ORNL-3014.txt

/G

[NEFANTD

3 445k 0361553 1

.u. . :“l led F

%

QOGUMER]

ORNL=-3014
UC-81 - Reactors = Power

MOLTEN-SALT REACTOR PROGRAM
QUARTERLY PROGRESS REPORT
FOR PERIOD ENDING JULY 31, 1960

CENTRAL RESEARCH LIBRARY
DOCUMENT COLLECTION

LIBRARY LOAN COPY
DO NOT TRANSFER TO ANOTHER PERSON

If you wish someone else to see this
document, send in name with document
and ‘the  library will arrange a loan.

OAK RIDGE NATIONAL LABORATORY

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

Printed in USA. Price .$_2£5__. Available from the

Office of Technical Services
Department of Commerce
Washington 25, D.C,

¥ LEGAL NOTICE

This report was prepared as an account of Government sponsored work. Neither the United States,

nor the Commission, nor any person acting on behalf of the Commission:

A. Makes ony warranty or representation, expressed or implied, with respect to the accuracy,
completeness, or usefulness of the information contained in this report, or that the use of
any information, apparatus, method, or process disclosed in this report may not infringe
privately owned rights; or

B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of
any information, apparatus, method, or process disclosed in this report.

As used in the obove, “person acting on behalf of the Commission'' includes any employee or

contractor of the Commission, or employee of such controctor, to the extent that such employee

or contractor of the Commission, or employee of such controctor prepares, disseminates, or
provides occess to, any information pursuant to his employment or contract with the Commission,
or his employment with such contractor.

ORNL-301kL
Uc-81 — Reactors — Power
TID-4500 (15th ed.)

Contract No. W-T405-eng-26

MOLTEN-SALT REACTOR PROGRAM

QUARTERLY PROGRESS REPORT
FOR PERIOD ENDING JULY 31, 1960

H. G. MacPherson, Project Coordinator

DATE ISSUED

15108

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

" Hfl‘{!flin "I’W ifl]flr‘i "fl‘ i"‘m""i‘vs‘m il

3 4456 0361553 1
SUMMARY
PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS

1. MSRE Design

The Molten-Salt Reactor Experiment is a logical extension of the program con-
ducted at ORNL during the past ten years in the investigation of molten fluoride
mixtures and container materials for circulating-fuel reactors. The major objec-
tives of the MSRE are to demonstrate the safety, dependability, and serviceability
of a molten-salt reactor and to obtain additional information about graphite in an
operating power reactor.

Conceptual designs of the MSRE have been made in which the core is construc-
ted of vertical graphite stringers in which fuel passages are machined. Flow of
fuel is two-pass through the core, with downflow through an annulus along the
core-vessel wall and upflow through the moderator passages.

The heat-removal cycle is from fuel salt to a secondary coclant salt to air.
No control rods are needed because the fuel has a temperature coefficient of re-
activity of approximately -4 x 1077 (&k/k)/°F.

A dual-purpose sampling and enriching system is provided for the removal of
semples for chemical analysis or the addition of enriched fuel. This sampler is
located at the pump expansion volume.

In the design of the core, consideration was given to the Poppendiek effect,
and the flow passages were designed to minimize the resulting temperature rise.

The primary heat exchanger is designed for 10-Mw duty. It is a conventional
U-tube heat exchanger in a 25% cut baffled shell. The fuel is in the shell and
the coolant salt is in the tubes.

The radiator is designed to fit the existing space in the air duct in Build-
ing 7503. It is a bare-tube radiator with large tubes to minimize the possibility
of salt freeze-up.

Four drain tanks are required for the MSRE, two for the fuel salt, one for
the flush salt, and one for the secondary salt. A steam boiler is provided for
after-heat removal in the fuel drain tanks. The system layout was designed to
facilitate remote maintenance in the primary circuit and in the fuel drain tank
pit.

A cover-gas system is being designed to remove fission products from the gas
stream and minimize the radiation damage effects on the pump-lubricating oil sys-
tem for bearings and seals.

A study of the requirements for a remote maintenance system for the MSRE is

in progress. Additional studies are being made of the modifications to Building
7503 required to accommodate the construction of the reactor.
2. Component Development

A program of component development and testing in support of the MSRE design
is designed to improve the reliability of the freeze flanges, freeze valves, sam-
pler and enricher, gas-handling system, and fuel and coolant pumps.

A facility for thermally cycling freeze flanges between room temperature and
1400°F was designed, and fabrication is nearing completion. This facility will
also be used to test improved gas seals for the flanges.

Two freeze valves, which have no moving parts, are being fabricated for test;
one of these is heated by electrical resistance heaters, and the second is heated
by a high-frequency induction coil.

The problem of the removal of fuel samples and the addition of enriched fuel
is being studied. Test equipment is being fabricated to test the feasibility of
using a metal freeze seal that can be broken and remade whenever the sampler and
enricher system is used.

Fabrication of a one-fifth-scale plastic model of the MSRE core is nearing
completion. The model is to operate with water at 40°C with a flow of 50 gpm.
This simulates the MSRE fluid velocity and Reynolds number.

Work is continuing with the remote-maintenance development facility. Follow-
ing operation with molten salt, disassembly, and reassembly, an evaluation of the
tools and procedures was made. These were found to be generally satisfactory, al-
though some suggestions for improvement have resulted.

Operation of forced-circulation corrosion loops continued. Nine INOR-8 loops
and two Inconel loops are in operation. Operating times range from 3000 to 18,000
hr.

An engineering test loop is being designed and will be constructed to evalu-
ate such components as the sampler and enricher, gas-handling system, dual drain
tank, freeze valves, level-indicating devices, and heaters. The system will
operate isothermally at temperatures up to expected maximum MSRE temperatures.

Design and testing work on the MSRE primary and secondary pumps continued.
Design of a water test was completed and fabrication started. Procurement was
started for some pump components.

A layout of the primary pump was completed and is being reviewed. Analyses
of the thermal stresses in the pump and the nuclear heating in the structural
material are being made. An experiment was designed to determine the extent of
back diffusion of fission gases up the pump shaft to the region occupied by the
oll seal.

Testing of in-salt bearings continues. The pump which incorporates this
bearing development has operated for 35000 hr and has undergone 63 stop-start
tests.

3., Reactor Engineering Analysis

Graphite undergoes shrinkage at MSRE temperatures in a neutron flux of ener-
gies greater than 0.3 Mev. Calculations were made to determine the effects of
this shrinkage in the MSRE core. Both axial and transverse shrinkages were taken
into consideration.

v

An analysis was made of the temperature effects in a graphite-moderated core
with round and flat fuel channels. The hot spots resulting from the Poppendiek
effect were found to be considerably reduced in the case of the flat channels.

The effects of completely blocking a fuel channel were analyzed; results indicated
that if one fuel channel in the region of greatest power density were completely
blocked, the fuel temperature in that channel would probably rise no more than
4YOO°F above the mixed-mean temperature in adjacent open fuel passages.

Reactor physics calculations were performed for the MSRE. For a cylindrical
core 54 in. in diameter by 66 in. high, graphite-moderated with 8 vol % fuel salt
containing 4 mole % ThF), the calculated critical loading was 0.76 mole % uranium
(93.3% US32); the associated critical mass in the core was 16 kg of U232, At a
reactor power of 5 Mw, the peak power density in the fuel salt wai 60 w/cc and the
average was 24 w/cc. The computed peak thermal flux was 3.6 x 102 neutrons/cme-
sec, and the average was 1.2 x 1010, Camma heating produced a power density of
0.1 w/cc in the core wall at midplane and 0.2 w/cc in the support grid at the
bottom of the core at the reactor center line. The fast flux (above 1 Mev) in the
center of the 2-in.-square graphite blocks was calculated to be about 2% less than
the value at the edge in contact with the fuel channel.

An analog-computer analysis was made of the loss of flow in the MSRE primary
system. For this study the primary flow was decreased exponentially on periods of
1.5, 3, 6, and 10 sec. No after-heat, convection cooling, or moderator tempera-
ture coefficient of reactivity was simulated in this analysis.

Using a temperature coefficient of reactivity of -9 x 10'5(£k/k)/°53 the flow
was decreased from 100% to zero. On the 1l0-sec period the maximum fuel tempera-
ture leveled off at 1375°F after 80 sec. Using a temperature coefficient of -4.5
x 1072 (8k/k)/°F, the temperature leveled off at 1445°F after 80 sec.

PART IT. MATERIALS STUDIES

L, Metallurgy

The last of three INOR-8 corrosion inserts was removed after 15,000 hr from
an INOR-8 forced-convection loop. Weight-loss evaluations indicated the insert
to have lost 1.7 mg/cm2, which corresponds to a wall thickness decrease of 0.08
mil if uniform removal of the wall i1s assumed. The compositions of thin corrosion
films found on several of the long-term INOR-8 corrosion loops were investigated
by means of an electron-beam microprobe analyzer. Results of analyses of the film
indicated an increase in molybdenum content and virtually complete depletion of
chromium and iron, compared to the composition of the base metal.

Two types of solidified metal seals have been developed for use with molten
fluorides at elevated temperatures one contalning an alloy sump with a tongue-and-
groove Jjoint design, the other having an alloy-impregnated metal-fiber compact.
Seals on components made according to both methods have been made and broken a
number of times in an argon atmosphere, and subsequent helium leak tests indicate
both seals to be leaktight. Several potential braze metals are being investigated
for use in these seals.

A method was devised and equipment was built for the leak testing of graphite-
to-metal braze joints. Initial testing to 60-psi pressures was done with various
braze alloys, utilizing isopropanol as the testing fluid. A technique was
vi

established to improve bonding, which calls for oxidizing the ends of the graphite
prior to brazing.

Welding and back-brazing procedures are being developed for tube-to-tube-
sheet joints for the MSRE heat exchanger. A seven-tube sample has been fabricated
successfully which contains trepanned areas on both the welded and brazed sides
and includes a braze-metal sump with feeder holes in the tube sheet. This mini -
mzes the effect of different heating rates for thick and thin metal sections.

A1l the mechanical-properties data for INOR-8 were reviewed in order to es-
tablish design values for the alloy. Design stresses for temperatures below
1050°F were selected on the basis of two-thirds of the 0.2% offset yield strength.
Above 1050°F, the design stresses were based on the stress to produce 1% creep
strain in lO5 hr. A table with the selected design strengths for temperatures up
to 1L400°F is included.

Tensile tests were performed to determine the effect of low creep strains on
the strength and ductility of INOR-8. No effects were observed that could in-
fluence the structural integrity of reactor components made from INOR-8.

Tests on specimens from selected locations and orientations in large pileces
of R-0025 and MHA4Im-82 graphite indicated relative uniformity within each grade of
(1) apparent densities and (2) permeation by molten fluorides at 150 psig in
100-hr exposures at 1300°F.

Permeation of S-h and AGOT graphite with LiF-BeFp-ThF)-UFy (67-18.5-14-0.5
mole %) at 1300°F at pressures of 25, 65, and 150 psig in 100-hr exposures indi-
cated that (1) there were small differences in salt permeation of these grades
for the different pressures used and (2) actual and theoretical salt permeations
were practically the same except for the 25-psig permeation of grade S-k. Grades
S-i and AGOT, respectively, are moderately low- and high-permeability grades of
graphite.

A single series of five precipitation tests was made with AGOT graphite and
molten LiF-BeFp-UF) (62-37-1 mole %); only the volume of the graphite was varied
in order to determine the relationship of graphite volume to uranium precipitation.
For volume ratios of graphite to fuel of 27:1 to 5:1, the uranium precipitated per
cubic centimeter of the bulk volume of the graphite remained approximately con-
stant and averaged (1.5 + 0.4) mg to (1.3 - 0.3) mg.

Additional tests were made in order to confirm data indicating that the
thermal decomposition of NH}F+«HF removes oxygen contamination from graphite to
such an extent thet it could contain molten LiF-BeFo-UF) (62-37-1 mole %) at
1300°F without causing the usual UO, precipitation from the fuel.

No carburization was detected on unstressed INOR-8 specimens after exposure
to LiF-BeF,-UF), (62-37-1 mole %) - graphite system for 12,000 hr at 1300°F.

5. Chemistry

A fuel composed of LiF-BeFo-ThF)-UF) (65-30-4-1 mole %; m.p. 450°C) has been
selected as representative for the MSRE design work, but there is a strong possi-
bility that 5 mole % of ZrFy will be included as an oxide scavenger in a revised
composition. The change is pending confirmation of current experiments which in-
dicate that ZrOp is more insoluble than UOp, and thus provides protection against
the precipitation of U0y as a result of accidental contaminations wilth oxide.

vii

The coolant composition, LiF-BeFp (66-34 mole %; m.p. 465°C), affords a low
viscosity in combination with a suitable melting point.

Treatment with hydrogen as a means of removing oxide from graphite proved
relatively ineffective. The rate of permeation of graphite immersed in a wetting
salt is slow, at least in the later stages, presumably because gas trapped in the
graphite voids can be replaced only as fast as it leaves by diffusion through the
salt.

6. Fngineering Research

The surface tensions of two NaF-BeF, (57-43 mole %) mixtures have been de-
termined to fall between 200 and 150 dynes/cm over the temperature range 500 to
300°C. A 6% discrepancy between the two measurements may relate to differences
in contaminant content of the two samples due to differences in exposure time in
the circulating loop from which the samples were drawn. The results are 1n rea-
sonable agreement (although somewhat lower) with data obtained with an NaF-BeFy
(63-37 mole %) mixture. An analysis of the precision of the measurements indi-
cates that residual effects due to errors in pressure and geometrical measurements
have been reduced to about *3%; however, a large uncertainty (as much as an addi-
tional *3%) still remains in the salt density as used in evaluating the data.

Corrections to the original data and the inclusion of more recent results
have yielded a revised correlation of the mean heat capacity of BeFE-containing
salt mixtures.

Heat-transfer studies with LiF-BeF,-UF)-ThF) (67-18.5-0.5-14 mole %) in
Inconel and INOR-8 tubes have been interrupted after 5560 hr of operation to re-
place the circulating pump. Damage appears to be restricted to the upper shaft-
bearing. Analysis of the salt (pre- and post-operational) shows a composition
differing from the nominal composition; this will necessitate a re-evaluation of
the data using corrected values of the thermal properties.

7. Fuel Processing

Preliminary studies indicated that ThFy in molten-salt reactor fuel may be
decontaminated from rare-earth fission products by dissolution of the rare-earth
fluorides in SbF5-HF. The LiF of the fuel must be removed first, by dissolution
in OF, to prevent precipitation of the antimony, probably as LiSbFg.
P v R D T T e

A e

b i o

CONTENTS

SL]WARY. ---------- P R T NI I I I I I R I T I R R B N I I I A [ A BB B Y R RS N N B R R R I N N N L iii

PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS

1. MSRE DESIGN ......... Cetsaseen fhreesensaas Chereeaaenas Creerassssenennn 1
1.1 Introduction seseevecsvassesesssansanns cesessscscsesenstesesenans 1

1.2 MSRE Objectives s.eeierecsascsesssncssanns cesesisesesesansesrenns 2

1.3 Conceptual Designs seeeseesasssascssascaancs cesateassearnsseanns 2

1.4 The Reactor Core and Vessel «e.eeeveeeses Ceeedracesenstersrnenns L

1.5 The Primary Heat Exchanger ......ceee... Ceteisasescasarasesnsanas 8

1.6 Radiator ......... Cietisensesanaesssanssasararecaaans B 1

1.7 Drain Tanks «eec... csesesessernstssas st nesetonerane cetesesesess 11

1.8 Fquipment Arrangement +.ceeeececesencacns S 24

1.9 Design of Cover-Gas SyStem .eeeeeseiecesccacan cerieccerscesensas 18

1.10 Design of the Remote-Maintenance System for MSRE «vieevsrsscsess 19

1.11 Modifications to Building 7503 ...ceceecennn =24

2. COMPONENT DEVELOPMENT ..vvvvevanernannas cieeens Cereesieieiaeees cevess 24
2.1 Freeze Flange Development su.eeseveeeacnns =

2.2 Ireeze Valves .cviveescerercscssnssasansns =)

2.3 Sampler-Enricher Development ...... 25

2.4 MSRE Core Development +..eeces.. -

2.5 Remote Maintenance Development Facility ..cceceeeen.. crscaesasnss 27

2.6 TForced-Circulation Corrosion LOODPS secsececcceccccasnens cevessess 2B

2.7 Pump Development ..ieeiviescersescrsnsscressnscescscnsons ciessese 29
2.7.1 MSRE Primary Pump ..ccesrecnccanetscansnnes cevssenssases 29

2.7.2 MSRE Secondary PUmp eecvsrerennarassanencns Ceteiesesanaas 30

2.7.3 Advanced Molten-5Salt Pumps «sevevecenn. cesseas teseneesrsss 30

2.7.4 MF-F Pump Performance LOOD eeeevevececnaanss cesesassenes 31

2.7.5 Frozen-Lead Pump Seal sieseveectciccccenannn ceveesesrees 31

2.7.6 MSRE Engineering Test LOOD svecsccacseecosenns cessessens 31

3. REACTOR ENGINEERING ANALYSIS ....... Gt eseerrseanaata s G -

3.1l Effects of Graphite Shrinkage Under Radlatlon in the MSRE Core . 32
3.2 Temperature-Rise Effects in MSRE Cores with Round and Flat

Fuel Channels cevecsoeseccescononsseanss G L

3.3 Temperature of Fuel in a Blocked Passage in the MSRE ....cc00nee 39
3.4 MSRE Reactor Physics ..e.veeevonn. O 'S |
3.4.1 Core CalcUlations seussecsssuesesosessasvsonesnssssennsns 41

3.4.,2 Gamma-Heating Calculations s.cevecenrconarsssnaacescnsans i

3.4,3 Drain Tank Criticality ....... Ceeeretecenaneasanannns ee.  Lh

3. 4,4 Pump-Bowl Fission Product Activities ...eeveneen. .

3,4.5 Cell Calculations sevesesseesssosssnsansaasnnass S [

3.5 Analog Computer Study of MSRE Primary-System Flow LoSS sveaeesse 45
3.5.1 Description of the System Simulated .....cviveeecsasesss k6

3.,5.2 Analog Computer Program .eeesescsesassaoes B ¥

3.,5.3 Simulator Operation «ersessesesescescancocsanasoannsns R [T

3.5.4 Conditions Used to Obtain CUIrveS cvevevvreanecenans ceanas L9
x

PART II. MATERIALS STUDIES

b, METALLURGY +eetevovceaatoooonostsonssssasnsonaancsssosnsssosaseonssnsnsas . 55
4,1 Dynemic-Corrosion Studies s.eeeieveesacasnn ettt teteeee e 55
4.1.1 Forced-Convection Loops «.ee... C e ieeaseresaieeaean 55
4.1.2 Microprobe Analyses of Surface Fllm .................... 56
4.2 Welding and Brazing StUdi€s +veeirrieeieerenensnennreeasconanonns 58
4,2.1 Solidified-Metal-Seal Development Ceteiater ey 58
4.,2.2 Brazing of Graphite ...... Ceeerseeseterarteceatatecennans 59
4,2.3 Heat Exchanger Fabrication ........ Ceenarsasecenaaeaneae 63
4.p.4 Mechanical Properties of INOR-8 .......... et teresaenean &l
4.3 Permeation and Apparent-Density Uniformity in Large
Pieces of Graphite iveeeisrentrennnas e e e s e n s eenneene s 67
4,35.1 Permeation of AGOT and S-4 Graphites by Molten
Salts at Different Pressures ....... Ceie i chaeasen 67
4.4 Precipitation from Molten Fluoride Fuel in Contact
with Various Volumes of Graphite .......... creerneesraasas e 69
k.4.1 Removal of Contamination from Graphite ......ccievveeean. 69
L.k,2 INOR-8 - Fuel - Graphite Carburization Tests ........... 70
4.5 Tn-Pile TeStS seeeereerenenanennans f et et ee ettt T1
5. CHEMISTRY vieevieerviestsatsosensssronsanassas teeensaranans Pesecescsnna T2
5.1 Phase Equilibrium Studies +evvoeereeonsesss B -
5.1.1 MSRE Fuel and Coolant seievsnenrntnnencscnna s et et es e T2
5.1.2 Systems Containing ThFy .......c.0v.n. Cereesrerseans veens T2
5.1.3 The System ZrFU-ThF) «.evtevreteretvestessnsssssrsavesnsnse Th
5.2 Effect of Tetravalent Fluorides on the Freezing Point
of Scdium Fluoride ............ Sececesiesitaeeretteennaens N )
5.2.1 Phase Diagram of Fluoride Systems .........0... Ceeeeases 7T
5.2.2 The System LiF-YFz .....v0vnnnns Ches et iesaressensanen s 78
5.2.3 Melting Point of NiFo et iiinitieereveenensnns 79
5.3 Oxide Behavior e.svevesee cheeeaes Pesesesencresontaonsacnn cheasas 79
5.3.1 Oxide Behavior in Fuels .cieieivntenneerenens Cerreecnens 79
5.3.2 Zirconium Oxyfluoride and Attempted Preparation
of Urancus Oxyfluoride .......cceinieeeicnersannrsncaas 80
5.4 Graphite Compatibility cveieieieieriereennaeaeasasannsaansnssanns 80

5.4.1 Removal of Oxide from Graphite by Treatment
wWith Hydrogen cvieveerinestiecssrienssecssssansassssnase 80
5.4.2 Behavior of Graphite when Wetted by a Molten Fluoride .. 81

5.5 Preparation of Purified Materials ............. e o1 1

6. ENGINEERING RESEARCH ...ivuivrieienenenenereenenenensnancnansenonnennnnns 83
6.1 Physical-Property Measurements ....c.eeieeeieneeeeenncnarnncnans 83
6.1.1 Surface Tension and Density «eveeeeeeeenercrenenennansas 83

6.1.2 Heat Capacity ce.eeeve.. e Ceeeenaaas Cetetrecireaae 85

6.2 Heat-Transfer STUGIES .eieeeiecererrasoroaoasnsesncsaesonaseaans 86

7! F[JEIIPROCESSING LR R R B N B I B D O R N N B R R Y R AN A L R R DR R N N B R R R B R DL R I I I I I B B A 88

PART I. MSRE DESIGN, COMPONENT DEVELOPMENT,
AND ENGINEERING ANALYSIS

1. MSRE DESIGN

1.1 INTRODUCTION

The concept of using molten fluorides, one of which is UFy with highly enriched
uranium, as a liquid fuel for reactors has been actively studied at ORNL for about
ten years. Many design studies have been made, and one reactor was constructed and
operated successfully as a high-temperature low-power ( X2.5 Mw) reactor experiment,
the ARE.

Much experimental work has been done on phase diagrams involving molten fluo-
ride fuel systems. The work included some in-pile as well as many out-of-pile
experiments in which test loops of circulating fluoride melts were operated for
thousands of hours with substantial imposed temperature differences in the loops.
This work, most of it under the ANP program, has developed fluoride-fuel technology
to a point where it is ready for serious consideration for high-temperature power
reactors.

Container material has also been the concern of an extensive metallurgical
program for about ten years. Stainless steels, Inconel, Hastelloy, and other
alloys have been studied for corrosion resistance to the molten fluorides; the
experimental reactor, which was run successfully, employed Inconel. While the
corrosion resistance of Inconel was adequate for a short-term experiment, it was
considered to be unsatisfactory for long-term service.

INOR-8, an alloy, was developed specifically for compatibility with the fluo-
rides. This alloy was made into components for pumped test loops, and thousands of
hours of corrosion tests were performed. Here also some in-pile as well as many
out-of-pile tests were run, and it was concluded that INOR-8 was a satisfactory
material for the construction of the components of a high-temperature, molten-salt-
fueled reactor.

Having established a satisfactory fuel and satisfactory container material,
the next question concerned the kind of moderator which would be acceptable for
such a system. Theoretical calculations, for which no experiments were performed,
indicated that the fluorine in the salt could be employed, thus making possible a
homogeneous molten-fluoride reactor system. Many design studies were made of
this concept.

Tt was obvious that, while a homogeneous system was possible, the neutron
economy of this type reactor was such as to preclude any serious attempt at thermal
breeding. Therefore some studies, both theoretical and experimental, were begun in
order to determine the possibility of employing unclad graphite as a moderator for
the fluoride-fuel reactor. An estimate of the breeding potential of the graphite-
moderated molten-salt reactor can be found in an ORNL report.l
2

Compatibility tests between graphite and the fluoride fuels were run, and no
chemical incompatibility between graphite and the fluorides was found. While some
fuel penetrated the graphite sample, it was also determined that(neglecting the
effects of radiation)not more than 1 vol % of the graphite would be permeated if
especially low permeability graphite were used.

In view of the favorable results of the long development program on materials,
the design and construction of a molten-salt reactor experiment has been initiated.

1.2 MSRE OBJECTIVES

The primary objectives of the MSRE are to determine whether a molten-fluoride,
circulating-fuel reactor system can be made to operate safely, dependably, and ser-
viceably. By this it is meant that no credible accident can endanger personnel,
that the system is capable of achieving a very high percentage of operational time,
and that any component can be removed from the system and a replacement made, per-
mitting return to normal operation.

The reactor will, however, provide answers to many questions of importance to
future molten-salt reactors. It will provide long-term irradiation tests of fuel,
INOR-8, and graphite under actual service conditions. From the behavior of graphite
with respect to the absorption of fuel and fission products, important questions
with regard to the feasibility of breeding in molten-salt reactors can be answered.

1.5 CONCEPTUAL DESIGNS

The reactor consists of a cylindrical vessel in which a graphite matrix con-
stitutes about 88% of the volume. Fuel enters the vessel at an annular volute
around the top of the cylinder and passes down between the graphite and the vessel
wall. A dished head at the bottom reverses the flow and directs it up through
rectangular passages in the graphite matrix into a dished head at the top, from
which it goes to the suction line of a sump-type pump mounted directly above and
concentric with the reactor vessel. Flow through the reactor is laminar, and the
passage width is narrow enough to prevent an excessive Poppendiek effect. The
reatangular passages were chosen because they gave, with the design flow, a much
lower hot-spot temperature than did cylindrical fuel passages. From the pump dis-
charge, the fuel flows through the shell side of a cross-baffled tube-and-shell
heat exchanger and thence to the reactor inlet.

The tube side of the primary heat exchanger contains the secondary coolant,
which is also a binary fluoride melt (LiF-BeFp, 66-34 mole %). This fluid is cir-
culated by a sump-type pump through the tubes of the heat exchanger and through the
tubes of an air-cooled radiator. Air is blown across the unfinned tubes of this
radiator and up a T70-ft stack by two axial blowers. A basic flow diagram of the
primary and secondary salt systems is shown in Fig. 1.1.

Control of this reactor is quite simple and does not call for internal control
rods of any kind. Fast control is effected by the volumetric temperature coeffi-
cient of the fuel, which produces a temperature reactivity coefficient of approxi-
mately -4 x 10-5 (ak/k)/OF. Slow control is accomplished by fuel enrichment, with
the provision of poison addition (e.g., ThF)) if desired.

Fuel addition in gross amounts for the original loading will take place in
the drain tank. Subsequent addition for burnup and fission-product poisoning will

UNCLASSIFIED
ORNL-LR-DWG 50409

COOLANT
PUMP

FUEL
850 GPM
PRIMARY SALT PUMP HEAT EXCHANGER 110Q °F
o - >
LiF-61% — I L
BeF-34% % 10251°F ) '
ThF- 4% ! i ‘
UR- 1% ” SECONDARY SALT
X LiF-66%
|
i BeF-34%
1225 °F ¥
1450 GPM D ]
——— L : m : ::
5 T wzs=F ¥ i
» 1 i
3 i REACTOR ": LJ
; CoRe 1 VESSEL ;5
l E:‘ AIR
e g b 160,000 RADIATOR
!
FJ/ ! E SCFM —» 300 °F
P {00 °F
L
REACTOR CELL noot
Iy
[
e
- FREEZE PLUG
SPARE FILL & DRAIN FLUSH TANK COOLANT
FILL & DRAIN TANK (60 cuft) DRAIN TANK
TANK (60 cuft) (30 cu ft)

(60 cuft)

Fig. 1.1. Basic Flow Diagram of Primary and Secaondary Salt Systems.

be made through an enricher assembly communicating with the gas space in the pump
bowl. This enricher assembly will also be used in essentially the reverse manner
for taking fuel samples at any time during operation of the systen.

The fuel system is to be an all-welded system except that each component will
employ a freeze-flange connection to the piping system. The rotary element of the
pump will also employ a conventional metal-gasketed flange with provision for inert-
gas buffering to ensure inleakage of gas in case of imperfect closure.

The secondary coolant system will be all-welded and will not employ the freeze
flanges. This is made possible because direct maintenance can be effected on this
circuit at all times.

The molten salt in both the primary and secondary circuits will be sealed off
from the respective drain tanks by means of freeze plugs in the drain lines. Pro-
bosed designs of the freeze flanges and the freeze plugs are now under construction,
and tests of these units will be made to determine the ultimate design.
Heating of the primary and secondary salt systems will be accomplished by means
of electrical heaters around all lines and components of these systems; several types
of commercial heaters have been studied. It now appears that the heaters will be
locally constructed, and Tnconel or stainless steel pipe will be used. Designs of
these units are being made, and tests will be run to verify the advisability of using
this kind of heater.

The off-gas system has been planned as a recirculating system, to minimize the
consumption of inert cover gas. Holdup capacity for radioactive decay and filters
for cleanup of the gas will be incorporated. Provision is made for continuous ex-
haust through charcoal beds to the stack, in case the recirculation system fails
at any time.

The fuel dump tank must be provided with heat for keeping the clean fuel molten
and must have cooling available to remove afterheat from radioactive fuel. The tank
is provided with electric heaters similar to those on other parts of the salt cir-
cuits. Cooling is provided by means of thimbles penetrating the tank through the
top; the thimbles have coolant tubes inserted in them which are spaced away from
the thimble wall. The coolant tubes will be filled with water which will be boiled
by the radiant transfer of heat from the thimble walls.

1.4 THE REACTOR CORE AND VESSEL

The physical structure of the reactor consists of a containment vessel; an
inner, open-ended TINOR-8 cylinder serving both as a separating baffle for the cool-
ing annulus and support for the graphite core; the composite graphite-moderator
matrix with positioning and support members; and various flow regulating devices.
Figure 1.2 shows the concept of the reactor. Significant geometrical and flow
characteristics are listed in Table 1l.l.

A heat-generation rate of 0.2 w/ce in the INOR-8 wall of the reactor vessel
requires a heat removal of 23 kw. The wall cooling will be accomplished by the fuel
flowing along the wall. Turbulent flow is desirable in the annulus in order to mini-
mize the Poppendiek effect. With the design flow rate of 1450 gpm in the 1l-in.-wide
cooling annulus, the Reynolds modulus is 13,500, and the temperature of the outside
wall surface is less than 5°F above the bulk stream temperature.

The moderator graphite of the core is built up from 2 x 2 x 66 in. stringers.
The size (area) of these matrix stringers is limited by the size of impervious
graphite available at present. The stringers are pinned in beams at the bottom
of the core. A graphite band will hold the matrix together, and an INOR-8 yoke
is provided for centering. A coarse screen prevents possible graphite fragments
from leaving the core.

If the graphite is packed tightly in the inner can at room temperature, the
differential expansion between the TNOR-8 and graphite will open a radial clear-
ance of 3/16 in. at operating temperature. It is expected that the fast-flux dis-
tribution in the core may tend to cause radial bowing in the graphite. In order
to reduce the nuclear effects of bowing and shrinkage, the graphite stringers will
be banded over the middle two quarters with molybdenum bands.

Flow passages in the matrix are provided through rectangular channels machined
into the faces of the graphite stringers (see Fig. 1.%). The tabulated channel
dimensions provide a fuel volume fraction of 12%. The specified channel configura-
tion is the product of intensive studies into the temperature effects asgociated
with the relatively slow flow through the core (about 2 fps). These effects are

UNCLASSIFIED
ORNL-LR-DWG 52034

FUEL OUTLET

GRAPHITE SAMPLE BLOCK

S
T
<]
5
o
QO
>z
e
MS
o
<
o
U]

FUEL INLET

()
(X

B0

%

()
)

)

5
g
o

0
e
¥

o
o

% ave
R
BERGEX )
J0e
)

OO

VAN
AV

A O

S

\ FUEL INLET VOLUTE

~~— REACTOR VESSEL

; GRAPHITE-MODERATOR
— STRINGER

B
%

0

Y
AV

()

"

(]

%

X
4
o
A

CCRE YOKE AND
SCREEN

REACTOR CORE GAN

FUEL PASSAGE

CORE-POSITIONING
GRAPHITE BEAMS

VESSEL DRAIN LINE

SWIRL VANES

ANTI

CORE GRID SUPPORT

MSRE Reactor.

Fig. 1.2.
6

Table 1.1. MSRE Reactor-Vessel Design Data

Inlet pipe 6 in., sched 40
Outlet pipe 8 in., sched 40

Core vessel

oD 58-3/8 in.
ID 57-1/4 in.
Wwall thickness 9/16 in.
Design pressure 50 psi
Design temperature 1300°F
Fuel inlet temperature 1175°F
Fuel outlet temperature 1225°F
Inlet Volute
Annulus ID 54-1/2 in.
Annulus OD 56-1/2 in.
Over-all height of core tank 8 ft
Head thickness 1l in.

Graphite core

Diameter 54 in.

Core blocks (rough cut) > x 2 x 67 in.
Number of fuel channels 1064
Fuel-channel size 1.2 x 0.200 x 63 in.
Effective reactor length ~ 65 in.
Fractional fuel volume 0.120

Core container

ID 54-1/8 in.
0D 54-5/8 in.
Wall 1/k in.
Length 73-1/b4 in.

a composite of the Poppendiek gradient and a less significant temperature gradient
in the graphite produced by the internal heat generation. With reasonable flow
rates in a once-through multichanneled core, the flows are laminar or at best un-
stably turbulent. Ior laminar-flow conditions, the Poppendiek effects become
rather severe with wide channels. The radial temperature gradient of a circular
channel providing 10% fuel volume in a 2- X o>-in. graphite section is 357 times
as large as that of a 0.1-in.-wide channel with equal fuel fraction. With the
selection of the present 0.200-in.-wide channels, the graphite temperature at the
midpoint will be AO°F above the mixed mean temperature at the nuclear center of
the core. This problem is treated more fully in Chap. 3.

UNCLASSIFIED
QORNL—LR-DWG 52035

PLAN VIEW

TYPICAL MODERATOR STRINGERS

SAMPLE PIECE

X
CROSS-COMMUNI -
CATING CHANNELS
83?

NOTE: NOT TO SCALE

‘\\\\!'A LI

Fig. 1.3. Typical Core-Block Arrangement.
It is expected that complete blocking of one fuel channel would raise the
temperature of the fuel in the channel about 6800F above the level of adjacent
channels.

Provision will be made for the removal of five graphite samples from the
center of the core. These full-length samples will be lifted out through the
suction line and the pump bowl.

1.5 THE PRIMARY HEAT EXCHANGER
The primary heat exchanger is being designed (Fig. 1.4) for a duty of 10 Mwto
be transferred from the fuel salt to the secondary coolant salt. Pertinent data

of the preferred design are given in Table 1.2,

The design of the heat exchanger follows the configuration of conventional
25%-cut, baffled shell-and-tube units, with greater emphasis on reliability and

UNCLASSIFIED
ORNL-LR-DWG 52036

FUEL INLET

U-TUBE BUNDLE

Y2-in ~0D HEAT
EXCHANGER TUBE

CROSS BAFFLES
{25 % CUT)

THERMAL-BARRIER PLATE

COOLANT INLET
L) -

COOLANT-STREAM
SEPARATING BAFFLE

COOLANT OUTLET

FUEL OUTLET

Fig. 1.4. Primary Heat Exchanger for MSRE.

9

Table 1.2. Primary-Heat-Exchanger Design Data

Heat load 10 Mw
Shell-side fluid Fuel salt
Tube-side fluid Coolant salt
Layout 25% cut, cross-baffled shell
and U-tubesgs
Baffle pitch 10 in.
Tube pitch 0.812 in., triangular
Active heat-transfer length of shell 5 ft 10 in.
Over-all length ~ 7 ft
Nozzles
Shell side 6 in. IPS
Tube side 5 in. IPS
Shell diameter 16-1/4 in. ID
Shell thickness 1/5 in.
Number of U-tubes 156
Tube-sheet thickness 1-1/2 in.
Heat-transfer surface area 250 ft2
Fuel holdup ~ 5.5 ft3
Terminal temperatures at design point
Fuel salt Inlet 1225°F; outlet 1075°F
Coolant salt Inlet 1025°F; outlet 1100°F

Effective LMDT 135%°F

simplicity of construction than on particularly high performance. The space limita-
tlons of the containment area call for a fairly short unit.

The heat-transfer and pressure-drop design are based partly on experimental
heat-transfer data of Amos, MacPherson, and Sennc (for the tube side) and partly
on methods suggested by Kern.? From the heat-transfer point of view it is prefer-
able to pass the larger flow of fuel salt through the shell side, and the smaller
flow of the coolant through the tubes. The shell side presents less opportunity
for retention of gas pockets during filling operations than does the tube side.
The fuel salt operates at lower pressure; thus thinner shell walls may be used.
The shell side, however, has slightly more liquid holdup.

The U-tube configuration results in a much shorter over-all length. The tem-
perature effectiveness of the unit is 97.5%, compared with true counterflow. The
1809 bend in the tubes minimizes the thermal expansion problem. The tube and baf-
fle pitches were chosen to give an even number of baffles in the shell side within
the range of baffle pitches of 0.2 to 1.0 shell diameter, where the methods of Kern
have good accuracy.

The stresses in the heat exchanger have been analyzed. Because of the low-
pressure operation, mechanical stresses are significant only in the tube sheet.
10

With a 50-psi pressure differential on a 16-in.-dia flat plate, 1.5 in. thick,

the estimated stress is 3000 psi. In order to minimize thermal stressing of the
plate, which would be additive to the mechanical stresses, a thermal baffle is
placed about 2 in. from the tube sheet. This baffle will provide a stagnant layer
of salt, reducing the thermal gradient across the tube sheet to ~20°F. The stif-
fening effects of the stream-separating baffle in the tube-side header would induce
high localized stresses if attached to the tube sheet. For this reason, a laby-
rinth seal is provided at the tube sheet, and the baffle is welded to the dished
head.

While mechanical stressing of the tubes is insignificant (only ~228 psi hoop
stress), stresses due to thermal gradients across the wall may reach 6500 psi at
the fuel inlet end. However, the combined stresses remain well below the T700-psi
level allowed for a metal temperature of 12250F, which is the maximum design tem-
perature for the fuel.

A tube-to-tube-sheet joint of the welded and back-brazed construction is being
tested to determine the tube-sheet thickness required in the present exchanger.
The first specimens appear to have produced sound joints.

Two other heat exchanger configurations were also investigated. An axial-
shell-flow exchanger having closely spaced tubes was rejected on the basis of
expected manufacturing difficulties. A second loosely spaced heat exchanger with
similar flow pattern would have resulted in a much longer unit, not suited to the
limited space in the reactor pit.

1.6 RADIATOR

In the salt-to-air radiator the thermal energy of the reactor is rejected to
the atmosphere. The radiator will occupy a part of the existing air duct in Build-
ing 7503, and its over-all dimensions are SO tailored that it can be installed in
the available space. The physical design concept is shown in Fig. 1.5, and the
design data are listed in Table 1.3.

The heat transfer design is based on Curves by Kays and London)“L for the air
gide and on data previously reported.2 The performance of the heat exchanger
under reduced load conditions is presently under study to determine the most effec~
tive method of power control.

Several features were incorporated in the design as protection against freez-
ing of salt in the radiator tubes:

1. Large-diameter tubes were used.

5. The heat rate per unit area was kept low by using bare tubes on the air
side to take most of the temperature drop in the air film.

%, The lowest design temperature of the bulk salt is 1859F above the freez-
ing point.

4, An elaborate headering system is proposed to ensure even flow distribu-
tion between tubes.

5, Studies are under way to provide adequate heating of the radiator should
one or more tubes freeze accidentally.

11

UNCLASSIFIED
ORNL-LR-DWG 52037

5-in. SECONDARY-SALT INLET

\k\\“%h4n-ODX()OTZ#anALL TUBING

B-in. HEADER
10 ROWS, 12 TUBES PER ROW

5-in. SECONDARY -
SALT OUTLET

2Y-in. TUBE
MANIFOLD

AIR FLOW

Fig. 1.5, 8Salt-to-Air Radiator.

The layout of the tube matrix is expected to allow movement of the tubes
under thermal expansion with a minimum of restraint. The sloping tube configura-

tion will promote drainability.

1.7 DRAIN TANKS

Four salt-carrying drain tanks are under design for the MSRE, as listed below.

1. a duplicate pair of fuel drain tanks,
2. one container for the flush salt,
5. one container for the secondary salt.

Items 2 and 3 are simple pressure vessels made of INOR-8, each supplied with
a dip-tube fill-and-drain line and gas connections for Pressure filling. Their

dimensions are listed in Table 1.L.
12

Radiator Design Data

Table 1.53.
Duty
Temperature differentials
Salt
Air
Air flow
Salt flow

Effective mean At

Over-all coefficient of heat transfer
Heat~transfer surface area
Tube diameter

Wall thickness

Tube matrix

Tube spacing

Subheaders per row

Main headers

Air-side AP

Salt-side AP

10 Mw

1025 to 1100°F

100 to 300°F

167,000 cfm at 15 in. 32
830 gpm at avg temperature
920°TF

53 Btu/hr-ft2-°F

685 £t°

0.750 in.

0.072 in.

O pressure

12 tubes per row; 10 rows deep
1-1/2 in., triangular

% in., IPS

8 in., IPS

11.6 in. Hy0

6.5 psi

The after-heat in the fuel salt requires cooling in the drain tanks if the

fuel is dumped a short time after shutdown.

Considering time lags in the drain-

ing procedure, it is unlikely that the tanks would receive the fuel less than

15 min after shutdown.

Accordingly, the heat removal rate of 100 kw would main-
tain bulk-salt temperatures below 1350°F during the decay period.

To make the

heat removal as nearly uniform as possible throughout the tank, 40 immersed

bayonet coolers are used, with boiling water as coolant.

Water cooling was

selected above gas, molten salt, or NaK because of its simplicity and relative
independence from utilities failures. To further enhance the reliability of the
fuel drain system, two identical tanks will be provided: one in use, the other
in standby. Double-wall separation is used between the salt and the water. (It
appears that the induced thermal stresses are not excessive, although further in-
vestigations will be made.) The thimble design is shown in Fig. 1.6, and the
fuel-drain-tank system is shown in Fig. 1.7.

The fuel drain tank, filled with salt but without cooling thimbles, would
have a multiplication constant of about O.44. Even if the cell is flooded with
water (e.g., as an emergency cooling method), the multiplication constant would
reach only O.77.

1.8 EQUIPMENT ARRANGEMENT

The equipment arrangement for the MSRE has been developed on the basis of the
following criteria:

1. All equipment that contains fuel is to be replaceable by remote mainten-
ance.

Table 1.k4.

13

Drain~Tank Design Data

l.

2.

3,

Fuel Drain Tank

Height
Diameter
Wall thickness
Vessel
Dished head
Capacity
Fuel
Gas blanket
Max. operating temp.
Cooling method
Cooling rate
Coolant thimbles
Number

Diameter

Secondary Drain Tank

Height

Diameter

Liquid capacity

Wall thickness
Vessel
Dished head

Cooling method

Flush Salt Tank

Height
Diameter
Capacity
Coolant
Gas blanket
Wall thickness
Vessel
Dished head
Cooling method

35 in. (without coolant headers)
48 in.

1/2 in.
3/4 in.

55 £t

b7 £t

1350°F

Boiling water in double-walled thimbles
100 kw

14

UNCLASSIFIED
ORNL—-LR-DWG 52038

COOLING-WATER INLET\ /STEAM QUTLET

ot

= JACKET BREATHER HOLES

A

TANK HEAD

T AT TTT
N

y ¥4

— QUTER CONTAINMENT TUBE

STEAM RETURN TUBE —

i 7SPACER FINS

WATER FEED TUBE — {1\ 7
4

Fig. 1.6. Cooling Thimble for Primary-Salt Drain Tanks.

UNCLASSIFIED
ORNL-LR-DWG 52039

STEAM CUTLET CONNECTION

STEAM DRUM

SAMPLING CONNECTICON
THROUGH DRUM

1

TYPICAL FOR

FLEXIBLE TUBING;
QUTER RING

FLEXIBLE TUBING; TYP-
[CAL FOR INSIDE RING

L
[}
ZT
-0
5 €T
5 @ E
| <1
| od
ic w
O N
=z Sm
< v -
2 Zs
& =
O \

B

//////////./////,////////I/

) |
S D
=2 =

o —
WW Wi o1
= S0
i >
=z < c

EU @
@ 2N
% i A

T —
Ya ao
= >
i

TYPICAL BAYONET BRACING,

FOR INNER AND OUTER RING

Primary Drain and Fill Tank.

Fig. 1.7.
16

2. Equipment that contains coolant salt is to be replaceable by direct
maintenance.

3, The primary fuel loop is to be located in the existing 24_ft-dia contain-
ment vessel in Building 7503.

4, The fuel storage tanks are to be located in separate containment.
5. Only vertical movement of the fuel-circulating pump should be permitted.

Several studies of possible layouts have been made, and the following conclu-
sions have been reached: From the standpoint of piping stresses, the most favor-
able arrangement is to mount the pump on top of the reactor, with a minimum length
of piping from the reactor to the pump suction. A minimum of 3 ft is required by
the hydrodynamics of the pump. It is also desirable to anchor the coolant piping
close to the pump so that expansion of the parts of the cooclant system outside the
primary-loop containment vessel does not affect the stresses inside the vessel and
so that the differential thermal-expansion stresses imposed on the fuel system by
the coolant piping are minimized.

The arrangement shown in Fig. 1.8 and 1.9 has been developed on the basis of
the above criteria. The reactor is placed in the southwest quadrant of the con-
tainment vessel, with the fuel pump mounted on top. The heat exchanger is aligned
with the pump discharge. The reactor is supported near the top, and the pump is

UNCLASSIFIED
ORNL-LR-DWG S0410A

TOP OF PENTHOUSE
Ibaliaibibinlttii N

CONCRETE SHIELDING, ; At e

STEEL BARRIER,
\

MOTOR
WEIGH =
| SCALES = PRIMARY PUMP.
4 A . 8-in. SUCTION
! . HEAT EXCHANGER

i R . S, R R CRBERGA Sk T — ‘-i
b , y - :
i &1+ Qin. ; .
B SHIELDING N —_— |
¥ -
| | I
D REACTOR 7 AIR PLENUM WALLS

. t¥2-in. DRAIN LINE

************ I 07774 .1 1%-in. DRAIN AND
T ; : FILL LINE. ! RADIATOR
‘ / i : RS W — PIT
P — e

S o

- EXISTING STEEL{~,
] WATER \
| SHIELD R

> : : e LOW POINT
FILL AND DRAIN | i OF TANK - ¢ TEST CELL

TANK NO. 2

_—FILL AND DRAIN
TANK NO. 3

FLUSH i
TANK i

R
T Y

-
i

¥

Fig. 1.8. General Arrangement of Primary, Secondary, and Drain Systens
(Elevation).

L ] [} ] L]
UNCLASSIFIED
ORNL—LR—DWG 52040
— NUCLEAR
WATER SHIELDING
s
v
///
/
m
i pi it
1 li ’t— B '| // AND PRIMARY PUMP
| /7
| | |l I “ I /// /PIPE SLEEVES
I | ! i L FILL AND DRAIN TANK NUMBER 3
I /
1 I L__ _/ —
{ 1\ L 4H|EAT EXCHANGER =
I
{ |
HOT STORAGE PITS } | l:
_— |
| N - !
FILL AND YIRS ]
| i DRAIN TANK /\g/ NEUTRON SOURCE TUBE RAg:fl OR H
I | NUMBER 2 —1 >\ \\\ \ p il
4 7
i | ll X N et /', /
N SR A T “f >
| | ? NN RS L s k
! I ' FILL AND DRAIN TANK NUMBER ' STACK AND AIR DUCT
Il J | W PLENUM
| FLUSH TANK— NG S s
T ¥
J—L CABLE AND PIPE TRAY !
TR |
T
H / i AR DUCT
[ W
/ ! ii PLENUM—/’%’:
! ANl 1 I
! o il il
; a - b
i S I 1
I h i I Iy
/ ; U L I
/ 7
{
/ /
/ EXISTING
FANS
——
= = -
ES BELLM
. ’
i G
% BLOWER HOUSE
@‘ - </
[%]

=

S

o

(=]
5 TRACK
2

g |

H

Fig. 1.9. General Arrangement of Primary, Secondary, and Drain Systems (Plan).

LL
18

supported above it, using spring hangers. The heat exchanger is mounted so that

it is free to move away from the reactor and is also free to move vertically. Pre-
liminary studies indicate that the above layout will minimize the thermal-expansion
stresses. Detailed stress analyses of the layout under several operational condi-
tions are to be completed before the layout within the containment vessel is
finally accepted.

The drain tanks are being located in a new pit placed northwest of the reactor
containment vessel. This location was selected after attempts to place the tanks
in the existing deep pit southeast of the reactor containment vessel indicated that
there was insufficient space. Connections to the tanks are being designed to per-
mit weighing the tanks.

1.9 DESIGN OF COVER-GAS SYSTEM
The functions of the MSRE cover-gas system will be to

1. remove fission-product gases from the pump-bowl vapor space on a
continuous basis;

2. help minimize radiation damage to the fuel-pump lubricating oil by
minimizing diffusion of fission-product gases into the oil system;

5. minimize contact of the fuel with oxidants by maintaining an inert-gas
blanket in fuel storage tanks, transfer tanks, sample systems, and
freeze flanges;

4, provide a supply of inert gas in sufficient quantity and at the required
pressure level so as to permit purging of the reactor fuel system prior
to startup and to permit the transfer of fuel by means of inert-gas
pressure;

5. determine the feasibility of an inert-gas recycle system;

6. minimize the release of radioactive gases into the cell enclosure when
the fuel loop is opened up for maintenance or inspections; and

7. provide an off-gas system similar to that at the HRT, whereby fission-
product gases may be dispcosed of directly if desirable or necessary.

Tentative specifications which have been established for the cover-gas
system are as follows:

1. meximum flow rate of inert gas to the pump bowl, 5000 liters/day, STP;

2. maximum 02 concentration in inert gas supply, 1 ppm by volume;

)

3. maximum activity in inert-gas supply, 0.59 x 1077 w/em”; and

4. minimum supply of inert gas available at all times, based on normal
purge rate, 72 hr.

Calculations based on work by Nestor5 indicate that the equilibrium activity
of the fission gases in the pump vapor space will be about 0.1 w/cm5 when opera-
ting at 10 Mw with a 5000-liter/day gas purge. Lindsey's data® indicate that a
decay holdup period of 16 days will be necessary in order to reduce the activity

19

of the inert gas to the point where it will be acceptable for re-use as purge

gas (i.e., 0.59 x 10-2 w/cm3). Charcoal adscorber beds have proved to be a simple
and effective means for dynamic holdup of fission gases at the HRT,7 and it is
expected that they will be utilized in the MSRE cover-gas system for recycle and
for biological disposal. However, due to the necessity for storing large volumes
of gas, initial concepts provide for splitting the holdup time so that about 70%
is taken care of by the charcoal beds and the remainder by the storage tanks.
Figure 1.10 presents a simple schematic diagram of the proposed system. The im-
portant components are a volume holdup and cooler, charccal beds, gas compressor,
and storage tanks. The system provides for a delay of 1 hr in the volume holdup
and cooler to permit a reduction of the rate of decay-heat emission before pass-
ing the material into the charcoal beds. From the charcoal beds, which are de-
signed for a 17-day xenon holdup, the gas is pumped into the storage tanks to a
maximum pressure of 500 psig. Ten tanks are provided, each capable of holding a
full day's supply of gas.

From the storage tanks, the inert gas may be fed to the pump bowl or to the
transfer and buffer systems. An oxygen-sample system and an oxygen-removal sys-
tem will be provided to control the oxygen inventory. If the oxygen concentra-
tion is held continuously at 1 ppm, & maximum of 10 g of uranium per year could
be oxidized to UO,. It is hoped to maintain the oxygen level at somewhat less
than 1 ppm.

1.10 DESIGN OF THE REMOTE-MAINTENANCE SYSTEM FOR MGRE

To date, the accomplishments of the Maintenance Equipment Design Group con-
sist principally of (1) gathering information about the job and what is expected
of the group, (2) developing a plan of attack on the maintenance problems, and
(3) assigning of specific jobs and scheduling the work of the group to suit the
over-all project schedule. During this pericd, work has been done on the plan-
ning and scoping of the maintenance for the MSRE.

The design of maintenance equipment for the reactor requires that the main-
tenance operations be carefully planned and that definite procedures be outlined
for the removal and replacement of any component in the primary system. This re-
quires, in addition to tools for specific purposes, that general tools be designed
to take care of certain possible unanticipated failures.

There are three zones (Fig. 1.11) in which components must be maintained,
although remote maintenance is required in zone I only. Zone I consists of two
hermetically sealed containment pits: (1) the primary-system pit, which con-
tains the reactor vessel, the fuel pump, and the primary heat exchanger; and (2)
the drain-tank pit which contains two drain tanks and the flush tank. Zone II
consists of a large stainless-steel-lined enclosure into which the zone I pits
open when thelr seals are broken and in which are located the remote handling
tools. Zone II1 consists of the remainder of the operating part of the building,
specifically, the radiator area and the main entrance area.

Outline of the planned maintenance of reactor components in zone I is in
progress. Briefly, it consists in removing and storing faulty components and in-
stalling new components. The planning for these operations also involves pro-
viding tools for unsealing the pits, disconnecting, bagging, and handling the
faulty components as well as providing an adequate storage space and the means
for moving the components into the storage space. Finally, it involves the de-
sign of the means for intrcduction of new components through air locks into zone
ITI and the means of handling, installing, and connecting the new components.
250 Ii’rers/day\\

UNCLASSIFIED
ORNL-LR-DWG 52041

4750 Iifers/dcy\

A

5000 liters/day STP
1 ppm 0; MAX
0.59 x 10 w/cm®> MAX

e

<t

{ hour VOLUME —
HOLDUP AND CCCLER o

c—-:b
J \MAIN FUEL A

* CIRCULATING PUMP TO

oy

CONTAINMENT
VESSEL

THROWAWAY HOLDUP SYSTEM T

SURGE
TANK

72 days XENON HOLDUP

RECYCLE HOLDUP SYSTEM
17 days XENON HOLDUP

FRESH HELIUM A
SUPPLY

L

COMPRESSOR

OXYGEN
REMOVAL

.

& SAMPLE

STATION

TO SALT TRANSFER SYSTEM

TO FLANGE BUFFER SYSTEM

X

X

X

10 EACH STORAGE TANKS AT 7.5 ft°
MAX OPERATING PRESSURE: 500 psia

Figl l.lO.

7 days HOLDUP

Off-Gas Handling System.

0C
21

UNCLASSIFIED
ORNL-LR-DWG 52042

ACCESS
| MAIN AREA ZONE I1 PRIMARY-
ENTRANCE SYSTEM PIT
ZONE III
AIR LOCK
ZONE III 1 I DRAIN-TANK ZONE I ™~
STORAGE PIT PiT RAE:#TOR
Il ZONE I
\ DECONTAMINATION
AREA
PLAN
ZONE III ZONE 11 ZONE
I
ZONEII _
3 ] zoner {1
g ZONE II | ZONE
E | =
SECTION

Fig. 1.11. Operating Area.

As presently conceived, the reactor-cell seal can be broken by men working
in direct contact in zone II, after the fuel has been drained into the drain tanks
and the system flushed with barren salt. After the cell seal is broken, the
shielding material located within the sealed zone must be handled by remotely
operated equipment. The disconnecting, bagging, removing to storage, installing
of new components, and replacing of shielding will be done remotely. The seal can
be replaced by direct control; however, decontamination equipment must be provided
for cleaning zone II before the welder re-enters to seal the pits.
22

As presently planned, the remotely controlled equipment for maintenance of
zone 1 components will consist principally of

1. a manipulator of the General Mills 300 type which sheall have several
locations from which it can be operated;

2. a crane {or cranes) located in zone II but with access to all pit com-
ponents from above (particular study is being made on how best to avoid
crane failure);

3. a remotely operated air-lock truck for moving components through the air
lock;

4. viewing equipment consisting of television cameras, direct-viewing
shielding windows, and periscope equipment;

5. special portable lighting equipment to augment the area lighting; and

6. some specific tools for definite jobs such as loosening the freeze-
flange connections and disconnecting electricel and gas lines.

The remote-handling equipment in zone II must also be maintained. Direct-
contact maintenance is planned for this. The air-lock area will be used for main-
tenance of all this equipment except the crane. Provision will be made in the air
lock for decontamination of the equipment before it is repaired. The crane will
be moved to the north end of its bay for maintenance.

Maintenance operations in zone III include handling of new system components
before they are introduced into the air lock and upkeep of the components in the
radistor area. At present it is planned that the existing 30-ton crane will re-
main in the building and will be confined to operation in the northern end of the
building. Incoming components can be handled by this crane and placed in the air-
lock truck. The radiator, pump drain tank, and piping for the radiator system can
be maintained by contact maintenance if a relatively short time is allowed to
elapse after the reactor is shut down. This area will require a hoist capable of
1ifting the components and moving them to the area just inside the existing door
in the south end of the building.

An outline of the arrangement of the operating area of the building and
specifications for the major handling tools to be purchased are receiving inmme-
diate attention.

1.11 MODIFICATIONS TCO BUILDING 7503

Several building modifications are required to provide for safe and depend-
able operation of the reactor. Interior modifications are of prime concern, al-
though new exterior facilities such as filter pit, cooling tower, holdup tanks,
and retention pond are also required.

The interior changes affect the existing reactor vessel, penthouse-radiator
area, air duct to the stack, and storage-pit areas. These areas will be modified
for the installation of the primary and secondary reactor components. Shielding
walls and roof slabs, new supports, and structural alterations are required. Ex-
cavation within the crane bay and next to the existing reactor vessel is required
in order to provide a new shielded pit for reactor drain tanks.

Soo it SR e TR
23

Secondary containment to provide Protection against the spread of contamina-
tion during maintenance operations will be accomplished by the construction of a
large air-tight enclosure within the existing crane bay and over the reactor ves-
sel, hot-storage pit, and drain-tank pit. supporting facilities include decontam-
ination, air-lock, change-room, and crane-maintenance areas. Painted sheet metal
on light steel framing (stainless steel floors and walls in certain areas) will be
considered for the enclosure.

Building services will be modified and expanded. A liquid-waste system will
be provided for all areas of possible contamination. This system will permit
liquid waste to be diverted to the retention pond, soil column, or holdup tank,
based on its activity level. The water activity of the hot-storage pit will be
regulated by a recirculating system capable of removing major contaminated par-
ticles and solubles.

Air-handling systems in the existing building will be modified, and a new
system will be installed in order to control the containment-enclosure environ-
ment. Cooling requirements for the building, containment enclosure, reactor ves-
sel, and radiator area present the need for a cooling tower. An air-drying system
will be provided to service the containment enclosure during certain maintenance
operations. Handling requirements for the contaimment air also call for a filter
pit for exhaust air and a small stack for discharge to the atmosphere. Reactor-
radiator cooling will be accomplished by modifying existing blower units.

Existing diesel units will be modified to provide emergency power to the
building. A new 13.8-kva power line, in addition to the existing building line,
will be installed to provide three power sources.

Several lighting modifications, as well as a new fire Protection system, will
be required for the building.

The design criteria and laboratory design schedule concerning the modifica-
tions and new construction mentioned are nearing completion. Detailed design on
the site modifications and building services will proceed as soon as criteria and
schedules are established.

REFERENCES

1. L. G. Alexander et al., Preliminary Report on Thermal Breeder Reactor Fvalua-
tion, ORNL CF-60-7-1 (July 1, 1960).

2. J. C. Amos, R. E. MacPherson, and R. L. Senn, Preliminary Report of Fused Salt
Mixture No. 150 Heat Transfer Coefficient Test, ORNL CF-58-L-23 (May 1Lk, 1958).

5. D. Q. Kern, Process Heat Transfer, p 138-48, McGraw-Hill, New York, 1950.

L. W. M. Kays and A. L. London, Compact Heat Exchangers, McGraw-Hill, New York,
1950.

5. C. W. Nestor, Jr., Reactor Physics Calculations for the MSRE, ORNL CF-60-7-96,
(July 26, 1960).

6. Memorandum from E. E. Lindsey to D. Scott, Jr., MSRE Recirculating Gas System.
Radiolytic Degradation by Purge Gas of Lubricating Oil in Oil Seal Bowl and

Salt Pump, (July 25, 1G60).

7. R. E. Adams and W. E. Browning, Fission Gas Holdup Tests on HRT Charcoal Beds,
ORNL CF-58-4-14 (Apr. 2, 1958).

2. COMPONENT DEVELOPMENT

»,1 FREEZE-FLANGE DEVELOPMENT

Analysisl of the thermal stresses generated in freeze flanges fabricated
from an existing design (ORNL-LR-DWG 31199) indicates that the flanges may fail
after a relatively small number of thermal cycles. Further analytical and ex-
perimental work, including photoelastic stress analysis, was initiated to produce
an improved design.

A facility for thermal cycling flanges between room temperature and 1400°F
was designed, and fabrication is nearing completion. The general arrangement of
the facility is shown in Fig. 2.1. Heaters on the two tanks will be energized at
all times during testing. At the start of a thermal cycle, with all the molten

UNCLASSIFIED
ORNL-LR-DWG 52043

SPRING-LOADED
RELIEF VALVE

VENT T~ g /

s& P

SYSTEM-OVERPRESSURE
REGULATING VALVE

UPFLOW PRESSURE
REGULATING VALVE

l< TEST FLANGES

N

Fig. 2.1. Freeze-Flange Thermal-Cycle Facility.

24

25

salt in the lower tank, heaters on the piping adjacent to the flanges will be
manually energized. When the flanges reach a predetermined temperature, the
molten salt is autometically forced by gas pressure and gravity to flow from one
tank to the other. This oscillating flow through the flanges is maintained for
a specified length of time, after which the pipe heaters are automatically de-
energized and the salt is drained to the lower tank. The flanges are then
allowed toc cool before starting the next cycle.

It is believed that the occurrence of excessive gas leakage during cool-down
of existing flanges is at least partially due to overstressing and subsequent
relaxation of the copper ring gasket during heatup and hot operation. Gold-
plated Inconel gaskets are on order for testing in the thermal-cycle facility.
These rings, with higher yield strength and a coefficient of thermal expansion
identical to that for the flange material, should alleviate the gas-sealing
problem. A more flexible clamping ring is being designed to reduce gasket-stress
cycling.

The cooling requirements for the test flange will be evaluated to determine
if forced air cooling can be eliminated without jeopardizing the frozen seal.
The resulting higher temperature at the outer edge of the flange will also reduce
the thermal stresses. Attempts will be made to design the flange to withstand
external loads consistent with the code for the pipe size. To facilitate main-
tenance, two sizes of flanges will be developed for use with all pipe sizes in
the MORE.

2.2 FREEZE VALVES

Two valves are being fabricated for test as fast-freeze or fast-melt valves
for use with molten-salt drain systems. These valves contain no moving parts.
A smell restriction is created by crimping a l—l/2-in. sched-40 INOR-8 pipe, in
which a solid mass or "plug” is formed by lowering the salt temperature at that
point. As shown in Figs. 2.2 and 2.3, two methods of heatjng the valve are being
examined.

One valve (Fig. 2.%) is heated by electrical resistance applied along the
section of pipe containing the V-restriction. Power requirement for a 3- to
5-min melt time is approximately 3 kw. Cooling is accomplished by passing air
over the pipe restriction through a 2-in. tube which is welded around the "crimp".
The air requirement has been estimated to be 20-30 cfm for a cooling time of
about 15 minutes.

The alternative valve (Fig. 2.2) is heated by a high-frequency induction
coil (150 to 400 kc). This valve is made of 1-1/2-in. sched-40 INCR-8 pipe, and
the restriction consists of a 2-in. flattened section formed by pressing the pipe.
A l/h—in. copper U-shaped induction coil is formed around the 2-in. plate for
ease in removal. Bench tests in air indicate that the melt may be accomplished
(AT of approximately 800°F) in 1/2 to 2 min. Cooling will be done by passing air
through a nonmetallic chamber which surrounds the flattened section of pipe.

2.5 SAMPLER-ENRICHER DEVELOPMENT

The design of the sampler-enricher for the primary loop of the MSRE in-
cludes provisions for removal of the entire assembly for maintenance. A reliable
valve and flange are required to prevent the escape of fission-product gases
during this operation.
26

UNCLASSIFIED
ORNL-LR-DWG 52044

Fig. 2.2. Induction-Heated Freeze Valve, 1,5 in.

UNCLASSIFIED
ORNL~ [ R-DWG 52045

Fig. 2.3. Resistance-Heated Freeze Valve, 1.5 in.

27

An apparatus is being fabricated to test the feasibility of metal freeze
flanges and valves for this application. The metal freeze seal is based on g
soldered joint which can be opened and closed by melting and refreezing a pool
of solder. Joints must be leak-tight to helium (<lO"7 cm3/sec) and must be
strong enough to withstand a pressure differential of 30 psi. The wetted portion
of the apparatus is fabricated from INOR-8. A 72% Ag - 28% Cu silver solder
(melting point 1434°F) will be used as the metal freeze seal. This alloy was
recomnended by Metallurgy Division as having the desired wetting characteristics
and thermal expansion properties compatible with nickel-based alloys. The high
melting point is considered an advantage in that no additional cooling is re-
quired beyond that necessary to remove the gaseous fission-product decay heat.
Diffusion of constituents between the molten solder and INOR-8 is considered a
problem and is being studied by the Metallurgy Division.

The unit being constructed will be used to test the leak-tightness of the
proposed type of joint, both initially and after repeated openings of the valve
(or flange). By monitoring the melting point of the solder, the effect of the
diffusion of nickel from the INOR-8 and of the copper into the INOR-8 can be de-
termined. The metal from the valve will also be examined for diffusion effects.
Strength and leak rates will be evaluated as a function of the time the solder
is molten and of the number of operations of the valve. Heating rates for
various types of heaters will be measured.

2.4 MSRE CORE DEVELOPMENT

A one~fifth-scale, geometrically similar, plastic model of the proposed
MSRE core (see Sec. 1.4) is being built. The model will be operated with water
at approximately LO°C and a flow of approximately 50 gpm; the following condi-
tions of flow similarity will exist: the linear dimensions, inlet fluid velocity,
and inlet Reynolds number for the model will be equal to those for the MSRE core,
and the relative surface roughness of the model will approximate that for the
MSRE core.

Components of the MSRE core that are being simulated are the inlet volute,
core-wall cooling annulus, and lower plenum with baffles for reducing swirl. An
orifice plate is used to simulate the pressure drop across the graphite lattice.

Measurements to be made include pressure and velocity distribution, and
local fluid residence times in the lower plenum. Fluid transport of solid par-
ticles through the model will be investigated.

Design was started also on a full-scale model of the MSRE core. This model
will be used for the final fluid dynamic test of the MSRE core.

A glycerol solution having a kinematic viscosity equal to that of the MSRE
fuel salt will be circulated through the model at a flow rate of 1250 gpm.

2.5 REMOTE-MAINTENANCE DEVELOPMENT FACILITY

As previously reported,2 the remote-maintenance facility was operated at
1200°F and then disassembled to demonstrate equipment and procedures previously
developed for maintenance. The following comments are a result of experience
gained from this operation.
28

Tools and technigues developed prior to molten-salt circulation were in
general satisfactory after the salt circulation for the remote removal and re-
placement of the salt circulating pump and motor, reactor vessel, east heat ex-
changer, and pipe preheater. Closed-circuit stereotelevision was used as the
only viewing medium for the maintenance operations and was satisfactory. Main-
tenance on the television equipment is considered excessive as one or more of the
four television sets required maintenance every 200 hr of continuous duty.

After the pump and freeze flanges were disconnected, it was found that a re-
mote tool was necessary to clean gasket grooves prior to reassembly. In addition,
a method is necessary to minimize the size of the salt cake formed in the freeze
flanges and to confine i1t and prevent it from falling on the floor when the
flanges are disassembled.

The pipe lines should be pitched a minimum of 3° for the molten salt to
drain completely from the pipe lines and flanges when the loop is drained for
maintenance. This was not done for all locations in the facility, and some of
the pipe lines were partially filled with salt after the loop was drained. The
drain tank should be the lowest point in the piping system soO that the molten
salt will flow by gravity into the drain tank. In this facility a pressure dump
was necessary as the drain-tank level was above the lower runs of piping. This
latter method leaves some salt in the fill and drain line. In addition, a better
method is required to prevent the freezing of salt in the freeze-flange area of
the fill and drain line when the line is full of stagnant salt. A complete de-
scription5 of this work is being prepared.

Final editing of the color motion picture with sound, "Remote Maintenance
of Molten-Salt Reactors," was completed and copies were issued.

2.6 FORCED-CIRCULATION CORROSION LOOPS

Operation of long-term forced-circulation corrosion test loops, nine of
INOR-8 and two of Inconel, was continued. The accunulated hours of operation
range from 3000 to 18,000. Nine loops have operated for a year or more. Table
2.1 gives a summary of the loops in operation during this period.

Operation of INOR-8 loop MSRP-12 was inadvertently terminated June 29 during
a routine pump-drive maintenance operation. The insulated cooler-coil support
had worn sufficiently to allow contact of coil and frame. Wwhen power was applied
+to the coil to allow maintenance, a hole was burned in the tubing wall causing a
leak. The loop was dumped and cooled immediately.

INOR-8 loop MSRP-16 had a pump failure on January 31. It was necessary for
a new batch of salt mixture BULT-14 to be prepared for filling the loop after
pump replacement. The loop was restarted June 28. The higher temperature of
operation (1500°F) probably contributed to the pump difficulty.

L

Examination of INOR-8 loop 9354 -4 " was completed and is reported in Sec.
4.1. The disassembly of the pump (Hastelloy B) revealed no metallic deposit or
attack. All salt-wetted surfaces were clean and bright.

Loop 9354-3 was drained temporarily on July 1, when an oil line ruptured on
an auxiliary system. It resumed operation after the oil-scaked insulation was
replaced.

29

Table 2.1. Molten-Salt Forced-Circulation Corrosion-Loop
Operations: Sumary as of July 31, 1960

Hours of
Loop Composition (mole %) High Wall Operation at
No.* LiF QWaF BeF2 UF),  ThF) Temp.(°F)  Conditions Comments
9354 -3 35 27 38 1200 18,520 Loop drained tempo-
rarily and restarted
MSRP-7 71 16 13 1300 16,920 Normal operation
MSRP-6 62 36.5 0.5 1 1300 16,774 Normal operation
MSRP-10 55 L5.5 0.5 1 1300 16,087 Normal operation
MSRP-11 5% 46 1 1300 15,813 Normal operation
9377-5 62 36.5 0 1300 14,536 Normal operation
MSRP-12 62 36.5 0 1300 14,408 Terminated 6-29-60
9377-6 71 16 13 1300 12,041 Normal operation
MSRP-15 67 18.5 0.5 1k 1400 6,696 Normal operation
MSRP-14 67 18.5 0.5 1k 1300 6,339 Normal operation
MSRP-16 67 16.5 0.5 1k 1500 2,869 Resumed operation
6-28-60 after pump
change

*All loops constructed of INOR-8, except 9377-5 and -6, which are of Inconel.

2.7 PUMP DEVELOPMENT

dork on the primary and secondary pumps for the MSRE received most emphasis
during the quarter. Testing was continued for the advanced molten-salt pump
program, and the long-term test of the MF-F pump (nearly three years of contin-
uous operation) was terminated.

2.7.1 MSRE Primary Pump

A cost estimate and schedule were prepared to cover design, fabrication,
and testing of primary pumps. Preliminary layout of the primary pump was com-
pleted and is being reviewed. Design of the water test of the primary pump was
completed, and fabrication was started. Procurement of INOR-8 castings of im-
peller and volute was initiated. A tentative specification for an electric drive
motor for the primary pump was written.

The selected impeller and volute for the primary pump will be tested in
water. Control of pump tank liquid and leakage flows, prevention of gas entrain-
ment, priming and liquid-level limits of operation, and hydraulic performance
will be investigated.

Design of the water test loop was completed, and fabrication and construc-
tion began. The design of the water-test-pump rotary element was completed and
fabrication started.
30

A layout of the hot test prototype of the primary pump was completed and is
being reviewed. The impeller and volute and the pump tank heads, to be fabri-
cated of INOR-8, were placed for procurement. Pipe for the hot test stand is on
hand.

Calculations of temperature gradient and thermal stress in the uninsulated
pump tank wall (exposed to gas on both sides) as a function of gamma-heat gener-
ation rate have been completed. The calculations included the radiation and
natural-convection heat transfer from the molten-salt surface and a constant
value of beta-heat generation from gas-borne radioactivity. A constant heat-sink
temperature of 200°F was assumed. The surface temperature of the outside of the
pump tank wall varied from 1000 to 1165°F, and the maxirmum thermal stress at the
outside surface of the wall varied_from 1920 to 2060 psi as gamma-heat generation
rate was varied from O to 1.0 w/cm’.

An experiment was designed to determine the extent of the back diffusion of
fission gases up the pump-shaft annulus to the region of the lower oil seal.
This diffusion rate will be determined as a function of flow rate down the shaft.
Krypton-85 will be added to the helium which normally occupies the pump tank.
Equipment is being fabricated for the experiment, and testing will be performed
in an existing NaK pump test stand. Conversion of this stand for molten-salt
operation is nearly complete.

Specifications are being prepared to cover the design and construction of
the primary-pump drive motor and containment shell. The same basic specification
will be applicable to the secondary-pump drive motors.

2.7.2 MSRE Secondary Pump

A cost estimate and schedule were prepared covering the design, fabrication,
and testing of the secondary pumps. FProcurement of the INOR-8 impellers and
volutes was initiated.

2.7.3 Advanced Molten-Salt Pumps

Hydrodynamic Journal Bearings.--One test was performed on the molten-salt-
lubricated hydrodynamic-journal-bearing tester. The test bearing and journal
(an. INOR-8 bearing, three helical-grooved, and a carburized INOR-8 journal)
experienced a seizure after 160 hr of operation at 1200°F, 1200 rpm, and 10 to
200 1be. Throughout the test, satisfactory performance was in doubt. Investi-
gations are under way to determine the cause of seizure.

Pump Egquipped with One Molten-Salt-Lubricated Journal Bearing.--The test5 of’
the pump eguipped with one molten-salt-lubricated journal bearing (test No. 6)
continued throughout the quarter. To date, the pump has operated for 3000 hr and
has been stopped and started 63 times. One stop was made as a result of the
molten salt being dumped inadvertently. Normally, when the molten-salt level in
the system is reduced, the pump 1is stopped automatically. In this particular
case, the pump-tank liquid-level probes were shorted and therefore did not trip
the automatic device which stops the pump. The pump continued rotating for ap-
proximately 20 min after the system was dumped. The pump was placed back in
operation 16 hr later, and there was no evidence of damage to the salt-lubricated
bearing; pump operation is continuing.

31

2.7.4% MF-F Pump Performance Loop

The MF-F pump test6 was terminated after a total of 25,500 hr of continuous
operation at 2700 rpm and 645 gpm. The last 23,700 hr of the test included in-
tentlonal operation of the impeller in a cavitation region with NaF-ZrFL-UF)
(50-46-4 mole %) at 1200°F (pump-tank gas pressure of 2.5 psig). Examination of
the pump parts revealed that the impeller had been damaged considerably by the
cavitation. There was evidence that oil and/or 0ll vapors entered the system by
traversing down the pump-shaft annulus from the lower-seal catch basin to the
bump expansion tank. A report of the test will be written. The impeller is
being sectioned for metallurgical examination.

2.7.5 Frozen-Lead Pump Seal

The small frozen-lead-sealed pump,T consisting of a centrifugal pump mounted
vertically over a fractional horsepower motor drive, has operated continuously
during this period to accumulate 18,708 hr of operation. The pump circulates
molten salt isothermally through a small loop at 1200°F.

2.7.6 MSRE Engineering Test Loop

An engineering test loop is being designed for construction in Building
920k-1 for the purpose of evaluating certain MSRE components and operations. The
facility will consist of a centrifugal pump circulating molten salt at 25 to 100
gpm through a 1-1/2-in. bipe loop. Prototype mechanisms and mockups which will
be included are: a remotely operated salt sampler and enricher, a graphite con-
tainer, freeze-type flanges, dual fuel and flush tanks with asgociated freeze
valves, controlled-atmosphere maintenance provisions, gas-handling system, level-
measuring device, and piping heaters. The system will operate isothermally at
temperatures up to the maximum expected MSRE temperature.

REFERENCES

1. ©Sturm-Krouse, Inc., Thermal Stress Analysis of Large Freeze Flanges for the
Molten Salt Reactor Project (July 6, 1960).

2. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 8.

3. C. K. McGlothlan and W. B. MecDonald, Remote Maintenance Development,
ORNL-2981 (to be issued).

k. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 12.

5. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 3.

6. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 7.

7. MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 23.

3. REACTOR ENGINEERING ANALYSIS

3,1 EFFECTS OF GRAPHITE SHRINKAGE UNDER RADIATION IN THE MSRE CORE

The amount of shrinkage of graphite under irradiation varies linearly with

1.6
UNCLASSIFIED ‘ ~
ORNL-LR-DWG 52047 N
1.6 1.4 I ‘
taf——t— 'g 1.2 | :
o .
1.2 £ .
— ™~ 1.0 '
2 0 |
7] =
o & \ '!
£ 10 &
~. Lju 0-8 \ i
2 L .
(o] —
A ; l l.\ ‘
= 2 o 5
o 0.6 © |III |
: A\
0.4
0 | |
. FLUX CURVE |
02— T— $=1652x10" cos Z/op g ] 0.2 ‘ i
1 J_JJQ N ‘
0 - | | o |
o 3 6 9 12 15 1B 21 24 27 30 33 o) 3 6 9 2 15 18 21 24 27
z, AXIAL DISTANCE FROM MIDPLANE {inJ RADIAL DISTANCE FROM CENTER LINE (in.)
Fig. 3.1, Axial Flux Distribution at Fig. 3.2. Radial Flux Distribution at
Center Line. > 1 Mev. Midplane.

flux for the exposures that have been tested, and differs between graphites and

between axes within a piece.

The more permeable graphites appear to have lower

shrinkage rates, by as much as a factor of 8, than the less permeable types

needed for use with molten salts.

For CEY graphite, typical
shrinkages for a neutron dose of 10
0.004 parallel and 0.002 normal to

of graphite expected to be used in the MGSRE,
nvt of energies of 1 Mev and greater were
the extrusion axis.

The plots of the neutron

flux of energies greater than 1 Mev in the MSRE are shown in Figs. 3.1 and 3.2.

The value at the center is 1.52 x 10
year at full power of 10 MwT.
tends to produce bowing, is 0.0735 x 10t
midplane.

32

nv, or a dose of 4.8 x 10
The maximum flux differentisl, radially,

nvt for a full
which
nv per inch, at a 20-in. radius on the

UNCLASSIFIED
ORNL-LR-DWG 52046

33

The shrinkage may produce four effects, which are considered separately.
These are:

1. tension at the surface, caused by the fast-flux depression in the center
of a graphite piece;

2. gross axial shrinkage of pieces, tending to shorten the reactor;

5. ‘transverse shrinkage, which tends to increase the fuel volume fraction
within the reactor;

4. bowing of the pieces, caused by flux gradient.

Calculations have been made on the assumption that there is no yielding or
annealing, although there is evidence that shrinkage-induced stresses may be re-
lieved. Results of the calculations are shown primarily as strain and second-
arily as stress, which is the product of strain and Young's modulus. The strain
to produce rupture has been shown to be about 0.001 in./in. for graphite made
from coarse particles, and increasing to about 0.003% for types made from fine
particles. It is estimated to be greater than 0.002 for CEY graphite. As
Young's modulus and strength at rupture both increase with irradiation, and as no
data are available, both the stress prgduced by a strain and the allowable stress
are very uncertain. However, 1.5 x 10° psi is used for Young's modulus to indi-
cate the value of the stresses that might be produced.

The analysis presented here is based on a core 66 in. long by 27 in. in
radius, made of 2- by 2- by 66-in. pieces, with axial fuel channels cut into each
face. Some effects, particularly in relation to bowing, would be entirely dif-
ferent for different-size pieces and a different arrangement.

The flux depression within a graphite piece is due to the Progressive ther-
malization of the neutrons and may be considered separately from the macroscopic
variation of the flux across the piece, and the results added. The variation be-
tween average and edge, which is proportionel to the tensile strain, amounts to
1.4% of the average flux (see Fig. 3.8). At the reactor center after a full year
at 10 MwT, the strain would be 0.00027. With E = 1.5 x 10¥ psi, the stress would
be 400 psi in tension.

The axial shrinksge of a 66-in.-long Piece at the reactor axis is calculated
to be 0.860 in.; for an edge piece the value is 0.122 in. (see Fig. 3.3).

UNCLASSIFIED
ORNL-LR-DWG 52048

0
//

c
= 0.5 / .
g —
=
£ 1.0 — —
I
)

1.5 :

0 4 8 12 16 20 24 28
RADIUS (in.)

Fig. 3.3. Axial Shrinkage vs Radius
After One Full-Power Year,
34

Transverse shrinkage, which is half of the axial shrinkage, is greater in
the center than at the surface of the reactor, as shown in Table 3.l.
This shrinkage would result in opening of the fuel passages, and to a greater
extent in the center than at the surface.

Table 3.1. Transverse Shrinkage
(inches per inch per full-power year)

Radial Location

Axial Location Center Edge
Center 0.0096 0.0010
End 0.0012 0.0002

The flux gradient across the graphite pieces would produce outward bowing at
the midplane, allowing an unrestrained reactor to become barrel-shaped. The
bowing effect reaches a maximum about three-fourths of the way out from the cen-
ter, with the outer pieces less bowed.

With the center of the reactor banded, the effect of transverse shrinkage of
the pieces must be considered. The bowing would cause inward packing at the ends
and outward packing, against the band, at the midplane. The ends would be keyed
against rows slipping, with the result that concentric rings would be locked to-
gether, with the rings inside relatively free. The pleces would bow outward at
the midplane, increasing the fuel fraction most at the center.

Deflections expected of unrestrained graphite pieces after a full-power year
are listed in Table 3.2 for a radial row. Also tabulated are the shrinkage per
5. x 2-in. piece at the midplane and the strain if the pieces are kept straight.
The hoop load to keep them straight, allowing for pieces bowing to teke up
shrinkage but not considering the expansion of the band under the load, is
11,000 1b.

3.2 TEMPERATURE-RISE FFFECTS IN MSRE CORES WITH ROUND AND FLAT FUEL CHANNELS

The temperature effects in cores with round and flat fuel channels, with the
fuel in laminar flow, have been investigated. It has been found that the differ-
ence between the fuel-graphite interface temperature and the mixed-mean fuel tem-
perature would be excessive if round channels were used but would not be if the
same fuel flow were provided in the form of flat channels.

The two core configurations considered are jllustrated in Fig. 3.4. In both
cases, it 1s assumed that the core is an asseumbly of graphite stringers of square
cross-section extending the full height of the core. The fuel flows vertically
through the core, making a single pass in both cases.

In the round-channel case, the fuel channels are produced by machining a
quarter-round of radius r at each corner of the graphite blocks. The blocks are
packed in as tightly as possible, but it is assumed that there will be cracks of
some width t between the blocks due to machining tolerances (these are exagger-
ated in the sketch).

35

Table 3.2. Deflections, Shrinkages, and Strains at Midplane
for a Radial Row After One Full-Power Year

Piece Deflection Shrinkage Bending Strain
No. (in.) (in.) (%)
0 (axis) 0 0.0191 0
1 0.057 0.0190 0.013
2 0.137 0.0186 0.030
3 0.192 0.0179 0.042
4 0.229 0.0170 0.050
5 0.272 0.0159 0.060
6 0.315 0.0145 0.069
7 0.367 0.0130 0.081
8 0.390 0.0113 0.086
9 0.407 0.0096 0.089
10 0.421 0.0078 0.092
11 0.412 0.0060 0.091
12 0.375 0.0042 0.082
13 0.315 0.0027 0.069

In the flat-channel case, the blocks are spaced spart a distance 2r in some
convenient manner {e.g., by providing bosses on the blocks).

In the comparison, it is assumed that the comparable cores contain equal
fractional fuel cross-section f and the same center-to-center block spacing s.

Poppendiek and Palm.er5’l'L have presented solutions for the difference between
the temperature at the channel wall and the mixed-mean fuel temperature for a
fluid in laminar flow containing a volume heat source. Their equations may be
rewritten in terms of the geometric parameters of these systems as follows:

2
Q, F f s 1-F
(t -t ) - R E { 0.229 [1 e
Round nk [f+ =2 (1- 2./f/n )]

rFr s

bo—= . 2t fl-eJfl&)J -oa&'}.

S Tor [ 2 { £(1- Fpp) }
. = == |2 (1-./1-T 0.485 |1 + - 0.400).
(t5-ty) kT [2 ( ) ] [ P (1-4/1-1)V/1-f }

In the round-channel case, the terms containing t represent the contribution
of the fuel trapped in the cracks.
36

NOTATION
a = shorter dimension of rectangular body used in Jakob's6 solution
for rectangular solid with uniform volume heat source, ft
b = longer dimension of rectangular body, ft

¢ = value for temperature in rectangle tabulated on page 179, ref 6,

dimensionless
f = fractional fuel channel volume, dimensionless
FPF = fraction of power released in fuel
F = wall conduction factor in round-channel equation,5 dimensionless
F/ = wall conduction factor in flat-channel equation,u dimensionless

H = core height, ft
k = thermal conductivity, Btu/hr(ft)(°F)
P = total reactor power, Btu/hr

p/pav = ratio of local to average power in core
g’/ = volume heat source term in Jekob's® solution for rectangular body
Qg = reactor power denmsity, Btu/hr(ftJ)
Q, = fuel power demsity, Btu/hr{ft”)
r = radius of round channel or half-width of square channel, ft

= center-to-center spacing of graphite blocks in core, ft

t = thickness of cracks between adjacent graphite blocks in round
channel case, ft

= temperature at channel wall

0
tm = mixed-mean fuel temperature
6 = local temperature in infinite rectangular body containing a
volume heat source, the temperature at the boundaries being zero
(6) = center-line temperature in infinite slab containing a volume heat

b = e« source, the temperature at the boundaries being zero

If we meke the assumptions that Fpp = 1 and t = 0 so that there is no heat
transfer across the fuel-grephite interface, we then find that:

(t -t )
°© ™ pound _ Lr (0.062)

(to_t ) n(l-\/l-f)2 (0.085)

T f1at

For £ = 0.1, the approximate value of interest in the present case,

(to-tm)
Round

357 -
(t,-t )
Flat

37

UNCLASSIFIED
ORNL-LR— DWG 52049

CHANNEL.

FUEL CHANNEL

{#) FLAT-CHANNEL CORE

Fig. 3.4. Core Assemblies,

Thus it is clear that the volume heat source produces a much less severe
temperature-rise problem in the flat-channel case. The simplifying assumptions
made are favorable to the round channels.

For the flat-channel core assumed for the MSRE, the following dimensions
were assumed:

s = 2in. = 0.167 ft,

r = 0.05 1n.,

from which f = 0.097,

W = flR]éDH (p_ffi ’

P = 10 M = 3.415 x 107 Btu/hr,
R = 2.5 ft,

H = 5.5 ft.
38

The nuclear calculations5 give the radial peak-to-average power as 2.08, and
if the axial power distribution is assumed to follow the cosine law, the axial
ratio is 1.57 and

1
_3:M13 %300 . 508 4 1.57 = 1.035 x 10° Bru/(hr)(ft0) .

(@ >
Max (2.5)° x 5.5

=)

il

Taking k of the fuel = 2.75 and Fpp 0.96,

6 - T2
o (1033 x 10 [0-167 ) [ 0.097 x 0.0k }
b <2.75 7 0.007/ | 2 (10095 {0'”85 1+ 57598 (0.05)(0.95), ~0+*°

5
(3.88 x 106) (L.17 x 10"3) %%-(1 + 0.043) - 0.4%0 } = T.2°F.

Jakob6 gives a solution applicable to the temperature rise in the graphite.
His solution takes the form

6= c (6)
O b ’

=
(9 ) _ 9/// a2
Ol - o T2k )
For our case, ¢/ = Qg (1- FPF) = 0.04 x 1.033 x 106 = 4.13 x 1oLF Btu/hr(ft5)
_ 2 -0.00
k = 12, a = 5095, < 0.083 f4,
(6 ) - 4.15 X loh‘ X (O-O(le2 = 11 80F
o’y L 2 x 12 : ?

and for b/a =1, y/b =0, x/a =0, ref6 gives ¢ = 0.59.
Thus 6 = 6.59°F.

Therefore it is clear that neither fuel nor graphite temperatures would be
excessive in the flat-channel core using laminar flow. For the round-channel
core, excessive fuel temperatures would be encountered in laminar flow. Calcu-
lations for turbulent flow show that the fuel temperature 1ls not excessive in
round channels in that case.

For the 0.2-in.-thick 1.25-in.-wide channels used in the actual core design,
the fuel fraction in the core is 0.1188. The difference between the fuel-graphite
interface temperature and the mixed-mean fuel temperature at the center of the
core increases to 23.6°F, while there should be a negligible change in the tem-
perature at the center of the graphite relative to that at the surface of the
graphite, and the center of the graphite will be approximately 30°F above the
mixed-mean fuel temperature at the center of the core.

39

3.3 TEMPERATURE OF FUEL IN A BLOCKED PASSAGE IN THE MSRE
An analysis was made to determine the temperature which the MSRE fuel may
attain 1f one of the fuel passages becomes blocked so that no fuel flow can
occur.
If unidirectional conduction is assumed, the problem is fairly straight-
forward, although the answers will be ultraconservative. BEssentially it is a
superposition of the following solutions:

(a) conduction through an infinite slab of fuel with a uniform volume heat
source,

(b) conduction of the heat released in the fuel across a graphite slab to
the adjacent fuel channel,

(¢) temperature rise in the graphite due to its volume heat source,

(d) temperature rise in the adjacent open fuel channel due to its volume
heat source.

Part a is the solution of the equation

with the boundary conditions

. . Wt
f(x=0) = 2k

For the MSRE core at the point of maximum power density,

g = 1.02 x 107 Btu/hr(rt’),

k = 2.75 Btu/hr(fit)(°F),

£ = 0.05 in. = 0.00417 ft,
and

t(x:O) = 52.1 F-

Part b is the solution of the same equation with boundary conditions:

0 X = £,
, 1/q )
tc X (A at x =0 .
40

From the solution to part a,

% = -k _t7 , t

!
£20) T

= -qf = 1.02 x 106 x 4.17 x 10'5 = L2.5 x 1o3 Btu/hr(ftg) :

=8
|

3
;1 (Q) _ Lo.5 x 107 3
tl= (A) = 5 = 3.54 x 10° °F/ft .

Solving the equation for part b,

2
3 qct
tc(g) = 3.5k x 107 + F3z
£ = 1.9 in. = 0.158 ft,
L 3
q, = k.56 x 10" Btu/hr(£t’) ,
k, = 12 Btu/hr(ft)(°F) ,
2
4 -1
) 3 -1 k.56 x 107 x (1.58 x 10 7)
t(4) = 3.5% x 10° x 1.56 x 10 ™ + 2 x 12

560 + k7.5 = 607.5°F.

Part & is the solution to the laminar-flow equations for flat channels:
2
Qr
-

£ L
Q) _ I 3z 4.56 x 10 x 0.158 )
. (dA = k_ <5.35 x 107 + = ) - 12 <5.55 x 10° + B ,

4.67 x 10" ,

]

Qv = 1.02 x 107 Btu/hr(ft3) for fuel at the center of the reactor,

H
!

0.05 in. = k.17 x 10° Tt,

Qr = 425 x 10% Btu/nr(rt?) .

. L.67 x 104
h.o5 x th

Thus F/ = 1

i b b 4 e B

4]

Qr2
£t = Y 17F - 14
o m =k 35 ’

2
1.02 x 107 x (4,17 x 103) < 35.7 - 14 )
2.75 35 ?

64l x 0.62 = 39.9°F.

Thus the temperature in the blocked channel will be
32.1 + 607.5 + 39.9 = 679.5°F
above the mixed-mean temperature in the adjacent open channel.

It is evident that the temperature rise across the graphite block is the
major factor in the above, and it seemed of interest to investigate the effect of
two-dimensional conduction in the graphite. Since no analytical solution was
found, this problem has been solved by a relaxation procedure, and the graphite
temperature drop was found to be 218°F or about a third of that for unidirectional
conduction. This answer is low since it is based on the assumption that the
graphite - flowing-fuel interface temperature is constant, while it doubtless is
somewhat higher near the blocked channel than elsewhere. Based on this solution,
the maximum temperature in the blocked channel would be about 290°F above that in
an adjacent flowing-fuel channel.

The above results suggest that if a single fuel channel were blocked, the
fuel temperature probably would be no more than 400°F above the mixed-mean tem-
perature in adjacent open fuel passages.

3.4 MSRE REACTOR PHYSICS

One-dimensional multigroup and two-dimensional two-group calculations have
been performed to obtain estimates of critical mass, flux and power-density dis-
tributions, and temperature coefficient of reactivity. Other calculations during
the period were concerned with gamma heating in the various INOR-8 structures
assoclated with the core, drain-tank criticality, and estimates of the heat de-
position and radiation dose due to fission products in the pump bowl.

3.4.1 Core Caleculations

For the purpose of survey calculations the reactor core was considered as a
bare right-circular cylinder 54 in. in diameter and 66 in. high. The total geo~
metric buckling of the bare reactor was then inserted in the input to GNUT as a
transverse buckling, and the reactor was treated as a slab with zero net current
on both faces. (This procedure eliminated the iteration on source shape and con-
sequently gave a considerable reduction in computing time.)

Calculated critical masses for fuel volume fractions of 0.08, 0.10, 0.12,
0.14, and 0.16 are shown in Fig. 3.5. Fuel-salt compositions are listed in
42

UNCLASSIFIED
ORNL—LR-DWG 52050
80 160

INVENTORY, COMPOSITION A

—_—
70 / 140
[ ]
A /

9 120

\ .

50 <

INVENTORY,COMPOSHW;;:;"\n\‘~‘
l\
A

CRITICAL MASS,
COMPOSITION A

30 i — 60

e J— 3

~

20 ,,””’ 40

CRITICAL MASS,
COMPOSITION B

.08 0.10 0.12 0.4 0.16
VOLUME FRACTION FUEL SALT

Fig. 3.5. Calculated Critical Masses
for Various Fuel Volume Fractions.

Table 3.3. Fuel-5alt Compositions
(Atomic densities in atoms per barn centimeter)

Constituent Composition A Composition B
Be 8.820 x 1077 1.002 x 1072
Li6 5.463 x 1077 6.20h x 1077
11! 1.621 x 10°° 2.068 x 1072

(99.997%)

F 4.154 x 1072 4,718 x 1072
Zr 0 1.292 x 1077
Th 1.138 x 1077 0
°3? 2.660 x 107" 5,021 x 107
P30 1.850 x 1077 £.100 x 1077

43

Similar calculations were performed for composition A with 8 vol % fuel salt
Tor reactor diameters of 3.5, 4.0, 4.5, and 5.0 ft and reactor heights of 5.5 and
10 ft. Calculated criticsl masses are shown in Fig. 3.6.

Multigroup one-dimensional calculations were also performed for the reactor
model illustrated in Fig. 3.7. These calculations, one along the midplane and
one along the center line, were employed to generate two-group conskants for use
in two-dimensional calculations with the IBM-704 program Equipoise. In general,
good agreement was obtained between the results of critical calculations by the
different methods.

Two-group flux and adjoint-function distributions were obtained from the
Equipoise® calculations and were used to estimate the neutron lifetime and the
reactivity loss associated with changing from isothermal startup conditions to
power operation. The calculated neutron lifetime was 290 usec; the perturbation-
theory estimate of the reactivity loss was 3.9 x 10~ 8k/k, compared to 5.0 x 10-%
as obtained from calculation of both the perturbed and unperturbed cases. Part
of the above discrepancy was due to neglecting the change in diffusion coeffi-
cient in the perturbation-theory calculations.

UNCLASSIFIED
ORNL—LR—DWG 50592A

50

) 10 ft H
o«
<
= \\\\\t * ®
|
b 55 ftH _
(@] g )
= 20 . 2 &
- Y QG
b el [=]
>
o
10 < - | &
\ ]
_.b
0 0
3.5 4.0 4.5 5.0

CORE DIAMETER (ft)

Fig. 3.6. Critical Masses and Concen-
trations for Various Diameters.
44

UNCLASSIFIEQ
ORNL—LR-DWG 50588

- -

L can (Vs INOR 8)

A ANNULUS

| | VESSEL (¥ INOR 8)

CORE

2}
° GRID |
{
2%
2] GRID 2
| HEADER

DIMENSIONS ARE IN INCHES

Fig. 3.7. Reactor Model.

An estimate of the quantity (8k/k)/(8M/M), where M is the mass of fuel in
the reactor, was cobtained from the GNU calculations. Assuming that the only ef-
fect of a temperature increase is the removal of fue% from the core, the estimated
temperature coefficient of reactivity was -2.4 x 107 (6k/fi)/°F based on a volu-
metric expansion coefficient of the fuel salt of 1.0 x 107'/°F and a (sk/k)/(eM/M)
value of Q.24 as obtained from the GNU calculations.

3.4.2 Gamma-Heating Calculations

The two-dimensional Equipoise calculations were used to provide source esti-
mates for the gamma-heating calculations, which were done with the IBM-T7O4 pro-
gram.Nightmare.9 Results of the gamma-heating calculations are given in Table
3.4 for a reactor 54 in. in diameter and 66 in. high, with 8 vol % fuel salt in
the core, operating at 5 Mw.

3.4.3 Drain-Tank Criticality

Criticality calculations were performed7 for a cylinder 5 ft long and 5 ft
high containing fuel salt. This cylinder, somewhat larger than the currently
proposed drain tank, had a calculated multiplication constant of O.4k4 when bare
and 0.77 when reflected with an essentially infinite layer of water.
45

Table 3.4. Results of Gamma-Heating Calculations for MSRE at 5 Mw

Heat Generation

Location (w/cc)
Core can 0.10
Pressure vessel (inside) ] 0.10
midplane
Pressure vessel (outside) 0.065
Pressure vessel (center line) 0.19
Upper support grid i . 1.12
} center line
Lower support grid 0.66

3.4.4 Pump-Bowl Fission-Product Activities

The equilibrium activity due to gaseous fission products in the pump bowl
was estimated by the method described by Stevenson.1® The heat generation in the
pump bowl amounted to about 6 kw for a reactor power of 5 Mw.

3.4.5 Cell Calculations
Three—firoup two-dimensional calculations were performed with the IBM-704
program PDg 1l t6 obtain an estimate of the flux distribution within a moderator
block. Sample traverses of the flux above 0.3 Mev are shown in Fig. 5.0,
3.5 ANALOG-COMPUTER STUDY OF MSRE PRIMARY-SYSTEM FLOW LOSS
An analog-computer analysis was made of a loss-of-flow accident in the pri-

mary system of the MSRE. The temperature changes and the maximum temperatures
attained in the system were of primary interest.

UNCLASSIFIED
ORNL-LR-DWG 52051

——_ .

. X ™~

~— FUEL SALT N

" S
— e 7

@)

0.5

o-bc)/ be (%)
Q
o
Se
2R
S
m

Z.os ,,,ff"!/////// _
@ L
-1.0 ‘ ///
@ (———”
s |

Fig. 3.8. Three-Group PDQ Cell Calculation.
46

3.5.1 Description of the System Simulated

Thermal System.--A preliminary analysis was made of the thermal system,
based on preliminary design information (Table 3.5). No after-heat, convection
cooling, or moderator temperature coefficient of reactivity was simulated in
this analysis. These phenomena will be simulated in subsequent analyses when

Table 3.5. Computer Information on MSRE

Reactor inlet temperature 1175°F
Reactor outlet temperature 1225°F
Mean graphite temperature 1250°F
Residence time in reactor 3.88 sec
Film drop from graphite to fuel Constant at all flows
Heat capacity of graphite Loko E%E (cy = 0.425 ffiqu)
Prompt 7 and neutron heating in graphite 9% of 10-Mw power
Residence time in piping from reactor outlet to H.E. inlet 0.75 sec
Residence time in H.E. l.3 sec
Heat capacity of metal in H.E. 200 Btu/°F
Average film drop between primary coolant and metal at design point 65°F
Average drop in metal at design point 66°F
Average film drop between metal and secondary coolant at design point 31°F

Film drop between primary coolant and metal as function of flow:
See graph, displace curve if necessary so that at 10 fps velocity, At = 65°F
Mean secondary-coolant temperature at design point 1038°F

Residence time in piping between H.E. outlet and inlet

(including coolant annulus) 2.88 sec
Total circulation time 8.81 sec
Temperature coefficlent of reactivity -9 x lO-5 °F‘l
Melting point of primary coolant 842°F
Melting point of secondary coolant QLO°F
Check points

. ft hr °F
Th 1
ermal resistances ( Bia )
In primary coolant film 3.28 x :I_O-l'L
In metal 3,32 X 10"1‘L
In secondary coolant £ilm 1.56 x lO“L‘L
Temperature differential t -t = 250°F at thermal convec-

core HE
tion heat removal of 10% of 10-Mw power

47

suitable design information is available. The secondary coolant system was simu-
lated merely as a heat dump since the main interest is in the primary system. A
very elementary schematic of the system is shown in Fig. 3.9.

It is understood that this was a simulation of primary flow stoppage and not
one of primary pump stoppage. Pump "run down" information is unavailable at this
time.

The primary flow rate was decreased exponentially on periods of 1.5, 3, 6,
and 10 sec.

The heat transfer coefficient between the primary coolant and the heat ex-
changer wall was made to vary with primary flow rate in accordance with the curve
shown in Fig. 3.10. This curve was derived from the curve supplied by the de-
signers, which is shown in Fig. 3.11.

Nuclear System.--The delayed neutrons were lumped into one weighted group.
Using the values from document LA-211& (dated 1957) for i and Bi, X for the one
group was found to be 0.0769 sec™t and B is 6.4 x 10°3. Using the given fuel
transit times and the curves in ORNL-LR-Dwg. 8919, B was found to be 3.4685 x 1072

6

where B’/ = ;{1 aiBi and @; is the ratio of the population of the ith group of

i=1

UNCLASSIFIED
ORNL—LR—DWG 46961

[ Umits oF smutation ]
REACTOR REACTOR
PUMP INLET, I OUTLET, %
— L
REACTOR

|
l
|
|
|
|
! PRIMARY LOOP
|
!
1
|
|
|
|

——
HE.%ETLEE ORIMARY HEAT H.E.INLET, 7,
PUMP i EXCHANGER
—/VWWWWWWWIWH——=
I 1
SECONDARY LOOP
—— -~

SECONDARY HEAT
EXCHANGER

Fig. 3.9. Schematic Drawing of MSRE,
UNCLASSIFIED
QRNL — LR— DWG 469634

/

/

HEAT TRANSFER COEFFICIENT

oy

0 2 4 6 8 10
PRIMARY FLOW VELOCITY (fps)

Fig, 3.10. MSRE Flow Rate vs Heat
Transfer Coefficlent.

delayed neutrons inside a circulating-fuel reactor to the population of this
group in an equivalent stagnant reactor.

In the simulation, the delayed-neutron contribution is a function of the
flow rate. The method used in the simulation is not precise; however, it appears
to be a good approximation for excursions with periods of the same order of mag-
nitude as those encountered in this analysis.

3.5.2 Analog-Computer Program

The analog-computer program of this simulation is filed as ORNL drawing
D-40325. (Copies may be obtained from the print files in the Engineering and
Mechanical Division.)

49

UNCLASSIFIED
ORNL — LR—DWG 46962A

o \\

60

40 Sl

AT =Teyer = The wae CF)

20

0 5 10 15 20
PRIMARY FLOW VELOCITY (fps)

Fig. 3.11, MSRE (Tf - Tw) vs Primary

Flow Rate for a Constant Heat Transfer Rate.

3.5.3 Gimulator Operation

With the simulator in steady-state operation at design-point conditions, the
switch was thrown to decrease the primary flow rate exponentially on a given
period. This procedure was repeated for various periods. Pertinent information
(temperatures, etc.) was recorded versus time. These curves are included in
Figs. 3.12, 3.13, 3%.14, 3%.15 and 3.16.

Loss of load was also simulated.
3.5.4 Conditions Used to Obtain Curves

The conditions used in the simulation to obtain the reported curves are de-
scribed below.

For Fig. 3.12, the primary temperature coefficient of reactivity is
-9 x 107 (5k/k)/°F. The primary flow rate was decreased exponentially from 100%
flow to zero flow on the following periods:

Curve No. Period (sec)
1 1.5
2 5
3 6
H 10

No after-heat or convection cooling was considered.
L ok o RT3

50

UNCLASSIFIED
ORNL-LR~-DWG 46964A

1500
n
o
l_
wl
-
l_
=)
O
o 1400 : - .
© ‘ \
<&> __._.—-———_——4\
o — ——
w L*%:‘:::—-______\\ 3
' ——
E /‘ : —— )
1
L 1300 /9/ - - i ' { T
o
=
Ll
'-7
> /
o
<L
=
1
o
1200
0 20 40 60 80 100 120 140 160 180 200 220 240

TIME ELAPSED AFTER INITIAL FLOW RATE DECREASE (sec)

Fig. 3.12. Temperature vs Time After Loss of Primary Flow in the MSRE.
(Curves 1 through 4.)

For Fig. 3.13%, the primary temperature coefficient is -9 x lOas(Bk/k)/°F.
The primary flow rate was decreased exponentially from 100% flow to a flow rate
sufficient to give a loop time of 400 sec. The curves corresponding to different
periods are as follows:

Curve No. Period (sec)
5 1.5
6 3
7 6
8 10

The 400-sec loop time is that calculated resulting from flow induced by convec-
tion cooling with a reactor-to-heat-exchanger temperature differential of 250°F,
No after-heat was considered.

For Fig. %.1l, the primary flow rate decreased exponentially to zero filow
rate. The primary temperature coefficient of reactivity used was -4.5 x 10~
(8k/k)/°F. No after-heat or convection cooling was considered. The correspond-
ing curves and exponential periods are as follows:

Curve No. Period (sec)
9 1.5
10 3
11 6
1?2 10

For Fig. 3.15, an instantaneous complete loss of load was simulated. Pri-
mary temperature ccefficient of reactivity 1s -9 x 10'5 (6k/k)/°F.

PRIMARY TEMPERATURE (°F)

PRIMARY TEMPERATURE, REACTCR OUTLET (°F}

1300

1200

1400

1300

1200

1500

1400

1300

1200

UNCLASSIFIED

ORNL-LR-DWG 469654

REACTOR MEAN
|

T—

REACTOR QUTLET

|

Fig. 3.13.
(Curves 5 through 8.)

60
TIME ELAPSED AFTER INITIAL FLOW RATE DECREASE (sec)

120 140

200 220 240

Temperature vs Time After Loss of Primary Flow in the MSRE.

UNCLASSIFIED
ORNL-LR-DWG 469664

\

/

'—-____1O

60

Fig. 3.14.

(Curves 9 through 12.)

80
TIME ELAPSED AFTER INITIAL FLOW RATE DECREASE (sec)

140

200 220 240

Temperature vs Time After Loss of Primary Flow in the MSRE.

TEMPERATURE {°F)

52

UNCLASSIFIED
ORNL-LR-DWG 469674

1300 | T ‘
|
/REACTOR OUTLET- PRIMIARY"\J\
] j J:
——a H " .
/
L~ ‘
|
|
~ SECONDARY MEAN i
1200 ; -
1100 - ‘
1000
0 20 40 60 80 100 120 140 160 180 200 220
TIME ELAPSED AFTER LOSS OF LOAD {sec)
Fig. 3.15. Temperature vs Time After Loss of Load in the MSRE.
UNCLASSIFIED
ORNL-LR-DWG 46968A
1475 T T
J
/J/
1425 : ///////
- e
o ‘
LLI I
o :
2 1375 i
<
c
Ll
a
=
()
'_
!
4
1325 —
1275 :
0 -5 -10 -15 -20 -25 ~30 -35 -40 -45 -50 -55
RECIPROCAL OF TEMPERATURE COEFFICIENT x10 >
Fig. 3.16. Primary Mean Temperature vs Reciprocal of Temperature Coef-

ficient.

53

For Fig. 3%.16, the primary flow rate was decreased from 100% flow to zero

flow exponentially on a period of 3 sec for all runs. A number of runs were made
with different primary-coolant temperature coefficients of reactivity. The maxi-
mum primary mean temperature attained in each run was plotted against the recip-
rocal of the primary temperature coefficient of reactivity for that run. Each
run comprises a point on this curve.

1.

2.

10.

11.

REFERENCES

D. R. DeHalas, HLO Graphite News Letter No. 2, HW 65642 (June 14, 1960), p kL.

J. H. W. Simmons, "The Effects of Irradiation on the Mechanical Properties
of Graphite,” Sections E and F, Proceedings of the Third Conference on Car-
bon, Pergamon Press, 1959.

H. F. Poppendiek and L. D. Palmer, Forced Convection Heat Transfer in Pipes
with Volume Heat Sources Within the Fiuids, ORNL-1395 (Dec. 2, 1952).

H. F. Poppendiek and L. D. Palmer, Forced Convection Heat Transfer Between
Parallel Plates and in Annuli with Volume Heat Sources Within the Fluids,
ORNL-1701 (May 11, 1954).

J. W. Miller, personal communication.

M. Jakob, Heat Transfer, vol. I, p 176-179, Wiley, New York, 1949.

C. L. Davis, J. M. Bookston, and B. E. Smith, GNU-IT - A Multigroup One-
Dimensional Diffusion Program for the IBM-704, General Motors Report GMR 101

(1957)

M. L. Tobias and T. B. Fowler, Equipoise - An IBM-704 Code for the Solution
of Two-Group Neutron Diffusion Equations in Cylindrical Geometry, ORNL-2967
(in press).

M. P. Lietzke and M. L. Tobias, personal communication.

R. B. Stevenson, Radiation Source Strengths in the Expansion Chamber and
Off-Gas System of the ART, ORNL CF-57-7-17 (July 1957).

G. G. Bilodeau et al., PDQ - An IBM-704 Code to Solve the Two-Dimensional
Few-Group Neutron-Diffusion Equations, WAPD-TM-70 (1957 .

ot Y

PART Il. MATERIALS STUDIES

4. METALLURGY

4.1 DYNAMIC-CORROSION STUDIES
4.1.1 Forced~Convection Loops

The final hot-leg insert was removed from INOR-8 forced-convection loop
93544 after 15,140 hr of operation. As previously discussed,l three inserts
were installed at the end of the hot-leg section of this loop to provide data on
the weight changes of INOR-8 in contact with a beryllium-base fluoride fuel
mixture. The operating schedule specified removal of one insert after 5000 hr
of exposure, another after 10,000, and the third after 15,000. Operation of the
loop began in July 1958 under the conditions:

Salt mixture LiF-BeF,UF) (62-37-1 mole %)
Max. salt-metal interface temp. 1300°F

Min. salt temp. 1100°F

AT 200°F

Reynolds number 1600

Flow rate 2 gpm

Examination of the third insert showed an average weight loss of 1.7 mg/cmg,
+6%, along its L-in. length. If uniform removal of surface metal 1s assumed, this
weight loss corresponds to a loss in wall thickness of 0.08 mil, *&%.

The first two inserts, which were removed from the loop after 5000 and 10,000
hr, showed respective weight losses of 1.8 mg/eme, *2%, and 2.1 mg/cm®, f3%. The
corresponding losses in wall thickness were calculated to be 0.08 and 0.09 mil,
respectively. The similarity of these values indicates that no significant weight
loss occurred after the first 5000 hr of operation. Om all three inserts no loss
in wall thickness could be detected from measurements of the inserts before and
after test.

As previously reported,2 Metallography of the 5000- and 10,000-hr inserts
revealed no evidence of attack other than the formation of a thin corrosion film
on the surface exposed to the salts. A zone comprised of unusually small grains
was also noted below the corrosion film on both inserts to a depth of 1 to 2 mils.

Metallography of a transverse section of the 15,000-hr insert revealed light
surface roughening, as shown in Fig. 4.1. As in the case of the 5000- and 10,000-
hr inserts, the thin corrosion f£ilm and the band of fine-grained material were
again found. The corrosion film was the same thickness as found on the 10,000-hr
insert; however, the band of fine-grained material had increased in size, compared
to the 10,000-hr insert.

35
56

UNCLASSIFIED
T-19159

- mat ! . " . / ’ <" 3 , 011

o1z

< ? ' 013

% ¢ 4 AN <~ 014

Fig. 4.1. Transverse Section of 15,000-hr Insert Removed from INOR-8
Forced-Convection Loop 9354-4. Etchant: 3 parts HCL, 2 parts H;0, 1 part
10% chromic acid. 250X.

The band of fine-grained material is attributed to recrystallization of the
matrix near the inside surfaces of the insert. This assumption is supported by
the fact that the inserts had been reamed prior to service, which probably induced
some cold work along the inside surfaces.

The status of two Inconel and nine INOR-8 forced-convection loops which are
in operation with various fluoride mixtures is summarized in Sec. 2.6.

4.1.2 Microprobe Analyses of Surface Film

The compositions of thin corrosion films found on several of the long-term
INOR-8 loop specimens after salt exposure have been investigated by means of an
electron-beam, microprobe analyzer. (These analyses were carried out by the
Ernest F. Fullam Corp., Schenectady, N. Y.) Included in these examinations were
hot-leg specimens from herma%-convection loop 1226iref 3) and from forced-
convection loops 9354-4lref 4) and 935&-5(T9f ). Pilms on these specimens,
which ranged from 1/4 to 1/3 mil in thickness, were first examined by aiming the
electron beam directly on the inner surface, which excited an area of 1.96 x 10-2
eme on the surface to a depth of approximately lu . A scan analysis was then made
along a cross section of the specimen from forced-convection test 9354-L4, This
latter specimen was sectioned along a plane tangent to the inside surface of the
specimen so as to increase the apparent cross section of the tube wall and,
correspondingly, the corrosion film. The results of the spectrographic analyses
obtained by using both methods are compared with as-received analyses of these
specimens in Tables 4.1 and 4.2.

Table 4.1. Spot Analyses Made of Surface Films

Source of Approx. Temp. Salt Spot Analysis (wt %)*
Specimen Location of Specimen Circulated
(°F) Ni Cr Fe Mo
As-received
(Heat SP-16) 71.66 6.99 4.85 15.82
Loop 1226 Hot leg 1248 No. 131 69.2 0 0.4 23.4
(section 2)
Loop 9354-4 End of second 1300 No. 130 67.6 0 1.2 23.6
heater leg (10,000-
hr insert)
Loop 9354-5 End of second 1296 No. 130 66.8 0 0.6 21.5
heater leg
(section 11)
Loop 9354-5 Recheck 67.0 21.6

* et +cdf
Estimated error = I5%.

LS
58

Table 4.2. YScan" Analysis Made of 10,00-hr Insert
Specimen from Forced-Convection Loop 9354-4

Distance from>’° Composition (wt %)°
Inner Surface
(Mils) Ni Cr Fe Mo
Surface 67.6 0 1.2 23.6
0.5 8.3 2.0 3.2 19.9
1.0 69,5 2.7 4,2 19.9
1.5 68.1 3.3 .2 17.5
2.0 66.1 4.0 .5 17.6
2.5 70.0 .7 4.5 17.6
3.0 69.4 Lh,7 .5 16.4
4,0 69.9 6.7 4.8 15.8
5.0 71.6 6.9 4.8 15.8

aMeasurements made along cross section tangent to inner surface.
b
Film ends at position just prior to 1.0-mil measutrement.

c
Estimated error = 15%.

As shown in Table 4.1, the spot analyses of the surfaces of all three samples
indicated an increase in the molybdenum content and virtually complete depletion
of chromium and iron. The single-scan analysis (Table 4.2) shows the depletion
of chromium and iron to have occurred to a depth of approximately 3 mils on the
"magnified” cross section. Nickel is seen to have decreased and molybdenum to
have increased over about the same depth. If these results are adjusted to
account for the magnification created by the specimen preparation, this concen-
tration gradient is found to occur in about a 1/2-mil-thick surface.

As can be seen in both tables, the percentages of Ni, Cr, Fe, and Mo do not
total 100%, although the deviation from 100% is within the estimated accuracy of
the instrument. To ensure that the deviation was not associated with residual
fluorides remaining on the surfaces of the specimens, the hot-leg specimen from
loop 9354-5 was cleaned with an ammonium oxalate solution and then re-analyzed.
(This solution dissolves fluorides quite effectively without disturbing alloys
composed predominantly of nickel.) The results of this re-analysis, as shown in
Table 4.1, showed no difference from the first analysis.

L.2 WELDING AND BRAZING STUDIES
Y.2.1 Solidified-Metal-Seal Development

The use of solidified metal seals for elevated-temperature, leak-tight,
quick~disconnect joints has been considered for a number of applicgtions includ-
ing high-vacuum valve and flange studies for the Sherwood Project. Such seals
appear to be especially suitable for molten-salt reactor applications since they
are readily applicable for remote handling and since relatively simple equipment
may be used for making and breaking the joint. Conventional types of seals
containing such materials as rubber, Teflon, and vacuum grease are obviously not
possible in view of the high temperature and highly corrosive environment.

59

In general, two designs have been considered, the first containing an alloy
sump and a tongue-and-groove joint design and the second containing an alloy-
impregnated metal-fiber compact which would be used similar to an O-ring. Three
ductile metals, corrosion-resistant to fused salt, were selected as potential
sealing materials, and two corrosion-resistant base metals were chosen for the
preliminary wetting and compatibility studies. A list of these materials is
shovn in Table 4.3.

Table 4.3. Materials Under Study for Solidified Metal Seals

Melting Base
Seal Materials Point (°F) Materials Remarks
Gold 1945 INOR-8 Reactor structural
Copper 1981 material
80 Au-20 Cu (wt %) 1625 Molybdenum Very limited solubility

in liquid Au or Cu

Compatibility specimens were made with each of the above combinations, and
aging studies are being conducted at 1292C°F and 90°F above the melting point of
the seal material. These conditions simulate reactor service temperature and a
typical seal opening or closing temperature. These specimens will be examined
for base-metal erosion and solid-state diffusion of seal-metal constituents.

In an effort to demonstrate feasibility of these types of solidified-metal
seals, several small components have been constructed and evaluated. A sump-type
seal, fabricated from INOR-8 and sealed with 80 Au-20 Cu, is shown in Fig. 4.2,
It was made and broken by induction heating in argon and was helium leak-tested
after every sealing. To date, this seal has been found helium leaktight after
each of ten cycles of breaking and sealing. Testing will be continued in an
effort to determine the lifetime.

In addition to the sump-type seal, an unsintered molybdenum-fiber compact
was prepared by the Metallurgy Division Powder Metallurgy Group and impregnated
with Au-Cu. This was used to make a seal between two molybdenum plates, as
shown in Fig. 4.3. The seal was spring loaded for opening and closing while the
component was induction heated under inert gas. A drawing showing this method
of making and breaking seals is shown in Fig. 4.4, After each cycle of breaking
and sealing, the seal was helium leak-checked. To date, the sample has been found
leaktight after each of five completed cycles. No solution of the molybdenum
prlates is evident, and no additional alloy has been added. Tests of this type
will also be continued to determine the lifetime of the Jjoint.

As a result of the above tests, it is believed that unsintered molybdenum-
fiber compacts impregnated with alloy have sufficient strength at sealing
temperature to resist tearing. To make this specimen more applicable to
proposed molten-salt operating conditions, a molybdenum-fiber compact will be
used between INOR-8 plates on the next seal instead of between molybdenum plates.

4.2.2 Brazing of Graphite
Since advanced molten-salt reactor designs will probably include heat

exchange systems of graphite, the development of techniques for fabricating such
systems has been a subject of study.7:8 Corrosion-resistant brazing alloys in
UNCLASSIFIED
Y-36196

Fig. 4.2. Test Rig with Sump-Type Seal,

the Au-Ni-Ta and Au-Ni-Mo systems have been developed for joining graphite to
itself and to metals, and a typical graphite-tube-to-INOR-B header mockup
assembly was constructed.

A method of leak testing these graphite-to-metal joints was developed and
preliminary results were obtained. The test consists of brazing molybdenum caps
on low-permeability graphite tubes and filling the assembly with isopropancl,
which has a room-temperature viscosity similar to that of the fused salts at
operating temperature. An inlet at the top of the assembly provides for a means
of internally pressurizing with argon, as shown in Fig. 4.5,

61

UNCLASSIFIED
Y-36360

'f(,']

’l ) -t \,g\'»\)‘- A

N AT \% ‘1\ 4\

e
5 O AR 8

Fig. 4.3. Unsintered Molybdenum-Fiber Compact Impregnated with Braze
Metal.

UNCLASSIFIED
ORNL- LR~ 0OWG 52052

MAKING BREAKING
SEAL SEAL

AFTER BREAKING SEAL

T

Fig. 4.4. Spring Loading of Impreg-
nated Fiber-Compact Seals.

62

UNCLASSIFIED
Y-35467

Fig. 4.5. Rig for Testing Graphite-to-Metal Joints.

On initial tests, leakage at the brazed joint was observed below 5 psig.
However, Jjoints that were more leaktight were obtained by oxidizing the ends of
the graphite tubes prior to brazing. It is believed that oxidizing makes the
surface of the graphite more porous and provides an additional keying action
with the brazing alloy. A similar keying action has been reported in brazing
tests on the porous-type AGOT graphite.® Table 4.4t summarizes the results of
three tests made in this manner,

Table 4.4. Results of Leak Tests on Graphite-to-Molybdenum Joints
Part g . Appearance
Test No. Tested Brazing Alloy (wt %) 5P PIret LRak
1 Top 70 Au-10 Ni-20 Ta None to 50 psig
Bottom 75 Au-10 Ni-15 Ta At 10 psig
2 Top 70 Au-20 Ni-10 Ta None to 35 psig
Bottom 70 Au-20 Ni-10 Ta At 20 psig
Bottom 75 Au-20 Ni-5 Ta At 35 psig
(rebrazed)
3 Top 60 Au-10 Ni-30 Ta None to 60 psig
Bottom 60 Au-10 Ni-30 Ta At 35 psig

63

4 .2,3 Heat Exchanger Fabrication

The main heat exchanger of the MSRE will consist of approximately 250 INOR-8
(1/2 in. in OD, and 0.045 in. in wall thickness) tubes joined to a 1-1/2-in.-
thick TNOR-8 header. Welding and back-brazing procedures are being developed
for fabricating the tube-to-header joints in this unit. The basic technigues
are those previously used for fabricating heat exchange components containing
thick tube sheets.9

Figure 4.6 is a cross-sectional view of the tube-to-header joint configura-
tion selected for the heat exchanger. Trepanning on the weld side serves to
ensure good heat distribution by providing a uniform weld geometry and to minimize
tube-sheet distortion and restraint in the vicinity of the weld Jjoint.

In back brazing these joints, there is a unique heating problem because of
the large differences in mass of the tube and tube sheet. As a result, the tube-
sheet temperature continuously lags the tube temperature during heating, and if

UNCLASSIFIED
ORNL—LR—-DWG 52053

S/TUBE

BRAZE SIDE N TREPAN
\ N

THREE
FEEDER HOLES

.

WELD 9DE//// TREPAN

(o)

%

(&)

Fig. 4.6, Tube-to-Tube-Sheet Joint. (a) Prior to welding and back
brazing; (b) after welding and back brazing.
64

brazing alloys are applied to the conventional location at the tube-to-tube-sheet
junction, the alloy preferentially flows to the hottest member, with consequent
poor flow on the tube sheet. Placing the braze metal in a trepanned sump
removes the problem by allowing the brazing alloy to remain at the temperature
of the tube sheet through the entire braze cycle. The sump acts as a reservoir
from which the alloy flows to fill the annulus between the tube and the tube
sheet and to form a fillet- where the tube meets the tube-sheet surface. The
fillet also aids in inspection since the capillary joint must be filled with
alloy to form a fillet. The use of back brazing in the fabrication of this unit
is unique, since high-temperature brazes are not normally used on tube sheets of
this thickness.

Optimum welding conditions for the tube-to-tube-sheet joints to obtain weld
penetrations of 1.5t and 1t (t = tube-wall thickness) were determined and are
listed below.

Conditions for Conditions for
1.5t Penetration 1t Penetration
Welding current, amp 55 45
Welding speed, in./min 6.3 8.5
Inert gas Argon Argon
Arc length, in. 0.050 0.050

Those conditions which give maximum weld penetration of 1.5t also cause
severe "roll over" of the weld and thus a constriction in the tube entrance.
Therefore, welds with 1t penetration, and less "roll over,' were considered to
be more suitable.

A seven-tube test assembly was then constructed, and a photograph showing
the braze side of the unit is shown in Fig. 4,7. Brazing was conducted in dry
hydrogen, using the 82 Au-18 Ni (wt %) alloy. The brazing temperature was
1830°F, and a rate-of-temperature rise of 3OOOF/hr was selected, since it
approximates that to be readily obtained with commercial facilities. Good
filleting was observed on all seven tubes, and the general visual appearance
of the welds and brazed joints was excellent. A metallographic evaluation of
several joints is being conducted.

A large assembly containing about 20 tubes is being planned to further
demonstrate reliability in making a large number of joints.

L.2.4 Mechanical Properties of INOR-8

Most of the mechanical-property tests on INOR-8 were completed in 1959,
and the results of this program are presented in a topical report.lO A critical
review of these data was made in order to establish design values. It was
decided that up to 10500F, the design stresses would be based on two-thirds of
the 0.2% offset yield strength, adjusted for the minimum specified yield
strength. This value was chosen to be 35,000 psi. Above 1050°F the design
stresses are based on the stress to produce 1% creep strain in 10 hr. This
value was chosen because several heats of material did not exhibit a minimum
creep rate. For heats which did exhibit a minimum creep rate, the stress
value for 0.1 CRU was above the values based on creep strain. Stresses in the
creep range were obtained from Larsen-Miller plots of data at 1100, 1200, 1250,
1300, 1Lk00, 1500, 1650, 1700, and 1800CF. Over 100 creep tests were performed,

65

UNCLASSIFIED
Y-35957

Fig. 4.7. Tube-to-Tube-Sheet Sample, Showing Back Brazing.

with times ranging from 0.1 to 25,000 hr; the majority were performed between

1100 and 1300°F. The values obtained from the Larsen-Miller curve are conservative
with respect to the values obtained by direct extrapolation of creep data.
Extrapolation of the rupture stresses to 102 hr revealed that they were well above
the creep limits and did not affect the design stresses. The allowable stresses
are shown in Table 4.5.

Preliminary steps have been taken to obtain approval of these stresses by
the ASME Council for The Boiler and Pressure Code. The stress values in Table

4.5 are therefore tentative and may be modified in accordance with the recommenda-
tions of the Boiler Code Committee.

A series of tensile tests at 1100 and 1300°F were performed to determine
the effect of creep strains of less than 2% on the strength and ductility of
INOR-8. A general pattern of the results is presented in Table 4.6. The
only deleterious effect is on the elongation which, in our case, was a loss of
50% after approximately 12,000 hr of exposure. In spite of this relative loss
in elongation, the minimum value was 20% in 2 in.

66

Table 4.5. Allowable Design Stresses for TNOR-8
Wrought and Annealed Sheet and Rod

Temperature Allowable Stress, Static
(°F) (psi)
100 23,300
200 20, 700
300 19, 100
400 17,900
500 16,900
600 16,200
700 15,600
800 15,300
850 14,900
900 14,700
950 14, 600

1000 14, 600
1050 14,000
1100 10, 40O
1150 7,200
1200 5,200
1225 4, 400
1250 3,700
1300 2,700
1350 2,050
1400 1,550

Teble L.6. Effect of Prior Creep on Tensile Properties of TNOR-8

Conditions Temp. of Change in Tensile Properties

Material of Creep Tensile
L
Exposure ?8;? vield Tensile
Strength Strength Elongation

TNOR-8 rod  1250°F in air 1100 No change No change 15% loss

up to 7500 hr

1250°F in air 1300 No change No change L0% loss

up to 8000 hr
INOR-8 sheet 1100°F in salt 1100 20% gain No change 50% loss

up to 12,000 hr

1300°F in salt 1300 25% gain No change 25% loss

up to 12,000 hr

1250°F in air 1300 No change No change 30% loss

up to 8500 hr

%
All strains less than 2%.

67

4,3 PERMEATION AND APPARENT-DENSITY UNIFORMITY IN LARGE PIECES OF GRAPHITE

Since there appears to be some advantage in using large pieces of graphite
in the design of the molten-salt-reactor moderator, grades R-0025 and MH4LIM-82,
which were the largest pieces of low-permeability graphite available, were
tested for uniformity of properties. Specimens from various locations and with
different orientations were subjected to the standard molten-salt permeation
test. The original sizes of both grades, the location and orientation of the
test specimens, and the test results are summarized in Table L4.7.

In the fabrication of the graphite cylinder of grade R-0025, 1.5-in.-dla
holes were drilled through one quadrant of the cylinder parallel to the axis.
The piece was impregnated with pitch and then graphitized. The center specimens
from this pilece were taken at the location of a hole. The other specimens were
taken from a quadrant opposite the one that was drilled.

It appeared that the pores of the R-0025 graphite may have been slightly
more continuous (accessible) and/or larger in the direction parallel to the
molding force (parallel to the cylinder wall) and that the lower porosity of
the center specimens extended only 3/h in. from the surface of the center hole.
In general, the results indicated that the large pieces of grades R-0025 and
MH4YIM-82 were uniform.

4.3.1 Permeation of AGOT and S-4 Graphites by Molten Salts at Different Pressures

To determine the effect of pressure ( <150 psig) on molten-salt permeation
in graphite, and to determine approximately the relationship between molten-salt
permeation and mercury permeation of graphite at room temperature, S-4 and AGOT

Table 4.7. Apparent Densities of Graphite Grades R-0025 and MHLIM-82
and Degree of Permeation by LiF-BeF,-ThF),-UF), (67-18.5-14-0.5 mole %)
in Relation to Specimen lLocation and Orientation

Exposure conditions: 100 hr at 1300°F and 150 psig
All values are averages of six; test specimens were
0.500 in. in diameter and 1.500 in. long

Specimen Apparent Bulk Density Bulk Volume Permeated
Orientation (g/cc) (%)
Location with Respect to * ¥ * '
Forming Force R-0025" MH)LM-82 R-0025 MH), LM-82
Center Parallel 1.89 1.82 4.4 7.7
Perpendicular 1.85 6.4
Midway on Parallel 1.88 1.82 5.5 8.0
radius Perpendicular 1.86 6.7
At outside Parallel 1.87 1.81 5.2 8.4
diameter Perpendicular 1.87 1.80 5.7 8.7

*
39 in. in diameter, ll% in. long; size as fabricated.
**L9 in. in diameter, L4 in. long; original length not given by vendor.
68

specimens were permeated with LiF-BeF,-ThF}-UF) (67-18.5-14-0.5 mole %) at 1300°F
at pressures of 25, 65, and 150 psig In 100-hr exposure periods. A good
correlation between the molten-salt and mercury permeation of graphite would
contribute to the general indications that the molten fluorides do not wet
graphite. Grade AGOT, because it is porous and has good uniformity, was included
in the test as a control.

Typical mercury permeation (intrusion) data for AGOT and R-0025 graphite
(s-4 is a laboratory designation for grade R-0025) were used to plot curves for
theoretical bulk permeation of graphite by molten salts.ll These curves were
used to obtain the theoretical permeation data that are compared with the actual
data in Table L.8.

The agreement of the actual permeation values with the typical theoretical
values is generally good except for grade S-4 in the 25-psig permeation.
Either there is a structural difference in the graphite tested or this is one
of the first indications of a slight wetting of graphite by a molten salt.
Further investigation is planned.

The molten-salt permeation data also show that pressure reduction from 150
to 25 psig does not appreciably decrease the salt permeation into these grades
of graphite even though S-U4 is a moderately low permeability grade of graphite.

Table 4.8. Comparison of Theoretical and Actual Molten-Fluoride Permeation
of S-4 and AGOT Craphites at Different Pressures

Test conditions:
Temperature, 1300°F
Exposure, 100 hr
Salt, LiF-BeF,-ThF)-UF),

(67-18.5-14-0.5 mole %)

Bulk Volume of Graphite
Permeated by Salt®

(%)

Permeation
Pressure AGOT gl
(psig)
Theoretical Aectual Theoretical®® Actual
o5 10.5 11.4 1.5 3.7
65 14,2 12.0 4,2 L4
150 15.0 13.2 5.4 h.9

*
Each value is an average of six.
*¥ These are based on the typical pore spectrum for grade R-0025;
however, S-4 1s the laboratory designation for grade R-0025.

69

4.4 PRECTPITATION FROM MOLTEN-FLUORTDE FUEL IN CONTACT WITH
VARTOUS VOLUMES OF CRAPHITE

A single series of five precipitation tests have been made with AGOT
graphite and LiF-BeFp-UF), (62-37-1 mole %) fuel in which only the volume of the
graphite was varied in order to determine the relationship of graphite volume to
uranium precipitated from the fuel. The test conditions and results are
summarized in Table 4.9. The quantity of uranium that precipitated as U0y was
determined chemically. The uranium precipitated per cubic centimeter of bulk
volume of the graphite averaged (1.3 + 0.4)mg to (1.3-0.3)mg. This would be
equivalent to approximately 9% by weight of the uranium in the fuel for a
reactor system with a volume ratio of graphite to fuel of 9:1. This estimate
is made with the assumptions that the volume of salt and the area of graphite
in contact with the fuel do not affect the quantity of precipitate that occurs.
Past data for volume ratios of graphite to fuel in the range of these tests
support these assumptions.

L.4.1 Removal of Contamination from Graphite

It has been reportedle that the thermal decomposition of NHyF.HF in the
presence of graphite removed its oxygen contamination to such an extent that the
graphite could subsequently contain LiF-BeFp-UF), (62-37-1 mole %) at 1300°F
without causing the usual UOo precipitation from the fuel. A duplicating test
has produced the same results. Work is in progress to determine if INCR-8 is
seriously attacked during the thermal decomposition of the NHMF-HF.

Table 4.9. Uranium Precipitation from Molten LiF-BeF,-UF),
(62-37-1 mole %) Exposed to ACOT Graphite

Test conditions:
Temperature, 1300°F
Test period, 100 hr
Atmosphere, vacuum

Ratio of the projected surface of the
graphite® to the volume of the fuel:
9.6 in.2/in.3 (3.8 cm2/cm3)

Volume of the fuel: 0.4334 in.3 (7.102 ce)

Ratio of Graphite Weight of Uranium Precipitation
Volume to Fuel Total Per Unit Volume of Graphite
Volume (mg) (mg/cc)
5:1 3k 0.97
10:1 83 1.2
15:1 177 1.7
20:1 202 1.4
27:17%% 203 1.1

*
This is the area of the graphite that the molten fuel would
be in direct contact with 1if the graphite were free of pores.

*3%

This is the ratio for the crucible size arbitrarily chosen
for the standard permeation tests.
70

L.4.2 TNOR-8 -- Fuel -- Graphite Carburization Tests

The tendency for TNOR-8 and Inconel to be carburized in LiF-BeFo-UF),
(67-32-1 mole %) -- graphite system at 1300°F has been investigated in a series
of tests conducted in multiples of 2000 hr up to 12,000 hr.

No carburization was detected metallographically on specimens from any of
the above tests. Sheet tensile specimens were included in each test system, and
these were tested after exposure to the fuel-graphite to determine the effect of
this exposure on their tensile properties.

These tensile-test results are listed in Table 4.10. After 4000 hr of
exposure, the corrosion on the Inconel specimens was heavy enough to mask any
carburization that might have occurred; consequently, only the results obtained
with INOR-8 tensile specimens are listed.

Comparing the tensile strengths and elongation values of the long-term
tested specimens with those of the short-term control specimens, it is evident
that no significant change has occurred.

Table L4.10. Tensile-Test Results on INOR-8 Specimens Exposed to Fuel-
Graphite Systems Compared with Control Specimens Exposed to Argon for
Various Times

Time of Test Control Specimens* Test Specimens**
(hr) Tensile Tensile
Strength Elongation Strength Elongation

(psi x 1073} (% in 2 in.) (psi x 10-3) (% in 2 in.)

Room-Temperature Tests

2,000 123.2 ho.5 124.8 41.0
Iy, 000 123.0 h2.0 12L.5 43.0
6,000 12k .7 39.0 12k.2 36.0
8,000 127.7 43.5 128.2 36.5
10,000 130.8 43.0 130.3 41.0
12,000 130.6 36.5 126.7 36.3
1250CF Tests
2,000 4.6 18.0 75.6 18.5
4,000 75.9,76.0 19.0,20.0 76.2,T4.3 18.5,18.5
6,000 72.6 16.5 .5 16.0
8,000 76.7 17.5 76.6 19.0
10,000 78.9 15.0 80.8 17.5
12,000 72.8 15.5 75.8 17.0

*
Control specimens were exposed to ARGON at 1300CF for the times
indicated.

%
Test specimens were exposed to LiF-BeF,-UF) (62-37-1 mole %) -
Graphite at 1300°F for the times indicated.

71

These data indicate that unstressed INOR-8 is not detectably carburized in
the fuel-graphite system (described above) at 1300°F for exposures as long as
12,000 hr.

4.5 IN-PILE TESTS
Two MSR graphite-fuel capsules (ORNL-MTR-47-1 and -2) were irradiated and
were then removed from the MTE on March 11 and June 20, respectively. The first
experiment was irradiated for 720 hr and the second for 1600 hr. No serious

difficulties were encountered. The two capsules were shipped to BMI for post-
irradiation examination. Disassembly is scheduled to start August 15.

REFERENCES

1. MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-217L4, p 31.

2. MSR Quar, Prog. Rep. Oct. 31, 1959, ORNL-2980, p 35.

3. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 33-38.

L. MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 55.

5. MSR Quar. Prog. Rep. Oct. 31, 1959, ORNL-2990, p 35.

6. J. W. Tackett, Progress Report — Bakable High-Vacuum Valve and Flange
Studies, ORNL CF-59-2-3 (Feb. 6, 1959).

7. MSR Quar. Prog. Rep. Oct. 31, 1959, ORNL-2890, p 33-36.

8. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2799, p 45.

9. R. L. Heestand, ORNL-24L0, p 159-62 (classified).

10. R. W. Swindeman, The Mechanical Properties of INOR-8, ORNL-2780 (to be
published).

ll1. Private communication from F. F. Blankenship and P. S. Spangler of the
Reactor Chemistry Division.

12. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 59.

5. CHEMISTRY

5.1 PHASE EQUILIBRTUM STUDIES
5.1.1 MSRE Fuel and Coolant

For design purposes, the MSRE fuel has been chosen at the composition 117F-
BeF,-ThF),-UF), (65-30-%-1 mole %; m. p. 4500C). Chemically, this composition can
be roughly approximated as Lip-BeFy to which 5 mole % of quadrivalent fluorides
are added; however, the exact proportions are based on carefully established phase
diagramsl showing the primary phase fields and lowest liquidus temperatures asso-
ciated with a given UF) and ThF), content. Fortunately, for concentration ranges
of interest as MSRE fuels, the quadrivalent fluorides are interchangeable, with
1ittle effect on the melting point. Furthermore, the LiF-BeFp solvent can hold as
much as 15 mole % quadrivalent fluorides in solution without requiring temperatures
in excess of 5000C. The preferred fuel compositions for MSRE-type reactors occur
just outside the primary phase field of LiF, and thus take advantage of both the
unusually large fiieezing- oint depression of LiF caused by strongly "acidice" cations
such as Bett, Th*+, and U + and the diminished stability ranges of 3LiF-MF, crystals
(as compared with 3:1 coumbinations containing other alkali fluorides).

The uranium concentration in the MSRE will be adjusted by adding concentrate
to a carrier. For example, the addition of LiF-UFy (73-27 mole %; m. p. 490°C) to
the carrier, IiF-ThFj-BeFs (64.75-4.15-31.1 mole %; m. p. 450°C), gives the final
fuel composition: 65-30-4-1 mole ¢%; m. p. 450°C. The concentrate is also used for
replenishment. During mixing, a complete range of intermediate compositions has at
least a transient existence before uniform blending is accomplished, and it is impor-
tant that no composition corresponding to a high-melting-point compound or insoluble
solid be encountered. Although the concentrate was chosen to conform with this re-~
quirement, to establish with certainty the temperature above which no solid appears,
equilibrated samples have been quenched from the neighborhood of the liguidus tem-
perature throughout the range between concentrate and carrier. The identified
phases in some of the quenched samples are listed in Table 5.1, and the liquidus
temperatures derived from these identifications are shown in Fig. 5.1. The absence
of any high-melting-point precipitate along the quaternary join conforms with ex-
pectations based on previously established ternary behavior.

The coolant for the MSRE, L17F—BeF2 (66-34 mole %; m. p- 415°C) closely resem-
bles the fuel in that LisBel) can be considered as the predominant constituent. As
a consequence, the fuel and coolant are compatible and no chemical interaction ensues,
merely dilution of the quadrivalent cations, in case of inadvertent mixing. The pro-
portions of LiF and BeF, were selected primarily as a compromise between increasing
melting points at greater IiF content and increasing viscosities at greater BeFo
contents. Also, a coolant with a higher freezing point than the fuel provides added
insurance against accidental freezing of the fuel by too rapid heat removal.

5.1.2 Systems Containing ThFu
A realization that the compounds previously reported as K¥:ThFj (ref 2) and

RbF-ThF), (ref 3) are actually TKF-6ThF), (ref 4) and TRbF+6ThFy (ref 4) led to a
review of KF-ThF), and RbF-ThF) phase relationships. Recent studies of these systems,

72

Table 5.1. Thermal Gradient Quenching Data for the System LiF-BeF2~UFh~ThFh

Composition Phase Change
(mole %) Temperature Phases Found Just Phases Found Just

LiF BeF, UF), ThF), (°c) Above Phase Change Below Phase Change

€5 20 1 4 4u8 T ox Liquid Iiquid + 2IiF-BeFs
+ the 5LiF-ThF)+ Ss¥*¥*

65 30 1 L yoz ¥ o Liquid + 2LiF-BeF, Liquid + 2LiF*BeF

+ the 5LiF-ThFlL ss + the 7LiF-6(U,Th§Fh
ss, with 9 mole % UFy

66.4 24,9 5.4 3.3 w6 T2 Liquid Liquid + the TIiF-6
(U,Th)F), ss, with 21
mole % UF),

68 18.7 10.8 2.5 w6 T o Tiquid Liquid + the TLiF-6
(U,Th)F), ss, with 34
mole % UF)

69.7 12.4 16.2 1.7 w1 ¥ o Liquid Iiquid + LiF + the
TLiF-6(U,Th)F), ss, with
38 mole % UFY

71.h 6.2 21.6 0.8 483 %1 Iiquid Liquid + the 7LiF-6
(U,Th)Fy ss, with 43
mole % UF)

1.4 6.2 21.6 0.8 580 I o Liquid + the TLiF-6 Liquid + IAF + the

(U,Th)F)y ss TLiF-6(U,Th)F) ss

¥ The uncertainties in temperature indicate the temperature differences between the gquenched
samples from which the values were obtained.

¥** So0l1id solution.

gL
UNCLASSIFIED
ORNL- LR - DWG 50121

600 | % ‘
| S
5501——L)4—~'4¢—-*#T—f'1
) o
i; 500 P—F—T ) ‘ - *L;
g | uoum;;14j2,,/¥"
§ 450 ¢o———Ia *JT),ét._ ,7‘7_,_‘,_‘
z \’ | | | |
| i
400 —%——l——-}——-l_——j
I
| | x | |
0 10 20 27
UE, {mole Fo)

Fig. 5.1. Quaternary Join Between LiF-
UF, (73-27 mole %) and LiF-BeF,-Th¥, (64.75-
31.1-4.15 mole %).

now nearing completion, are leading to significant modifications in the phase
diagrams.

The diagram portraying the phases in the system KF-ThF), based on new data
for the composition region 30 to 90 mole % Th¥F), is shown in Fig. 5.2. Invariant
equilibria are listed in Table 5.2. The formulas and crystal structures of the
jntermediate compounds in this system were reported originally by Zachariasen.
The cubic phases, SNaF-3UFy and 3NaF-2ThFy, are structural analogs of a phase
which Zachariasen identified as 0-2KF+ThFy; attempts to isolate and identify this
phase are continuing. Demonstration of a homogeneous solid phase at TS5 mole % ThF'),
corroborates the view of Asker, Segnit and Wylie's that the compound of highest
ThFu/KF ratio in the system is KF-3ThF) rather than KF-6ThF), as was previously
claimed.2 In fact, it now appears that KF-+3ThF) melts incongruently to ThF) ss
and liquid, that it undergoes solid state transitions which account for the exis-
tence of three crystal polymorphs, and that it is unstable with respect to KF-2ThF)
and ThF) below T775°C.

Although relatively detailed work had been made on the phases in the system
KF-ThF) before the current studies, the system RbF-ThF) had been investigated in
only a cursory fashion.”? Results of thermal gradient quenching experiments in the
system RbF-ThF) now reveal five intermediate compounds, 3RbF+ThF), 2RbF-ThFy,
7RbF-6ThF),, RbF-3ThFy, and RbF+6ThFy, in contrast with ZRbF+ThF,, RbF«ThFy, and
RbF«3ThF) as reported by Dergunov and Bergman. Elucidation of the solld state
relationships among the compounds RbF, 3RbF+ThF)y, and 2RbF:ThF) is continuing.

5.1.3 The System ZrFu'ThFu

In conjunction with problems arising from the presence of ZrFy in ThF)-contain-
ing fuels, thermal gradient quenching experiments of binary ZrF) -Th¥F), mixtures have
shown that the system ZrF)-ThFy forms continuous solid solutions with a minimum at
approximately 23 mole % ThF), and at 858°. Although the four tetravalent heavy-metal

75

UNCLASSIFIED
ORNL-LR-0WG 513914

1200
5:1= 5KF-ThF,
3:4== 3KF-ThF,
4= 2KF-Th
1mohflfié—§EPG$F ~7
o= p / /
1:2 = KF-2ThE, /
1:3 = KF-3ThF, LIQUID //
ThF455+ /
1000 Pd LQuio /|
at3+uow0\\\ ////, /
. ¢/ ThE, ss
BL3+L@UM\§"’C— - _::f;-
o 716+ LIQUID - \‘at3+Thass
_ ////r N~ Lo pr:3+ BY:3+ThE,
< .
o adi + ///’
% k\\\\uouu)\ // \\\\ ///
= 7.6+
5 BOO LIQUID
oy add + hed 1:2+ThE
> KF -+ E%Lg LIQUID i3 ’ 4
- LIQUID \\\ l
700 \V I e M
KF + a5:1 3:
| 764142 o
T -
B[5:1+ 3+
600 34 B2
- 1:2+ThE,
M A+7:
KF+35:4  _ BeA+T6
500 0 - 2
BS54+ R2:|
400
KF 10 20 30 40 50 60 70 80 90 ThE,
Thl-;(mole%)

Fig. 5.2. The Binary System KF-ThF,.

fluorides, ZrFh, HfF), ThF), and UFy are all monocliniec, the unit cell parameters
indicate a close similarity between ZrF) and HfF), and between ThF}y and UFy. This
relationship is reflected by the temperature minimum in the continuous solid solu-
tions between ZrF)-UF) and ZrF)-ThF), while ThF)-UF) forms continuous solid solutions
without a temperature minimum.

5.2 EFFECT OF TETRAVALENT FLUORIDES ON THE
FREEZING POINT OF SODIUM FLUORIDE

Freezing-point temperatures of NalF were measured for solutions containing ZrF),
HfF), ThF), and UF}, at concentrations up to 15 mole % MF). This investigation was
undertaken to further understand molten fluoride solution behavior, in particular
to see if tetravalent fluorides followed the same correlations as divalent solutes.’

Freezing points with a precision of ¥20c were obtainead by conventional thermal
analyses techniques, using calibrated thermocouples. Melts were stirred by means
of an argon gas stream.
Table 5.2. Invariant Equilibria in the System KF-ThF),

ThF, in Invariant
Ligquid Temperature Type of Phase Reaction at
(mole %) (°c) Equilibrium Invariant Temperature
1k 69k Eutectic Liquid <= KF + p-5KF-ThF)

635 Inversion a-SKF+ThF), 3= B~5KF-ThF)

712 Peritectic Liquid + §KF-ThFh -— a-SKF'ThFh
25 875 Congruent Liquid 3= 3KF-ThF)

melting point

570 Decomposition §K.'F-ThFh -— B-SKF-ThFh + QKF-ThF)_L
32 652 Eutectic Liquid == 3KF-ThF, + 2KF - ThF)
32.5 658 Peritectic Liquid + 7KF-6ThFh -— 2KF * ThF)
46,2 899 Congruent Liquid == TKF-6th

melting point

52 870 Eutectic Liquid 3= TKF+6ThF) + KF-2ThF)
65 929 Peritectic Liquid + Q-KF.3ThF, _— KF - 2ThF)
67 938 Peritectic Liquid + ThF) ss = O-KF-3ThF)

934 Inversion Cx—KF-jThFh -— oz-KF-fiThFu

808 Inversion B-KF+3ThF) -— y-KF - 3ThF,

T22 Decomposition 7-KF-5ThFh -— KF-EThFh + ThFh

9L
77

The results, summarized in Table 5.3, show that the values for Zth and Hth
agree within experimental error and that these two solutes give higher freezing
points than UFy, which in turn gives a higher value than ThFy. Thus, the magnitude
of the freezing point depression of NaF increases with increasing radius of the
quadrivalent solute cation, in contrast with divalent fluoride solutes for which a
decrease in radius is associated with an increase in freezing-point depression.
However, a significant difference between the two types of solutes is the stronger
electric field around the tetravalent ion and the accompanying tendency toward
higher coordination numbers in the liquid state. Assuming eightfold coordination,
and according to the crystal radii, only the thorium cation is sufficiently large
for all eight fluorides to be in simultaneous contact with the cation. Possibly
the smaller cations cannot effectively complex as many fluorides as thorium and, as
a consequence, exert a smaller influence on the activity of NaF. In other words,
zri+ and Hfu+, having ionie radii that are smaller and nearly identical, have a
smaller capacity for fluoride ions, would thus be less acid, and would show smaller
but equal negative deviation with the basic NaF.

The notion of eightfold coordination of quadrivalent ions in solution is sup-
ported by the absorption spectra of UF), in LiF-NaF-KF eutectic6 where the predomi-
nant absorption maximum corresponds closely to the maximum for solid UFu;7 the
positions of eight fluorides surrounding each uranium in the UFY crystal has been
established by x-ray diffraction. Activity coefficients less than unity can re-
sult from negative heats of solution or from positive excess entropies of mixing.
The heats of solution probably follow previously established correlations with
cation radii, but the entropies apparently have more influence in the case of the
quadrivalent cations. It appears that NaF is more disordered by ThFy (big cations)
as a solute than by ZrF) (little cations), and that this effect is large enough to
reverse the sequence of increasing negative deviations expected from only electro-
static considerations.

5.2.1 Phase Diagrams of Fluoride Systems
A decennial index to the ORNL reports (1950 to 1960) which contain data on

fluoride, chloride, and hydroxide phase-equilibrium studies has been issued.?

Table 5.3. NaF Liquidus Temperatures (°C) in NaPF-MF) Systems

Solute Solute

(mole %) ZrF), HEF), ThF), UFy,
0.0 995 995 995 995
2.0 98 985 985
5.0 968 967 96 967
10.0 924 925 911 916
15.0 854 854 823 834

* For ZrF) the measurement was taken at 2.1 mole % and was
983.3. The value given is an extrapolation to 2.0 mole %.
78

5.2.2 The System LiF-YF3z

A recently completed phase diagram of the system 1iF-YFz (Fig. 5.3) differs
substantially from that reported by Dergunov.10 Results of 8RNL phase studies have
shown that the system contains the single binary compound LiF-YFz, melting incon-
gruently to YFz and liquid at 815°C. The peritectic invariant point for this re-
action occurs at 49 mole % YFz. A single eutectic occurs in the system LiF-YFz

at 19 mole % YFz and at 695°C. The structure of the compound LiF-YFz, determined
from single crystal x-ray diffraction measurements, has been shown to be tetra-
gonal.ll Optical properties of the compound have been established.

In conjunction with the ghase studies, the melting point of YFz has been de-
termined to be (1144 * 3)°C.le A large exothermic effect in the YFxz heating and
cooling curves indicates that YFz undergoes a solid state transition at 10600C.

The temperature of this transition has been confirmed by the results of thermal
gradient quenching experiments, but it has not been possible to preserve the high-
temperature modification of YFz by quenching from temperatures above 10600C. Char-
acterization of this form will probably be possible only with high-temperature x-ray
diffraction measurements.

UNCLASSIFIED
ORNL- LR~ DWG 381164
1200

1100 ,////!//,//”
1000 ///////

900 /

3
&
L /
< 800
i N
o
= \ /
W
=
700 N
600
L.I_m
bl
u.
3
500
400
LiF 10 20 30 40 50 60 70 80 20 YF3
YF3 (mole 75)

Fig. 5.3. ‘The Binary System LiF-YFj.

79

5.2.3 Melting Point of NiFo

Apparently NiFp melts somewhere between 1400 ang 1500°C, probably about 1425
or 1450°C, but attempts to measure the melting point have been thwarted by the vola-
tilization of NiFy, at the requisite temperatures and by some poorly understood
phenomena. Previous reported values for the melting point, usually estimated,15'15
have ranged from 1027 to 1450°C.16 The growth of single crystals of NiFs by zone
melting is alleged to have occurred with a hot-zone temperature of lh2000.1¥

The necessary temperatures were obtained in a platinum-wound resistance furnace
or in a controlled-atmosphere Globar furnace. The furnace had three compartments:
a sample heating area, thermocouple well, and a heating-element container. Separate
control of the atmosphere in each of these compartments was possible. The enclosed
heating elements were made of silicon carbide. Temperatures are measured with a
Pt, Pt-10% Rh thermocouple located in a tantalum thermocouple well.

Since the temperature is above the melting point of nickel, containment of
molten NiF, presents a problem. Sealed platinum capsules of 5-mil wall thickness
ruptured from internal pressure developed by NiFp at temperatures of 1400 to 15500C.
(The extrapolated vapor pressure of solid NiFp is 650 mm at 1450°C.) In one experi-
ment, it was observed that the rupturing of the capsule was accompanied by a flash
of light. This effect might be an indication of dissociation to vield Fp. If, as
in the case of copper(II) fluoride, fluorine is produced,l8 then a nickel(I) fluo-
ride must exist, which has not, as yet, been isolated nor characterized. Graphite
was unsuitable as container material because NiFp was reduced to nickel in the
inert atmosphere of the high temperature furnace when a graphite liner was tried.

Although the equilibrium products of the reactions involved in the dispropor-
tionation of chromium(II) fluoride and iron(II) fluoridel9 have been identified, it
has not yet been possible to identify all the reaction products from the dissocia-
tion reactions of nickel(II) fluoride.

5.5 OXIDE BEHAVIOR
5.3.1 Oxide Behavior in Fuels

High-melting-point oxides are generally very sparingly soluble in fluoride
fuels, but when compared with each other at reactor temperatures the apparent solu-
bilities may differ by several orders of magnitude. The differences provide the
basis for metathetical reactions between oxides and fluorides, many of which are
potentially useful for reprocessing fuels.20 However, the very low solubility of
UO2 leads to hazards associated with unintended contamination by oxide.

Recent experiments provide a tentative ranfiing for solubfilities, in solvents
like the MSRE fuels, corresponding to Pa't < zrt < UM << ThH¥+ < Bett << Li*, in
order of increasing solubility of the oxide.

The possibility that ZrOs may be preferentially precipitated in the presence
of UFy is under investigation. Quantitative evaluations are not yet available
because of unexplained discrepancies in some of the analytical results, but there
are plausible indications that the inclusion of ZrF) in the MSRE fuel, at a concen-
tration of about 5 mole of ZrF) for 1 mole UFy, results in the solubility product
for ZrOp being exceeded at an oxide concentration too low to precipitate UOp. If
these results are borne out, the hazard of oxide precipitation in the MSRE could
involve ZrO; ratner than UOp, and this much more favorable situation could be
80

achieved without causing a noticeable increase in vapor pressure OTr important
changes in other physical properties.

Demonstrations of the metathetical reactions involving oxides in fuel mixtures
have been carried out in Pyrex apparatus in a furnace equipped for visual observa-
tion. The oxide from Pyrex contaminates the melt sufficiently slowly that the oxide
content can be considered constant over short ranges of time.

The oxide present in clear solvents such as LiF-BeFp, mixtures can be titrated
with UFy. A black precipitate of UOp settles to the bottom until the end point is
reached, when the green color of UFy (detectible at 300 ppm) becomes apparent.

Either ZrFy or ThFy added to LiF--BeF2 containing precipitated UOp caused a
dissolution of at least part of the precipitate, and ZrOo seemed to be more insolu-
ble than UOp. Analyses of samples removed with a glass filter stick from molten
LioBeF), saturated with ThOp gave a solubility of 2.03 wt % or 8.05 x 10-3 moles of
ThO» dissolved in 100 g, and the solubility of BeO appears to be about 0.22 mole/lOO g

5.%.2 Zirconium Oxyfluoride and Attempted Preparation of Uranous Oxyfluoride

Oxyfluorides were studied because they may result from the contamination of the
molten reactor fuel with oxides. Although some early workers claimed to have pre-
pared UOFp, the compound has not been substantiated.2l Wright and Warf22 found that
U0, and UFy did not combine on heating to 1000°C.

The synthesis of ZrOFp from ZrF) and ZrOo was performed and crystallographic
data were obtained. Zirconium oxyfluoride also sometimes forms when ZrOo 1s reacted
with molten NHpHF5.

Although UOF, is not known, an attempt was made to form a mixed uranous-
zirconium oxyfluoride. The ionic radius of ut* (0.97A) is sufficiently close to
that of Zr+% (0.80A) that a solid-solid solution was considered likely. Reaction
of ZrF), with UOp at 1000°C yielded UF), and ZrOp rather than the hypothetical
UZr(OF5)p. If a solid solution does form, it is very limited, as UF) was still a
major constituent of a fusion corresponding to the composition 47100 . 4ZrF) - UF), - UOs.

5.% GRAPHITE COMPATIBILITY
5.4.1 Removal of Oxide from Graphite by Treatment with Hydrogen

The feasibility of a hydrogen treatment for removing oxide from the MSRE
moderator graphite appears to be unsatisfactory. A specimen (1—1/2 in. in diameter
and 2 in. long) of low-permeability R-0025 graphite obtained from the National Car-
von Co. was treated with hydrogen at 6000C for 6 hr at a flow rate of ~2 liters/min.
Following this treatment, the specimen was degassed at 1000°C for 21 hr. The results
are given in Table 5.4 along with data obtained from two similar specimens of R-0025
graphite which had been degassed thoroughly at 600°C prior to degassing at 1000°C.

Because of the variation of the gas content from sample to sample, it is diffi-
cult to establish whether the hydrogen treatment had any beneficial effect. At best
the treatment removed less than half the surface oxide and, since this would be the
portion most readily removed, a satisfactory cleanup would require a prohibitively
long period of treatment.

81

Table 5.4. Gases Removed from R-0025 Graphite at 1000°C

Volume Composition of Evolved Gas
Evolved (vol %)
(em>/100 emd Hydro-
of graphite) Ho Hs0 carbons co No CO,
Ho-fired 6.7 66 1 6 23 3 0.5
Not Ho-fired T.5 55 1 2 34 1 5
Not Hp-fired 8.7 Sk 0.3 2 3l L L

A sample of AGOT-NC6, which 1s more pervious, is being hydrogen-treated. Pre-
liminary indications here also point toward a removal by the Ho treatment of only
about half the oxide in the graphite.

5.4.2 Behavior of Graphite when Wetted by a Molten Fluoride

Among the molten fluorides which have been observed to soak through graphite
containers rather rapidly, and are therefore presumed to wet graphite, are CsF,
PoFp, and SnFp. Of these, SnFp (m. p. 213°C) has been selected as the most con-
venient for study. The presumptive evidence for wetting by SnFo was confirmed by
the appearance of a thin film of salt on graphite which has been exposed to SnFo
melts. In a rough measurement with a DeNouy tensiometer the surface tension of Snkoy

at 290°C was found to be about S0 dynes/cm, compared with about 190 to 200 dynes/cm
for the MSRE fuel.

As an exploratory test, the rate of permeation by SnFp was followed by examin-
ing l-in.-long cylinders (1/2 in. in diameter) of AGOT graphite which had been
immersed in molten SnF2 for various times at about 230°C. Weight increases, as
obtained from density measurements after removing the adherent salt, varied approxi-
mately linearly with the logarithm of time in the interval between 10 and 1000 wmin
(10, 20, and 27% at 10, 100, and 1000 min, respectively). The rate during this
interval was conjectured to have been controlled chiefly by the diffusion of trapped
gas, occupying the graphite voids.

5.5 PREPARATION OF PURIFIED MATERTIALS

S1x 100-kg batches of LiF-BeFp-ThF)-UF) (65.0-30.0-4.0-1.0 mole %) have been
purified in addition to numerous small-scale preparations of a variety of melts.
The effectiveness of strong reducing agents, such as zirconium metal, for pretreat-
ing coolant mixtures to exhaust oxidizing impurities, seems to be hampered by
deposits, as yet unidentified, on the surface of the reducing metal. Experiments
along this line are continuing.

REFERENCES

1. R. E. Thoma, Phase Diagrams of Nuclear Reactor Materials, ORNL-2548, pp 81,
109 (Nov. 6, 1959),

2. W. H. Zachariasen, J. Am. Chem. Soc. 70, 2147 (1948).
22.

82

E. P. Dergunov and A. G. Bergman, Doklady Akad. Nauk. 5.8.8.R. 60, 391 (1948).

R. E. Thoma, Crystal Structure of Some Compounds of UF4 and ThFh with Alkali
Fluorides, ORNL CF-58-12-40 (Dec. 11, 1953).

S. Cantor, Reactor Chem. Div. Ann. Prog. Rep. Jan. 31, 1960, ORNL-2931, p 48-L9.

J. P. Young and J. C. White, Anal. Chem. 32, 793 (1960); also, personal commu-
nication, J. P. Young.

J. G. Conway, The Absorption Spectrum of UFL and the Energy levels of Uranium V,
UCRL-8613 (Feb. 3, 1959).

R. D. Burbank, The Crystal Structure of Uranium Tetrafluoride, K-769 (June 6,
1951).

R. E. Thoma, Fused Salt Phase Equilibria--A Decennial Index to ORNL Progress
Reports, ORNL CF-60-7-52 (July 20, 1960).

E. P. Dergunov, Doklady Akad. Nauk, S.5.5.R. 60, 1185 (1948).

Met. Ann. Prog. Rep. Sept. 1, 1959, ORNL-2839, p 288.

3. Cantor and T. S. Carlton, personal communication.

L. L. Quill (ed.), Chemistry and Metallurgy of Miscellaneous Materials:
Thermodynemics, p 202 of Natl. Nuclear Energy Ser. Div. IV, vol. 19B, McGraw-
Hill, New York, 1950.

M. Blander and F. F. Blankenship, ANP Quar. Prog. Rep. Sept. 30, 1957,
ORNL-23%87, p 121 (classified).

C. Poulenc, Ann. Chem. Phys. 2, b4l (189L4).

L. M. Matarrese, The Magnetic Anisotropy of Iron Group Fluorides, Thesis,
Univ. Chicago, Aug. 3, 195k.

H. Guggenheim, J. Phys. Chem. gi, 9%8 (1960).

H. V. Wartenberg, Z. anorg. Chem. 241, 381 (19%9) .

J. D. Redman, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p U45.

MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 81.

J. L. Katz and E. Rabinowitch, The Chemistry of Uranium, Part I, The Element
Tts Binary and Related Compounds, lst ed., p 576-T7, McGraw-Hil1l, New York, 1951.

J. Wright and J. Warf, "Attempted Preparation of Uranous Oxyfluoride,” p 19 in
See. V of Technical Report, Ames Projects, Chemical Research, Analytical Chem-
istry, Report for Period Feb. 1 to Mar. 10, 1944, by C. A. Thomas, F. A. Sped-
ding, and H. A. Wilhelm.

6. ENGINEERING RESEARCH

6.1 PHYSICAL-PROPERTY MEASUREMENTS

6.1.1 Surface Tension and Density

The surface tensions of two additional NaF-BeFp (57-43 mole %) mixtures were
determdined in the temperature range of 500 to 800°C using the maximum-bubble-
The results are shown in Fig, 6.1 in comparison with the
previously reported measurement® for a relsted mixture (NaF-BeFz; 63-37 mole %).°
A least-squares analysis of the three sets of data yielded the linear correla-
tions given in Table 6.1; similar equations for two mixtures containing small
additions of UF4 and ThF4 to the NaF-BeF> base salt are also tabulated.

pressure technique.1

SURFACE TENSION (dynes /cm)

250

UNCLASSIFIED
ORNL-LR-DWG 52054

200 -

8OO

—“-—-_._____. ‘ A-'—':-'_ —.-‘:"_m::. :‘._._._
| T\“"—-—-—_ —
150 — - SAMPLE NO. SYMBOL COMPOSITION (mole %)
NaF BeF,
1 . 571 429 ’
2 4 57.5 42.5
00 3 0 63.0 37.0
500 550 600 650 700 750
TEMPERATURE (°C)
Fig. 6.1. Comparison of Surface Ten-

sions of Several NaF-BeF, Mixtures,

Table 6.1.

Surface Tensions of Seversl

BeFa~Containing Fluoride Mixtures

*
Surface Tension

Composition

Sample (mole %) o (dynes/cm)

ve- NaF BeFz  UFy  ThFg a b

1 57.1  42.9 - - au6.h 0,116

2 57.5 42,5 - - 269.4 0.136

3 63.0 37.0 - - 293.2  0.165

4 62.0 36,5 0.5 1.0 272.2  0.143

5 62.0  37.0 1.0 - 235.5 0,090
*s=a-bt for t in °C.

83
84

The results for the two NaF-BeFz (57-43 mole %) mixtures show somewhat lower
surface tensions than that obtained with the NaF-BeFg (63-37 mole %) salt. Thus,
at 650°C, o for sample 1 (see Fig. 6.1) is 8.5% below and ¢ for sample 2 is 2.5%
below the value for sample 3. The discrepancy of 6% between samples 1 and 2
(having essentially identical compositions) may relate to the longer operational
exposure of sample 1 in the circulating loop from which the specimens were drawn
and hence to the increased smount of corrosion-product contaminents. An analysis
of the samples for Fe, Ni, and Cr will be made. The effect of impurities on the
surface tension of a fluid is illustrated by recent results on o for two distilled-
water samples using the meximum-bubble-pressure method. For the first of these, a
sample obtained directly from the laboratory distilled-B20 supply line, 0 was meas-
ured as 67 to 68 dynes/cm; while for the second (triply distilled and doubly de-
jonized), ¢ was found to be T2 to T3 dynes/cm.

The surface tension was calculeted from the Schrodinger equa.tion,4
1 1.2
o =2 [too), - )] v -5 0, (1)

written so as to make apparent the correction to the maximum bubble-pressure (as
indicated by & precision water manometer) arising from the "hydrostatic" pressure
due to the immersion of the capillary tip below the salt surface. In Eq. (1), r
is the capillary radius, h, the fluid head, and p, the fluid density; the sub-
scripts m and s designate the manometer fluid and the salt, respectively.
Terms of the Schrodinger equation higher than the second degree in r have been
neglected in that they represent an adjustment of only 0.1% in the value of the
surface tension.

Examination of Eq. (1) shows that the precision of the g-measurement may be
related to errors in the pressure measurement, in the determinetion of the cap-
illary radius, and in the value of the salt density. It is believed that the
residual errors associated with the pressure and geometrical measurements have
been reduced to a total of *3%. Thus the biggest remaining source of error is in
the salt density; e.g., an uncertainty of +10% in the salt density leads to an
pdditional *3% variation in o. Accordingly, an experimental determination of the
densities of these mixtures is planned.

A comperison of predicted and experimental densities for the NaF-BeFz (63-37
mole %) mixture is given in Fig. 6.2. It is seen from this figure that the curve

UNCLASSIFIED
ORNL-LR-DWG 52055

2.2 : \
74_(‘?!NTERPOLATED,r«u_
- 2 ‘*afij\\\‘~a |
" _ —~ (A) EXPERIMENTAL, ORNL
Q -'---L-__-__._-_ T —
o U ———
o0 b= h‘ —+===
t ‘--...__“\ \ | ‘j-—-
T ~~(C) PREDICTED, ML
B ~
< 1.9 (D) PREDICTED, ORNL- = ==md -
.~
.8
600 650 700 750 800 850

TEMPERATURE {°C)

Fig. 6.2. Comparison of Experimental
and Predicted Densities for NaF-BeF, (63-
37 mole %).

etk i

85

(&) representing the mean of the experimental measurements® at ORNL lies about 7%
above the line (D) obtained from the Cohen and Jones® correlation of the measured
densities of 15 fluoride mixtures of diversified composition. This correlation
is reported as showing a maximum deviation of +6%. In similar fashion a general
density correlation based on more than 60 measurements by Blanke et al.® at the
Mound Laboratory on compositions in the NaF-BeFz-UF4 system yielded curve C. If
the density data for NaF-BeFz mixtures alone are abstracted from the total of the
Mound data and cross-plotted as a function of the NaF content (for various temper-
atures), curve B results. In the calculation of the surface tensions reported in
Table 6.1, the correlation developed from the data of Blanke et al, was used to
estimate the salt densities,

6.1.2 Heat Capacity

The correlation of the heat capacities of BeFa-containing salt mixtures has
been revised to correct errors in the original presentation’ and to include addi-
tional experimental data, The results are summarized in Fig., 6.3 in terms of the
average heat capacity over the temperature range 550 to 800°C and the "mean molec-
ular weight" (the ratio of the molecular weight to the average number of atom
species in the mixture, M/fi). In obtaining the correlating line,

¢, = kel (/RO T (2)
the data point for LiF-BeFa-UFq (70-10-20 mole %) was omitted, The maximum devia-

tion of the remaining data from Eq. (2) is %8.5%; LiF-BeFa2-UFq (70-10-20 mole %)
falls 16% below the correlation and its heat capacity will be redetermined.

UNCLASSIFIED
ORNL-LR-DWG 52056

1.0

o0 COMPOSITION , (mole %)
-. 08 - {7 12| — |12 |0.325 | 30.34
o — 1 25| 80| — |15 |0.315 | 28 .14
. 07 - | 53| 46| - (1.0 |0.459!18.88
o 53 —1 46| — [1.0 |0.513 | 15.46
S o6 62| — | 37 — [1.0 |0.522 | 15.25
2 7t -1 16 {13 | — lo.312 | 25.87
a N 62| — | 365/1.0 | 0.5/0.491 |15.71
5> % 700 —| 10| — [20|0.246| 3.72
£ 05 . 700 —| 14 16 | — |0.323] 28.27
g \ 70.3| - 1 15.3/13.8| 0.6/0.335| 27.00
o
<X
© 04 AN
= -
x & =421M4fivr°7“::>\
u © '\.

e ®

= [ ] '\\
3 0.3
=
T ° \

0.2

10 15 20 25 30 35 40 50 60 70 80 S0 100

A&u AVERAGE MOLECULAR WEIGHT
V' AVERAGE NUMBER OF ATOMS

Fig. 6.3. Correlation of Heat Capacities of Bel2-
Containing Salt Mixtures.
86
6.2 HEAT-TRANSFER STUDIES

The experimental studya of the heat-transfer coefficients for the salt mix-
ture LiF-BeFp-UF4-ThF4 (67-18.5-0.5-14 mole %) flowing in heated Inconel and
INOR-8 tubes was interrupted after 5560 hr of circulation to replace the LFB
pump. Damage to the pump seems to be restricted to one of the shaft bearings;
visual observation of the pump bowl and impeller showed no obvious effects from
exposure to the salt. A new pump and belt drive have been installed. Since the
flow characteristic of this pump can be expected to be somewhat different from
that of the original pump, a recalibration with respect to the turbine flowmeter
will be necessary.

A large semple of salt was found in the vicinity of the ring seal at the
top of the pump bowl. This sample was submitted for chemical analysis; the re-
sults are listed in Table 6.2 in comparison with the nominal and preoperational
compositions for the mixture, The extent of circulation of the ring-seal sample
is unknown, since the salt could have been deposited and frozen in this location
at any time during the 5560 hr of operation. A representative semple will be
obtained from the sump for analysis. Although no measure of the magnesium con-
tent of the original salt is available for comparison, it is believed that the
presence of magnesium in the ring-seal semple results from the introduction of
a small smount of magnesium-bearing thermal insulation into the system during
the replacement of a liquid-level probe located in the pump.

Table 6.2. Composition of LiF-BeFz-UF4-ThF4 Salt Mixture Semples

Composition (mole %)

Constituent Prior to Ring-Seal

Nominal Operation Sample
LiF 67.0 70.8 69.8
BeFo 18.5 16.6 16.0
ThF4 14.0 12.1 13.7
UF4 0.5 0.5 0.5

ppm
Ni | 30 95
Cr 115 510
Fe 305 165
Mg * 2825
Mo * <10
*

Not measured.

87

The effect of this uncertainty in composition on the analysis of the data

is significant. For example, the use of the analysis for the preoperational
salt rather than the nominal in estimating the properties results in a 3% in-
crease in the mean heat capacity of the mixture (see Fig. 6.3). Variations

could also be expected in the other thermal properties of interest — the viscos-
ity and the thermal conductivity — though the changes will probably be of lesser

magnitude than indicated for the heat capacity.

REFERENCES

MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 38.

MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 26.

Chemical analysis indicates a 63-37 mole % (NaF-BeFz) composition rather
than the 6L4-36 mole % constitution previously designated for this sample,

J. R. Partington, An Advanced Treatise on Physical Chemistry, Vol. 2: The

Properties of Liquids, p 185, Longmans, Green and Co. (1951).

S. I, Cohen and T. N. Jones, A Summary of Density Measurements on Molten
Fluoride Mixtures and a Correlation for Predicting Densities of Fluoride
Mixtures, CRNL-1702 (July 195k).

B. C. Blanke et al., Densities of Fused Mixtures of Sodium Fluoride,
Beryllium Fluoride, and Uranium Fluoride, Mound Laboratory Memorandum
(MIM) -1076 (July 1958).

MSR Quar, Prog. Rep. July 31, 1959, ORNL-2799, p 37.

MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 46; MSR Quar. Prog. Rep.
July 31, 1959, ORNL-2799, p 39; MSR Quar. Prog. Rep. Apr. 30, 1960,
ORNL-2973, p 27-29.

7. FUEL PROCESSING

Thorium fluoride will be present in MSR fuels either intentionally in a oOne-
region reactor or as a result of leakage of blanket salt into the fuel in a two-
region reactor. Preliminary investigations indicate that SbFg-HF solutions may be
useful for decontaminating the ThF) from rare-earth fission products. Since the
1iF carrier of the fuel salt interferes by precipitating the SbFg, it must be re-
moved from the fuel by HF dissolution first, leaching the HF-insoluble residue
with SbFg-HF solution would probably decontaminate the ThF), sufficiently for re-use
in the blanket. The behavior of fission products other than rare earths has not
yet been investigated. This particular solvent was considered because it has been
reportedl that SbFg is the strongest "acid" in HF of a number of compounds investi-
gated, and that rare-earth fluorides are soluble in SbF:--HF solutions. Other HF
dissolution systems for MSR fuel processing are generally basic in nature.

In the initial work, HF containing about 20% SbFg was used for the solvent.
Rare-earth fluorides dissolved rapidly to give clear brown solutions; the absence
of precipitation on partial evaporation indicated a high rare-earth scolubility,
probably as the compound RE(SbF6)5. This formula was deduced from the results of
adding liquid SbFc to dry powdered rare-earth fluorides. Considerable heat was
evolved, and the liquid appeared to wet and darken the color of the powder. These
effects were incomplete until 3 moles of SbFg had been added per mole of rare-earth
fluoride; above this ratio some excess liquid SbF5 was present. The compound was
not appreciably soluble in liquid SbFg but was extremely soluble in HF, to perhaps
200 to 300 g of rare-earth fluoride per liter. If less than three SbFsg molecules
are present per rare-earth fluoride molecule, solution is incomplete, but excess
SbF5 does not interfere. Addition of small amounts of water to the HF solutions
resulted in a precipitate.

Thorium tetrafluoride is probably less basic in HF than the rare earths. It
did not react with SbFg or dissolve in SbFS-HF solutions. From mixtures of the rare-
earth fluorides and ThF)y, SbFg5-HF solutions dissolved the rare earths but not the ThF).

The 20% SbF: - HF reacted with LiF, forming an insoluble compound. Addition
of SbFg to an HF solution of LiF resulted in precipitation of a white solid, pro-
bably LiSbFg. Addition of LiF to an HF solution of SbFS resulted in formation of
a white solid, apparently different from the LiF.

A1l observations reported here are qualitative since analytical results have
not yet been obtained.

REFERENCE

1. A. F. Clifford et al., J. Inorg. Nucl. Chem. 5, 27 (1957).

88

OAK RIDGE NATIONAL LABORATORY
MOLTEN SALT REACTOR PROGRAM

OCTOBER 1, 1960

R. B. BRIGGS, DIRECTOR
W. D. REEL, PROGRAM EDITOR

D
Tl

MOLTEN SALT

REACTOR EXPERIMENT
A. L. BOCH, PROJECT ENGINEER R

SPECIAL PROBLEMS

W. B. MCDONALD*

ADYANCED DEYELOPMENT

ENGINEERING RESEARCH COMPONENT DEVELOPMENT FUEL REPROCESSING METALLURGY REACTOR CHEMISTRY
H. W. HOFFMAN R . SPIEWAK™ R D. E. FERGUSON cT W. D. MANLY M W. R. GRIMES RC
D. O. CAMPBELL cT R. G. DONNELLY M R. E. THOMA RC
R. W. SWINDEMAN M G. M. HEBERT RC
T. K. ROCHE M W. F. SCHAFFER RC
R. G. SCHOOSTER M J. E. EORGAN RC
C. M, BLOOD RC
W. M. JOHNSON RC
B. F. HITCH RC
OESIGN METALLURGY REACTOR CHEMISTRY RADIATION TESTING COMPONENT DEVELOPMENT CONSTRUCTION OPERATIONS
E.S. BETTIS R A. TABOADA M F. F. BLANKENSHIP RC 1. A. CONLIN R I. SPIEWAK* R W. B. MCDONALD* R S. E. BEALL
C. H. WODTKE M B. 5. WEAVER RC W. L. SCOTT R D. SCOTT, JR, R
1. H. DEVAN M S. S. KIRSLIS RC
W. H. COOK M F. A. DOSS RC
M. A. REDDEN M . K. R. FINNELL RC
J. L. GRIFFITH M R. G. WILEY RC
W. JENNINGS RC
*DUAL CAPACITY
€T CHEMICAL TECHNOLOGY DIVISION
E&M ENGINEERING AND MECHANICAL DIVISION
I&C INSTRUMENTATION AND CONTROLS DIVISION
M METALLURGY DIVISION
BUILDING AND FACILITIES COMPONENTS SYSTEMS INSTRUMENTATION AND CONTROLS ANALYSIS RC REACTOR CHEMISTRY DIVISION
R REACTOR DIVISION
L. N. HOWELL E&M 3. H. WESTSIK R L. F. PARSLEY R J. R. TALLACKSON I&C B. W. KINYON R
C. E. BETTIS E&M C. H. GABBARD R A. N, SMETH® R R. L. MOORE 14C _C. W. NESTOR R B&R BURNS AND ROE
T. E. NORTHUP E&M A. G. GRINDELL R R. C. ROBERTSON R 0. W. BURKE 1&C D. W. VROOM R L LOCKHEED
0. C. WATKIN E&M P. G. SMITH R S. E.80LT R J. R. BROWN 1&C W. H. FORD L .
T, WATSON E&M L. V. WILSON R R. E. RAMSEY BaR B. SQUIRES B&R W.L.BREAZEALE L D DIRECTOR'S DIVISION
R. W. KNIGHT E&M H. R. PAYNE R R. P. CAPPELLO B&R TI TECHNICAL INFORMATION DIVISION
C. B. HAYNIE E&M W. C. ULRICH R A. J. MADAIO B&R
S. 0. LEWIS E&M J, C. MOYERS* R 5. 5. SCHILKE B&R
0. T. GIBSON E&M D. E. KLION B&R L. J. SCIRICA B&R
G. H. LLEWELLYN E&M C. N. SRSICH B&R
W. E. SALLEE E&M R. E. ANDERSEN B&R
G. W. RENFRO E&M H. N. GORECKI B&R
M. ROSENBAUM E&M $. E. SPERRY B&R
J. T. COX E&M R. E. FRANKLIN R
J. W, KIDD E&M
F. J. UNDERWCOD E&M
W. D. HOYLE E&M
FLOW MODELS REMOTE MAINTENANCE FUEL HANDLING LOOPS AND TANKS GAS SYSTEMS
R. J. KEDL R E. . HISE % AND HEATERS J. L. CROWLEY R AN, SMITH R
T. L. CASTILE R C. K. MCGLOTHLAN R 1. C. MOYERS" R R. B. GALLAHER R
P. P. HOLZ R M. RICHARDSON R J. R.DICKIE R
J. R. SHUGART R G. A. WINDLE R W. H. DUCKWORTH R
C. JONES R F. EISEN B&R

89

-

ORNL-3014
UC-81 — Reactors-Power
TTD-4500 (15th ed.)

INTERNAT, DISTRIBUTION

1. G. M. Adamson 46, A, S. Householder

2. L. G. Alexander 47. L. N. Howell

3. S. E. Beall 48. W. H. Jordan

4, C,. E. Bettis 49, P. R. Kasten

5. E. S. Bettis 50. R. J. Kedl

6. D. 8. Billington 51. G. W. Keilholtz

7. F. F. Blankenship 52, C. P. Keinm

g§. E. P. Blizard 53, M. T. Kelley

9. A. L, Boch 54. B. W. Kinyon
10. 8. E. Bolt 55. R. W. Knight
11. C. J. Borkowski 56. J. A. Lane

12. G. E. Boyd 57. M. T. Lundin
13. W. L. Breazeale 58, H. G. MacPherson
14. TE. J. Breeding 59. W. D. Manly
15. R. B. Briggs 60. E. R. Mann
16. F. R. Bruce 6l. W. B, McDonald
17. 0. W. Burke 62. C. K. McGlothlan
18, D. 0. Campbell 63. E. C. Miller
19, C. E. Center (K-25) 64. R. L. Moore

20, R. A. Charpie 65. K. Z. Morgan
21, W. G. Cobb 66, J. C. Moyers
22. R. S. Cockreham 67. J. P. Murray (Y-12)
23, J. A, Conlin 68. M. L. Nelson
24. W, H. Cook 69. C. W. Nestor
25. G. A. Cristy 70, T. E. Northup
26, J. L. Crowley 71l. W. R. Osborn
27. F. L. Culler 72. L. F. Parsly
28. D. A. Douglas 73. P. Patriarca
29. L. B. Emlet (K-25) 74. H. R. Payne
30. E. P. Epler 75. D. Phillips

31l. W. K. Ergen 76. W. B. Pike
32. W. H. Ford 77. R. E. Ramsey
33. A. P, Fraas 78. W. D. Reel

34. J. H. Frye, dJr. 79. P. M. Reyling
35, C. H. Gabbard 80. M. Richardson
36. W. R. Gall 8l. R. C. Robertson
37. R. B. Gallaher 82. T. K. Roche
38, W. R. Grimes 83. H. W. Savage
39, A. G. Grindell 84. D. Scott
40, C. S. Harrill 85. W. L. Scott
41, M. R. Hill 86. H. E. Seagren
42, E. C. Hise 87. E. D. Shipley
43, H. W. Hoffman 88. O. Sisman
44. P. P. Holz 89. M, J. Skinner
45. A, Hollaender 90, G. M. Slaughter

1
91. A,
92. P.
93. A.
9%. T.
95. C.
2. J.
97. R.
98. A.
9e. J.
100. E.
101. D.
102. W.
163.
164.
165-"741.

N. Smith

G. Smith

H. Snell
Spiewak

D. Susano

A, Swartout
W. Swindeman
Taboadsa

R. Tallackson
H. Taylor

B. Trauger
G. Ulrich

92

:]_03.
104,

105,

106.
107.
108,
109,

110-113.

114-158.

159.

160- :].62 L]

D. C. Watkin

A. M. Weinberg

J. H. Westsik

G. C. Williams

L. V. Wilson

C. E. Winters

C. H. Wodtke

ORNL — Y-12 Technical Library,
Document Reference Section
Laboratory Records Department
Laboratory Records, ORNL R.C.
Central Research Library

EXTERNAL DISTRTBUTION

D. H. Groelsema, AEC, Washington

Division of Research and Development, AEC, ORO
Given distribution as shown in TID-4500 {15th ed.) under
Reactors-Power category (75 copies — OTS)