; 5 5 [ e

2 { 5 i£; fi” = . | £ R l: 1fi," ,,,,, -

r—m
Lo

REACTOR

 

____b_mgfl_p.

¢
BLANKET HEAT EXCHANGER’

   

STORAGE v 26

30 FUEL PROGESSING

    
 
   
 

CONTROL AREA

ALL DIMENSIONS ARE IN FEET DECONTAMINATION
i.....—_. o e e e ,180 [, — - U

 

Fig. 6.2. Plan View of MOSBR Cell Arrangement.
175

ORNL-DWG G&6-733

      
 
 
 
 

COOLANT FUEL CIRCULATING CONTROL ROD DRIVE
/" saLT PUMP
! eumies BLANKET CIRCULATING PUMP

 

 

 

-

 

 

 

CONTROL ROOM
LEVEL —

 

 

 

 

58 ft

 

ANALYTICAL
LAB LEVEI—"

 

 

 

 

 

 

 

 

 

 

 

 

BLLANKET HEAT EXCHANGER (F?EHEATERS

Fig. 6.3. EBElevabtion View of MSBR Reactor Arrangement.

ORNL.-DWG 66-4787

 

 

 

 

    
  
 
 
   
    
   
 
  
  
   
 
  
  
     

STEAM GENERATOR
CELL

BLANKET HEAT
EXCHANGER -

— BLANKET
REGICN

| _—— CORE REGION

 
        
 

 

e T DL

CONTROL ¢f
ROD #
DRIVES [
A

’ii

   

 

     

9

 
  

  

 

 
 

c‘COOLANT SALT
TO SUPERHEATERS
AND REHEATERS

‘ ; S EAT EXCHANGER
-COOLANT SALT SUPPORT STRUCTURE

RETURN LINE

 

 

    

 

   

     

PRIMARY HEAT
EXCHANGER

36-ft-diom REACTOR &
CELL

A

 

  

Fig. 6.4. Plan View of MSBR Reactor Cell Showing Location of Re~
actor Equipment.
176

CHNL.-DWG 86-794
FUEL PUMP MOTOR \ » CONTROL ROD DRIVES /BLANKET PUMP MOTOR
i /

3

 

/
-

   

 

 

CONSTANT SUPPORT
HANGERS .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- A0 -ft-digm CCRE

— REACTOR VESSEL

 

 

 

— FUEL SALT DISTRIBUTION
PLENUMS

 

 

FUEL DUMP TANK ‘

WITH COOLING I3
COILS FOR AFTER
HEAT REMOVAL — 1.

 

 

 

 

 

 

 

42 1 —

 

 

 

T~ BLANKET HEAT
EXCHANGER

PRIMARY HEAT J
EXCHANGER-unfigTJW

-5 in, -

 

 

 

 

 

 

 

 

— 19 ft

 

 

 

 

 

 

 

 

\ REACTOR CELL HEATERS

Fig. 6.5. Elevation of Reactor Cell.

The reactor vessel, about 14 £t in diameter by about 15 ft high,
contains a 10-ft-dlam core assembly composed of reentry-type graphite
fuel cells. The graphite tubes are attached to the two plenum chambers
at the bottom of the reactor with graphite-to-metal transition sleeves.
Fuel from the entrance plenum flows up fuel passages in the ouber region
of the Tuel cell and down through a single central passage to the exit
plenum. The fuel flows from the exit plenum to the heat exchangers, then
to the pump, and back to the reactor. A l«l/2-ft—thick molten-salt blanket
plus a 1/7~ft—thick grapnite reflector surround the core. The blanket salt
also permeates the interstices of the core lattice so that fertile material
flows through the core without mixing with the fisgile fuel salt.

The MSBR requires structural integrity of the graphite fuel cell. In
order to reduce the effect of radiation damage, the fuel cells have bee
177

made small to reduce the fast-flux gradient across the graphite wall.
Also, the cells are anchored only at one end to permit axial movement.
The core volume has been made large in order to reduce the flux level in
the core. In addition, the reactor i1s designed to permit replacement of
the entire graphite core by remote means if required.

Figure 6.6 shows a cross section of a fuel cell. Fuel fiuid flows
upward through the small passages and dowmward through the large central
passage. The outside diameter of a fuel cell tube is 3.5 in.; there are
534 of these tubes spaced on a 4.8-in. trisngular pitch. The tube assem-
blies are surrounded by hexagonal blocks of moderator graphite with blanket
salt filling the interstices. The nominal core composition is 75% graphite,
18% fuel salt, and 7% blanket salt by volume.

A gummary of parameter values chosen for the MSBR design is given
in Table 6.1.

Fuel Processing

 

The primary objectives of fuel processing are to purify and recycle
fissile and carrier components and to minimize fissile inventory while
holding losses to a low value. The [luoride volatility—vacuum distilla-
tion process fulfills these obJjectives through simple operations.

The core fuel is conveniently processed by fluocride volatility and
vacuum distillation. DBlanket processing is accomplished by fluoride
volatility alone, and the processing cycle time is short enough to main-
tain a very low concentration of [issile material, The effluent UFg is
absorbed by fuel salt and reduced to UF, by treatment with hydrogen to
reconstitute a fuel-galt mixture of the desired composition.

Molten-salt reactors are inherently sulted to the design of process-
ing facilities integral with the reactor plant; these facilities require
only a small amount of cell space adjacent to the reactor cell. Because
all services and equipment avallable to the reactor are available to the
rrocessing plant and because shipping and storage charges are eliminated,
integral processing facilities permit significant savings in capital and
operating costs. Also, the processing plant inventory of fissile material
is greatly reduced, resulting in low fTuel inventory charges and improved
fuel utilization characteristics for the reactor.

The principal steps in core and blanket stream processing of the
MSER are shown in Fig. 6.7. A small side stream of each fluid is con-
tinuously withdrawn from the Tuel and blanket circulating loops and is
circulated through the processing system. After processing, the decon-
taminated fluids are returned to the reactor at some convenient point —
for example, via the fuel and fertile stream storage tanks.

Fuel inventories retained in the processing plant are estimated to
be aboulb lO% of the reactor system inventory for core processing and less
than l% for blanket processing.
178

ORNL-DWG 66-4788

  
  

FUEL PASSAGE (DOWN)
3Y%-in. OD FUEL TUBE

MODERATOR (GRAPHITE )
FUEL PASSAGE (UP)

—-—MODERATOR HQLD DOWN NUT
(GRAPHITE)

 

 

 

 

REACTOR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P:_;’- . 6:
r J,//T‘ SPACER
L L ,_L/~f—_—' “““ “METAL TO GRAPHITE SLIP-JOINT
i%jjr’ fifi/e#”‘""' METAL TO GRAPHITE BRAZED JOINT

 

 

 

[ E.u H =TT -BRAZED JOINT

 

 

  

 

 

 

221z2)

\ FUEL INLET PLENUM

 

 

 

 

 

 

 

 

~-w FUEL QUTLET PLENUM

 

=

 

77 ]

Fig. 6.6. Cross Section of a Fuel Cell.
 

S

SORBERS /
e

 

~ COLD
e FR‘AP
ey

ORNL-DWG 65-6194 3

 

 

 

IOO 400

 

 

NaF /MgF, /F2

NGF/;/ // MoF‘/ /—/70" //
g/ 7 /44 {_/4?

  
  
  

 

EXCESS

(LS

C’F\'ODJCTION

/‘
4

 

UF, + <
VOLATILE FP
MAKE UP
LiF/Bes,/ThE,

<
%

 

 

 

FERTILE /
MAKE up/

/ CONTINUOUS
. FLUORIDE
VOI_MI_.TY

 

/ %501;//

/ -

 

 

LiF/BeF, / Tni, /FP

[~
DiSCARD FOR
FP REMOVAL
s

;o |

 

FERTILE STREAM RECYCLE

 

 

]

Fig.

6.7,

    
  

MSBR Core

 

 

 

 

 

  

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

UR RECYCLE TO REACTOR
y SORBEDS
70 //’
//// o ///
N e
a VL MygF, S =-70°C
o
2 Mgra
100-200°C /E// v
WASTE
STORAGE
NoF/MgF,/ o
UF, +
VOLATILE FP MAKE U
LiF/BeF,
v
o o
POLDJP//J CONTINUOUS |/ i |/ /
FOR 7t FLUORIDE // VAZUUM / DISTILLATE | ", U= UR ///
SPENT . :D DC\;AY//VOLATI._ATY 7 // DisT |=_;_ATfO'\|/ ~ LJi-/BeFZ/// /t\(’ED\JC~I ION / FlLTP\AT]ON
71 ~B50°C /// 1000°c/ ~500 ’// 550- 600°C/////
U+, =UF, jflmmhg //// S /
o) i’ / A
h &
LiF + RARE —H,
EARTH FP -
REDUCED METALS
Cr, Fe, N

 

 

LiF/Befy/UFq RECYCLE

and Blaunket Stream Processing Scheme.

 

641
180

Table 6.1.. Parameter Values of MSBR Design

 

Power, Mw
Thermal 2220
Blectrical 1000
Thermal efficiency 0.45
Plant factor 0.80
Dimensions, ft
Core height 12.5
Core diametler 10.0
Blanket thickness
Radial 1.5
Axial 2.0
Reflector thickness 0.25
Volumes, £t>
Core 982
Blanket 1120
Volume fractions
Core
Fuel salt 0.169
Fertile salt 0.0735
Moderator Q.7575
Blanket
Fertile salt 1.0
Salt volumes, £t°
Fuel
Core 166
Blanket 26
Plenums 147
Heat exchanger and piping 345
Processing 33
Total 717
Fertile
Core 72
Blanket 1121
Heat exchanger and piping 100
Storage (protactinium.decay) 2066
Processing 24
Total 3383
Salt compositions, mole %
Fuel
LiF 63.6
Bel'o 36.2
UF, (fissile) 0.22
Fertile
LiF 71.0
Bel's 2.0
Thi'y, 27.0

Ul (fissile) 0.0005
181

Tmne6J,hmmflmmd)

 

Core atom ratios

Th/U 41,7
c/u 5800
Fissile inventory, kg 769
Fertile inventory, thousands of kilograms 260

Processing by fluoride volatility

Fuel Fertile
Stream stream
Cycle time, days 47 23
Rate, ©t2/day 14.5 14
Unit processing cost, $/ft3 183 6.85

 

Heat Exchange and Steam Systems

 

The structural material for all components contacted by molten salt
in the fuel, blanket, and coolant systems, inecluding the reactor vessel,
pumps, heat exchangers, and piping and storage tanks, is Hastelloy N.

The primary heat exchangers are of the tube-~and-shell type. Hach
shell contains two concentric tube bundles connected in series and at-
tached to fixed tube sheets. The fuel salt flows downward in the outer
section of tubes, enters a plenum at the bottom of the exchanger, and
then flows upward to the pump through the center section of tubes. hn-
tering at the top, the coolant salt flows on the baffled shell side of
the exchanger down the central core, uander the barrier that separates the
two sections, and up the outer annular section.

Since a large temperature difference exists in the two tube sec-
tions, the tube sheets at the bottom of the exchanger are not attached
to the shell. The design permits differential tube growth between the
two sections without creating troublesome stress problems. To accom~
plish this, the tube sheets are connected at the bottom of the exchanger
by a bellows-type joint. This arrangement, essentially a floaling plenum,
permits enough relative motion between the central and outer tube sheets
to compensate for differential tube growth withoubt creating intolerable
stresses in the joint, the tubes, or the pump.

The blanket heat exchangers increase the temperature of the coolant
eaving the primary core heat exchangers. BSince the coolant-salt temper-
ature rise through the blanket exchangers is small and the flow rate is
relatively high, the exchangers are designed for a single shell-gide pass
for the coolant salt, although two-pass flow is retained for the blanket
salt in the tubes. Straight tubes with two tube sheets are used.

The superheater is a U-tube U-shell exchanger using disk and donut
baffles with varying spacing. It is a long, slender exchanger having
182

relatively large baffle spacing. The baffle spacing is established by
the shell-gide pressure drop and by the tewmperature gradient across the
tube wall and is greatest in the central portion of the exchanger, where
the temperature difference between the flulds is high. The gsupercritical
fluid enters the tuve side of the superheater at 700°F and 3800 psi and
leaves at 1000°F and 3600 psi.

The reheaters transfer energy from the coolant salt to the working
Tfluid before its use 1n the intermediate-pressure turbine. A shell-tube
exchanger is used, producing steam at 1000°F and 540 psi.

Since the freezing temperature of {he secondary salt coolant is about
700°F, a high working-fluid inlet temperature is reguired. Preheaters,
along with prime fluid, are used in raising the tempersture of the work-
ing fluid entering the superheaters. Prime fluid goes through a preheater
exchanger and leaves at a pressure of 3550 psi and a temperature of about
g70°F. It is then injected into the feedwater in a mixing tee, producing
fluid at 700°F and 3500 psi. The pressure is then increased to about 3800
psi by a pressurizer (feedwater pump) before the fiuid enters the super-
heater.

Capital Cost Hstimates

 

Reactor Power Plant

 

Preliminary estimates of the capital cost of a 1000-Mw (electrical)
MSBR power station indicabe a direct construction cost of about $80.4
miliion. After applying the indirect cost factors used in the advanced
converter evaluation,l an esbimated total plant cost of $113.6 million
is obtained. A summary of plant costs is given in Table 6.2. The con-
ceptual design was not sufficiently detailed to permit a completely re-
1iable estimate; however, the design and estimates were studied thoroughly
enough to make meaningiul comparisons with previous converter-reactor plant
cost studies. The relatively low capital cost estimabe obtained results
from the small physical size of the MSBR and the simple control require-
ments. The results of the study encourage the belief that the cost of
an MGBR power station will be as low as for stations utilizing other re-
actor concepls.

The operating and maintenance costs of the MSBR were not estimated.
Baged on the ground rules used in ref. 1, these costs would be about 0.3
mill/kvhr (electrical).

uel Recyecle Plant

 

The capital costs associated with fuel recycle equipment were ob-
tained by itemizing and costing the major process eguipment required and
by estimating the costs of site, bulldings, instrumentation, waste dis-
posal, and building services associated with fuel recycle.
183

Table 6.2. Preliminary Cost-Estimate Summarya for a
1000-Mw (Blectrical) MSBR Power Station

 

 

Federal Power Commission Account (thousandgozgsdollars)
20 Land and land rightsb 360
21 Structures and improvementis
211 Ground improvements 866
212 Buildings and structgres
.1 Reactor building 4,181
.2 Turbine building, auxiliary building, 2,832
and feedwalter heater space
.3 Offices, shops, and laboratories 1,160
4 Waste disposal building 150
.5 BStack 76
.6 Warehouse 40
.7 Miscellaneous 30
Subtotal account 212 8,469
Total account 21 9,335

22 Reactor plant equipment
221 Reactor equipment

.1 Reactor vessel 1,610
.2 Control rods 250
.3 Shielding and containment 1,477
4 Heating-cooling systems and 1,200
vapor-suppression system

.5 Moderator and reflector 1,089
.6 Reactor plant crane 265
Subtotal account 221 5,801

222 Heat transfer systems
.1 Reactor coolant system 6,732
.2 Intermediate cooling system 1,947
.3 Steam generator and reheatersg 92,853
.4 Coolant supply and treatment™ 300
.5 Coolant salt inventory 354
Subtotal account 222 19,186
223 Nuclear fuel handling and storage 1,700

(drain tanks)
225 Radioactive waste treatment and disposal 450
(off—gas system)

226 Instrumentation and controls 4,500
227 Feedwater supply and treatment 4,051
228 Steam, condensate, and FW piping 4,069
229 Other reactor plant eguipment 52000e

(remote maintenance)
Total account 22 bty , B4T
184

Tsble 6.2 (continued)

 

23 Turbine-generator units

231 Turbine-generator units 19,174
232 Cilrculating water system 1,243
233 Condensers and auxiliaries 1,690
234 Central lube oil system 80
235 Turbine plant instrumentation 25
236 Turbine plant piping 220t
237 Auxiliary equipment for generator 66
238 Other turbine plant equipment 25
Total account 23 22,523
24 Accessory electrical
241  Switchgear, main and station service 550
242 Bwitchboards 128
243 Station service transformers 169
244 Auxiliary generatox 50
245 Distributed items 2,000
Total account 24 2,897
25 Miscellaneous 800
Total direct construction cost?® 80,402
Total indirect costs 33,181
Total plant cost 113,583

 

YEstimates are based on 1966 costs, assuming an established molten-
salt nuclear power plant industry.

b . . . ;
Land costs are not included in total direct construction costs.

°MSER containment cost is ineluded in account 221.3.
Yssumed as $300,000 on the basis of MSRE experience.

“The ample MGBR allowance for remote mainbtenance may be too high, and
some of the included replacement equipment allowances could more logically
be classified as operating expenses rather than first capital costs.

fBased_ on Bull Run plant cost of $160,000 plus ~37% for uncertainties.

®Does not include account 20, land costs. This is included in the in-
direct cosis.

Table 0.3 summarizes the direct construction costs, the indirect
costs, and total costs associated with the integrated processing facility
having approximately the required capacity.

The operating and maintenance costs for the fuel recycle facility
inelude labor, labor overhead, chemicals, utilities, and maintenance
materials. The tobal annual cost for the capacity considered here (15
62 of fuel salt per day and 105 712 of fertile salt per day) is esti-~
mated to be $721,230, which is equivalent to about 0.1 will/kwhr (elec-
trical).® A breakdown of these charges is given in Teble 6.4.
185

Table 6.3. Summary of Processing-FPlant Costs

1000-Myw {Electrical) MSBR

 

 

Processing-FPlant Expendituresg Costs
Installed process equipment $ 853,760
Structures and improvements 556,770
Waste storage 387,970
Process piping 155,800
Process instrumentation 272,100
Electrical auxiliaries 84,300
Sampling commections 20,000
Service and ubtility piping 128,060
Insulation 50,510

Radiation monitoring
Total direct costs
Construction overhead (30% of direct costs)
Total construction cost

Engineering and inspection (25% of total
construction cost)

Ssubtotal plant cost
Conbingency (25% of subtotal plant cost)
Total plant cost

100,000
$2,609,270

782,780
3,392,050

848,010
4,240,060

1,060,020
$5,300,080

 

Table 6.4. Summary of Operating and Malntenance Charges
for Fuel Recycle in a 1000-Mw (BElectrical) MSBR

 

Operation and Maintenance

Annual Charges

 

Expenditures

Direct labor $222,000
Labor overhead 177,600
Chemicals 14,640
Waste containers 28,270
Utilities 80, 300
Maintenance materials

Site 2,500

Services and utilities 35,880

Process equipment 160, 040

Total annual charges $721,230

 
186

Nuclear Performance and Fuel Cycle Analyses

 

The fuel cycle cost and the fuel yield are closely related, yet
independent in the sense that two nuclear designs can have similar costs
but significantly different yields. The objective of the nuclear design
calculations was primarily to find the conditions that gave the lowest
fuel cycle cost, and then, without appreciably increasing this cost, the
highest fTuel yield.

Analysis Procedures

 

Calculation Method. The caleculations were performed with OPIIMERC,
a combinaticon of an optimization code with the MERC multigroup, diffusion,
equilibrium reactor code. The program MERC? calculates the nuclear per-
Tormance, the equilibrium concentrations of the variocus nuclides, in-
cluding fission products, and the fuel cycle cost for a given set of con-
ditions., OPTIMERC permits wp to 20 reactor parameters to be varied, within
limits, in order to determine an optimum, by the method of steepest ascent.
The designs were opbimized essentially for minimum fuel cycle cost, with
lesser welght given to maximizing the annual fuel yield. Typical param-~
eters varied were the reactor dimensions, blanket thickness, fracticns
of Tuel and fertile salts in the core, and fuel and Tertile stream proc-
esging rates.

 

beveral equations were included in the code for approximating cer-
tain capital and operating costs that vary with the design parameters
(e.g,, capital cost of the reactor vessel, which varies with the reactor
dimensions). These costs were automatically added to the fuel cycle cost
in the optimization routine so that the optimization search would take
into account all known economic factors. However, only the fuel cycle
cost itself 1s reported in the results.

Modified GAM-1-[HERMOS cross-section libraries were used to compute
the broad group cross sections for these calculations. It was assumed
that all nuclides in the reactor system are at their equilibrium con-
centrations. To check this assumption, a typical reactor design was
examined o determine the operating time required for the various uranium
isotopes to approach their equilibrium concentrations from a startup with
235y, It was found that 222U and 227U were within 95% of their equilibrium
concentrations in less than two years. Uranium-234 was withia 95% of equi-
librium after elght years, while 2367 was within 80% after ten yecars.
oince the breeding performance depends mainly on the ratio of 233y 4o
2357 in the fuel, the equilibrium calculation appears to be a good rep-~
resentation of the lifetime performance of these reactors, even for start-
Up on 2337,

Basic Assumptions

 

Economic. The basic economic assuuptions employed in the calcula-
tions are given in Table 6.5. The values of the fissile isotopes were
taken from the current AEC price schedule.
187

The procesgssing coghs are based on those given in the seclion entitled
"Capital Cost Estimates" and are included in the fuel cycle costs. The
capital and operating costs were estimated separately for each stream as
a function of plant throughput, based on the volume of salt processed.

The total processing cost is assumed to be a function of the throughput
to some fractional power called the scale Tactor.

Processing. The processing scheme is that indicated in Fig. 6.7.
A Tissile material loss of O.l% per pass through processing was assumed.

In addition to the basic processing scheme employed, results were
also obtained for the case where protactinium can be removed directly
from the blanket stream. The improvement in performance under these
circumstances is a measure of the incenbive to develop protactinium re-
moval ability.

Fission Product Behavior. The disposition of the various fission
products was assumed as shown in Table 6.6. The behavior of 135%e and
other fission gases has a significant influence on nuclear performance.
A gas stripping system is provided to remove these gases from the fuel
salt. However, part of the xenon could diffuse into the moderator graph-
ite. In the calculations reported here, an 135%¢ poison fraction of 0,005
was assunmed.

 

Corrosion Product Behavior. The control of corrosion products in
molten-salt fuels does not appear to be a signiflcant problem, and the
effect of corrosion products was neglected in the nuclear calculations.
The processing method considered here can conbrol corrosion product
buildup in the fuel.

Table 6.5. Basic Economic Assumptions

 

Reactor power, Mw (electrical) 1000
Thermal efficiency, % 45
Load factor .80
Cost assumptions
Value of 222U and 22°Pa, $/g 14
Value of *3°U, $/g 12
Value of thorium, $/kg 12
Value of carrier salt, $/kg 26
Capital charge, annual rave, %
Plant 12
Nondepreciating capital, in- 10

cluding fissile inventory
Processing cost, dollars per
cubic foot of salt

Fuel {at 10 ft3/aa§) 228
Blanket (at 100 f£t2/day) 8.47
Processing cost scale factor 0.4

(exponent )

 
188

Table 6.6. Disposition of Fission Products in MSBR Reactor

and Processing Systems

 

Flements present as gases; assumed to be partly
absorbed by graphite and partly removed by

gas stripping (1/2% polsoning assumed)

Elements which plate out on metal surfaces; as-

sumed to be removed instantaneousl.
5

Elements which form volatile fluorides; assumed

to be removed in the fluoride volatility
process

Elements winich form stable fluorides less vol-

atile than LiF; assumed to be separated by
vacuum distillation

Elements which are not separated from the car-
rier salt; assumed to be removed only by salt

discard

Kr, Xe

Ru, Bh, Pd, Ag, In

e, Br, Nb, Mo, Tec,
Te, T

sr, Y, Ba, la, Ce,
Pr, Na, Pm, Snm,
Eu, G4, Tb

Rb, Cd, Sn, Cs, Zr

 

Table 6.7. MSBR Performance

 

Fuel yield, %/year
Breeding ratic
Fissile losses in processing, atoms/fissile
absorption
Neutron production per fissile absorption,
ne
Specilic inventory, kilograms of fissile
material per Mw (electrical)
Specific power, Mw (thermal) per kilogram
of fissile material
Power density, core average, kw/liter
Gross
In fuel salt

Neubtron flux, core average, 10*% neutrons
e ? sec™t
Thermal
fast
Fast, over 100 kev

Thermal flux factors, core, peak/mean
Radial
Axial

Fraction of fissions in fuel stream

Fraction of fissions in thermal neutron
group

Mean n of 233y

Mean n of 235y

4 .86
1.049
0.0057

2.221

0.769

2.89

80
473

6.7
12.1
3.1

2.22
1.37

0.987
0.806

2.221
1.958

ed

 
189

Nuclear Design Analysis

The important parameters describing the MSBR design are given in
Table 6.,1. Many of the parameters were basically fixed by the ground
rules for the evaluation or by the engineering design. These include
the thermal efficiency, plant factor, capital charge rate, maximum fuel
velocity, size of fuel tubes, processing costs and fissile loss rate,
and the out-of-core fuel inventory. The parameters which were optimized
by OPTIMERC were the reactor dimensions, the power density, the core
composition, including the C/fi and Th/U ratics, and the processing rates.

Nuclear Performance. The results of the caleculations for the MSBR
design are given in Table 6.7, and the neutron balance is given in Table
6.8, The basic design has the inherent advantage of no neutron losses

 

Table 6.8. MSBR Neubron Balance

 

Neutrons per Fissile Absorption

 

Material

 

Absorbed Absorbed Produced
Total by Fission

232, 0.9710 0.0025 0.0059
233p, 0.0079
233y 0.9119 0.8020 2.0233
234y 0.0936 0.0004 0.0010
2335y 0.0881 0.0708 0.1721
2361; 0.0115 0.0001 0.0001
237y 0.0014
238y 0.0009
Carrier salt (except 611) 0.0623 0.0185
614 0.0030
Graphite 0.0300
135ye 0.0050
149gm 0.0069
151qy 0.0018
Other fission products 0.0196
Delayed neutrons lost® 0.0050
Leakageb 0.0012 o L

Total 2.2209 0.8828 2.2209

 

aDelayed neutrons emitlted ocutside the core.

bleakage, including neutrons absorbed in the reflector.
190

to structural materials cother than the moderator. Except for some un-
avoidable loss of delayed neutrons in the external fuel circult, there
is almost zero neutron leakage from the reactor because of the thick
blanket. The neubtron losses o Tission proeoducts are minimized by the
availability of rapid and inexpenslve integrated processing.

Fuel Cycle Cost. The components of the fuel cycle cost for the MoBR
are given in Table ©6.92. The main components are the fissile inventory
and processing costs. The inventory costs are rather rigid for a given
reactor design, since they are largely determined by the assumed external
fuel volume. The processing costs are, of course, a function of the

processing cycle times, one of the chiefl parameters optimized in this

study .

 

SBR Performance with Protactinium Removal Scheme. The ability to
remove protactinium directly from the blanket of the MSER has a marked
effect on fuel yield and fuel cycle cost. This isg due primarily to the
marked decrease in protactinium necutron absorptions when protactinium is
removed from the blaonket region. A simple and inexpensive scheme for the
removal. of protactinium from the blanket would give the MSBR the perfor-
rmance indicated under MSBR (Pa) in Table 6.10; for comparison, the results
without protactinium removal are also given in the table.

 

 

 

 

Table 6.9. Fuel Cycle Cost for MSBR
Costs (mills/kwhr)
Fuel Fertile Total Grand
Stream Stream obe Total
Inventory
Fissile” 0.1180 0.0324 0.1504
Fertile (0.0000 0.0459 0.0459
Salt 0.0146 0.0580 0.0726
Total 0.2690
Replacement
Fertile 0.0000 0.0185 0.0185
Salt 0.0565 0.0217 0.0782
Total 0.0967
Processing 0.1102 0.0411 0.1513
Total 0.1513
Produetion credit 0.0718
Net fuel cycle cost 0.4452

 

a'I:acluding 233Pa, 233U, and 237U,
191

Power Cost and Fuel Utilization Characteristics

 

Based on the above, the power cost, specific fissile inventory, and
fuel doubling time Tor the MSBR and MSBER (Pa) are summarized in Table 6.11.

Table 6.11 illustrates the economic advantage of MSBR's as nuclear
power plants. Also, the fuel utilization characteristics as measured by
the product of the specific inventory and the square of the doubling time?
are excellent. On this basis the MSBR is comparable to a fast breeder
with a specific inventory of 3 kg/Mw (electrical) and s doubling time
of 10.5 years, while the MSBR (Pa) is comparable to the same fast breeder
with a doubling time of 6 years.

Table 6.10, Comparison of MSER Performance With and
Without Protactinium Removal

 

 

MSBR, Without MSBR (Pa),
Protactinium with Protactinium
Process Removal
Fuel yield, %/year 4 .86 7.95
Breeding ratio 1.049 1.071
Fuel cycle cost, mills/kwhr 0.45 0.33
Specific inventory, kg/Mi 0.769 0.681
(electrical
Specific power, Mw (thermal) /kg 2.89 3.26
Neutron production per fissile 2.221 2.227
absorption, mne
Volume fractions, core
Fuel 0.169 0.169
Fertile 0.0745 0.0735
Moderator G.7565 0.7575
oalt volumes, 3
Fuel
Core 166 166
External 547 551
Total 713 717
Fertile
Total 3383 1317
Core atom ratios
Th /U 39.7 41.7
¢/u 5440 5800

 
192

Table 6.11. Power Cost and Fuel Utilization Characteristics
of the MSPR and the MSBR (Pa)

 

 

 

Cost
[mills/kwhr (electrical)]
MSBR MSER (Pa)
Capital cost™ 1.95 1.95
Operating and maintenance costb 0.30 0.30
Fuel cycle cost® 0.45 0.33
Total power costh 2.70 2.58
Specific fissile inventory, Q.77 0.68
kg/Mw (electrical)
Fuel doubling time, years 20.6 12.6

 

a’.]_2% fixed charge rate, 80% load factor, 1000-Mr (electrical)
plant.

b \ . . -
Nominal value used in advanced converter evaluation (see ref. l).

C. . . | . . . .
Costs of on-site integrated processing plant are included in this
value.

References

1. M. W. Rosenthal et al., A Comparative Evaluation of Advanced Con-
verters, ORNL-3686 (January 1965).

2. C. D, Scott and W. L. Carter, Preliminary Design Study of a Con-
tinuous Fluorination—Vacuum Distillation System for Regenerating
Puel and Fertile Streams in a Molten Salt Breeder Reactor, ORNL-
3791 (January 1966).

3. T. W. Kerlin, Jr., et al., The MERC-1 Equilibrium Code, ORNL-TM-847
(Apr. 22, 1964).

4. P. R. Kasten, "Nuclear Fuel Utilization and Eccnomic Incentives,”
paper presented at the American Nuclear Society Meeting, Nov. 1518,
1965, Washington, D.C.
193

7. MOLTEN-SALT REACTOR PROCESSING STUDIES

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. The processing
plant will operate on a side stream withdrawn from the fuel stream, which
circulates through the reactor core and the primary heat exchanger. For
a 1000-Mw (electrical) MSBR, approximately 14.1 ft? of salt per day will
be processed, which will result in a fuel-salt cycle time of approxi-
mately 40 days.t

The probable method for fuel-stream and fertile-stream processing
is shown in Fig. 7.1. The salt will first be contacted with F, for re-
moval of uranium as volatile UFg. Purified UFg will be obtained from the
fluorinator off-gas (consisting of UFg, excess Fp, and volatile fission
product fluorides) by use of NaF sorption. A semicontinuous vacuum
distillation will then be carried out on the remaining salt for the re-
moval of the rare earths, barium, stroontium, and yttrium. These fission
products will be removed from the still in a salt volume equivalent to
0.5% of the stream. A small fraction of salt may also have to be dis-
carded at some stage in the process for removal of fission products such

CRN|-DWG 65-1801R24A

Fp RECYCLE e

 
      
 
 

Eongs
LD MgF,
AP SORBER

8! #tday

NoF SORBER
27 -day CYCLE

100-400¢C

      
 
  

    

BLANKET SALT

SPENT

   

 

 
  
   

 

  

     
 
 

 
 

 

 

 

NaF +MgF, l"’ NaF + MgF,
Ufg F,
LiF - BeF,- UF,
15 f1%/day
WASTE .
FERTILE SALT LiF - BeF,~UF, Lif MAKE-UP -—— UFe
046 #t3duy
‘ LiF+BsF,
PRODUCT
U= Z?jdg/day
WASTE = 0059 f1%/day

 

TLi=078 kg May
FP's

 

 

 

 

Fig. 7.1. MSBER Fuel and Fertile Stream Processing.
194

as zirconium, rubidium, and cesium. The barren salt, the purified U¥g,
and the makeup salt will then be recombined. This step involves re-
duction of UFg to UF,, mixing of these streams, and sparging the result-
ant material with an H,-HF stream. Finally, the salt mixture will be
filtered before return to the reactor.

Semicontinuous Distillation

 

The present concept of the distillation step in the MEBR processing
plant will use a continuous feed stream and vapor removal; however, there
will be a buildup of less volatile fission products (FP) in a static pool
of liguid in the still, with perlodic discard.’ Fission products will be
allowed to build up in the still liquid until the heat generation rate
becomes excesslive, or until the liquid FP concentration becomes too large
for useful decontsmination.

One measure of the decontamination achieved in the distillation
process is the relative volatility of the nonveolatile fission products
as compared with the carrier salt. The relative volatility of a non-
volatile component A compared with a more volatile component B in & mix-
ture of A and B is defined as

Va/%,

0 =
@ 5%, v

where

QhB = relative volatility of A compared with B,
= vapor-phase mole fraction,

x = liguid-phase mole fraction.
To achieve good decontamination from the less volatile P, the relative

volatility must be small. For systems in which the FP concentration is
small, the relative volatility can be approximated by

o 3
which is the Henry's law constant, HA’ used in the expression
Yo = g% - (3)

Thus, determination of the Henry's law constant or relative volatiility
for each of the nonvolatile FP will be sufficient for determining the
size and operating conditions for the distillation step.

The importance of relative volatility in determining the operating

characteristics of a distillation system is shown by the following calcu-
lation. Consider a material balance of an FP in the proposed distillation
195

process. The amount of FP fed into the still per unit time (I ) must
equal the FP leaving the still per unit time (Dy), plus the rate of change
of FP in the still liquid a(vx)/at:

 

a{Vx) (4)

where
x = mole fraction of FP in liquid,
Xo = injet mole fraction of FP,
= megs feed rate in moles/unit time,
= vaporization rate in moles/unit time,
mass of 1liquid holdup in moles,

= mole fraction of FP in vapor,

¢ <N < oS
1

= time.

In the above material balance, vapor noldup is assumed to be negligible
compared with the liguid holdup.

Substituting Eq. (3) into the material balance gives, after rearrange-
ment,

a(vx)
at

Solution of Eq. (5) with the boundary condition

+ DOx = Fxg . (5)

 

x=0att =0
and the condition that the liquid holdup V is consgtant ylelds

X -G /v
X:_O.g.(l..efl/). (6)

Iinally, the average vapor concentration at time €t is given by the re-

lation
v, = % {1~ [i%{' (1 — o-FO/7) ] } _ ()

Figure 7.2 shows the effect of relative volatility on the fraction
of FP retained in the liguid for different values of & as a function of
time for a feed rate of 14.1 £t2 of salt per day and a liquid volume of
4 £43. TFor retention of 90% of the FP introduced to the system during
60 days of continuous distillation, a relative volatility of ~0.001 is
needed.

In order to establish the relative volatility of an FP in the MSBER
carrier salt, 1t is necessary to determine the concentration of the FP
in the vapor at egquilibrium with a known liquid composition. At this
196

pnase of development, exact operating condlitions in the still have not
been set, but the following ranges appear reasonable: temperature, 900-
13100°C; maximum FP liquid concentration, 0.1-1.0 mole %; pressure, vapor
pressure of mixture. Since the MSBR carrier salt is a mixture of T4F and
Bel's, the exact composition of the carrier salt or solvent in the still
liquid will be dependent on operating temperature and pressure; however,
in the temperature range of interest, the major portion of the still
carrier salt will be the less volatile Tit. Therefore, all experimental
tests were made with binary mixtures of LIF and FP fluoride. In later
tests, multlcomponent systems will be used.

ORNL-DWG 66-644A

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a = 0.001
N2 |
|
S
_
-
:(,I:J
lad
I
'_
=
o
L
=
g
w
c
6 \“ii
)
O |
Q _
e _ s
o
o
43)
w
L
L
O
S \~\Q\ZL8;02
— \
(&)
2 N
o
e @ = 0.04
LIQUID VOLUME: a 2 h\“%““w»- “H“"T““
04 b 3 — S eiicii
FEED RATE: 14.1 ft3/day =505
| )
O | ______________
0 10 20 30 40 50 60 70

Fig. 7.2. Fraction of Fission Product Retained and the Still as a
Function of Time for Several Values of Relative Volatility.
197

The FP contaminants of main concern are the rare earths. Of these,
the ones which will be pregent in the largest amount or which will present
the largest neutron losses are Nd, Sm, Pm, Pr, Eu, Ia, and Ce. These
fission products will probably be present as the trifluoride, with the
excepltion of cerium, which might be present partially as the tetravalent
fiuoride.

The equipment used for determining relative volatilities was a simple
equilibrium still with a cold finger in the vapor phase. This still was
constructed from 1l-in. nickel tubing for the liquid and disengagement
space, and S/Swin. tubing for the vapor space through which a 3/8-in.
nickel cold finger was inserted (Fig. 7.3). The top of the vapor section
was cennected to vacuum and inert gas. The entire assembly was placed in
a 5-in.-diam tube furnace. Thermocouples were ingerted in the still
through thermowells din the ligquid phase and at three points in the wvapor
rhase. During a test, the temperature measured at the three lower points
was maintained within 5°C of a predetermined value.

The cold finger could be cooled by air, water, or a combination of
both. This was done by introducing the coolant through a center 1/8-in.-
diam tube and removing 1t through the annular space betwesen the 1/8- and
3/8-iu.~diam.tubes. By using a coolant rate of 0.2 std liter of H,0 per
minute and 11.4 std liters of alr per minute, the tip of the cold finger
could be cooled from the maximum still temperature (1075°C) to less than
the melting point of I4F (847°C) within 2 sec. This rapid cooling would
prevent preferential condensation of the less volatile component during
the cold-finger operation.

The experimental procedure used was to first charge the still with
a known mixture of IiT—FP Tluoride; then, after the still was filled with
an inert gas (purified helium or argon), it was brought to the desired
temperature. At this point the still was subjected to a vacuum pump with
the capability of reducing the pressures to less than 50 p Hg. After the
temperature and pressure reached an apparent steady state, the cold finger
was cooled for 2 to 4 min for collection of a vapor sample. The inert gas
was again introduced, the cold [inger was removed and replaced with a
clean one, and the procedure was repeated.

The solid accumulated on the cold Tinger was scraped off, and it
constituted the equilibrium vapor-phase sample. Approximately 0.01 g
wag collected Tor each sample, and two samples were collected at each
get of conditions. For each still liquid mixture, tests were made at
four temperatures, 925, 975, 1025, and 1075°C.

Experimental tests have been made on all the important rars-earth
fluorides with the exception of promethium. Results are given in Table
7.1. They are generally consistent and compare favorably with results
obtained by Kelly? in work described in the subsection "Evaporative-
Distillation Studies on Molten-Salt Fuel Components" of this report.
198

ORNL—DWG 65— 78B47RA

g—'HQO AND AIR N

N

[%§~fingo AND AIR QUT

<oty
TR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

— || 3uaCOMPRESSION FITTING
| 11 AR DRILLED THROUGH
3/e"‘ln. é H [gzzzg
NICKEL TUBE 1|0 VACUUM AND INERT GAS
s 7
¥l O H
INSULAflON\\Q\ g; s 3/16 —in. STAINLESS STEEL TUBE
‘ f: H
, 48104
12 1in. A; A é [ /.
< A i 1
10|41 TUBE FURNACE
101 |87
i é?wqu# > — THERMOCOUPLE
101160 E§§§§1
\ Y in. A et — 5/g—in. NICKEL TUBE
L R
I B _E§§§§\
/ 'lg 7 1-in. NICKEL TUBE
; ; E 5 |/4i|"l.
RYAERE
I
C
< b
7 S THERMOWELLS
/ 4 N2
, 2
‘ ”
i 9
7 in. ;
| /
,
4
/
~—THERMOWELL
[ -
1
1]
2
]
/

 

 

 

 

 

 

 

AN

 

 

\\\WNSULAHON

Fig. 7.3. BEquilibrium Still with Cold Finger.
199

Table 7.1. Relative Volatilities of Rare-FKarth luorides in
ILithium Fluoride

 

 

 

Rare~Farth Tiguid Mole Average Relative Volatillities

Fluoride Fraction 900°C 950°C 1000°¢C 1050°C
Cely 0.0067 0.133 0.167 0.208
DmE 5 0.01 0.033 0.009
NdF 3 0.01 0.025 0.016
PrF 3 0,001 0.038 0.020 0.014
ks 0,001 0.041 0.037 0.028 0.012
CeF3 0.01 0.043 0.033 0.018
Lafs 0.001 0.035 0.024
LaF3 0.01 0.051 0.027 0.011 0.008

 

Fuel Reconstitution

 

A necessary step in the processing of the MSBR fuel is the recombl-
nation of the purified uranium hexafluoride with the purified carrier
salt, which includes reduction of UFg, the product of the [luorination
step, to UF,. The usual method for reducing Uy to UFy uses excess
hydrogen in an Hp-Fp flame which produces hydrogen flucride ag a by-
product. The resulting UF, powder is collected at the base of a tall
reaction vessel. Although this operation has veen reduced to rouvine
production, it appears undesirable for radiochemical application because
of the inherent solids handling problem. An alternative 1ls the reduction
of Uf'g to UF, in a molten salt, involving only gases and liquids. When
UF; ig contacted with a molten flvoride salt containing UF,, it is absorbed
with reaction to form intermediate fluorides of uranium such as UFs. These
intermediate fluorides can then be reduced to UF, by contacting the salt
with hydrogen. Initial but definitive tests have sghown this alternative
to be quite feasible. A tower which might serve well to conduct this
gequence of reactions is shown in Fig. 7.4.

Questions of feasibility are raised concerning the eguilibriunm
distribution of the various species of uranium fluorides and the rate
at which the reactions proceed. It is believed that the addition of Ulg
to a molten salt containing UM, results in the formation of dissolved
FTluorides of uranium with a valence intermediate between 4+ and 6+. This
pehavior 1s indicated by the fact that quantities of F, sufficlent for
the Formatlon of UFs can be absorbed by molten salt containing UF, with-
cut the evcolution of UFg. Similar behavior 1s also noted in reactions
between UF, and UFg in the absence of molten salt to yield intermediate
200

fluorides such as UgF3,. It could be expected that the homogeneous re~
action rate would be very rapid and that the absorption rate would proba-
bly depend upcn diffusion to and from the interface. Previous data on
the reduction of uranium fluorides intermediate between UF, and Uy in
molten salts do not exist; however, rate data may be inferred from the
reduction of UF% with hydrogen in molten mixtures of IiF and Bel', per-
formed by Long. He observed thal the ratio of the concentrations of
hydrogen and HF in gas bubbles rising through the molten salt reached
equilibrium in only a few inches. His data also indicate only l% re-
duction of UF, to Ul; by a gas stream containing 1% HF in hydrogen at
pressures of 1 atm at 600°C.

The experimental equipment consisted of a reaction vessel in which
molten salt containing UF, could be contacted with a metered stream of
UFg, HF, Hp, or N, and Na¥ traps to collect UFg and/or HF in the off-gas
(Fig. 7.5). Two NaF traps were provided downstream of the vessel; one

ORML~DWG 65— 4794RA

 

 

 

 

 

H, + HF
SALT + UF,
s g FUELL SALT
I “"fi'-'-"-—HE
5
MAKE~ UP 2
SALT o ; UFe
BARRENSALT”“J“”““J——"—££;2:}-—_——}
PUMP

Fig. 7.4. Continuous Reductilon of Ulfg by Hs in a Molten Salt.

ORNL-DWG 65— 30984

Ho No

UFG vNZ

 

- OFF - GAS

    

4 —-in. NICKEL PIPE NaF BEDS

 

 

REDUGCTION VESSEL

Fig. 7.5. Equipment Used in Reduction of U¥g to UF,; 1n a Molten Salt.
201

trap was used only for trapping UFg from the vessel off-gas during UFg
addition to the molten salt, and the other trap was used for all other
HF or UFg absorption. The reduction vessel was constructed from 4-in.-
diam sched-40 nickel pipe and was 26 in. long. A 3/8-in. nickel inlet
line was located in the center of the vessel and terminated 1/4 in. from
the bottom of the vessel. A 3/4—in. fitting on the top flange allowed
the insertion of a cold, 3/8-in. nickel rod, which was used for sampling
the salt. A 3/8-in. off-gas line was connected to the top flange. The
vessel was heated by two Nichrome-wire resistance furnaces.

Three experiments were carried out at 600°C in which UFg was intro-
duced at the rate of 1.5 g/min at a polnt 12 in. below the surface of a
molten IiF-ZrF, mixture containing ~0.5 mole % UF,. The initial salt
charge consisted of 5320 g of ZrF,, 863 g of LiF, and 61.8 g of UFy
(0.197 g-mole of UF,) and had a melting point of approximabely 510°C.
Complete absorption of the UFg was observed during each of the tests,
which resulted in the sgbsorption of a total of 147 g of UFg during a
period of 98 wmin.

During a typical run, the salt charge from the previous run was
heated to 600°C and sparged with Np for 15 min at the rate of 100 cm?/min
(51P), after which a salt sample was taken. The salt was then sparged
with HF at the rate of 0.5 1b/hr for 1 hr and with N, for 15 min, after
which a second salt sample was taken. Uranium hexafluoride was then
bubbled into the salt at a rate of 1.5 g/min for a specified length of
time, with the vessel off-gas passing through an Nal bed used exclusively
during this period. The salt was then sparged with Np for 12 min and
sampled. The salt was then sparged with Hp at the rate of 95 cmB/min
(8TP) for 30 min and sampled, after which the Hp sparge was continued for
an additional 30 min.

Two questions related to the experimental work are of primary in-
terest. These are (1) the fraction of UFg which was absorbed by the
molten salt and (2) the valence of the uranium in the resulting mixture.
It was concluded that, within the accuracy of the experimental data,
complete absorption of the UFg by the molten salt had occurred. TNo
uranium was found on the NaF trap used during the UFg addition period.
The concentration of Uy in the salt sample taken after UpF; addition was
below the limlt of detection of 0.05 wt %. Reductlon of the uranium to
UF, probably occurred during the addition of UFg by the reaction of the
intermediate {luorides with nickel from the vessel wall.

From these tests it appears that Ul could be rapldly absorbed by
molten fluoride salt containing about 1 wt % Uf, at 600°C and thal the
intermediate fluoride formed could be reduced with hydrogen to UF,. Sub-
sequent studies are recommended to provide more guantitative data for
engineering design; however, this form of recombination of UFg with the
purified carrier salt will be indicated on all subsequent [lowsheets.
202

Continuous Fluorination of a Molten Salt

 

Since the presence of uranium in the distillation step to separate
the carrier galt from the Tission products would cause unnecessary compli-
cations, it is removed continuously in a prior fluorination step. Pre-
vious experience with the removal of uranium from molten salt by fluori-
nation includes the operation of the Molten-Salt Fluoride Volatility
Pilot Plant at ORNL.% In this facility, batch fluorinations completely
volatilized the uranium as UFg, which allowed its subsequent purification
and recovery by absorption and cold trapping. Observed corrosion in this
facillity was severe but acceptable in a batch process of this sort. How-
ever, it would be intolerable in a continuous unit, or in any unit with
enough capacity to handle the processing stream for an MSBR. A possible
solution to the corrosion problem is the cperation of the fluorination
vessel, presently envisioned as a tower, with a layer of frozen salt on
the vessel wall. FExperience with this type of system was obtained with
batch fluorinations made at Argonne National Laboratory5 in support of
the molten-salt fluoride volatility process. Successful tesis were also
made at ORNL using ohmic heating to provide the internal heat generation,
where it was found that a gas flow could be maintained through an unheated
line which entered the Tluorinator vessel at a point below the molten-
salt surface when a frozen salt layer was present on the fluorinator
wall. 1In application to the MSBR fluorinator, internal heat generation
will be provided by the fission product decay heat.

Experimental studies of continuous fluorination of molten salt are
being made in a l-Iin.~diam nickel column with a salt depth of 48 in. MNo
provision is being made in the preseanl experimental work for corrosion
protection by a frozen layer of salt (Fig. 7.6). Fluorination tests in
which 15 cmB/min of molten salt (NaF-IiF-ZrF,) containing 0.5 wt % UFg
was contacted countercurrently with 70 cm®/min of F, {STP) at 600°C
showed removal of uranium from the salt at 96 to 92.4% efficiency during
a l-hr period of continuons operation. Material balances were compli-
cated by the inevitable corrosion of the nickel vessel. Complete removal
of uranium from the salt with no corrosion would yield, for the above corn~
ditions, a UFg concentration of 17.56 mole % in the off-gas. Observed con-
centrations ranged as high as 35 mole % UF.

These results indicate that subsequent development can be expected
to produce acceptab.le recoveries of uraniwn by continuous fluorination.

Chromivm Fluoride Trapping

 

At the conclusion of tests on the MSRE, uranium will be recovered
from the fuel salt as Ulg by sparging the salt with F,. Fluorides of
chromium will be present in the fuel salt as a result of corrosion of
reactor plping and of equipment used for hydrofluorination or fluorination
of the salt. A potential problem associated with the recovery of the
uranium is the presence of volatile fluorides of chromiuvm {(Cr¥, and CrFs)
203

ORN(. - DWG 65—-9354A

INFRARED ANALYZER
FOR UFg ANALYSIS

SALT SAMPLING

METERED L

 

 

VESSEL SODA LIME TRAP
{.5=in. diam MaF TRAP 7 kg
e 1.25 kg
METERED Ny FOR
SALT DISPLACEMENT 1t 1
’ ’ b oFF-cas

 

SALT RECEIVER

 

 

 

 

 

 

 

SALT FEED TANK 14 liters
14 liters
NICKEL FLUQRINATOR
1—in. diam
72-in iong

Fig. 7.6. DEguipment for Removal of Uranium from Molten Salt by Con-
tinuous Fluorination.

in the fluorinator off-gas; these fluorides cannot only contaminate the
U product but also render equipment inoperative by deposition in lines,
valves, etc. A study has been completed which will permit the design of
a trapping system for removing these fluorides from the oiff-gas, which
will also contain Ufg and Fso.

Ixperiments were carried out in which 1 1iter/min of Fs (STP) was
sparged through a molten NaF-1iF-ZrF, mixture at 650°C which contained
0.5 to 4 wt % Crf3. The resulbing off-gas contalning fluorine and vola-
tile fluorides of chromium then passed through beds of pelleted Nal' at
400°C for removal of chromium fluorides. In some tests, a UFg flow of
100 em?/min was added to the Fi.

It can be concluded that (1) fixed beds of NaF at 400°C are effective
in removing fluorides of chromium from a gas stream which also contains
UFe and Fo; (2) pelleted NaF having a surface area of 0.074 w?/g and a
void fraction of 0.277 is superior to material having a surface area of
1 m?/g and a void fraction of 0.45, and has an effective capacity of about
20 g of chromium per 100 g of NaF; (3) uranium losses to the 400°C NaF
bed of less than 0.01% are achievable when working with a gas stream that
containg 0.4 mole of CrFs per mole of UFg dn Fp.

Deslgn and Evaluation Study

 

A preliminary design study has been made of a conceptual processing
plant to treat irradiated fuel and fertile streams from the 1000-Mw
(electrical) MSBR described in the section "Molten-Salt Breeder Reactor
204

Design Studies"” of this report. The study evaluated the engineering
Teasibility and costs for a plant that operated continucusly as an inte-
gral part of the reactor system, belng located in two cells adjacent to
the reactor cell. The plant was designed to treat 15 ftB/day of fuel
salt and 105 ft3/day of fertile salt. The fuel salt was an Lil-Bel's
(69-31 mole %) mixture containing the fissionable 272UF,; fertile salt
was a 71-29 mole % mixture of Ii¥~Thf,. The lithium component of each
stream was enriched to about 99.995 at. % 7Ii. The processing cycle was
selected to give the optimum combination of fuel cycle cost and breeding
gain.

Description of Fuel Process

 

The primary obJective of the fuel process is to recover uranlium and
carrier salls sufficiently decontaminated from fission and corrosion
products so that the reactor has an attractive breeding potential. The
recoverad materials are recycled to the reactor, and the fission products
are discarded. Only four major operations are required to accomplish
this for the fuel stream: fluorination, sorption of UFg, vacvum distil-
lation, and salt reconstitution. These operations are shown schematically
in Fig. 7.1.

As it enters the processing cell, fuel salt is only a few seconds
removed from the Tission zone and is extremely radiocactive. The strean
is delayed for about 36 hr before fluorination to allow the heat genera-
tion rate to decrease to a point that temperature control in the fluori-
nator is made easier. The curve in Fig. 7.7 shows the gross heat gener-
ation rate of the fuel salt. The molten salt flows into the top of =a
column anc is contacted by a countercurrent stream of fluorine, which
strips out the uranium according to the reaction

500~550°C
Uy, + Fp —————=> Ul'g .

figsion products Ru, Tc, Nb, Cs, Mo, and Te are also volatilized and ac-
company the Ulg.

The system, consisting of molten LiF-BeF,-UF,, fission products, and
clemental fluorine, is extremely corrosive to the walls of the fluorinator,
requiring clever design if a significant lifetime is to be obtained. It
is proposed to Jacket the fluorinator with a coolant that will maintain
a 0.5- to 0.75-in.-thick layer of frozen sallt on the inner surface of the
column to shield the wall from the molten salt.” A schematic diagram of
the fluorinator is shown in Fig. 7.8.

The gas stream leaving the fluorinator passes through a sorption
system composed of temperature-controlled beds of Nal and Mgl, pellets.
The first section of the Nab bed is held at about 400°C and sorbs most
of the rission products; the secound seection of the bed at about 100°C
e
O
R

GRNL —~DOWG 65 - 9394 RA

|

 

 

 

 

 

 

HEAT GENERATION RATE {8tu/ hr-ft>)

 

 

 

 

 

 

 

10 100
MINUTES o lsgr bt il ]
i 10 i 1iC 20
| HOURS l YEARS
100 | (RN L rre Foi Lot L L Lo oL
1072 1072 o+ 10° 10 102 to” i?

TIME AFTER DISCHARGE FROM REAGCTOR (days)

Fig. 7.7. Fission Product Decay Heat in MSBR Fuel Stream for a 1000-
My (Electrical) Reactor.

sorbs technetium, part of the molybdenum, and UFg, and allows the re-
maining fission products to pass. Upon heating from 100 to 400°C, the
second section of the sorber releases molybdenum, technetium, and UFg,
which passes through Mgls for retention of technetium while allowing UFg
to pass. Uranium hexafluoride is frozen in cold traps and retained for
recycele to the reactor.

Uranium-free salt flows from the fluorinator into a continucus :
distillation unit, which is operated at about 1 mm Hg pressure and 1000°C.
Under these conditions, it is possible to distill IiF and Bel¥, from the
bulk of the fissicn products.6 Rare-earth fission products are much less
volatile than lithium or beryllium fluoride, allowing a good separation
to be achieved. Zirconium fluoride, however, is sufficlently wvolatile
that this fission product will contaminate the Lif-Bel's product.

In this study the vacuuwn s5till was a 2.5-ft-diam by 4-ft-high vessel
containing a bank of cooling tubes over most of its height. A condensing
surface at the top condensed and collected the overhead product. To
initiate the operation the interior of the still is charged with 4 £t of
molten ILiF; the still is evacuated and brought to temperature, and salt
from the fluorinator is allowed to flow into the pool of molten LiF.
Temperalure 1s controlled so that liguid is vaporized at the same rate at
which 1t enters the still. There is no bottom discharge, so the still
volume remains constant. Accordingly, the concentration of fission
ORNL—-DWG 65—-3037TRA

f}sz,UFG, F.P. TO SORPTION SYSTEM

  
  
     
   

DE-ENTRAINMENT SECTION

COOLANT

 

FUEL

FROZEN
WALL OF SALT —

1/2 TO 3/4 in. THICK

GRAVITY

 

 

 

%\‘“‘FUEL SALT
DRAIN

 

 

 

NaK COOLANT::;b( =
&

Fa

Fig. 7.8. Continuous #luorinator with Frozen Salt Wall for Cor-
rosion Protection.
207

products steadily increases in the 4 ft? of LiF. After about 67 days®
operation the accumulated heat generation rate (see Fig. 7.9) has become
so great that the heat removal capablility of the cooling system is
reached; the still contents are then drained to waste storage, and the
operation 1s repeated. Heat 1s removed by forced circulation of NaK on
the shell side of the tubes.

The concentration factor for rare~earth fission products 1ln the
gtil1l is about 250. The fraction of the process stream, which 1s almost
entirely “LiF, discarded at this point is slightly lesas than 0.4%. Be-
cause of the volatility of Zrf., an additional discard of the distiliate
ig required to purge this fission product. As much as a 5% throwaway
might be necessary in this type of operation. At the time of this szstudy,
data were not available to assess the effect of increasing fission
product concentration in the stlll on relative volatilities. Conse~
guently, the overall decontamination factor (DF') of the distillate cannot
be predicted accurately, but it is believed that a DF of at least 100 can
ve attained.

ORNL~DWG €65 ~ 1821 R2A

1 O6 _________ e L L LA .

HEAT GENERATION RATE (Btu/hr)

 

0 AT 1o’ 10? 10*
255 days  TIME AFTER DISGHARGE FROM REACTOR (days)

Fig. 7.9. Healt Generation Rate in the LiF Pool Resulting from
Fission Product Accumilation in the Still.
208

The final step in fuel processing is reconstitution to make a
suitable feed for the reactor. The LiF-BeF, distillate 1s admitted to
a reduction column containing molten (~600°C) IiF-BeF,-UF, that is
approximately the correct fuel composition. Concurrently, gaseous UFg
from the cold traps is introduced near the bottom of the column, and
hydrogen gas is admitlted at a point a little farther up the column. The
UFg absorovs in the molten salt to form an intermediate Tluoride of
nranium such as UFs, which reacts with H; according to the reaction

UF's +~%H2-——-——-——9’ UF, + 1F .

Makeup UFg from the blanket process and makeup IiF and BeF; are added at

this point. The reconstituted fuel is sent to the reactor core to com-
plete the fuel processing cycle.

Degcription of Tertile Process

 

The fertile stream process consists only of continuous fluorination
and UFg purification by sorption. The operation is analogous Lo the
corresponding fuel stream operatiocn but at a higher volumetric rate. The
cycle time of the fertile stream is purposely kept short (20 to 25 days)
to keep a low uraniuvm concentration in the blanket, thereby keeping the
fission rate low. The low fission rate cnsures a low fission product
accumuiation rate so that it is unnecessary to remove them on the same
cycle as uranium. In fact, a 30-year discard cycle of the barren fertile
stream is a sufficient purge rate for Tisslon products.

Excess UlFg over thal required to refuel the core 1is sold.

Waste Treatment

 

Four waste streams requiring storage leave the processing facility:
(1) aqueous waste from the KOH scrubber, (2) NaF and MgF, sorbent from
the Ul'g purification system, (3) molten-salt residue from the distillation
unit, and (4) molten salt from the fertile-stream discard. The aquecus
waste comes from vent-gas scrubbing and is small in volume; it was as-
sumed that this stream could be combined with reactor agueocus wastes for
storage. The Uwo molten-salt wastes are stored in underground tanks, and
the pelletized sorbents are stored in cylindrical containers in an under-
ground vault. Forced draft cooling is provided for these three storage
areas.

This study includes a charge for 30-year interim storage of the
molten-salt wastes and for 5-year interim storage of the solld waste.
Perpetual storage beyond these times was not considered.
209

Off-Gas Treatment

 

Most of the off-gas from the process comes from the continuous
fluorinators. Although fluorine is recycled, a small amount is bled
of f to purge gaseous Tission products. The off-gas is scrubbed with
an agueous caustic solution, filtered, and discharged 0 the atmosphere.

cummary of Capital and Operating Costs

 

The desglign study included an estimation of capital and operating
costs for the integrated processing plant. Space requirements and costs
were estimated Tor a typical layout (Fig. 7.10) adjacent to the reactor
system. FBach item of major equipment was designed to the extent that a
reascnably accurate estimate of its cost could be made; the costs of
auxiliary items, such az utilitles, piping, instrumentation, electrical
connections, insulation, and sampling, were estimated by applying appro-
priate factors to process equipment costs.

Direct operating costs were estimated for labor and supervision,
e

consumed chemicals, utilitlies, and maintenance materials. These costs
and capital costs are summarized in Table 7.2.

ORNL — DW5 85 ~41804RA

%

 

 

 
 
  

 

 

 

 
  

"SORBERS
Naf - Mgf, WASTE

Lfi?@??@YlfiNfi
\ TWASTE '
RECEIVER
FLUORINATOR

 
  

 

 
  

 

------

 

 

 
 

 

 

     

 

 

 

 

HEAT EXCHANGER

 

REACTOR POWER
GENERATION SYSTEM

 

REACTOR

69 ft

 

 

 

 

 

 

 

 

 

   

\

SFERTILE MAKE-UP

   

FLUORINATOR
— SORBERS
—COLD TRAPS

Tk A T L e I ek

   

 

   

 

 

 

 

 

   

Fig. 7.10. Reactor Integrated Processing Plant FPreliminary Layoutb.

 
210

Processing Cost

 

The costs summarized in Table 7.2 contribute about 0.2 mill/kwhr to
the fuel cycle cost when the fixed charges are amortized at lE%/year and
the plant factor is taken at 80%. The amortization charge includes
10%/year for depreciztion, 1%/year for taxes, and 1%/year for insurance.

Table 7.2. Cost of an Integrated Processing Plaant for a
1000-Mw (Electrical) Molten-Salt Breeder Reactor

 

Fixed Capital Costs ($)

Building space 1,130,200
Process equipment 1,734,200
Interim waste storage 785,100
Services and utilities 1,648,300

Total 5,301,500

Taventory Costs® (§)

Fuel salt carrier 89,460
Fertile salt £9,200
NaX coolant 40,000

Total 198,660

Direct Operating Costs ($/year)

Supervision and labor 399,600
Chemicals 70,390
Waste containers 28,270
Utilities 80,300
Maintenance materials 209,230

Total 787,790

 

a . . .
Ixcludes fissile material.
211

References

C. D. Scott and W. L. Carter, Preliminary Design Study of a Con~
tinuous Fluorination—Vacuum Distillation System for Regenerating
Fuel and Fertile Streams in a Molten Salt Breeder Reactor, ORNI-3791
(January 1966).

M. J. Kelly, ORNL, unpublished data (May 21, 1965).

G. Long, Stebility of UFs (in preparation).

 

R. P. Milford et al., Ind. Eng. Chem. 53, 357 (1961).

 

R. W. Kessie et al., Process Vessel Design for Frozen-Wall Contain-
ment of Fused Salt, ANI-6377 (1961).

 

M. J. Kelly, Removal of Rare Earth Flssion Products from Molten Saltb
Reactor Puels by Distillation, paper presented at 1lth annual meeting

 

of American Nuclear Society, Gatlinburg, Tenn. (June 21=24, 1965).
 
213

OAK RIDGE NATIONAL LABORATORY
MOLTEN-SALT REACTOR PROGRAM

FEBRUARY i, 1966

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R. B. BRIGGS, DIRECTOR D
P. R. KASTEN,* DEPUTY DIRECTOR R
W. B. MCDONALD,** ASST. IRECTOR R
MSBR STUDIES MSRE OPERATIONS COMPONENT AND 5YSTEMS DEVELOPMERT J8C DESIGN AND DEYELOPMENT REACTOR CHEMISTRY METALLURGY
PAUL R. KASTEN* R ‘ i W, R. GRIMES* RC A, TABOADA" M&.C
- R.L. M
W, 3. M DONALD: ¢ R P. N, HAUBENREICH R D. SCOTT R L. MOORE 18C Bl ANKENSHIP® RC
£ G BOHLMANN® RC D. A. CANIOCO™ MaC
H F. McDUEFIE* ac W, H. CO0K** MAC
. <DUFFIE " W. W. DAVIS® MAC
i R. G. DONNELLY* MAC
MSRE ON-SITE CHEMISTRY . JONES” wae
R. E. THOMA** RC C. R. KENNEDY" M&C
REACTOR DESIGN METALLURGY COORDINATION NUCLEAR AND MECHANICAL ANALYSIS ENGINEERING ANALYSIS A. H. ANDERSON 18C 5 5. XIRSLIS ™" RC W. R, MARTIN® M&C
5. 1. BALL™ 1&C 2. D. NEUMANN RC H. E. MCCOY” M&C
E. 5. BETTIS R A, TABOADA® M&C €. H. GABBARD"" R J. R. ENGEL R R. 1. KEDL* R D. 6. DAVIS 18C G. M. TOLSOR” M&C
- , . C. H. GABBARD" " R s HERN . J. L. GRIFFITH* M&C
Ho F. BAUMAN R W. H. COOK MAC P. G. HERMDON 1&C {RRADIATION PROGRAM - L .
€. £. BETTIS” GEAC C. R. KENNEDY MEC H. B. PIPER R J. W. KREWSON &G A E. J. LAWRENCE M&C
R. ). BRAATZ" GE&C H T MCCOYT MRC 1. A WATTS R T M. CATE e 1 W BAKER ac
W, L. CARTER® T . B. J. JONES* 1&C -
G, M. TOLSON M&C -
G. A, CRISTY~ GE&C P. E. SMITH 1&C S' JL‘IEEO]TAS;R‘ :E
A. G. GRINDEELL™* R C. £, STEVENSON " 1&C M C. SAVAGE RC
T. W. PICKEL™ GE&C .C. 3E
R. C. ROBERTSON™ R OPERATION REMOTE MAINTENANCE 1. R. HART RC
0. SCOTT R ] 1. ¥. MYERS RC
’j‘ S aPEé#;i?‘ GE&S R. j fiu;gngEY g IMSTRUMENTATION AND CONTROLS R. BLUMBERG 2
- H. - K. : J. R. SHUGART R Y3
Ly WILSON R CHEMISTRY B RARLEY R 1 8. TALLACKSON 18C ; PHYSICAL AND INORGANIC CHEMIST RY
W, TERRY R - AMALYTICAL CHEMISTRY V. D, HOLT R F. K. ADAMS* 18C \
C. JONSS* R T. L. HUDSON R G.H. BURGER 18C o r Sars. IR i
W. KRICK" R J. C. WH:TE* AC A, | KRAKGYIAK R ©.N. FRAY® 18 o M. BURNS
H. R. PAYNE R C. D MARTINT &C b E{AEJ?EFER*W sg
REACTOR CHEMISTRY H, C. ROLLER R C. B, MATHEWS 1&C | J. 1. _J \
R, STEFFY ] 2. L. REDFORD 18C NUCLEAR ANALYSIS R.E. THOMA' - RC
W. R. GRIMES* RC W, C. ULRICH R J, E. SCHMITH™ 1&C COMPONENTS AND LOOPS B. E. PRINCE R G. BRUNTOM RC
H. 7. McDUFFIE” RC T. ARNWINE R R. W, TUCKER 1&C T W KERLIN" Q F. A DOSS RC
i ) - s E rLLISn i, FRiEDMAN RC
C. 5. BARTON® RC W. H. DUCKWORTH R 8 E.ELLIS 18C H. A, FRIE
y P . A, N, SMITH R HEBERT RC
: . F. BLANKENSHIP* 1, 0. EMCH R C. K. KiIRKWOOD 13C ~ G. Mt
CAPITAL COST F CFAS;O;;“ ;g i GUINN R T e LETELLIERT 12 R. B. CALLAHER R K. A. RCMBERGER RC
— > R, T el " D. D. RAY™ 18C M. RICHARDSON R 0. R. SEARS RC
R. C. OLSON* R 5. 5. KIRSLIS RC LGUHILL s T R. E. CARNES R L
e ~ccper SLC HURTT R V. R.SILER” 1&C - e H. H. STONE RC
M. L. MYERS™ R J.H. SHAFFER RC s C. A. GIEFGRD R
R ~ . R. E. THOMA® " RC J. C. JORDAN R R. #ARD™ 1&C - A . W. T. WARD RC
. C. ROBERTSON R i1 MILLER R C. JONES R W. K. R. FINNELL RE
L. E. PENTON R 8. F.HITCH <
W. E. RAMSEY R W. P, TEICHERT C
R. SMITH, JIR. R
J. L. STEPP R FUEL FROCESSIMG DEVELOPMENT
J. L. UNDERWOOD R I
5. R, WEST R ! CHEMISTRY T M. E. WHATLEY? cT
) E.WOLFE R - i
! R.E. THOMA*! RC ! W, L. CARTER™ cT
; J. R HIGHTOWER cT
AC AMALYTICAL CHEMISTRY DIVISION ! R. B, LINDAUER ™ 7
CT CHEMICAL TECHNOLOGY DIVISION PUMP DESIGN AND DEVELOPMENT L. £. McNEESE cT r
. T L C. £ SCHLLING cT
D D|RECTORS» DIVISICN A. G. GRINDELL " R . 0. 5C0TT CT E ANALYTICAL CHEMISTRY
GZ&C GENERAL ENGINEERING AND CONSTRUCTION DIVISION DESIGH LIAISON o o omith N S BEAMS it i o
1&C IKSTRUMENTATION AND CONTROLS DIVISION ] L. V. WILSON * R €. J. SHIPMAN cT Ijg wmii-l‘N :g
MRC METALS AND CERAMICS DIVISIDN €. K. McGLOTHLAN R CHEMICAL PROCESSING : W. G. SISSON: T o
- R. F. APPLE AC
R REACTOR DIVISION R. 8. LINDAUER"" R J. M. DALE AC
RC REACTOR CHEMISTRY DIVISION £. K. HEACKER AC
" PART TIME ON MSRP C. E. LAMB* AC
v DUAL CAPACIT | W, L.MADDOX" AC
v Y i D.L.MANNING AC
i
MAINTENANCE
B. H. WEBSTER R MATERIALS
A. F. CILLEN R fl
L. P. PUGH R W, H. COOK** M&C

 

 

 

 

 

 
215

ORNL-393%6
UC-80 — Reactor Technology

INTERNAL DISTRIBUTION

1. G. M. Adamson 57. A. G. Grindell
2. L. G. Alexander 58. R. H. Guymon
3. C. F. Baes 59. P. H. Harley
4. J. M. Baker 60. C. S. Harrill
5. 8. E. Beall 6l. P. N. Haubenreich
6. E. S. Bettis 62. P. G. Herndon
7. D. 8. Billington 63. R. F. Hivbs (Y-12)
8. F. F. Blankenship 64. M. R. Hill
9. R. Blumberg 65. E. C. Hise
10. H. F. Bauman 66. H. W. Hoffman
11. A. L. Boch 67. V. D. Holt
12. E. G. Bohlmann 68. P. P. Holz
13. C. J. Borkowski 69. A. Hollaender
14. G. E. Boyd 70. A. 5. Householder
15. M. A. Bredig 71. T. L. Hudson
16. E. J. Breeding 72. H. Inouye
17-26. R. B. Briggs 73. W. H. Jordan
27. F. R. Bruce 74-83. P. R. Kasten
28. G. H. Burger 84. R. J. Kedl
29. 8. Cantor 85. M. T. Kelley
30. D. W. Cardwell 86. M. J. Kelly
31. W. L. Carter g7. C. R. Kennedy
32. E. L. Compere 88. T. W. Kerlin
33. J. A. Conlin 89. H. T. Kerr
34. W. H. Cook 90. A. I. Kraskoviak
35. L. T. Corbin 91. J. W. Krewson
36. G. A. Cristy 92. C. E. Lamb
37. J. L. Crowley 93. C. E. Larson
38. F. L. Culler 9%. T. A. Lincoln
39. J. M. Dale 95. R. B. Lindauer
40. D. G. Davis 96. R. S. Livingston
41, W. W. Davis 97. M. I. TLundin
42. J. H. DeVan 9¢. H. G. MacPherson
43, R. G. Donnelly 99. F. C. Maienschein
44. D. A. Douglas 100. W. R. Martin
45, N. BE. Dunwoody 101. H. E. McCoy
46. J. R. Engel 102. W. B. McDonald
47. E. P. Epler 103. H., F. McDhuffie
48. W. K. Ergen 104. C. K. McGlothlan
49. D. E. Ferguson 105. E. C. Miller
50. A. P. Fraas 106. C. A. Mills
51. J. H ¥rye, dJr. 107. W. R. Mixon
52. C. H. Gabbard 108. R. L. Moore
53. W. R. Gall 109. K. Z. Morgan
54. R. B. Gallaher 110. J. C. Moyers
55. R. G. Gilliland 111. M. L. Nelson
56. W, R. Grimes 112. W. R. Osborn
113-114.
115.
11e.
117.
118.
119.
120.
121.
122.
123.
124
125.
126.
127.
128.
129.
130.
131.
132.
1373,
134.
135.
136.
137.
138.
139.
140.
141.

209-210.
211.
212.
213.
214
215.
216
217.
218.

219-549.

R. B. Parker

L. F. Parsly
Patriarca

R. Payne
Phillips

B. Pike

B. Piper

E. Prince

L. Redford
Richardson
C. Robertson
C. Roller

W. Rosenthal
C. Savage

W. Savolainen
Scott
Seagren
Shaffer
Shipley
Skinner
Slaughter
Smith
Smith
Snell
Spencer
Splcwak
Steffy

E. Stevenson

QUHZedrRHGEHUPPOA R R R G EENE T DT
@jflzqi?fg:citfifiim

. F. Cope,
. B. Deering,

Z?UOU

142.
143.
144
145.

146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.

161.
162.
163-164.
165-169.

170-172.
173-207.
208.

nEEHmxdg

-
-

-

N H YR Y Li$=EiC2C12 WG W o Gy

t+<:c:czhzmimiz m=E Qo

G J

. Susano
. Tallackson
. Taylor

Thoma
Tolson
Trauger

. Tucker

Ulrich

. Watkin

Watson

. Webster

Weinberg

. Weir
. Wegtsik

Whatley
Williams

. White
. Wilson

Yost
Young

Biology Library

Reactor Division Library

ORNL — Y-12 Technical T.ibrary
Document Reference Section
Central Research Library
TLaboratory Records Department
Laboratory Records, ORNL R.C.

EXTERNAL DISTRIBUI'ION

Atomic Energy Commission, ORO

Atomic Energy Commission, ORO

G. Garrison, Atomic FEnergy Commission, Washington
Shaw, Atomic Energy Commission, Washington

F E. Slnclalr Atomic Energy Commission, Washington
W. L. Smalley, Atomic Energy Commission, ORO

J. A. Swartout,

270 Park Avenue, New York 17, New York

M. J. Whitman, Atomic Energy Commission, Washington

Research and Development Division, AEC, ORO

Given distribution as shown in TID-4500 under Reactor Technology
category (75 copies — CFSTI)