ORNL-1649.txt

 
 

ORNL-1649

This document consists of 152 pages.
Copy “?éof 2646 coples. Series A,

Contract No. We7405~enge24

AIRCRAFT NUCLEAR PRGPULSION PROJECT
QUARTERLY PRGGRESS REPORT
For Period Ending December 10, 1953

R. C;:Briuht, Director
A. J. Miller, Assistant Director
W. B. Cottrell, Editor

DATE {SSUED

SAN 18 1954

 

OAK RIDGE NATIONAL LABORATORY
:Operated by
CARBIDE AND CARBON CHEMICALS COMPANY
A Divisien of Union Carbide and Carbon Corporation
Post Office Box P :
Ock Ridge, Tennessee

 

IR

3 4456 03L0OLAI &

 
FO Oz

MEZOQAILPPEIPECENPIEIOOLROCTPIREMADMO Qo

x 1 T =,

 

. Adamson

. Affel
. Baldock
Barton

. Blizafd,

. Briant
. Bruce

. Callihan
. Cardwell

Cathcart
. Center

. Charpie
. Cisar

. Clewett
. Clifford
. Cottrell
. Cochran
. Cowen

. Culler

. Eister

. Emlet (Y-12)

Ergen
. Fraas

||qe er

  

. Fiowe

gcl—imIJUE‘U?'CUJ?R‘F‘OOUJFHI§>m.<=EUZJO>'U-nE,,.'.§:J;:

. Johnson

. Keilholtz

. Keim
. Kelley
ertesz

 

INTERNAL DISTRIBUTION

. Blenshlp

. Bredig

. Grigorief] {consultant)

. Grlmesm
e seholder

# Humes (K-25)

    
 
   
  
  
 

%, 64
”g 65 .

6?‘*

68. %A
69.
70.
71.
72.
73.

74,

 

EgflmmIbUFI'flmfliigflg;Nt_f—055:553?«

l-..

ORNL-1649

Progress

» M. King.

- A. Lane

. E. Ldrsgfi:

. S. Livingston
. N. L};Brn

. D Manly

. Mann

. McDonald
. Meem

. Miller

. Morgan

. Murphy

. Nessle

. Poppendiek
. Reyling

. Savage

. Savolainen
. Shipley
isman

. Smith (consultant)
. Snell

. Storrs

. Susano

. Swartout

. Taylor

. Trice

. Van Artsdalen
. YonderLage

. Warde

. Weinberg

. White

G. E&Whlfman
E. P. fllgner (consultant)
G. C. Wfilmms
J. C. Wllsem
C E. Wmfer&
75-84. ANP Library #,

85. B lology Li brar;fi?:_;;,_.

iii
\

 

86-&’6. Laboratory Records Department 94. Reactor Experimental
9]\:\\€_quoratory Records, ORNL R.C. Engineering Library
92. Health Physics L.ibrary ' 95-97. Central Research Library

93. Méfq__‘l'lurgy Library

98-100.
101.
102.
103.
104.

105-116.
117.
118-126.
127.
128-133.
134.
135.
136-137.
138-145.
146.
147.
148.
149.
150.
151.
152-155.
156.
157.
158-160-
161-164.
165.
166.
167.
168-175.
176-177.
178-179.
180.
181.
182.
183-185.

 

EXTERNAL DISTRIBUTION

Air Force Engineering Office, Oak Ridge

Air Force Plant Representative, Burbank

Air Force Plant Representative, Seattle

Air Force Plant Representative, Wood-Ridge

ANP Project Office, Fort Worth

Argonne National Laboratory (1 copy to Kermit Andersan)
Armed Forces Special Weapons Project (Sandia)
Atomic Energy Commission, Washington (Lt. Col. M. J. Nielsen)
Baottelle Memorial lnstitute

Brookhaven Natianal Laboratory

Bureau of Aeronautics {(Grant)

Bureau of Ships

California Research’ ‘and Development Company
Carbide ond Carbon Cf'xg,mnccnls Company (Y-12 Plant)
Chicago Patent Group -

Chief of Naval Research

Commonwealth Edison Compqny

Convair, San Diego (C. H. Heims)

Department of the Navy — Op- 36'2

Detroit Edison Company :

duPont Company, Augusta

duPont Company, Wilmington

Foster Wheeler. Corporation

General Eiegj‘f!c Company, ANPP

General Electric Company, Richland

Glen L. Martin Co. (T. F. Nagey)
Hanfordfbperations Office

lowa State College

Knolls Atomic Power Laborotory

|Lockland Area Office

Los Alamos Scientific Laboratory :
Massachusetts Institute of Technology (Kdufmann)
Materials Laboratory (WADC) (Coi. P. L. Hill) "
Monsanto Chemical Company

Mound Laboratory
 

'''
Qi

 
 
 
 
 
  
  
   
 
 
   
         
  

186-189. i i i utics, Cleveland

190. i A i *Aeronautics, Washington
191. ' i
192-193. >
194-195. North American AV4
196-201. Nuclear Developmen ociates, Inc.
202. Patent Branch, Was}
203-210. - Compdiyy
211-220. Pratt and Whitp#y Aircraft Diyision (Fox Project)
221-232. Powerplant L##b % (2copies to B. Beaman)
233-234. Rand Corp G,
235. San Frangisco Operations Office %,
236, Sylvan lectric Products, Inc.

  

 
   
  
    
   
 
 
  
 
  

237. U. S.d8aval Radiological Defense Lab"fi: 1tory
238. eadquarters ’ '~
239-240. rsity of California Radiation Laboraf’%y, Berkeley
241-242. Hiversity of California Radiation Luboratorg, Livermore
243, A alter Kidde Nuc!ear Laborotorles, inc. %

244-249.5
250-264; Technical Information Service, Oak Ridge, Tennessee

, Curtiss-Wright Corp., Wright Aeronautical Division (K. Campbell)
Armed Forces Special Weapons Project, Washington

 
vi

ORNL-528
ORNL-629
ORNL-768
ORNL-858
ORNL-919
ANP-60
ANP-65
ORNL-1154
ORNL-1170
ORNL.-1227
ORNL-1294
ORNL-1375
ORNL.-1439
ORNL-1515
ORNL-1556
ORNL-160%

 

Reports previously issued in this series are as follows:

Period Ending November 30, 1949
Period Ending February 28, 1950
Period Ending May 31, 1950
Period Ending August 31, 1950
Period Ending December 10, 1950
Period Ending March 10, 1951
Period Ending June 10, 1951
Period Ending September 10, 1951
Period Ending December 10, 1951
Period Ending March 10, 1952
Period Ending June 10, 1952
Period Ending September 10, 1952
Period Ending December 10, 1952
Period Ending Merch 10, 1953
Period Ending June 10, 1953
Period Ending September 10, 1953

 
 

CONTENTS

FOREWORD

PART I. REACTOR THEORY AND DESIGN

INTRODUCTION AND SUMMARY

1.

CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT
The Experimental Reactor
Control rod sleeves
Reactor control
Flow in fuel tubes
Fuel tube cleaning
Reactor Physics
Critical mass
Effect of delayed neutrons
Xenon poisoning
The NaF-ZrF ;-UF ; Fuel
Physical properties
Fuel production
Corrosion of Inconel
Pumps
Prototype sodium pump test
Pump auxiliaries
impeller fabrication
Acceptance tests
Fluid Circuits
Fuel system
Sodium system
Valves
Auxiliary systems
{nstrumentation
Reactor and Fluid Circuit Cleaning
Fuel Recovery and Reprocessing
Transportation of fuel
Rate of dissolution of ARE fuel
Molten fuel dissolution

EXPERIMENTAL REACTOR ENGINEERING

Pumps for High Temperature Liquids
Frozen-sodium-sealed pump for sodium
Gas-sealed sump pump for in-pile loop test

 

Covil
 

Magnstic-torque-transmitter pump
Rotary-Shaft and Yalve-Stem Seals for Fluorides
Graphite-packed seal for spiral-grooved shaft
Graphite-BeF , packed seals
V-ring seal
Bronze-wool, graphite, and MoSz-packed frozen seal
Packed seals for valve stems
High-Temperature Bearing Development
Materials compatibility tests
Bearing characteristics
Heat Exchanger Test
Forced-Circulation Corrosion Loop

. REFLECTOR-MODERATED REACTOR DESIGN STUDIES

Reactor Physics

Fuel Composition and Properties
Moderator Regions

Hydrodynamics of the Fuel Circuit
Pump Design

Pressure Shell

Fuel-to-NaK Heat Exchanger
Reactor Controls

Shielding

Filling and Draining of the Reactor

CRITICAL EXPERIMENTS

PART Il. MATERIALS RESEARCH

INTRODUCTION AND SUMMARY

5.

vifi

CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS
Thermal Analysis of Fluoride Systems
NaF-ZrF ,-UF,
RbF-LiF-UF,
NaF-ThF,
LiF-RbiF-BeF,
RbF-BeF ,-ZrF
Therma! Analysis of Chloride Systems
LiCl-UClH,
NaCl-UCI,
KCI-UCl,
RbCl-UCI

 

24
24
24
24
24
25
25
25
25
26
28
29

31
31
33
33
38
39
4]
42
42
43
43

44

A7

49
49
49
50
S0
51
51
51
51
51
$1
52
60

 

CsCl-UCl,
KCI-LiCl-UCl,
RbCI-UCI,
CsCI-UCl
KCI-LiCl-UCl,
NaCI-LiCl-UCI
Quenching Experiments with the NaF-ZrF , System
Differential Thermoanalysis of the NaF-ZrF , System
Filtration Analysis of Fluorides
Mixtures with 53 mole % NaF
Mixtures with 50 mole % NafF
Mixtures with 75 mole % Naf
Fundamental Chemistry _
Spectrophofometry of supercooled fused salts
EMF measurements in fused salts
Physical chemistry of fused salts
Production of Purified Fluorides
L aboratory-scale production of molten fluorides
Purification of smali fluoride samples for phase studies
Production of enriched material for in-pile loop experiment
Experimental production facilities
Reduction of Na,UF , by hydrogen
Treatment of molten NaZrF . with strong reducing agents
Reduction of Ni!f:2 by hydrogen
Preparation of various fluorides
Purification and Properties of Hydroxides
Purification of hydroxides
Reaction of sodium hydroxide with carbonaceous matter

CORROSION RESEARCH

Fluoride Corrosion in Static and Seesaw Tests

Inconel corrosion by fluorides with metal fluoride odditives

Corrosion of various metal combinations

Corrosion of cermets

nconel with oil and trichloroethylene additives

Inconel corrosion by fluorides with MoS, additive

Screening tests of metallic bearing materials
Fluoride Corrosion of Inconel in Thermal Convection Loops

Effect of fluoride batch purity

Effect of chromium additives

Pretreatment of fluoride with inconel

Effect of exposure time

Effect of temperature

 

52

- 52

52
52
53
33

55
55
55
55
56
56

-7

58
59
59
59
59
60
60
60
61
62
63
63
63

64
65
65
65
66
67
&7
67
67
&7
71
72
72
72
 

EHect of surface-to-volume ratio
Fluoride with 6.5 mole % UF,
The fluoride NGZrF5
Fluoride Corrosion of Nickel and Stainless Steel Loops
Liquid Metal Corrosion
Mass transfer in liquid lead
Static tests of BeD in sodium, lithium, and lead
Spinner tests of Inconel and type 405 stainless steel in sodium
Static tests of beoring materials in sodium, lithium, and lead
Static tests of stainless steels in lithium
Static tests of solid fuel elements in sodium ond sodium hydroxide
Fundamental Corrosion Research
Oxidizing power of hydroxide corrosion products
Equilibrium pressure of hydrogen in hydroxide-metal systems
Mass transfer of chromium in Inconel-fluoride systems

. METALLURGY AND CERAMICS

Welding and Brazing Research
Brazing of radiator assemblies
Brazing of high-conductivity radiator fins
“*Electroless’’ preplating of brazing alloys
Inert-arc welding of solid fuel elements
Mechanical Properties of Incone!
Stress-rupture of Inconel in fluoride fuel
Tube-burst tests of triaxially stressed tubes
Environmental effects on creep of Inconel
High-Conductivity Metals for Radiator Fins
Diffusion barriers
Clod copper
Electroplated copper
Solid phase bonding
Fabrication of Special Materials
Extrusion of Inconel-type alloys
Rolling of chromium-cobalt alloy
Rolling of cobalt
Rolling of iron-chromium-nicke! alloy
Cold-rolled columbium alloy
Tubular Fuel Elements
Ceramic Research
Glass-type pump seals
Ceramic container for fuel
High-density graphite
Combustion of Sodium Alloys

 

72
72
73
73
74
74
75
75
76
76
76
80
80
80
82

83
84
84
86
89
89
21
91
91
92
92
93
93
93
93
93
93
94
94
94
24
94
96
96
96
96
726
8.

10.

 

HEAT TRANSFER AND PHYSICAL PROPERTIES RESEARCH
Heat Capacity

Thermal Conductivity of Solidified Salts

Density and Viscosity of Fluorides

Vapor Pressures of Fluorides

Electrical Conductivity of Fluorides

Forced Convection Heat Transfer with NaF-KF-LiF Eutectic
Flow in Thermal-Convection Loops

Fivid Flow in an Annulus

Transient Surface-Boiling Studies

Circulating-Fuel Heat Transfer

. RADIATION DAMAGE

Irradiation of Fuel Capsules
Creep Under lrradiation
In-Pile Circulating Loops

ANALYTICAL STUDIES OF REACTOR MATERIALS
Analytical Chemistry of Reactor Materials
Oxidation states of iron
Oxidation stotes of chromium
Determination of UF 4 and "U‘Fd
Reducing power of NaZrFs with zirconium addition
Dissolution of fluoride mixtures containing zirconium
Petrographic Examination of Fluorides
Mass Spectrometer Investigations of irradiated Fluoride Fuels
Calculation of UF 4 in unirradiated fuels
Caleulation of U235 lost from irradiated fuel
Determination of U232 burnup
Summary of Service Chemical Analyses

PART 1il. SHIELDING RESEARCH

INTRODUCTION AND SUMMARY

11.

BULK SHIELDING REACTOR

Spectrum of Gamma Rays Emitted by the BSR
Spectrometer arrangement
Results

Fast-Neutron Leakage Spectra of the BSR

12, LID TANK FACILITY

Slant Penetration of Neutrons Through Water
Air Duct Tests

 

97
97
98
98
99

100
101
102
102
102

103
103
104
106

107
107
107
108
108
109
109
109
110
110
11
11
112

115

116
116
116
116
117

120
120
121

xi
13.
14,

15,
16,

 

Removal Cross Sections
Survey of Lid Tank Experiments

TOWER SHIELDING FACILITY

SHIELDING ANALYSIS
Visible Light from a Nuclear Power Plant
Neutron Reflection Coefficient for Water

PART V. APPENDIXES

LIST OF REPORTS ISSUED DURING THE QUARTER
DIRECTORY OF ACTIVE ANP RESEARCH PROJECTS AT ORNL

ORGAMIZATION CHART

xii

 

122
123

125

126
126
126

129
131
139
ANP PROJECT QUARTERLY PROGRESS REPORT

FOREWGORD

This quarterly progress report of the Aircraft Nuclear Propulsion Project ot ORNL
records the technical progress of the research on the circulating-fuel reactor and all other
ANP research at the Laboratory under its Contract W-7405-eng-26. The report is divided
into three major parts: |. Reactor Theory and Design, Il. Materials Research, and Iil.
Shielding Research. Each part has a separate introduction and summary.

The ANP Project is comprised of about 300 technical and scientific personnel engaged
in many phases of research directed toward the achievement of nuclear propulsion of
aircraft. A considerable portien of this research is performed in support of the work
of other organizations participating in the national ANP effort. However, the bulk of the
ANP research at ORNL is directed toward the development of a circulating-fuel type of
reactor.

The nucleus of the effort on circulating-fuel reactors is now centered upon the Aircraft
Reactor Experiment — a high-temperature prototype of a circulating-fuel reactor for the
propulsion of aircraft. The equipment for this reactor experiment is now being assembled;
the current status of the experiment is summarized in Section 1 of Part 1. The supporting
research on materials and problems peculiar to the ARE ~ previously included in the
subject sections — is now included in this ARE section, where convenient. The few
exceptions are referenced to the specific section of the report where more detailed infor-
mation may be found.

The ANP research, in addition to that for the Aircraft Reactor Experiment, falls into
three general categories: (1) studies of aircraft-size circulating-fuel reactors, (2) materials
problems associated with advanced reactor designs, and (3) studies of shields for nuclear
aircraft, These three phases of research are covered in Parts |, H, and llI, respectively,
of this report.
 
INTRODUCTION AND SUMMARY

Assembly of the Aircraft Reactor Experiment

{sec. 1) is nearing completion; all but a few of the
components have been received and installed. The

items still missing include the new control rod
sleeves and parts of the fuel and sodium gas-sealed
pumps. As the installation progresses, various
auxiliary systems, such as the helium, water, and
hydraulic off-gas systems, are being subjected to
operational tests. Tests of a prototype pump have
established the design criteria for the pump cooling
and lubricating systems, as well as the operating
characteristics of the pump.
sleeves will effect a reduction in the structural
poison in the core and hence reduce the critical
mass.
the uranium requirement inside the pressure shell
will be well under 40 Ib of U235, Other physics
calculations on reactor kinetics reveal that reactor
operation and control will not be adversely affected

by either xenon poisoning or the loss of delayed

neutrons in the circulating fuel. The fuel, which
will be obtained by the addition of d concentrate,
NazUFé, to a solvent mixture, NaZrF_, has been
shown to be reasonably compatible with the Inconel
container metal for the temperatures and times
required. Production of both the solvent and con-
centrate are essentially complete and the impurities
in each are well below acceptable levels. Pro-
cedures have been established to assure that the
fuel and sedium systems in the experiment will be
adequately cleaned prior to being filled with the
solvent and sodium, respectively.  Additional
studies of the fuel recovery and processing problem
have established dissolution rates for both molten

and solid fuel in batches containing 4 kg of U235,

With the near-completion of the Aircraft Reactor
Experiment the emphasis of the experimental work
has shifted to the development of components for

general aircraft reactor application and supporting

research. Valves, pumps, bearings and other
components of high-temperature fluoride and liquid
metal systems are being developed for these studies
(sec. 2). Satisfactory, short, frozen-sodium shaft
seals, with length-to-inside-diameter ratios of 1 to
5, have been developed for sodium pumps. A small
gas-seaied pump for an in-pile loop and a canned

The new control rod

Recent physics calculations indicate that

magnetic-torque transmitter for g high-temperature
gircraft pump are being developed. Packed seals
for fluoride pumps and valves have achieved {imited
success under controlled conditions, but none of
the seals are sufficiently reliable for service
conditions. Equipment for high-temperature bearing
tests is being assembled, and a number of potential
bearing materials are being screened on the basis
of wear. A fluoride-to-sodium heat exchange tube
bundle and a high-velocity forced-convection cor-
rosion loop are being operated, but results are not
yet available on either test.

The designs of a family of 50- to 300-megawatt
reflector-moderated reactors have been prepared to
permit an Air Force evaluation of over-all aircraft
performance (sec. 3). The four reactors considered
for 50-, 100-, 200-, ond 300-megawatt power out-
puts are all of the same general species, having
all been extrapolated from the same base., The
designs are based upon presumed fuel and were
made with insufficient physics information for the
establishment of over-all design limitations. In
other respects, however, the designs are on a firmer
basis, since considerable shielding and heat
exchanger data are availableand the hydrodynamics
of the fuel circuit and fuel pump have been con-
firmed by hydraulic mockups. The gas-sealed
centrifugal pump is designed with special baffling
above the impeller so that it will cperate con-
tinuously with its shaft 80 deg from the vertical,
and it will even operate for 1 min when inverted
without “‘gassing®® the pump. Other aspects of the
aircraft  reactor, including moderator cooling,
assembly of the reactor, control, and fuel filling
and draining, are discussed.

The Critical Experiment Facility was used during
this quarter to determine the static physics charac-
teristics of a mockup of an air-cooled water-
moderated reactor for the General Electric Aircraft
Nuclear Propulsion Project (sec. 4). At the same
time, preparations were made for measurements of o

“mockup of a Pratt and Whitney supercritical-water

reactor and a Nuclear Development Associates
sodium-cooled reactor, as well as the Laboratory’s
reflector-moderated reactor. '
ANP QUARTERLY PROGRESS REPORT

1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT

E. 5 Bettis

Jo L.n Mcem

ANP Division

There were no modifications in the ARE design
or critical developments in the ARE program during
the quarter, and therefore installation of equipment
received maximum attention. The installation is
proceeding satisfactorily, ond there was no mojor
holdup occasioned by lack of components. For the
first time since the project began, it is not neces-
sary to explain either the cause for some delay or
some change in plan, and this is the most signifi-
cant remark that can be made to indicate the status
the project attained during this quarter. [t is fairly
clear now that completion of the installation of the
ARE can be effected in an uneventful manner,

One fuel pump sump tank ond one sodium pump
sump tank were received, and therefore the instal-
lation of piping for the fuel and sodium circuits
could proceed in a relatively uninhibited manner,
even though the spare fuel and spare sodium tanks
have not yet been received. Tests on a prototype
pump have established the requirements of the
auxiliary systems for the pumps, as well as the
pump's operating characteristics and performance.

Since all major components of the experiment are
now on hand, all crafts have been able to proceed
with relatively little interference. Installotion of
elecirical heaters ond insulation is progressing
rapidly, The heaters are approximately 65% in-
stalled, but, since the locations for the remaining
heaters are more accessible, the job is 75% com-
plete with regard to man-hours. Insulation work is
following closely behind heater installation. The
most difficult insulating job, that of insulating
tanks and lines in the dump tank pit, is 95% fin-
ished, and cbout 15% of the other insulation has
been installed.

Additional daota hove been obtained from calcu-
lations of the reactor statics and dynamics. An
evaluation of the effects of xenon poisoning und of
the loss of delayed neutrons indicated that neither
will be of much consequence to reactor operation.
Caleulations of the critical mass by two different
methods gove values which show that the uranium
requirement in the pressure shell will be well be-
low the 40 lb permissible from the viewpoint of
vranium allocation and ARE technology. A soving
in mass of about 4 |Ib was realized by a reduction
in the peisoning from the control rod sleeves.

The new Inconel-clad control rod sleeves, which
will replace the three concentric rod sleeves thai
were to be used around each rod, have been fabri-
cated. The spoce outside each new sleeve will be
filled with beryllium oxide strips. The reactor,
which is complete except for installation of the
new rod sleeves and the beryllium oxide strips, has
been subjected to preliminory cleaning tests.
Tests with simulated reactor fuel tubes indicate
that fully established turbulent flow will be real-
ized at a Reyneolds number of 5000, which is wel!

under the ARE design value of 14,000,

The ultimate fuel will be obtained by the addi-
tion of sufficient fuel concentrate, N02UF6, to the
fuel solvent, NaZrF_, to make the reactor critical.
The production of the 3300 Ib of solvent was com-
pleted some time ago, and conversion of the 115

kg of U233 into Na,UF is nearing completion,

Initial tests of the concentrate show that, with the
exception of one batch which was rejected, the
impurities in the concentrate are gratifyingly low.
Density and viscosity data have been obtained for
the concentrate,

Methods for recovering and reprocessing the fuel
after completion of the experiment have received
increased attention. It now appears that the used
tuel may be transported from the experiment to the
reprocessing area in aluminum containers contain-
ing 4 kg of U?33 and that about 20 such transfers
will be required. While the reprocessing procedures
had been established previously, additional dota
have been obtained on dissolution rates for the
fuel in both liquid and molten states.

THE EXPERIMENTAL REACTOR

The reactor is completely assembled (Fig. 1.1)
except for the Inconel-clad stainless steel control
rod sleeves, which have not yet been received, and
the beryllium oxide strips that will be ploced
around these sleeves, The fuel tubes have been
cleaned according to the established cleaning pro-
cedures. Additional measurements of flow phe-
nomena in glass mockups of the fuel tube indicate
that fully developed turbulent flow is realized at
Reynolds numbers abeve 5000, that is, well ynder
the reactor design value of 14,000.
PERICD ENDING DECEMBER 10, 1953

UNCLASSIFIED
PHOTO 11991

SODIUM QUTLET LINE

CONTROL ROD S T " 3 REGULATING-ROD SLEEVE

FISSION C
N :

 

 

Fig. 1.1. Photograph of the Experimental Reactor.
ANP QUARTERLY PROGRESS REPORT

Control Rod Sleeves

In order to reduce the structural poisons in the
reactor, it was decided to replace the three con-
centric Inconel sleeves around each of the four
control rods (one regulating and three safety) with
a single Inconel-clad stainless steel sleeve in
each hole. Fabrication of the new control rod
sleeves has been finished by the International
Nickel Company, but they have not yet been de-
livered.

To further reduce the critical mass of the reactor,
the volume outside the rod sleeve was to be packed
with beryllium oxide pellets. However, since the
beryllium oxide pellets would occupy only about
50% of the volume, it was recently decided to fill
this volume with strips from unused beryllium oxide
moderator blocks. It is anticipated that approxi-
mately 90% of the void volume can be filled with
the beryllium oxide strips.

Reactor Control

A. Lo SOUthern Jo Kn LeSIie
ANP Division

There has been only one chonge (on addition) to
the reactor control equipment. An auxiliary low-
speed drive has been designed for the regulating
rod. This addition was made to provide a means
of manually operating the regulating rod when the
reactor is operating at zero power during rod cali-
bration.

In the event of fission chamber failure or in-
adequate Ak in the regulating rod, it may be neces-
sary to change these components during the course
of the experiment. Consequently, some of the
drive mechanism has been relocated to facilitate
replacement of both the regulating rod and the
fission chambers. No other changes have been
made in the reactor controls and this phase of the
work is complete insofar as preliminary checking
can determine completeness,

Flow in Fuel Tubes
J. Lang G. M. Winn

Reactor Experimental Engineering Division

The experimental relationship between the friction
factor and the Reynolds modulus obtained in a
glass replica of the reactor fuel tube system

—-—

indicated that the transition region (that region
which lies between laminar and fully developed
turbulent flow) lies approximately between Reynolds
moduli of 2000 and 5000. Recent photographs,
Figs. 1.2, 1.3, and 1.4, show the diffusion of dye
filaments in the second 180-deg return bend in a
fuel tube under three Reynolds number flow con-
ditions: 600, 1300, and 2000, respectively. Un-
fortunately, there was more than one dye source
present in the entrance region of the fuel tube
system, and thus several filament trajectories
were superimposed. In general, the sharp, distinct
filaments shown in Figs. 1.2 and 1.3 are character-
istic of laminar flow, whereas the diffuse filaments
indicate some turbulence, Fig. 1.4.

Fuel Tube Cileaning
G. D. Whitman, ANP Division

The cleaning procedures for the fuel and sodium
systems inside aond outside the reactor are de-
scribed in a following subsection (cf., ‘*Reactor
and Fluid Circuit Cleaning’’). However, a de-
scription of the application of these procedures to
the actual cleaning of the reactor fuel tubes follows.

The six parallel fuel circuits in the reactor were
first rinsed, individually, with tap water. This
flushing was accomplished by valving off the out-
lets of all but one circuit and passing water at a
pressure of 60 psi into the fuel inlet header.
Following this pressure flushing, each circuit was
flushed with distilled water, and the water was
blown out as completely as possible with helium.
After the water that could be blown out had been
removed, a vacuum system was connected through
a dry-ice cold trap to the fuel tubes, and the re-
maining water was evaporated from the fuel circuit
at room temperature, During the evaporation pro-
cess 2 cth of helium was bled into the system inlet
to sweep out all moisture. The residual moisture
was determined by the Karl Fisher and dew-point
methods.

The reactor will be complete when the control
rod sleeves and the bottom fuel outlet manifold
have been installed, and it will then be ready for
inclusion in the system.

 

TORNL-1609, ANP Quar. Prog. Rep. Sept. 19, 1953,
J. l. Lang and G. M. Winn.
PERIOD ENDING DECEMBER 10, 1953

e L

UNCLASSIFIED

PHOTO 20724

 

 

Fig. 1.2. Diffusion of Dye Filaments in a Mockup of a Fuel Tube at a Reynolds Number of 600 (Second
180-deg Bend).
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIE
PHOTO 20738

0 SR

 

 

 

Fig. 1.3. Diffusion of Dye Filaments in a Mockup of a Fuel Tube at a Reynolds Number of 1300 (Second

180-deg Bend).

10
PERIOD ENDING DECEMBER 10, 1953

 

WA

UNCLASSIFIED |
PHOTO 20725

 

Re—-2000

 

Fig. 1.4, Diffusion of Dye Filaments in a Mockup of a Fuel Tube at a Reynolds Number of 2000 (Second
180-deg Bend).

11
REACTOR PHYSICS

W. K. Ergen J. Bengston
Cl Bo Mi”s, ANP DiViSion

The UNIVAC multigroup calculations and the
locally performed two-group calculations gave
values which show that the reduction of the poison-
ing by the control rod sleeves has brought the ARE
uranium requirement in the pressure shell to well
below the 40 |b permissible from the viewpoint of
uranium allocation and ARE technology., The de-
layed neutrons lost from the reactor by fuel circu-
lation are predominantly those of long decay
constants, and hence a small amount of excess
reactivity makes it unnecessary for the reactor to
wait for the long-delayed neutrons. Small excess
reactivities will bring the reactor to relatively
short-time constants (10 sec or so), but when the
reactor is at such a time constant it is still sub-
stantially below the prompt-critical condition be-
cause the short-delayed neutrons are not appreciably
reduced by fuel circulation. Speculation as to the
xenon poisoning shows that the poisoning effect
would be easily observable if the xenon were not
lost from the fuel and if the xenon cross section
were as large for the ARE as it is for low-tempera-
ture reactors. Failure to observe the poisoning
might mean that Xe!3® or its parent iodine had
evaporated from the fuel, or it might mean that the
effective xenon cross section was reduced by the
high moderator and neutron temperature. In the
latter case, xenon poisoning would still be a prob-
lem in high-powered reactors of the ARE type.

Critical Mass

The multigroup calculations made by using the
UNIVAC at New York University and essentially
the method previously employed? by the ANP Phys-
ics Group have now yielded fairly reliable, though
not yet final, data regarding the uranium requirement
of the ARE. These data are in agreement with data
obtained by the two-group method. They indicate
that the elimination of much of the poison origi-
nally incorporated into the control rod sleeves has
brought the mass requirement to about 15% below
the 40 Ib which would be permissible from the
viewpoint of phase diagrams,
allocation.

corrosion, and

 

2Reporf of the Technical Advisory Board to the Techni-
cal Committee of the Aircraft Nuclear Propulsion
Program, ANP-52 (Aug. 4, 1950).

35. Glasstone and M. C. Edlund, The Elements of

Nuclear Reactor Theory, p. 329 ff., Van Nostrand,
New York, 1952,

12

Effect of Delayed Neutrons

About one half the delayed neutrons are given
off outside the reactor and lost from the chain re-
action. These delayed neutrons are predominantly
those with long-lived precursors. Most of the short-
lived precursors decay before they are swept out of
the reactor, and their delayed neutrons are hence
not appreciably reduced by the fuel circulation.
Since the neutrons with the long delay times are
not very effective, it will take only a small excess
reactivity to allow the chain reaction to proceed
without ‘‘waiting’’ for these neutrons. A small
excess reactivity will hence bring the reactor to a
relatively short time constant of about 10 seconds.
When the reactor is operating at full power, it will
probably exhibit small reactivity fluctuations, and
this may temporarily bring the reactor from just
critical (infinite period) to a period as low as 10
seconds. Hence, it might be difficult to keep the
reactor steadily at a period longer than 10 seconds.
It should be noted that reactivity fluctuations are
not expected at low power operation so that it
should be possible to operate the reactor on a
30-sec period during the preliminary, low-power
experiments. Because of the only slightly reduced
effectiveness of neutrons with short delay times,
the amount of additional reactivity required to
bring the reactor from a 10-sec period to prompt
critical is quite substantial, and the fuel circu-
lation of the ARE does not appreciably reduce this
safety feature below its value familiar in stationary-
fuel reactors.

Xenon Poisoning

Renewed consideration was given to the problem
of xenon poisoning of the ARE. There is reason to
believe that the Xe!33, and probably also its parent
iodine, will evaporate from the fuel so that the
xenon poisoning effect will not occur in a reactor
of the circulating-fluoride-fuel type. Confirmation
of this would be a major accomplishment of the
ARE. To investigate the problem experimentally,
the reactor would have to be run for a time some-
what shorter than the half life of 135 at full
power and then reduced to very low power. |[f the
xenon poisoning is appreciable, the reactivity
will exhibit the characteristic behavior of first
dropping because of xenon formation by 1135 decay
and of then increasing because of Xe!35 decay.
Calculations performed by standard methods® show
that this behavior should be easily detectable if
the iodine and xenon do not evaporate from the
tuel. If the behavior is not found it could mean
that the poison is given off. However, it couid also
mean that the xenon poisoning is largely reduced
by the high moderator and neutron temperature and
that the xenon cross section is rather low for
neutrons that are somewhat faster than room-temper-
ature thermal neutrons. If the latter explanation
were correct, even with reduced cross
section, would be a poison for reactors of higher
power than the ARE.

There is the interesting possibility that the
xenon, and maybe the iodine, is released from the
fuel at a rate comparable to the rate of radiocactive
decay. If so, the build-up and decay of the xenon
poisoning could still be observed, but the time
constants would be different. A quantitative
analysis of the reactivity following a full-power run
would disclose this situation. There is also the
theoretical possibility that other fission fragments
have large resonances at
energies above those corresponding to the xenon
resononce. in high-temperature reactors such as
the ARE, these poisons might conceivably be
observable.

xenon,

absorption neutron

THE NaF-ZrF4-UF4 FUEL

The reactor fuel will be obtained by the addition
of a sufficient amount of the fuel concentrate,
Na2UF6, to the fue! solvent, NaZrF_, to make the
reactor critical. [t is expected that the final fuel
composition will contain approximately 5.5 mole %
UF,. In lieu of more precise knowledge of the
ultimate fuel extensive material
research and physical property measurements have
been conducted on the fuel composition with 4 and
with 6.5 mole % UF ,, as well as on the carrier and
solvent.,  About 3300 [b of the carrier hasbeen
produced and is being held under helium pending
its use in the experiment. The batch production of
the fuel concentrate is nearing completion. The
impurities in all but one of these batches were
exceptionally low, averaging (not including the one
poor batch) 48 ppm of Fe, 16 ppm of Cr, and 46 ppm
of Ni.

composition,

Physical Properties
H. F. Poppendiek

Reactor Experimental Engineering Division

The densities and viscosities of the fuel solvent,
NaZrFs, and of the fuel concentrate, Na UF
being measured; these properties of the fuef wnth
6.5 mole % UF, were reported previously.? Pre-

PERIOD ENDING DECEMBER 10, 1953

liminary density and viscosity values for Ma, UF
are: viscosity, 17.5 cp at 725°C, 9.9 cp at 975°C;
density, 5.598 — 0.00119 T g/cm?®, where 660°C <
T < 1000°C.

Fuel Production

G. J. Nessle J. E. Eorgan
F. A. Doss

Materials Chemistry Division

Production of the required 3300 ib of NaZrF
now complete, and the material is being held under
helium pending its use in the ARE. Processing of
the 253 Ib of U235 to produce the fuel concenfmf&,
No U¥F g+ 1S nearing completion.

e fuel concentrate is being produced by the
Y-12 Production Division to permit control of the
large quantities of enriched uranium (93.5% U233)
involved. Several weeks were spent training
operators in the handling of the equipment, as well
as in testing the equipment. A few necessary
changes and repairs were made at that time. The
production equipment and the processing technique
are essentially the same as those previously
described® for fuel solvent production.

The concentrate will be produced in 15 batches
of approximately 30 |b each. It has proved possible,
after some simple operational changes, to process
three batches of material per week. However, three
weeks of down-time have been necessary during
this period because of analytical difficulties and
lack of approval of an additicnal uranium allotment.
To date, 10 of the required 15 batches have been
processed, and it is estimoted that the production
operation will be completed before December 1.

As the material was produced by the Y-12
Production Division for use in the ARE, analyses
of the material were made by both the Y-12 and
the ORNL groups involved. Any differences in
these analyses were to be explained and corrected
by wusing, if necessary, the triplicate sample
originally taken before production operations
proceeded. The analytical situation appears quite
satisfactory at present, as production
control and fissionable material accountability is
concerned. Some discrepancies between the two
laboratories have occurred in determinations of the
uranium content of the UFd; these difficulties have

insofar

 

4{\NP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609,
p. 15,

51bid,, p. 15.

13
ANP QUARTERLY PROGRESS REPORT

been shown to be due to foulty sampling of some
poorly milled UF,. Agreement between the labo-
ratories on uranium content of the Na UF, has
been excellent; tentative plans call for accounta-
bility transfer of the material on the basis of these
analyses.

A comparison of the analytical data for eight of
the ten batches for which results are available is
shown in Table 1.1. It should be noted here
that the Y-12 laboratory used spectrographic
analysis for determining the metallic impurities,
while the ANP laboratory used chemical metheds
of analysis. In generol, the agreement is ex-
ceptionally good. One batch (AREU-4) must, on

the basis of these figures, be reprocessed to lower
the impurities to an acceptable level.

Corrosion of Incone!

G. M. Adamson W. D. Manly
Metallurgy Division

The comrosion of Inconel by fluoride fuels has
been discussed at length in this and preceding
quarterly reports. In loop tests in which the
fluoride is circulated at 4 gpm by thermal con-
vection, the maximum depth of attack is less than
10 mils after 500 hr at 1500°F. The recent cor-
rosion dota emphasize the value of minimizing the
impurities in the fluoride and, to a lesser extent,

TABLE 1.1. COMPARISON OF ANALYTICAL RESULTS FOR SAMPLES OF ARE FUEL CONCENTRATE

 

 

 

 

COMPOSITION ANALYSIS
BATCH NO. LABORATORY Weight Fraction (g/g) Impurities (ppm)
U Na F Fe Cr Ni B
AREUA Y-12* 0.59534 0.1142 0.2881 25 25 35 1
{310423) ANP** 0.59521 0.1225 0.2830 20 20
AREU-2 Y-12 0.59529 0.1158 0.2865 23 11 53 0.2
{310430) ANP 0.59598 80 20 90
AREL-3 Y-12 0.59702 0.1168 0.2879 30 20 50 0.2
(310431) ANP 0.59661 75 20 20
AREU-4 Y-12 0.59470 0.1160 0.2876 150 12 7400 0.2
(310432) ANP 0.59577 220 25 475
AREU-5 Y-12 0.59637 0.1171 0.2883 14 6 21 0.2
(310433) ANP 0.59649 60 20 35
AREU-6 Y-12 0.59548 0.1174 0.2905 21 9 42 0.2
{310434) ANP 0.59582 60 20 45
AREU-7 Y-12 0.59513 35 6 67 0.2
(310435) ANP 0.59583 100 20 65
AREL-8 Y-12 0.59530 32 6 56 0.2
(310437) ANP 0.59544 75 20 50

 

 

 

 

 

 

 

 

14

*
Fe, Cr, and Ni contents were obtained by spectrographic analysis.

* K
ANP gnalyses were made by wet chemical means.

 
the value of clean container systems. Additives
to the fluoride, such as ZrH_, which would serve
to reduce the impurities in the fluoride mixtures
and hence the fluoride corrosion, are being
evaluated. For a detailed report of the investi-
gation of the effect of various parameters on cor-
rosion in the fluoride-Inconel system, as well as
of the postulated mechanism, see sec. 6, “‘Cor-
rosion Research.’’

PUMPS
H. W. Savage W. R, Huntley
W. G. Cobb A. G. Grindel!
ANP Division

Developmental work on the gas-sealed pumps and
auxiliaries for the ARE has progressed to the point
where acceptance tests of the pumps for the ARE
and the mockup reliability test of a spare pump and
auxiliaries are being performed.

Parts for the two fuel and two sodium gas-sealed
sump-type pumps are being received; one pump
has been completely assembled and two others are
almost complete. The pumps for the two systems
will be identical except for minor differences in
the impeller designs and the volumes of the sump
tanks, since the fuel pump sump tank must provide
volume for the addition of the fuel concentrate, as
well as for the thermal expansion of the fuel. The
operating characteristics of these pumps, as well
as of the components, and the design of the pump
auxiliary systems for oil and dibutylcarbitol
distribution are being determined on a prototype
sodium pump. The fuel pump auxiliary system
design will be the same as that determined for the
sodium pump, since the requirements of the latter
will be the more severe.

Prototype Sodium Pump Test

The tests of the gas-sealed sump pump with
NGF:«Zr1":“--UF4 (50-46-4 mole %) were reported
previously.® This pump is a prototype of those
being assembied for the ARE. The termination of
these tests was caused by bearing noises which
were determined to be the result of slippage of the
inner bearing race on the shaft. The bearing race
and the shaft were altered to give a 0.0002- to
0.0004-in. interference fit between them
attempt to eliminate the slippage.

in an
Subsequent

 

SANP Quar, Prog. Rep, Sept. 10, 1953, ORNL-1609,
p. 21,

PERIOD ENDING DECEMBER 10, 1933

operation of the pump with sedium was terminated
after more than 300 hr ot 1200°F, because a
gasket leak at the parting face of the pump casing
allowed sodium to jet against the top flange near
the primary gos seal. No further bearing trouble
was experienced during the tests, and postrun
examination revealed no bearing or shaft wear.
it is therefore concluded that altering the inter-
ference fit between the bedring and the shaft did
successfully eliminate the slippage that occurred
in the previous tests.

Postrun examination showed that sodium vapor
had no deleterious effect on the silver-impregnated
graphite seal material. To confirm this, silver-
impregnated graphite was immersed in sodium at
300°F for 150 hr in a laboratory test, and again no
harmful effects could be observed. The test to
determine the reliability of the secondary gas sedl
at high system pressure (50 psig) was not completed
because of the gasket leak which brought about
early termination of pump operation. This test is
now being made without sodium in the pump.

It was found that gas entrained during pump start-
up could be rejected in ]/2 min at a pumping rate of
35 gpm and in 4 min at a pumping rate of 125 gpm.
Since there is no apparent demand for higher
pumping rates during startup, the degassing charac-
teristic appears to be satisfactory.

Pump Auxiliaries

The guxiliary systems for the ARE pumps are
(1) a system for circulating light spindle oil as
shaft coolant and seal lubricant and (2) a system
tor circulating dibutylcarbitol, which acts as a
heat barrier between the sump tank liquid and the
gas seal. With the prototype sodium pump operating
at 1500°F, heat loads of 11,000 Btu/hr in the
dibutylcarbitol from the seal-cavity cooling annulus
and 7,000 Btu/hr in the light spindle oil used as
shaft and bearing [ubricant were found. This heat
was removed by water in external heat exchangers.

The tests conducted with the prototype pump to
determine heat loads of the shati- and seal-cooling
circuits showed the need for redesign of these
circuits to eliminate the rotary union in the shaft-
cooling circuit at the upper end of the pump shaft.
The cooling circuits have been modified, and the
cooling and lubricating oil now enters the bearing
housing between the upper bearing and the lubricant-
to-air face seal. Upon entry, the oil is divided;
some of it frickles over the bearings for lubrication

15
and heat removal and the other portion passes into
the hollow shaft for heat removal. All the oil
collects below the bottom bearing in a cavity out-
side the primary face seal. It flows from the cavity
by gravity through nine Y-in.-dia connections at
the bottom of the bearing T’lousing to the lubricant
reservoir for cooling and recirculation.

Heat exchanger tests indicate that eight pumps
will be required for the pump lubrication (oil)
system, thut is, one operating and one spare pump
for each of the sodium and fuel pumps. One pump
and its spare will suffice for the dibutylcarbitol
system for all sodium and fuel pumps. These
auxiliary lubrication and coolant pumps have been
checked and found to be satisfactory.

Impeller Fabrication

All the Inconel impeller castings made by the
vendor have been rejected because of imperfections
such as sand inclusions, cold shots in the cast
metal, and porosity. Each of these imperfections
could result in pump failure because of structural
failure of the impeller vanes or becouse the re-
moval of sand inclusions and metal particles in
the cold shots would cause intolerable unbalance
of the impeller. As a consequence of these re-
jections, impellers for ARE pumps are being
fabricated from Inconel plate stock. Impellers for
the sodium pumps (high flow) will have curved
vanes to provide heat and flow characteristics
similar to those of the cast impellers. Impellers
for the fuel pumps (low flow)} will be drilled, and,
although somewhat less efficient than the curved-
vane impellers, they will be easier to fabricate.

Acceptance Tests

Test equipment has been designed and some of
it has been placed in operation for running ac-
ceptance tests on all ARE pumps. These tests
will include (1) water tests to determine head,
flow, and efficiency of each pump, (2) short-time,
cold, run-in tests of seals at 65 psig, and (3) hot
tests at ARE temperature, flow,
conditions. The water test equipment is now in
operation, and an all-lnconel loop is being fabri-
cated for the hot tests. The hot test loop will
include an ARE-type gas conirol system and the
cooling and lubricating circuits described above.
One ARE pump has been assembled and checked
with cald water. The performance of the pump in
this preliminary test was acceptable.

and pressure

16

FLUID CIRCUITS

G. D. Whitman
ANP Division

G. A. Cristy

Fngineering and Maintenance Division

Except for the installation of the pumps, the
external fuel and sodium systems are essentially
completed. Installation of heaters and insulation
on these systems is about 65% complete, Additional
tests of the bellows-seal and the frangible-disk
valves have demonstrated the reliability -of each
under ARE operating conditions. A flow diagram
showing the major fluid circuits and design-point
operating conditions is presented in Fig. 1.5.

Fuel System

One fuel pump casing has been received and
installed in the fuel system; the other is expected
early in December, Only seven welds remain to be
made before the system, including the spare pump,
is completed; some operational tests may be
undertaken before installation of the spare pump.
The fuel fill-and-flush tanks have been thoroughly
cleaned by hot-water flushing and dried by prolonged
evacuation at slightly elevated

(around 200°C).

temperatures

Sodium System

The status of the sodium system is comparable
to that of the fuel system. Assembly of the sodium
piping is essentially complete, except for 17 welds
that are in the vicinity of either the pump casings
or of the reactor.

Valves

The valve tests reported previously indicated
that the valves to be used in the ARE would give
satisfactory performance if held at a temperature
of 1200°F or below. The only additional infor-
mation on bellows-seal valves involves operational
tests at 1100 to 1275°F after the valve had first
been held for some time at 1300°F in contact with
the fluoride Ncfl:fiZrF“‘-Uf"4 (50-46-4 mole %). In
these latter tests, there was no leakage but some
sticking; the operationat 1300°F apparently softened
the sealing surfaces. The pressure required to
open the stuck valves increased rapidly and then
leveled off as the valve-closed time was increased;
at no time, however, did the pressure exceed 30
psig.
D#G. 22047

TG STACK -_...8_29_.

     

 

 

 

   
 

  

 

 

 

 

 
   

    

 

 

  

 

 

  

 
      

    

     

 

   
 

  

 

 

    
     
  

 
 

 

 

 

 

 

 

 

 

 

  
 
      
   

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 
 

 

      

 

   

 

 

 

 

 
 
   
  

 

 

    
 

 

 

 

 

    

 

 

 

 

  

 

  

 

 
   
 

 

 

 

 

 

 
 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

pas
- %
N 7 SURGE 82A :L
; \_ TANX 855
M
zmGaTe  SwWitH avin. SATES v Iy o GUreoLs e h !
e ; T T LT PG 858 p30p : 7 @
P —— 1 I} \. - i ol ) :g @A =f . : Par o
: Yaoin.LINE } Yo | g 5l reEa x g5z 1o ssa N4 l 5
_______ . i — r ! o 4 la o) L\ AMAAL L - 3
’ A b 4-in LINE | IOTEST ey | W ' - a \ VYWY » | g
r+_RESERVOIR lr‘"D'fl“;"“—*“ beem—dE T 2| S o3 , g @ 3 ! oI cm z
A — 2 6-in. GATE p——— [ ‘ | YIR: iyl H ) 851 ot |, | ms3 . BASIN B
| | Lo rt! & |l =g i@ 5 WA
| i I %5 X D £ % ’e, o 4
& 3-in. VALVES | ] Nak VAPOR “lp 27 by vent 2
e —— : TRAP — &
Foy NO.5,8,7, NO. 1 A &
; AND 8 T AN | N
: ; P50 b g
: : na )
16-in LINE _@33,(}% ‘ w) N8
i 141-144 145148 1472143 'T:j"wux 201
WA'}ER 511 i VAPOR TRAP SUMP J —[ |
i
SUPPLY 505 - W) | ND.2 g!
| N\ 877 i 875 2
i & = =
T T se—{ 2 o LAPL i
509 Pigg [/ P85 Pi8E INEEY 92 0 33 5
- o
o ul36 . i Lo Yugz g «jug3 7
/8ol 801 =
Y 3
fi P193
o L8 5
o 2 &
~
w 7 u
N84,
! PUMP NO 1 PUMP NO. 2
x
2 \35/ l20{68
g FILTER | 19
\ P4t -
F32 ECONCMIZER
= COOLER (o
™ o
23,0 2
=
L v
& & 2= {108
E @ ax i
n g I3 0= Q:
o g Wz ; ; i
7 z 2 =R TR I Frow rueL
g z | HaX \ 4?*/ \ | ENRICHMENT
4 gl »iZ% 34
Lt ] 41 FES L
3 & T aEe —" Ppag
> ol nl A2/
o 2lE Il -
B
=z T
7 z TO DRAIN
s (He 2 2 @ e
5 ;\:r) EE 937 532
T2 2 {he =z 8
=1
| 5 /o
| 2
|
|
/
I

  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
      

 

 

 

 

 

 

 

 

 

 

 

/ Z
< Y 87y v
< o (O |
BULK- o) 106 " 101 8
- READ - | L {
L 70 | | i : 3
[ | | ; | \ £5
- FUEL o = :
- ] CARRIER I - —>—x3, 570 :
2 Do il S Loy \ o
! ' jo2 ==
: ! I o 28 5 DUMP /a
TANK NO.3 TANK NO.2 |5 B TANK o
'Papz 56/ P20
B9 /T @ T
e68
Pl FILL AND DUMPLINE 125
858 1 { 85
BLEsD Ling LJ
e FLOW IN gpm FUEL L RELIEF VALVE

 

REFLECTOR COOLANT —==3~ FREEZE VALVE

EI PRESSURE IN psig

 

a7 837 .y,
et 70_’ .
FCa ;
875 [ B70 gscz | 100 110120001

 

 

- = ' L . -3UT 7
~ . : FROM He SUPPLY TRAILER . HELIUM {s) PUSH-BUTTON CONTROL
862 gz oBn | g, ' A TEMPERATURE IN °F WATER
; H‘L RESERVE MANIFOLD
' I |  PRESSURE !N INCHES OF WATER, GAGE COOLANT FLOW CONDITIONS ARE BASED ON THE ASSUMPTION

THAY THE COQLANT HAS THEZ FOLLOWING PROPERTIES:

CEEOOG

VENT SYSTEM

 

v ) * FLOW N ¢fm

! s
12 He BOTTLES £ =240 o/ ¥t

 

_M__

—p—

—_

VALVE NORMALLY QPEN

VALVE NORMALLY CLOSED

THROTTLING VALVE

CHECK VALVE

X—D—@ REMOTE MANUAL VALVE

HREMOTE MANUAL ANGLE VALVE

5

¥— MANUAL VALVE

gfl SOLENOID VALVE

Fig. 1.5. Aircraft Reactor Experiment Flow Diagram.

¢ =7 TO 13 cp AT OPERATING CONDITIONS
Cp=0.23 Btu/Ib°F

APPROXIMATE VOLUME OF MAIN SYSTEM:
INTERNAL 0.3 13

 

INITIAL-EXTERMAL 3.5 f+2
ENRICHING FLUID (.7 £+ mMaX,
TOTAL 6.5 t3 MAX,

17
Additional data are also available from tests of
the frangible-disk valves operated in the fluoride
fuel. The frangible-disk valve was held in contact
with molten fuel at a temperature of about 1300°F
for 160 hr and was immediately by the
actuator at the end of this period. The valve operation
was normal in every way. These tests on ARE-type
valves in fluoride mixtures have produced con-
fidence in the valves that are to be used in the
ARE, and no serious valve problems are anticipated
in the operation of the experiment. All frangible-
disk valves have been completely fabricated,
tested, and installed. All the Fulton-Sylphon
valves were thoroughly flushed and tested with
helium for leak tightness across the valve seat.

opened

Auxiliary Systems

The helium heat exchanger loops and the as-
sociated hydraulic systems have been completed
and tested in air. These tests have shown that the
systems have lower pressure drop than calculated,
which will result in lower operational speeds and
consequently greater factors of safety (or greater
heat transfer capacity).

The pit for the secondary off-gas system has
been completed, and about 75% of the piping for
the system has been installed. The planned
program for inspecting and festing of individual
systems is being put info operation as rapidly as
the instaliations are completed.

INSTRUMENTATION
R. G. Affel
ANP Division
The detailing of all instrumentation is essentially
complete. System changes have necessitated the

 

TEST

TEMPERATURE

ordering of several small components which will
arrive soon.  Approximately two man-weeks of
labor will be required before the control room is
ready for complete testing. The flux servo system
was installed and tested with simulated error
signals. Ninety per cent of the needed copper
tubing (12,000% ft) is installed. All this tubing
is to be checked for leaks. The major jobs re-
maining to be done are checking and testing,

REACTOR AND FLUID CIRCUIT CLEANING

.. A. Mann
ANP Division

F. F. Blankenship

Materials Chemistry Division

G. A. Adamson
Metallurgy Division

G. A. Cristy

Engineering and Maintenance Division

Corrosion tests have repeatedly demonstrated
the deleterious effects of surface contamination
of container metals and of impurities in the fluoride
mixture on the corrosive attack on the container.
Accordingly, both the sodium and the fuel systems
will be cleaned. The methods and procedures to
be followed in cleaning these systems have been
established by a group consisting of representatives
from the ANP chemical, metallurgical, engineering,
and ARE groups. The complete, final report of
this committee has been issued, and a brief outline
of the procedures is given below in the order in
which they will be used.

PURPOSE

Reactor Fuel Tubes

. Helium pressure test

Flush with hot, high-velocity water
Flush with distilled water

Purge with helium

N W Ry -

. Dry by evacuation

Room Leak test

150 to 180°F Remove foreign material
Room Replace dirty water

Room Remove water

Room Complete removal of water

Fuel System Without Reactor

Flush components with hot water
Rinse with distilled water
Helium pressure test

Circulate hot water with de!‘ergent7

hobs o N

Drain and evacuate

150 to 180°F

 

Remove foreign material

Room Replace dirty water
Room Leak test

150 to 180°F Remove greuse
Room Remove dirty water

7 The detergent to be used is “*Kon Kleen;'" active ingredient, sodium metasilicate,

19
ANP QUARTERLY PROGRESS REPORT

TEST

TEMPERATURE

PURPOSE

Sodium System Without Reactor

1. Fiush components with hot water 150 to 180°F Remove fareign moterial
2. Rinse with distilled water Roam Replace dirty water
3. Helium pressure test Roam Leak test
4. Circulate hot water with defergent7 150 to 180°F Remove grease
5. Drain Room Remove dirty water
6. Rinse with distilled water Room Remove all traces of dirt
7. Alternate evacuation and helium
purging Up to 600°F Removs maoisture

Fuel System with Reactor, Part 18

1. Pressure test Room Leak test
2. Circulate hot water with cle'rergem7 150 to 180°F Remove grease
3. Rinse with tap water Room Rermove cleaning water
4. Rinse with distilled water Reom Remove all contamination
5. Alternate evacuation and helium
purging Up to 600°F Remove moisture
6. Vacuum rate of rise Up to 600°F ! eak test
7. Hot gas test: argon in fuel tubes,
helium in annulus Up to 600°F l_eak test

Sodium System with Reactor

1. Alternate evacuation and helium
purging
2. Circulate sodium

600°F

Up to 600°F

Remove moisture

Clean and leck test

Fuel System with Reactor, Part 1

1. Heat reactor and fuel piping 1200°F
. Evacuate 1200°F
3. Circulate solvent, NGZrFS 1100°F
Replace with clean solvent 1100°F

FUEL RECOVERY AND REPROCESSING

D. E. Ferguson F. N. Browder
G. I. Cathers

Chemical Technology Division

Transportation of Fuel

Although the basic outline of an aqueous pro-
cessing method to be used for ARE fuel recovery
has been established,® the method of transporting
the fuel from the recctor to the processing site

 

8Clecming of the ""Fuel System with Reactor'' is di-
vided into two parts because the high-temperature phase
is dependent upon preheating the reactor with sodium.

D, E. ferguson, G. I, Cathers, and O, K. Tallent,
ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556,
p. 102.

20

Above fluoride mixture melting point
To fill system
Clean and leak test

Ready for operation

will somewhat offect the detoils of the dissolution
procedure. A comparative study indicated that the
most satisfactory method would be transportation of
the solid fuel in partially filled open-top aluminum
A measured amount of the molien fuel would
be poured into the cans and allowed to solidify,
and the cans would be transported one at a time to
the Metal Recovery Building where both can and
contents would be dissolved. The use of open-top
instead of sealed cans would simplify the remote-
filling operation and decrease the dissolution time.
None of the valuable material should be lost as o
result of fragmentation of the charge if there is
sufficient freeboard in the aluminum container.

To illustrate the feasibility of canning molten
fluorides in metallic aluminum, three 10-kg batches

cans.
of NGF'ZFF4'UF4 (50-46-4 mole %) were transferred
to air-cooled aluminum containers. Each of the
cans, two with 0.065-in. walls and one with a
0.25-in. wall, withstood the test in which the
fluoride temperature was 600°C. It was discovered
that the thin-walled cans should not be subjected
to more than 5-in. vacuum or 5-psig pressure,

Previously, cans approximately 5 in, in diameter
for transporting 6.6 kg of fuel plus flush material,
which would contain 0.5 kg of U235 if the mixture
wiere homogeneous (or a maximum of 1 kg of U233
if the mixture were not homogeneous), were con-
sidered.'® This amount of fuel would dissolve
to give about 35 gal of solution.

The process appears to be most attractive if 4 kg
of uranium per container, which amounts to 53 kg
of fuel plus flush (assuming a homogeneous mix-
ture), can be transported and charged to the 500-ga!
dissolver to make one 265-gal batch of solution.
For this size batch, the containers would be about
9 in. in diameter and 18 in. long. This quantity
would be more compatible with the size of the
Metal Recovery Building equipment, and the
number of containers required to transport the 80 to
100 kg of U235 would be reduced from 100 to about
25. Processing of 4 kg of U235 per batch in the
existing dissolver is subject to approval by the
Criticality Hazards Committee.

Another method of dissolving the fuel involves
introducing it to the dissolver in the molten state.
If this method were used, the fuel could be trans-
ported in a special can that would hold about 25 kg
of U235, The advantages of this method are the

 

Y pid., p. 103,

PERIOD ENDING DECEMBER 10, 1953

rapidity of the dissolution and the decrease in the
number of truck trips required for the transportation.
The disadvantages are the difficulty of providing
shielding suitable for use in the handling of the
molten material and the excessive development work
required on equipment that would have to be built
into the cans, for example, heaters for remeliting
the fuel if it solidified before the dissolution was
carried out, temperature-measuring
and devices for controlling flow. Construction of
the dissolver at the ARE site would eliminate the
need for constructing transportable shielding and
would permit the use of less rugged heaters, but it
would require too large a capital expenditure. It
would also split the locations of the chemical
operations, and additional operators
required.

instruments,

would be

Rate of Dissolution of ARE Fuel

Based on the results of laboratory dissolution
experiments, a dissolution time of 6 to 12 hr is
estimated for a 53-kg charge of ARE fuel. For
a 6.6-kg batch, @ dissolution time of 3 to 6 hr is
estimated. A reliable value for the dissolution
time cannot be calculated from the data obtained
in these smali-scale experiments, since the dis-
solution rate is not constant throughout the dis-
solving.

With a 1-in.-dia cylindrical charge weighing
approximately 50 g, an initial penetration rate of
0.003 in./min was observed over o 4-min period
(Table 1.2). Over the next 3 min, the penetration
rate increased to (0.006 in./min, based on the
initial dimensions. This increased rate is due to
uneven penetration and cracking, which exposes

TABLE 1.2. DISSOLUTION RATE OF SOLID ARE FUEL

Cylindrical samples 1 in. in diameter tested inboiling 4 M nitric acid plus 0.67 M aluminum nitrate solution

 

 

INITIAL AREA TIME IN AMOUNT CALCULATED
OF SAMPLE DISSOL VANT DISSOLVED PENETRATION RATE
(in.2) {min) (wt %) (in./min)
First immersion 3.5 4 8.7 0.003
Second immersion 3.5 3 23 0.006
First immersion 5.2 5 13 0.005
First immersion* 3.9 2 2.2 0.002

 

 

 

 

 

+ ) ’
Solution already contained the normal complement of dissolved fuel.

21
ANP QUARTERLY PROGRESS REPORT

ln @ second experiment, a
rate of 0.005 in./min was observed
After the 5-min period, the
sample had broken up into several smaller pieces.

more surface area.
penetration
over a 5-min interval.

Another fresh sample tested in a solution that
already contained a normal complement of dis-
solved fuel gave a penetration rate of 0,002 in./min;
however, in the final stage of an actual dissolution,
the charge would probably have broken up to expose
a very large surface area per unit weight. Based on
these results, the average penetration rate will
probably exceed 0,006 in./min.

The sample charges for these experiments were
prepared by pouring molten fuel (NaF-ZrF -UF,,
50-48-2 mole %) into a split graphite mo?d and
cooling. The pit formed by contraction on cooling
was removed by pressing the cylinder against a hot
metal plate. The sample was placed in the dis-
solvant (boiling4 M nitricacid and 0.67 M aluminum
nitrate solution) and then removed and weighed
after a measured interval to determine the amount
dissolved.

Since the scale-up of these laboratory dota is
inconclusive, three charges will be poured into
aluminum cans, 6 in. in diameter and 9 in. long, and

dissolved to estoblish the dissolution time more
accurately.

Molten Fue! Dissolution

No excessive violence was observed in several
preliminary tests in which molten ARE fuel, at
temperatures of over 550°C, was poured into the
aqueous solvent (4 M nitric acid plus 0.67 M
aluminum nitrate). On contact with the aqueous
phase, the fuel solidified in the form of small
irregular particles 1 to 2 mm in diameter and then
dissolved at a rate of about 50 g in 10 min, which
is comparable with the rate of 50 g in 5 min that
had been observed with finely ground fuel. Ap-
parently the only precaution required is te make
sure that the addition is slow enocugh to prevent
excessive overheating and overboiling. However,
colloidal material has been present in all dissolver
solutions prepared todate from molten fuel, possibly
as a result of the instability or hydrolysis of
zirconium compounds in agueous solutions at high
temperatures., Some modification of the dissolver
solution to prevent formation of this material would
probably be required if molten fuel dissolution
were to be used.

2. EXPERIMENTAL REACTOR ENGINEERING

H. W. Savage, ANF Division

The completion of the developmental work on the
gas-sealed sump pumps for the ARE has made
possible a change in emphasis of the experimental
work, and increased effort has been devoted to the
development of a hydrodynamic type of bearing and
a heat exchanger for the reflector-mederated reactor,

The developmental work on the frozen-sodium-
sealed pump for sodiuvim during this quarter included
tests of a frozen-sodium shaft seal with a single
]/2-in.~|ong frozen-sodium sealing region backed up
with helium pressure in the sealing cavity., The
short seal is being developed in an effort to lower
the power consumption of the seal and yet maintain
the low leckage obtained with the 5]3’/]6-in.-long
seal previously described, In tests, the new short
seal dissipated opproximately 3:2 hp, with no power
fluctuagtions, ond it was found that the sodium
feakage could be made negligible if back-up pres-

22

sure was used to reduce the pressure differential
across the seal, Water tests on the gas~sealed
sump pump for the in-pile loop showed fluid gassing
to be a problem, and, accordingly, chonges are
being made in an effort to eliminate this condition,
In other tests, the problems associated with de-
veloping a conned magnetic-torque-iransmitter for a
high-temperature pump for aircraft opplication are
being explored,

Since the adoption of the gas-sealed sump pump
for ARE operation, seal developmental work has
been carried on at a reduced rote; however, tests
for rotary shaft and valve stem seals have con-
tinved with materials which showed promise in
earlier tests, These include graphite pocked around
a spiral-grooved shaft, graphite and BeF, mixtures,
a V-ring type seal, and a bronze-wool, graphite,
and MoS, mixture, Although each of these seals
has performed well under laboratory controlled con-
ditions, none has demonstrated sufficient reliability
to warrant adoption in a pump for reactor operation,
Also, none of the packing materials tested thus far
for valve.stem seals has demonstrated sufficient
reliakility to replace the bellows as a valve-stem
seal, ‘

A hydrodynamic type of bearing which will operate
in liquid metals, fused salts, or other fluids af
temperatures of up to 1500°F is being developed.
Materials for this application are being screened
in a compatibility testing device, and a test progrom
is being planned for obtaining the bearing charac-
teristics, which are usually represented as friction
factor vs. Sommerfeld number.

A heat exchanger incorporating some design
features proposed for the reflector-moderated re-
actor is currently being tested. The fluoride system
has operated .at above 1200°F for gver 600 hr, but
performance data are not yet available, :

A small forced=convection loop is being operated
to test corrosion of Inconel in circulating fluoride
mixture NaF-ZrF ,-UF , at high fluid velocities and
high temperature differences, The fluid velocities
obtained in this loop are a factor of 50 or 60 greater
than those of thermal-convection loops, but the
Reynolds modulus remains in the laminar region
at approximately 1700.

PUMPS FOR HIGH TEMPERATURE LIQUIDS

W. G. Cobb W. R. Huntley
A. G, Grindeli W. B. McDonald
D. R. Word
ANP Division

Frozen-Sodium-Sealed Pump for Sodium

A frozen-sodivm shaft seal with a ],r:’z-in.-l'ong
frozen-sodium region backed up with helium pres-
sure in the sealing cavity was tested. This seal
differed from the previously tested! short seal in
that only one Y%-in.-long sealing region was used
and the blanket gas was under pressure, The short
seal is being developed in an effort to lower the
power consumption of the seal and yet maintain
the low leakage obtained with the 5‘3/16-in.-!ong
seal previously described.?

'ANP Quar. Prog. Rep. Septf. 10, 1953, ORNL-1609,
po 20- :

2ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556,
p. 11. '

3bid,, p. 17.

 

PERIOD ENDING DECEMBER 10, 1953

For the first 168 hr of operation with the new
short seal, the sodium temperature at the pump
was 960°F, the suction pressure was 20 psig, and
the back-up pressure was 15 psig. The total sodium
leakage post the seal during this period was 32
em3 or an average of 4.8 cm?® per day, Approxi-
mately %:4 hp was dissipated in the seal and the
power trace was smooth,

The Buna N rubber O ring used fo seal the back-
up gas was destroyed during the application of
heat to remove the sodium leakage at the end of
168 hr, and the remainder of the testing was ac-
complished without the use of back-up pressure,
Operation was continued for a total of 980 hr with
a shaft speed of 1800 rpm, a sodium temperature
of 960°F, and a suction pressure of 12 psig. A
slight flow of helium was maintained across the
back of the seal, The increase in the pressure
differential across the seal from 5 psig during the
first part of the test to approximately 12 psig
during the remainder of the test resulted in an
increase in sodium leakage past the seal to ap-
proximately 20 cm? per day.

Erratic power fluctuations caused by sodium
oxide migrating to the cold section of the seal
became negligible when a standard, cold-irap,
bypass filter was installed in the loop. The filter
bypasses approximately 0.25 gpm of sodium, Ex-
cessive heat loss in the filter (approximately 3

kw) necessitafed its redesign to incorporate o heat

economizer designed to waste only 0.3 kw. Transit
time through the new filter is approximately 6
minutes, :

The results of these tests, together with the
previously reported results of tests with longer
frozen-sodium seals, have led to the conclusion
that a frozen-sodium shaft seal is manageable for
a sodium pump, A satisfactory length-to-diameter
ratio for such a seal is approximately 1 1o 5,
Leakage of sodium past the seal can be made
negligible by using back-up pressure to reduce the
pressure differential across the seal, Water cooling
is an efficient means of freezing the seal, and
several thousand hours of operation have demon-
stroted that the hazard in' using water with ¢
properly designed seal is frivial,

Gas-Secied Sump Pump for In-Pile Loop Tesit

A pump is being developed for in-pile loop tests
which is similar to the lcboratory-sized gas-sealed
pump? but which has been scaled down to reduce

23
ANP QUARTERLY PROGRESS REPORT

the fluid holdup in the pump. Water tests of the
first pump constructed show that the pump head
and flow are within design specifications; however,
a fluid gassing problem exists which indicates
that some design changes will be necessary,

While some gassing of the fluid is evident under
all operating conditions, its magnitude increases
rapidly as the height of the free fluid surface above
the suction is decreased or the shaft speed is
increased. This indicates that the difficulty is
caused by vortexing of the free fluid surface be-
cause of shaft rotation, As an initial aftempt to
suppress vortexing, chaonges are being made in
the antiswirl baffles in the pump, and the opening
through which the rotating shaft enters the suction
chamber is being moved to a point well below the
free surface of the fluid,

Magnetic-Torque-Transmitter Pump

High-temperature pumps for aircraft application
may include o canned magnetic-torque-transmitter
pump with the external shaft driven by suitable
means, for example, a gas turbine. The problems
encountered in building such a torque transmitter
are being investigated.

been

Reduced-scale tests have
conducted for dstermining the expected
efficiency of a magnetic torgue transmitter at
elevated temperatures, the geometry of such a
transmitter, and to what extent slippage occurs as
a function of temperature.

The first transmitter
permanent magnet outer rotor and a squirrel cage
(from a

tested consisted of a
]/"m-hp motor) inner rotor, A nonmagnetic
stainless steel can separated the driving and
driven rotors, The entire assembly (excluding the
outboard bearings) was heated to above 1100°F,
and the slippage rate was appreciable in the tem-
perature range of 600 to 1100°F, However, the full
zu-hp torque was transmitted from room tempera-
ture to 600°F, but the efficiency then dropped off
rapidly at higher temperatures, It is yet to be
determined whether the drop in efficiency is due
to the drop of the flux of the permanent magnef or
to lowered resistance of the windings of the rotor.
A permanent magnet inner rotor is presently being
constructed which will resolve this question,

 

YANP Quar, Prog. Rep. Sept, 10, 1953, ORNL-1609,
p. 23.

Sfbid., p. 24.

SANP Quuar. Prog. Rep. June 10, 1953, ORNL-1556,
p. 19.

24

ROTARY-SHAFT AND VALVE-S5TEM
SEALS FOR FLUORIDES

R. N. Mason P. G, Smith
W, C. Tunnell
ANP Division

Graphite-Packed Seal for Spiral-Grooved Shaft

Operation of the spiral-grooved shaft with a
powdered-graphite-packed seal? has continued for
over 3000 hr with no detectable leakage of fused
fluorides. During this periocd of operation, heater
failure occurred on two occasions and caused the
shaft rotation to stop. When repairs were mude,
operation was resumed with no difficulty.

Graphite-BeF, Pocked Seal

The previously reported® test of a seal packed
with @ mixture of 80% graphite powder and 20%
BeF, has now operated for over 3700 hours, The
leakage for the past 3500 hr was less than 15 in,3.
Power fluctuations have occurred, and some of
them were accompanied by squecking, which
indicated possible metal-to-metal contact between
the shaft and the gland,

When this pocking material was installed in a
tester having a 2]/4-in.-dia shaft, excessive graphite
and BeF2 leakage occurred, Investigation dis-
closed that the temperature during pretreatment
was insufficient to fuse the Ber, and thus the
seal contained loosely packed material rather than
the solidly fused material desired, For the second
test, the pretreatment temperature was raised to
1800°F,

solidly, During the dry run of this seal, there wus

and the BeF-graphite mixture fused

no leakage of powder. The seal region has been
operated at a higher temperature (above 975°F) in
this test than in previous tests to reduce power
fluctuations, There has been some fluoride leckage.

V-Ring Seal

Operation of the V-ring seal test® was continued,
and o heater was installed inside the 21,6-inmadiq
rotating shaft, During on extended dry run in
which the seal was gradually heated to about
1200° F,
tained, The inert gas blanket and seal gas supply
was inadvertently shut off for about 20 min during

satisfactory sealing of helium was at-

operation, and, since some damage to the seal
rings occuited, satisfactory sealing could no
[onger be obtained, Examination of the rings after
disassembly of the seal showed that the outside
ring was 75% destroyed and the other rings had
either oxidized or worn about 0.0005 to 0.0015 in,
on the inside diameter; the shaft was undomaged.
The outside ring has been replaced and another
dry run will be made in an attempt to seal helivm
at above 1000°F before fluorides are introduced.

Bronze-Wool, Graphite, and Mo$2~Pocked

Frozen Seal

The test’ of the bronze-wool, graphite, and MoS -
packed frozen seal continued 'until operation wcs
terminated at the end of 1198 hr because of leakage
of the fluoride mixture NaF-ZrF -UF  from a Hange.
The leakage rate was less than 1 cm? per day with
the fluoride pot at 1175°F ond pressurized to 10
psi, :

An odditional test was made with the packing
annulus reduced from ¥ to ¥ inch. This test wos
unsuccessful and was terminated at the end of
211 hours., The failure was attributed to faulty
alignment between the shaft and the packing
housing.

Packed Seals for Valve Stems

A series of tests is in progress to determine
whether the packing materials which sealed well
in the packing penetration tests® will seal high-
temperature fused salts under simulated valve
operating conditions. None of the materials tested
has demonstrated sufficient reliability to replqce
the beliows as a valve-stem seal.

The test apparatus consists of a standard valve
stem and bonnet placed in a vertical position with
the stem down, The pot in which the molten fluo-
ride mixture ‘is pressurized against the packing in
the bonnet is welded to the top of the bonnet. The
stem is rotated at frequent intervals to simulate
valve operation, Three tests have been completed
in which the valve stem was cycled once every 24
hr, and the pressure on the packing was 5 psi,
The packings tested were Asbury graphite 805 ot
1100°F, 90% Asbury graphite 805 plus 10% Bsz
at 1100°F, and 50% graphite (from the Y-12 Carbon
Shop) plus 50% BeF, ot 1350°F. Each of these

tests was terminated because of fluoride leokage.

 

72ANP Quar, Prog. Rep. Sept. 10, 1953, ORNL- 1609,
p. 25

8/bid., p. 28.

PERIOD ENDING DECEMBER 10, 1953

HIGH-TEMPERATURE BEARING DEVELOPMENT

R. N. Mason P. G. Smith
W. C. Tunnell W. K. Stair
ANP Division

Developmental work is under way on a hydro-
dynamic type of bearing which will operate in
liquid metals, fused salts, or other fluids at tem-
peratures of up to 1500°F. This developmental
program will probably be carried out in three phases:
(1) test programs to find materials which have low
wear and low corrosion rates in the fluids of
interest at 1500°F, (2) o test program to establish
the bearing characteristics, which are usually
represented os friction foctor vs. Sommerfeld
number, and (3) the design and test of a bearing for
a particular application using the information ob-
tained in phases 1 and 2. Effort to date has been
confined to materials compatibility fests and to an
evaluation of the accuracy which may be expected
in carrying out the program for obtaining bearing
characteristics,

Materials Compatibility Tests

Initial screening tests for finding materials that
possess satisfactorily low wear and low corrosion
rates are being made in apparatus in which a
23/’-m -dia, / in.-thick plote specimen of one ma-
fenai rofafes against a Y.in.-dia stationary pin of
another or the same material, The end surface of
the pin is ground to a Y-in.-radius cylinder to
approximate line contact along o radial line of the
rotating plate at a mean radius of 1 in, from its
center (Fig. 2.1). Contact pressure between the
plate and pin is maintained by means of an external
coil spring and a sliding shaft, '

A test was run of the equipment made with a
Graphitar plate and an Inconel pin. A Hertz stress
(estimated stress in the line of contact) of about
10,000 psi, or a 5-lb force load between the plate
and pin, was used. With the test specimens sub-
merged in acetylene tetrabromide to simulate the
fluoride melt, it was found that the 5-lb load was
insufficient for breaking through the hydrodynamic
film and that the film persisted until a load of
about 10 1b was applied to give o Hertz stress of
about 15,000 psi. The rate of rotation was 850
rpm, which gave a sliding speed of about 7.4 fps.
Having determined the load requirement, a series
of correlation tests was conducted with plates of
type 416 stainless steel rotating ogainst stationary

25
ANP QUARTERLY PROGRESS REFPORT

UNCLASSIFIED
DWG 22338

      

LOADING SPRING —— =3

DRIVE WHEEL

| —BALL BEARINGS
i[5 .
e,

SPINDLE SHAFT

 

 

Fig. 2.1. Bearing Materials Compatibility Tester.

pins made of babbit, a die steel, a high-speed tool
steel, Stellite 6, Stellite Star J, ond Superoilite.
In each case, the stationary pin and the rotating
plate were submerged in a regulor-grade, Texaco,
Regal “'A'" oil, without additives, that has a
viscosity of SUS 40-44 ot 210°F, The results of
this series of tests conformed with those obtained
earlier at KAPL.?

The first test at a high temperature was made
with Graphitar rotating against Inconel in Nal-

 

D, 8. Vail, Compatibility of Materials in Liquid
Metals, KAPL-589 (Aug. 18, 1951).

26

ZrF UF , ot 1200°F for 2 hours. The Graphitar
appeared to be scratched lightly and it was warped,
while the Inconel specimen appeared to be burred
and built up at the rubbing surface.

The following materials are being screened in
seesaw corrosion tests: B, ZrC + e, Mg0Q + Ni,
BeO + Ni, BeQO, hot-pressed Al O/, high-density
graphite, TiC + Ni, TiC + Co, GC, B,C, WC + Co,
and WC + Ni. Plates and pins of most of the above
materials either have been ordered or will be
fabricated.

Combinations of materials which possess both
satisfactorily low wear and low corrosion rates
based on the qualitative results of the screening
tests will subsequently be tested in equipment
which simulates a conventional journal and bearing.

Bearing Characteristics

An evoluation has been made of the accuracy of
the data which will be obtained in the test program
for determining bearing characteristics. Two ex-
perimental approaches for obtaining the data have
been reviewed. First, the test may be simply a
load-life experimental study to establish reliability
of a single bearing design in fluoride melts,
Information gained from this approach might or
might not be useful in predicting the behavior of
other bearings; certainly, the information could be
used only with limited assuyrance., On the other
hand, the experimental study may be a performance
test in which an effort is made to establish pa-
rameters which will describe the bearing charac-
teristics in such a way that they may be employed
to determine design configurations other than the
one tested,

The information to be obtained by using the
second approach is usually presenied as a plot
of fy vs. S, where {, is the bearing friction factor
(dimensionless) and §_ is the Sommerfeld variable
(dimensionless):

uN [ 1 ’
0 TR\
and
,
(2) fo = 1i=)
where
i = absolute viscosity of lubricating fluid,

Ibf-sec/i n.2,
= journal speed, rps,

unit bearing load, psi,

= journal radius, in.,

= radial ¢learance, in.,

= coefficient of bearing friction.

—_— 0 1"’32
i

The significance of these relationships indesign
work will depend on their accuracy, which can be
defined in terms of the accuracy fo be associated
with each of the variables. The fractional pre-
cision of § .will depend upon the fractional pre-
cision of y, P, N, r, and ¢, and it may be shown
that

as \ as \

2 2 o 2

(3) Isn = E P+ N Py
85” ’ asn ‘

st P2 e p?

aP or T

as ¥

" .
r? ,

where P, P+ . represent the probable errors

 

 

 

n
of § , pu ... . Taking the partial derivative: of
§, with respect to g, N, ... and substituting in
Eq. 3 gives ‘

\2 2
SN, s, 2
e [) + - —

S i # N N

2
S ? S
n 2 n 2
+ - Py, + 4 - P;

 

 

 

 

P 2 2 2
s P P P
s, p N P/

 

PERIOD ENDING DECEMBER 10, 1953

As an example, suppose that it is desired to
evaluate §  with a precision of 210%. If it is
assumed, for the first approximation, that each of
the quantities y, N, P, r, and ¢ may be measured
with the some precision, then

]
= N7 (0.1) = 0.0302,

 

 

and P_/x is the fractional precision of y, N, P, r,
or ¢, The variables N, P, and r can be measured
to even greater precision than that required for the
example given above, but it is doubtful whether
i or ¢ can be measured to even this fractional
precision, In this case, it may be assumed that
N, P, and r are measured without error, and Pp/,u
and P_/c are then found to have a fractional pre-
cision of :

5 (C'l) :'EII: -

 

X

It is possible to measure ¢ with this degree of
precision, but p is likely to involve a much larger
error; thus, if it were necessary to evaluate S, to
+10%, the measuring technique for p would have
to be much refined, '

Equation 5 may be used to indicate the resulting
precision of S for the case in which the probable
errors in y, N, P, r, and ¢ are known, As on ex-
ample, a journal bearing in a fluoride mixture under
the following typical operating conditions will be
considered: speed, 1800 rpm; unit load, 20 psi;
journal radius, 1 in.; fluoride temperature, 1200°F;
radial clearance between bearing and sleeve,
0.0015 inch. At 1200°F, the viscosity of the fluo-
ride mixture NaF-ZrfF -UF, is reported to be
11.8 cp plus or minus approximately 30%, or

p=11.8x 145 x 1077 = 1.72 » 107¢(lb;~sec/in.?)

+ 30%. The expected fractional precisions for the
variables are then:
Px/x (Px/x)z

I 0.3 0.09

N 0.0025 0.0000063

P 0.025 0.00063

r 0.001 0.000001

c 0.067 0.00449

27
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

and
PS
= ’\/(0.09) + (0.0000063) + (0.00063) + (0.000004) + (0.01796)
7
= 40,1086 = 0.3296 or 32.96%
Gnd o

 

2
S 1.72 x 107% x 30 ]
no 2 0.0015

= 1,146 £ 33%

If /, (bearing friction factor) is considered in a
similar manner,

 

 

For the assumed operating conditions stated

 

 

 

7,2 above, the estimated values of the guantities
f, = ; measured for evaluating f, are: 7, 1in.; R = 6 in.;
RWe W, 180 Ib; ¢, 0.0015 in.; ¥, 0.5 Ib. The expected
where frqcfiona{precisions for the variables are then
r = radius of journal, in., P /x (Px/x)2
= frictional force of torque arm length, R, Ibf, F 0.20 0.04
R = torque arm length, in., R 0.10 0.01
W = total bearing load, Ib, , 0:00] 0.000001
c = rc:diql 'cieamnce between bearing and W 0.1 0.01
sleeve, in. 0.067 0.0049
The expression for obtaining the fractional pre- ‘
cision can then be written as and
P
/e — -
""f*— = \[(0.04) 4+ (0.000004) + (0.01) + (0.01) + (0.0049) = V0.0649 = 0,255 or 25.5%
b
and T e
and the derivations represent ideal conditions,

0.5 x 12

o = Y80 % 6 » 0.0015

 

0.308 + 25.5% .

Thus it appears that the precision of the ex-
perimental effort on bearing characteristics is
limited by the precision of the data available for
the viscosity of the fluoride and the ability to
accurately measure the radial clearance and fric-
tional force. If consideration is given to the fact
that the Ap/AT for the fluoride is quite large ond
that the lubricant film temperature is extremely
difficult to accurately measure, the precision of
the experimental results will be even lower. Be-
couse errors of this type are usually cumulative

28

reality may result in errors in the relationships as
high as 75 or 100%, which would indicate that
refinements in measuring all critical parameters
may be needed.

HEAT EXCHANGER TEST

R. E. MacPherson H. J. Stumpf
ANP Division
A heat exchanger some design
features proposed for the reflector-moderated re-
actor is currently being tested. This exchanger,
which is enclosed in a 5-in. tube furnace for pre-
heating, has the molten fluoride mixture NaF-ZrF
(53-47 mole %) circulating in the she!! ond high-

incorporating
SURGE TANK

    

EM PUMP

PERIOD ENDING DECEMBER 10, 1953

UNCLASSIFIED
DWG. 22339

 

 

Fig. 2.2. Heat Exchanger Test Apparatus.

temperature sodium circulating through the tube
bundies (Fig. 2.2}. A sump-type pump continuously
circulates the fluoride mixture; the sodium flow
is maintained by an electromagnetic pump. The
sodium is resistance heated, and it enters the
tubes in one half the exchanger to heat the fluo-
rides. It is then cooled in a sodium-to-air radiator
{100 kw, interrupted-fin type) from which it enters
the tubes in the other half of the heat exchanger
to cool the fluorides and then returns to the sodium
heating coil. Thus both the fluoride mixture and
the sodium are alternately cooled and heated. The
heat exchanger is designed to transfer 1.5 mega-
wafts, but performance to date has been limited
by an inability to supply sufficient power to the
resistance heater for the sodium. The fluoride
system has been in operation at above 1200°F for

over 600 hr, and the completed bifluid system has
been operating at 1200 to 1400°F for 300 hours.

Performance data are not yet available.
FORCED.CIRCULATIOM CORROSION LOOP

D. F. Salmon
ANP Division

A smail forced-circulation corrosion loop was
operated with NaF~ZrF4-UF4 to extend the cor-
rosion data being obtained from low-velocity thermal
convection loops. The aim is to eventually obtain
a fluid velocity of 5 fps or greoter with o temper-
ature difference of 200°F. Circulation is obtained
with an Eastern Industries, E-100, centrifugal
pump that was redesigned for sump-type operation
and fabricated of lInconel. . A gas seal is main-
tained in a water-cooled, Teflon-packed gland on

29
ANP QUARTERLY PROGRESS REPORT

the shaft. The loop was fabricated of 0.1-in.-OD,
0.025-in.-wall Inconel tubing., Heoting is accom-
plished by passing an electric current through 17 ft
of the tubing, and cooling is by natural convection
to air from 9 ft of tubing. The tubing is coiled to
reduce the space requirement.

The fluid velocities obtained in this loop are
a factor of 50 or 60 greater than those of thermal
convection loops, but the Reynolds modulus re-
mains in the lominar region at approximately 1700.
The loop operated approximately 200 hr with o

gradual slowing down of flow, which indicated
possible plugging. A summary of the results is
presented graphically in Fig. 2.3, which shows
velocity and temperature difference as a function
of elapsed time. The speed of the pump, which
was checked periodically, did not vary appreciably.
The velocity of the fluoride mixture was calcu-
lated, rather than measured, from heat balance data
on the loop heating section,
nated becouse of the indicated plugging ond is
being examined metallurgically. A second loop is

The loop was termi-

being construcied.

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DWG. 22340
3 —
| | MAXIMUM TEMPERATURE, 1500°F
O] |
T WM’O\\ VELOCITY | 420
! o1 ‘-0."\,"“0\
2 | e O . 400
o
o —— 380
E
(&)
8 ¢
oo 360 —
bes
1 _________ i / ] q
A — gl o
E— AT S \ AT ——1 340
\/
//‘MMM“A
;#;//A —1 320
0 | L 300
0 12 24 36 48 60 72 84 96 108 120 132 144

OPERATING TIME (hr)

Fig. 2.3. Velocity and Temperature Difference as a Function of Operating Time for NaF-ZrFA-UFd in an

Incone! Forced-Circulation Corrosion Loop.

30
PERICD ENDING DECEMBER 10, 1953

3. REFLECTOR.MODERATED REACTOR DESIGN STUDIES

R. W. Bussard

A. P. ; Fraas

ANP Division

In May 1953, the U, 5. Air Force requested that
a series of design studies be prepared to outline
the characteristics of 100-, 300-, oand 600-megowatt
circulating-fuel reactors. Unfortunctely, the re-
actor physics work essential for establishing the
over-all design limitations has lagged far behind
schedule, and mony materials questions are not
yet solved. However, even though detailed designs
cannot be prepared at this time, it has been recog-
nized that design information is urgently needed,
and therefore the preliminary layouts presented in
this report were prepared to give the best estimate
of preliminary designs that can be made now.
While interest was expressed in reactor power
outputs of as much as 600 megawatts, a careful
examination of the basic {imiting factors indicated
that it would not be prudent to extrapolate the
information currently ovailable beyond a reactor
size of 300 megawatts. The design for the 600-
megawatt reactor was therefore dropped, ond studies
for 50- and 200-megowatt reactors, in addition to
the 100- ond 300-megawatt reactor designs re-
quested, were prepared. |

The more important parameters for the deSigns
are presented in Tables 3.1 and 3.2. Layouts for
50-, 100-, and 300-megawatt reactors are presented
in Figs. 3.1, 3.2, and 3.3, respectively. Detail
drawings for the main heot exchanger, pumps, and
core shells have been prepared for the 50-megawatt
reactor. The designs for the 50- and 100-megawatt
reactors appear to be quite conservative so that it
seems probable that their ratings could be increased
to 75 and 150 megawatts, respectively., Parameters
such as flow rates and temperature drops would, of
course, increase accordingly.

An explanation and qualification of the more
important features of the principal components and
the considerations that determined them has been
prepared to supplement the drawings. The static
physics, moderator cooling, pumps, hydrodynamics,
heat exchangers, pressure shell, control, and fuel
addition and drainage systems are each dlscussed,
in turn. An effort has been made to include in
these discussions an indication of the status of
the design ot this time aond of the experimental
work planned to provide badly needed design data.

A famthrxtty with information published previously
is presumed. 172

REACTOR PHYSICS

A wide variety of reactor types and geometries
has been considered for circulating fluoride fuels.
Of these, the most promising ot this time appears to
be a reflector-moderated reactor with a central
island. Sodium-cooled beryllium seems to be the
best choice as the moderating material. However,
the picture is very incomplete becouse calculations
for only about 25 reflector-moderated reactors had
been completed until recently because of lack of
computing facilities. Calculations for approxi-
mately 100 two-region reactors have just been
completed, and the results ore being analyzed;
calculations for about 200 three- to nine-region
reactors have been started.. By April 1954, this
work should be far enough along to give a more
complete picture of the effects of reactor geometry
and materiols of construction on parameters such
as critical mass and power distribution for every
type of circulating-fuel reactor that has appeared
promising. A particular effort will be made ta find
an arrangement that will give the most rugged,
reliable, and simple reactor construction consistent
with limitations imposed by corrosion and gommo-
heating effects.

Five critical assemblies have been run to date
on reflector-moderated reactors. All these have
been rather rough approximations to what currently
appears to be a promisingfull-scale aircraft reactor,
The wedkest element in all of these approximations
has been the fuel region composition. In the first
critical assembly, the fuel region was simulated
with cans’of sodium interspersed with disks of
metallic urenium, while in the subsequent assem-
blies, cans of powdered NaF-ZrF, -UQ, were
employed. - The first assembly suffered from severe
inhomogeneity and self-shielding of the fuel, and
the subsequent assemblies hod the shortcoming

 

A P, Froas, C. B. Mills, and A, D. Callihan, ANP
Quar. Prog. Rep. March 10, 1953, ORNL-1515, p. 41,

25, P. Fraas, ANP Quar, Prog. Rep. Sept. 10, 1953,
ORNL-1609, p. 31.

31
ANP QUARTERLY PROGRESS REPORT

TABLE 3.1. PRINCIPAL DIMENSIONS OF A SERIES OF REFLECTOR-MODERATED
CIRCULATING-FUEL REACTORS

Power, megawatts 50 109 200
Core diamater, in. 18 18 20
Power density in fuel, kw/cm> 1.35 2.7 3.9
Pressure she!l outside diemeter, in. 48.5 50.6 56.4
Fuel System
Fuel volume in core, 3 1.3 1.3 1.8
Core inlet outside diameter, in. 10 10 11
Core inlet inside diameter, in. 7 7 7.7
Core inlet araa, in.2 40 40 49
Fuel volume in inlet and outlet ducts, £13 0.4 0.4 0.5
FFusl volume in heat exchanger, £ 1.25 2.5 5
Fuel volume in pump and plenum, fr3 0.3 0.3 0.5
Total fuel volume cireulating, £43 3.25 4.5 7.8
Fuel expansion tank volume, f13 (8% of system volume) 0.26 0.36 0.62

Fuel Pumps

Fuel pump impeller diameter, in. 5.75 7 8.5
Fuel pump impeller inlet diameter, in. 3.5 4,5 5.5
Fuel pump impeller discharge height, in. 1.1 1.5 1.8
Fuel pump shaft center line to center line spacing, in. 20 21 22.5
Plenum chamber width, in. 14.5 15 15.5
Plenum and volute chamber fength, in. 30 31 33
Plenum and volute chamber height, in. 2.0 2.0 2.4
Impeller rpm 2700 2700 2500
Estimated impeller weight, b 8 12 17
Impeller shoft diameter, in. 1.5 1.75 2
Impeller overhang, in. 12 13 14
Critical speed, rpm 6000 6000 5200
Sodium Pump

Na pump impeller diometer, in. 3.4 4,1 5.0
Na pump impeller inlet diameter, in, 2.4 2.9 3.5
Na pump impeller discharge height, in. 0.75 0.9 1.1
Na expansion tank volume, f13 (10% of system volume) 0.08 6.09

Na in Be passages, 3 0.43 0.47

Na in pressure shell, £ 0.15 0.15

Na in pump and heat exchanger, £43 0.20 0.25

Fuel-to-NaK Heat Exchanger

Heat exchanger thickness, in. 1.7 2.75 4,65
Heat exchanger inside diameter, in. 42 42 44
Heat exchanger outside diameter, in, 45,4 47.5 53.3
Heat exchanger volume, £ 6 10 20
Angle between tubes and equatorial plane, deg 27 27 27
Number of tubes 2304 3744 6600
Tube diameter, in. 0.1875 0.1875 0.1875
Tube spacing, in. 0.2097 0.2097 0.2119
Number of tube bundles 12 12 12
Tuke arrangement in each bundle 8x 24 13 x 24 22x 25

32

300
23
3.9
62.0

2.7
12.8

67
0.7
7.5
1.0

11.9
0.95

10
6.75
3.2
27
17.5
37
3.0
2300
24
2.25
15
5000

5.9
4,2
1.2

5.9

47

58.8

30

27
9072
0.1875
0.2094
12

28 x 27
PERIOD ENDING DECEMBER 10, 1953

TABLE 3.1. (continued)

Moderator Region
Volume of Be plus fuel, fr3
Velume of Be only, £43
Mo. of coolant holes in reflector
No. of coolant holes in island
Na coolant tube inside diameter, in,
Na coolant tube wall thickness, in.

Mo pressure shell annulus thickness, in.

that the atomic density in the fuel region was only
about 33% of that of liquid fluoride fuel., These
shortcomings will be largely eliminated in a new
critical experiment that is being planned. Sheet
Teflon and 0.002-in.-thick uranium foil will be
stacked in laminge to simulate the fluoride fuel,
both from the standpoint of atomic density and of
neutron-scattéring and absorption cross sections.

Another wedk point in the critical assemblies
that have been run has been that the geometric
approximation left much to be desired. It is planned
that the sheet Teflon and the wranium foil will be
cut fo appropriate lengths to permit the construction
of a configuration essentially the some as that
envisioned for the full-scale reactor.

FUEL COMPOSITION AND PROPERTIES

It was assumed in the design work that the
physical properties of the fluoride fuel at 1500°F
would be similar to those of Naf-KF-LiF-UF,
(10.9-43,5-44.5-1.1 mole %), which are, opproxi-
mately, the following: heat capacity, 0.45 Btu/1b°F;
thermal conductivity, 2.6 Btu/hrft:°F; viscosity,
3 cp; density 2.0 g/cm®. Heat transfer considera
tions show thot such a fuel would be much better
than the sodium-zirconium fuel that has been pre-
pored for the ARE. If such a fuel were not de-
veloped, it would be necessary to increase the
heat exchanger size given in the accompanying
designs by a foctor of 2 or 3, This would mean an
increase in shield weight of about 6% and o serious
increase in heat exchanger header complexity,

There are two possibie ways of obtaining a
satisfactory fuel. The first is to use the above-
mentioned fluoride, or some modification of it, in a
container fabricated of an alloy that has no chro-
miuvm. This should eliminate the serious short-
coming of this particular fluoride mixture, that is,
its tendency to form low-conductivity films of a
highly insoluble potassium-chromium-flyoride mix-

22,4 22.4 25.8 31,5
21.1 21,1 24.0 28.8
208 208 554 554
86 86 210 210
0.155 0.187 0.187 0.218
0.016 0,016 0.016 0.016
0.125 0.125 0.187 0,200

ture on surfaces in Inconel systems. Initial tests
show Hastelloy B (80% Ni—20% Mo) to be a possi-
ble container material for this fluoride, and it will
be tested in the coming months,

A second way to obtain a fuel with satisfactory
physical properties is to test various melts that
have compositions that seem promising, and such
o program is now under way, Melts of seven dif-
ferent compositions containing lithium, sodium,
rubidium, and beryllium are being prepared, and
their physical properties are being determined.
The best of these melts will be selected for cor-
rosion test work,

MODERATOR REGIONS

Heat tronsfer analyses indicate that a sodium-
cooled beryllium moderator. is superior from the
heat transfer standpoint to any other arrangement
considered thus far.? Static and seesaw Inconel
capsule tests show virtually no signs of corrosion
or mass transfer at 1500°F for beryllium specimens
immersed in sodium. However, it is felt that mass
transfer in a larger-scale test loop operating with
a substantial temperature differential and with the
beryllium presenting the hottest surface in the
system {(as in a full-scale reactor) is very likely
to be serious. A test loop to investigate this
possibility has been designed, and it is scheduled
to operate about February 1954.
transfer prove to be serious, canned beryllium will
be tested. .

It is expected that the results of the multigroup
reactor calculations will indicate what further
moderator designs should be prepared and analyzed
from the heat transfer standpoint to investigate
the possibilities of exploiting lead and bismuth
as moderator coclants. lead and bismuth are

Should mass

 

3R. W. Bussard et al., The Moderator Cooling System
for the Reflector-Moderated Reactor, ORNL.-1517 (to be
published).

33
ANP QUARTERLY PROGRESS REPORT

interesting because of their high densities. Some
metallurgical work is currently under woy in on
attempt to find a system that is satisfactory with
respect to corrosion and mass transfer.

Some structural problems exist in designing the
suppoit of the reflector and the island. At first
glance, it would appear that thess assemblies

hang from the flat decks that span the pump region.
This would not be the case when the fluoride
system had been filled, however, because the be-
ryllium density (1.84 g/cm®) would be somewhat
less than the expected fuel density (about 2.0
a/cm®) and hence the moderator region would tend
to float in the fuel. With the moderator region

TABLE 3.2. HEAT TRANSFER SYSTEM CHARACTERISTICS FOR A SERIES OF
REFLECTOR-MODERATED CIRCULATING-FUEL REACTORS

REACTUR POWER, megawotts

Fuelsto-Nak Heat Exchanger and Related Systems

Fuel temperature drop, °F

NoK temperature rise, °F

Fuel AP in heat exchanger, psi

NaK AP in heat exchanger, psi

Fue! flew rate, |b/sec

NaK flow rate, tb/sec

Fuel flow rate, cfs

NaK flow rate, cfs

Fuel velocity in heat exchanger, fps

Fuel flow Reynolds number in heat exchanger
NaK velocity in heat axchanger, fps

Over-all hect transfer coefficient, Btu/hr 2O F

Fue!-NaK temperoture difference, °F

Sndiumsto-Nak Heot Exchanger and Relatad Systems

Na temperature drop in heat exchanger, °F
NaK temperature rise in heat exchanger, °F
Na AP in heat exchanger, psi

NaK AP in heat exchanger; psi

Power generoted in islond, kw

Power generated in reflectar, kw

Power generated in pressure shell, kw

Na flow rate in reflector, lb/sec

Na flow rate in island and pressure shell, lb/sec
Total Na flow rate, |b/sec

Ma temperature rise in pressure shell, °F
Na AP in pressure shell, psi

Na temperature rise in island, °F

Na AP in island, psi

Na temperature rise in reflector, °F

Ma AP in reflector, psi

Me-Mak temperature difference, °F

Shield Cooling System

Power ganerated in 6-in. lead layer, kw

Power generated in 24-in. HZO layer, low

34

50 100 200 300
400 400 400 400
400 400 400 400
35 51 61 75
40 58 69 85
263 527 1,053 1,580
474 948 1,896 2,844
2.1 4,2 8.4 12.6
10.5 21.0 42.0 63.0
8.0 9.9 1.2 12,2
4,600 5,700 6,700 7,000
36.4 44.9 50.9 55.4
3,150 3,500 3,700 3,850
95 100 110 115
160 150 150 150
100 150 150 150
7 7 7 7
7 7 7 7
500 1,000 2,000 3,000
1,700 3,400 7,500 11,200
190 350 500 620
53 72 154 234
22 28 51 76
75 100 205 310
28 39 30 26
4 6 6 6
72 ni 120 124
32 21 12 12
100 150 150 15C
36 27 18 18

A3

110 210 300 350
<3 <6 <12 <18
 

 

 

      
  

 

 

 

   

 

 

 

 

 

 

   
  

    
   
   

 

_ g ‘ R U
o ”\ g i

 
   

 

 

 

  

   

v %
A
£ Fog

  

 

'y / !
f/) Na~TO-NoK
i/ { HEAT EXCHANGER -

/ }
il HEADER FOR Na TO /
REFLECTOR

. ~ REFLECTOR Neo

  

 

 

Be ISLAND (%

OUTER CORE SHELL -1

 

 

INNER CORE SHELL -7

 

 

\\ QUTLET HEADER
s_\.

 

 

 

Fig. 3.1. 50-Meguwati Reflector-Moderated Reactor.

 

%A

i AT

PERIOD ENDING DECEMBER 10, 1953

   

p

L a” W A

 

 

 

 

g

"

e — FOR PUMP ASSEMEBLY
SEE FIG. 3.4

 
   
   
  
 
 
 
    
   
 
 
   
    
   
 
 
   
   
 
   
  
  
      
  
 

 

—-8'% LavER
-~ INCONEL JACKET

RN Be REFLECTOR

----—-MODERATOR
- COOLING TuBE

A~ COOLING TURE
1 CONNECTOR

--—TUBE BUNDLE
SPACERS

LAYER
10
8™ LAYER

  

- PRESSURE SHELL
LINER

35
ANP QUARTERLY PROGRESS REPORY

   

 

  
 
  
   
         
 
  
 
   

 

 

 

 

 
  

"/ / HEADER FCR Na TO

 

QUTER CORE SHELL
‘ :
\

 
 

' REFLECTOR Na |,
OUTLET HEADER-

 
  

 

I oM T E R w 3

 

IT S
o £ lgé‘
S

=
o

 

 

/7 REFLECTOR - ]

* . INNER CORE SHELL ~ l»fi ‘
I

 

 

 

 

 

 

 

 

 

 

 

 

  

 

A
DWG. 22342

— FOR PUMP 25SEMBLY

¥~ SEE FIG 3.4

 

 

 

= |--UFPER PUMP DECK

_-LOWER PUMP DECK

LA .- - B'C LAYER

Al
AN

W A~ INCONEL JACKET
\

 

- COOLING TUBE
CONNECGTER

/T JACKET £OR B0
; LAYER

g0 | AYER

T PRESSURE SHELL
LINER

T Na PASSAGE

-

Fig. 3.2, 100-Megawait Reflector-Moderated Reactor.

36
PERIOD ENDING DECEMBER 10, 1953

LWG 22343

  
   

T FOR PUMP ASSEMALY
SEE FIG. 3.4

/U PPER PUMP DECK

~LOWER PUMP DECK

~REFLECTOR SHELL

Na-TO- NoK 0
B LAYER

HEAT EXCHANGER™

  
  

     

TUBE.

    
 

-
v

SINCONEL JACKET

     
 
 
 
     
 
  

 

HEADER FOR Na
TO REFLECTOR

 
 
 
 
     

.-Be REFLECTOR

 

CONTROL )
ROD PASSAGE - TUBE BUNDLE

SPACERS

MODERATGR COOLING

TUBE CONNECTOR

 
  
  

Be ISLAND-— TUBE BUNDOLE

  

INNER ~JACKET FOR

 
 
 

CORE SHELL~— i
B"Y LAYER
OUTER - u'0 L avER

      
   

WA\ CORE SHELL-.
\ T PRESSURE SHELL
LINER

“Na PASSAGE

 
 
 

REFLECTOR Na
OUTLET HEADER- - PRESSURE SHELL

NoK OUTLET

Fig. 3.3. 300-Megawatt Reflector-Moderated Reactor,

37
ANF QUARTERLY PROGRESS REPORT

buoyed up by the fuel in on almost ideal ‘‘shock
mounting,”’ occelerations that would occur under
crash londing conditions (in which vertical ac-
celerations of as much as 10 g should be antici-
pated) would yield a net upward force on the
reflector assembly that would be about equal to
its weight, In the designs illustrated, the re-
flector assemblies are attached to the lower deck
of the pump region by welds to the walls of the
sodium return passages.

The sodium pump und heat exchanger subos-
semblies are positioned around the top edge of
the upper pump deck. The pipe from the sodium
pump discharge makes a slip fit into the reflector
sodium inlet tube. The leakage through this slip
fit into the sodium return passage simply re-
circulates with no penalty other than o small
increase in the required pump capacity. The two
sodium pump and heat exchanger assemblies could
be welded in place by passing a bead around the
lower edge at the outer periphery and up and
across the top where they adjoin the plenum
chamber between the two fuel pumps.

The reactor designs illustrated in Figs. 3.1, 3.2,
and 3.3 presume that canning of the beryllium will
be required but that trace leaks in the Inconel-
to-can connections can be tolerated. Therefore
the tubes passing through the rifle-drilled holes
in the reflector are designed to be driven into
tapered bores in the fittings shown at the equator,
and the outer ends of these same tubes are to be
rolled into their respective header sheets at the
top and botiom. The tube connecting fittings at
the egquator also serve as dowels to locate the
two beryllium hemispheres relative to each other.

HYDRODYNAMICS OF THE FUEL CIRCUIT

Hydrodynamic considerations impose a number
of resiraints on the geometry of the fuel passages
in the reactor fuel circuit. These restraints are
not sharply defined, but they are sufficiently rigid
that they limit the fuel circuit proportions for a
well-balanced design to a considerably narrower
range than would at first be expected. The pres-
sure drop through the hecot exchanger goes up very
rapidly with fuel flow rate —~ so much so that there
is little to be gained by going to fuel velocities
which correspond to pressure drops in the heat
exchanger of greater than 80 psi. This is particu-
larly true if on effort is made to keep the stresses
in the core shells and heat exchanger tube walls

38

to 500 psi or less, If the rest of the fuel system is
examined carefully, it is evident that it should
involve a pressure drop very much less than that
ocross the heat exchanger in a well-balanced
design. Thus, since o bleed-off-air turbine-pump-
drive arrangement seems to be at once the lightest,
simplest, and most flexible and since such a pump
drive system should be designed to require less
than 3% air bleed-off from the compressors, there
seems to be o strong incentive to limit the pressure
drop through the pump inlet and discharge passages
and the core inlet end outlet ducts to not more than
15 psi.

If it is assumed that the reactor core diameter is
determined by a compromise between shielding and
other considerations and that the islond diometer
will be determined by fission density distribution
considerations, it becomes evident that these two
basic dimensions for the fuel circuit moy be token
as o base for determining the proportions for the
rest of the system. The inlet and outlet duct
passage arrangements should be as small as possi-
ble to simplify the shielding problem and to reduce
shield weight. However, it is of paramount im-
portance that boundary-layer thickening and flow
separation be avoided in the diffuser region at
the inlet to the core. This, in turn, imposes a
lower limit on the flow passage area that may be
used ot the core inlet. An asymmetric design
might be employed in that this flow limitation
would not apply to the outlet duct, particularly
since a diffuser giving a reasonably high effi-
ciency can probably be provided between that duct
and the inlet to the heat exchanger. It is quite
evident from this stondpoint alone that further
shielding tests should be made to investigate the
effect of fuel duct size on over-all shield weight.
While the problem is not clearly defined, it does
appear from flow test work, as well as from analy-
ses of the adverse pressure gradients along the
diffuser walls in the core inlet region, that the
outer diameter of the core inlet duct should be
roughly 60% of the core diameter, while the inner
diameter of that duct should be between 60 and
70% of the outer diameter,

The plenum chamber between the volute dis-
charges and the core inlet ducts should provide for
smooth fransition and uniform velocity disiribution
at the inlet to the core. One promising way of
accomplishing this would be to provide a plenum
chamber having a height somewhat greater than the
thickness of the core inlet duct and a diameter at
least half again as large as that of the inlet duct.
The pump volutes could then be designed to give
a fluid velocity entering the plenum chamber about
equal to that in the core inlet duct, This would
mean that for high power density reactors, for
example, a 200-megawatt reactor with an 18-in.-dia
core, the di scharge annulus of a mixed flow impeller
should be made as wide as possible so that it
would approach the height of the plenum chamber.
For low power density reactors, for example, a
50-megawatt reactor with an 18-in.-dia core, the
disparity in these dimensions is inherently large.
A careful examination of the magnitude of the
eddy losses to be expected of the pump inlet,
through the impeller, in the pump volute, and in
the plenum chamber leads to the conclusion that
the pump should be designed for on inlet velocity
of not much more than about 30 fps, that is, «a
velocity giving a dynamic head of about 12 psi.

On the basis of these considerations, it is
possible to specify dimensions for the principal
fuel flow possages for a wide range of reactor
power outputs. While the dimensions are given in
Table 3.2, it should be noted that there is some
latitude in the selection of these dimensions.
Furthermore, it is clear that passages of larger
size would require substantially heavier shields,
while passages of substantially smaller size
would be likely to give some sort of difficulty from
the hydrodynamic standpoint.

Some preliminary flow tests were carried out a
year ago,‘s- and the results were encouraging.
Additional flow tests ore being made to determine
whether the proportions indicated in Table 3.2 lead
to a satisfactory configuration for a reactor de-
signed for an output of 50 to 100 megawatts.

PUMP DESIGN

In examining the over-all system requirements,
it becomes evident that it would be very desirable
if the fuel pumps could perform several functions
in addition to that of pumping the fuel. It seems
likely that most of the xenon and possibly some of
the other fission-product poisons might be removed
from the fluoride mixture if helium could be bubbled
through the fuel while it was being thoroughly
agitated. Furthermore, it would be advantageous

 

4R. E. Ball, Investigation of the Fluid Flow Pattern
in a Model of the *Fireball"’ Reactor, ORNL Y-F15-11
(Sept. 4, 1952). .

'

PERIOD ENDING DECEMBER 10, 1953

to have the pumps serve as mixing chambers in
which high-uranium-content fuel could be added 1o
the main fuel stream to enrich the mixture and
compensate for burnup. Finally, it is essentidl
that on expansion tank be provided sc that en-
trainment. of bubbles would not prove fo be a
problem in maneuvers or in “bumpy’’ flight.

The pump arrangement shown in Fig. 3.4 waos
worked out to provide for all the points mentioned
above. The pump design embodies o varigtion of
an idea tested three years ago.® The back of the
impeller is used to centrifuge out the gas bubbles.
The design also uses the overhung impeller, the
bearing arrangement, and the face-type gas seal of
the previous pump, along with several features that
have proved to be successful in sump pump test
work during the past year, The principles of opera-
tion are quite simple and have been demonstrated
with a plastic model. The fuel in the expansion
tank above the impeller tends to swirl in the same
direction as the impeller (but at a lower speed) so
that it forms a vortex with o roughly conical free
surface. A relatively large amount of Fluid leaks
from the high-pressure region at the impeller outlet
through the labyrinth seal ot the periphery of the
impeller into the swirl chamber immediately above
the impeller. Since this swirl chamber is connected
through a bypass line to the impeller iniet, the
leakage is simply recirculated at no cost other
than an increase in pumping power. Some fluid
lecks upward through the clearance between the
top lip of the swirl chamber and the outside of
the centrifuge cup on the back face of the impeller,
but this leckage rate is small because there is
little pressure differential across this gap. The
leakage that does occur is made up by bleed flow
from small holes in the periphery of the centrifuge
cup on the back of the impeller. Thus there is a
small, continuous flow from the fluid vortex in
the expansion tank into the ceatrifuge cup and
thence into the swirl chamber, The fluid region
close to the surface of the vortex in the expansion
tank is highly turbulent ond full of bubbles so
that it should be effective in degassing the fuel,

In initial tests of o water pump with the above-
mentioned features, the pump continued to pump
when it was inverted for as much as a minute
before the liquid supply in the centrifuge cup and

 

Pump Reworked to Test Special Features for Operation
with Liquid Metals, ORNL Y-F15-6 (Feb, 1, 1951},

39
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIED
DWG E-16919A

   
  
 
 
 
 
   
 
 
 
 
  
  
 
 
 
   
      
     
 
  
 
  

 

SECTION A-A
PUMP HOUSING -—— 4

11— - OlL SUPPLY PASSAGE

 

L3 -~ Ol BATH

 

 

LOWER BEARING ANC SEAL SUPPORT‘n\i s

 

 

 

IMPELLER SHAFT LOWER BEARING -~ LA

 

 

 

 

 

OiL BATH-—----

FACE-TYPE GAS SEAL.- ——-¥§ A= - PUMP WELL CASING

 

 

 

 

 

 

 

 

 

 

 

 

4™ COQLING CIL PASSAGE

 

——NATURAL URANIUM

 

 

 

 

 

 

LA

o

 

 

T ¥ TUBE ——-ri
FUEL PELLET SUPPL , “CSLINGER

 
   

Ik EXPANSION TANK

 

 

CENTRIFUGE CUP

 

LABYRINTH SEAle

 

 

  

 

SWIRL CHAMBER - —m [#

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

=7 PUMP INL

 

 

 

 

 

Fig. 3.4, Gas-Sealed Aircraft Pump.

40
swirl chamber was depleted by fluid ledkage from
the swirl chamber inte the expansion tank. The
slinger ring on the impeller shaft at the top of the
expansion tank acts to throw the fiuid away from
the shaft when the pump is inverted and thus pre-
vent flow into the clearance around the shaft
between the slinger ond the face-type gas seal.

A number of other special features have been
included in the pump design to adapt it to the full-
scale reactor shield, The pump has been designed
so that it con be removed or installed as a sub-
assembly with the impeller, shaft, seals, and
bearings in a single compact unit. As can be seen
from Figs. 3.1 and 3.4, this assembly fits into the
bore of a cylindrical casing welded to the top of
the reactor pressure shell, A 3-in. fayer of uranium
just above a'l’z-in. layer of B'? around the lower
part of the impeller shaft is at the same level as
the reactor gomma shield just outside the pressure
shell. The steel seal and bearing support above
the loyer of uranium gives additional gomma shield-
ing, while the oil-filled cavity between the bearings
acts as neutron shielding. An air-driven turbine
and geor box could be mounted on the flange at
the upper end of the pump casing. Since the outer
surface of this casing was designed to be about
8 in. inside the outer surface of the shield, there
need be no protuberances from the shield surface
in the final installation. The long, slender,
“vill supply’’ tube for fuel enrichment is suffi-
ciently smail thot it should not permit much gamma
edkage.

The large amount of impeller shaft overhang
inherent in the design ofthe pump could be avoided
if a journai bearing could be operated with the
molten fluoride mixture acting as the lubricant.
The pump impeller and volute could be designed
so that the radial load on such a bearing would
be very small, perhaps not more than 25 psi. Since
the viscosity of the fluoride mixture is greater than
that of water, the arrangement should be completely
satisfactory after the pump was storted. Pickup
and scuffing of the bearing or shaft during startup
or shutdown might prove to be problems. A test
rig for investigating the compatibility of different
materials in sodium at high temperature has been
built and is being used to evaluate the compati-
bility of various material combinations in fluorides
(cf., sec. 2). Preliminary results are encouraging.

PERIOD ENDING DECEMBER 10, 1953

PRESSURE SHELL

As much as 1/‘2% of the reactor output may go into
gamma heating of the pressure shell. The provi-
sion for the removal of this heat presents several
design choices, none of which is entirely satis-
factory. If no provision is made for cooling the
pressure shell, it will be much hotter than the
fluoride mixture that contacts its inner surface.
Unfortunately, this would give the highest pressure
shell temperatures in precisely the zone in which
the stresses are the highest, that is, in the ligo-
ments between the NaK ducts in the lower header
region. The pressure shell might be cooled with
the coolant system that serves the lead gamma
shield, but this would make temperature control
difficult over a wide range of power outputs because
of the inherently large temperature differential.
Furthermore, it would make necessary the dumping
of large amounts of heat at a low temperature level
through the shield coolantradiator — a costly matter
in an airplane. Cooling of the pressure shell with
sodium from the island moderator cooling system
has the disadvantage of complicating the pressure
shell construction, but it has many advantages.
It provides a simple way of returning the sodium
that has passed through the island cocling passages
to the sodium-to-NaK heat exchanger. It not only
serves to cool the pressure shell in the critically
stressed fuel discharge end of the reactor so that
the creep strength of the Inconel there becomes
mote than odequate, but it makes it possible to
maintain the entire pressure shell at an essentially
vniform temperature. This should markedly reduce
both thermal distortion and clearance buildups
resulting from differential thermal expansion be-
tween the pressure shell, heat exchanger tube
bundles, and the reflector.

The consfruction envisioned is shown in Figs.
3.1, 3.2, and 3.3. The inner liner assembly con-
sists of a %-in.-thick Inconel shell, a %-in. hot-
pressed B'? layer, and an immer 0.062-in.-thick
Inconel can. Inconel is also used to cover the
interior of the pressure shell between the two end
headers. Closely spaced grooves, 1@ in. deep and
2 in. wide, milled into the outer surface of the
liner provide omple flow possage areo for the
sodium to flow upward. Four transfer tubes carry
the sodium from the top of the pressure shell to the
outer face of the pressure shell ot the sodium-to-
NaK heat ‘exchanger inlets. The hot-pressed

4]
ANP QUARTERLY PROGRESS REPORTY

blocks are diamond-shaped, with 60-deg
angles at their vertexes, and they have rabbeted
edges. This design permits the covering of a
spherical surface with a single block size and
shape,

The construction in the vicinity of the end header
regions was deliberately varied in Figs. 3.1, 3.2,
and 3.3 to show three different ways in which the
detail design may be handled. Another design would
involve the use of a pressure shell with forged
annular protuberances to receive the headers, while
yet another possibility would be a shell sufficiently
thick (obout 2.5 in.) to permit machining annular
grooves for the headers. This latter arrangement
would have the advantage of a more regular ex-
terior surface on the pressure shell.

boron

FUEL-TO-NoK HEAT EXCHAMGER

The heat exchanger envisioned for the reflector-
moderoted reactors is similar to that described
previously.® The major unknowns in the fuel-to-
NaK heat exchanger are concerned with fabrication.
It is presumed that the tube-to-header connections
wouldbe heliarc welded by hand, as in the 210-tube
six-tube-bundle fluoride-to-NaK heat exchanger
presently being tested. It is expected that a large
amount of welding research and component testing
will be required to obtain a satisfactory solution
to this tube-to-header welding problem.

The tube bundles are separated by long cans that
are triangular in cross section and parallel the
tubes. The cans would be filled with hot-pressed
B'0 to assist in inhibiting the low-energy neutron
flux in the heat exchanger region.

REACTOR CONTROLS

One of the salient features of the circulating-fue!
reactor is its inkerent stability. It appears that
no fast-moving control rods will be required to cope
with fast transients. No large amounts of shim
control are required to compensate for burnup, be-
cause a wuranium-rich flueride can be added at
intervals during the course of operation. The only
important control needed is a means of varying the
mean operating temperature. |ln examining over-oll
power plont considerations, it appearad that provi-
sion for effecting a temperature change of 200°F
should be included in the design (that is, about

 

6A. P. Fraas and M. E. LaVerne, Heat Exchanger
Design Charts, ORMNL-1330 (Dec. 7, 1952).

42

3% Ak/R). |t appears from critical experiments
that this could be provided by « 3/-m.-OD tube
filled with B9 and placed to give a Q?-lnn stroke
from the pump deck down through the center of the
island. While the rod ond its actuating mechenism
are not shown in designs illustrated, a 0.50-in.-ID
tube has been provided through the center of the
island to receive the rod. In the design contem-
plated, the rod wculd be machined from a thick-
walled tube to give three equally spaced ]/1 “in.-
high ridges along its entire length. These ridges
would serve tc center the rod in the passage pro-
vided and thus give uniform flow of sodium coolant
around the rod periphery. The ridges would also
engage a grooved guide at the upper end to prevent
rotation. A short length of s/s-in,,-'IO thread ot the
top of the rod would be engaged by an internally
threaded tube enclosed in g casing extending up
through the shield. The tube would be rotated
by a drive at the top end through a bellows szal.
This seal would be brazed to the center of o
bar with three spherical bearings, one at each
end ond one in the center. This bar would be
inclined ot on angle of about 15 deg to the axis
of the control rod drive tube. The center bearing
would be fixed relative to the rod casing, while a
crank attached to the upper end of the control rod
drive tube would be attached to the lower spherical
bearing. As the rod drive tube rotated, the axis of
the drive bor would generate o cone, with the
centers of the spherical beorings at its upper and
The drive actuator
would be an electric motor and gear box connected
to the upper spherical becring. The closed end
of the bellows at the center bearing would describe
a wobble-plate type of motion.

lower ends generating circles.

The fue! enrichment system thot appears most
attractive is based on the use of cold-prassed disks
of No,UF . about ]/ in. in diameter and / in. high.
These dlsks could be fed to the tops of the fue!
addition tubes in the pumps by meens of arelatively
simple “’pill vending'’ machine. Several promising
detail designs of such machines have been pre-
pared from which it appears that this phase of the
development should not prove to be a serious
problem. However, it may be difficult to prevent
the fue! inthe expansion tank from entering, freezing,
and blocking the fuel enrichment tube during in-
advertent inverted flight., The pump design has
been worked out so that this should not happen,
but the effectiveness of this feature of the swirl
expansion tank design has not yet been checked.

SHIELDING

A substantial amount of fundamental work on the
shielding problems of the reflector-moderoted
reactor has been covered in other reports,?+7.8
Three important factors remain undetermined:
namely, the energy spectrumof the short half-fived
fission-product gammas, the effect of variations in
endduct and in fuel pump andheader tank geometry,
andthe possibilities inthe use of unusual materiols
such as lithiumhydride. Tests have been carefully
planned to provide the essential design data neces-
sary for determining the first two of these factors
during the coming months, and a long-range program
is contemplated for investigating unysual materials.
it is hoped that a shielding experiment that is to
be carried out in conjunction with a critical experi-
ment will yield data thot will validate the use of
multigroup, multiregion calculations as an aid 'in
shield analysis. ' '

Analyses of the effects of the neutron-to-gamma
dose ratio and of varying proportions of gamma
shielding on shield weight are being made. The
strength of lead and bismuth alloys at temperatures
of 250 to 350°F, shield design for jet fuels in
place of water, shield heat removal, and the dose
from the shield after shutdown are also subjects
currently being investigated,

FILLING AND DRAINING OF THE REACTOR

After 100 hr or more of full-power operation, the
fuel charge will require roughly 10 in. of lead
shielding to bring the rodiation from it down to a
level of 1 r/hr at 5 feet. This will mean that a
tank to contdin the roughly 4 3 of fuel required
for a 50-megawatt reactor will require about 10 tons

 

7 Report of the 1953 Summer Shielding Session, ORNL.-
1575 {to be published).

8k, H, Abernathy et af., Lid Tank Shielding Tests of
the ReflectorsModerated Reactor, ORNL-1616 (to be
published).

PERIOD ENDING DECEMBER 10, 1953

of lead shielding. This could be mounted on ©
heavily constructed dolly or on an elevator de-
signed to raise it up from a pit under the runway.
A cooling system to remove the afterheat from the
fuel will also be required. Preliminary estimates
indicate that the aofterheat can be removed most
readily by circulating a liquid metal through a
coiled tube in the fluoride drain tank and then
through a radiator to dump the heat to air forced
through the radiator by a blower. Such a system
should be reasonably light, compact, and mobile.

The filling and draining operation could be carried
out by carefully positioning the airplone over the
tank ofter removing a vertical plug, perhaps 5 in.
in diameter, from the bottom of the reactor shield.
The shielded tank assembly with a pipe extending
vertically upward from it could be raised so that
the pipe would project into the hole in the shield.
A coupling at the upper end of the pipe would
engage o corresponding coupling at the bottom of
the reactor pressure shell. The space between the
drain valve at the bottom of the reactor and the
valve ot the top of the pipe from the drain tank
could then be evacuated, the drain tank valve and
thereactor drain valve opened, and the fuel drained.
A blast of helium directed through appropriate
fittings could be used to blow out the fuel droplets
between the valves. The valves could then be
closed, the tank disconnected, lowered, and dropped
into its pit or removed to a suitable location. . The
reverse procedure could be followed in the filling
operation. Helium pressure on the ligquid in the
shielded tank would serve to force the fuel up
into the reactor, |

The biggest problem in the design of a system
of this sort is the detail design of the valves and
the remotely operated coupling to get the high
degree of leak tightness and exceptional reli-
ability required. Several rough preliminary layouts
have been sketched and work is proceeding in an
effort to develop valves and couplings that could
be used for routine filling and draining operations
in current ORNL expetimental laboratory work.

43
4. CRITICAL EXPERIMENTS

A, D. Callihan

D. V. P. Williams

H. Lynn

Physics Division

J. W. Noaks, Pratt and Whitney Aircraft Division

The Critical Experiment Facility at the Oak
Ridge National Laboratory may be used to de-
termine the static physics characteristics of
a wide variety of reactor mockups. During the
past quarter; measurements were completed on a
mockup of an air-cooled, water-moderated reactor
(AC-1) for the General Electric Aircraft Nuclear
Propulsion Project, and preparations were made
for measurements on a mockup of a Pratt and
Whitney supercritical-water reactor, a Nuclear
Development Associates sodium-cooled reactor,
and the ORNL's reflector-mederated reactor (cf.,
sec. 3, '‘The Reflector-Moderated Reactor’’).

The air-cycle reactor (AC.1) mockup was an
approximated cylinder 27 in, in length, 30 in. in
diameter, surrounded by a 7]/2-ft-thick beryllium
reflector. The fuel elements and the uranium dis-

tribution were simulated by properly spaced urenium
disks sandwiched between mild steel which repre-
sented the structural materials of the reactor.
Each fuel element was surrounded by a 1-in.-thick
layer of Plexiglas to simulate the woter moderator.
The originally designed assembly was an octagonal
array of 37 fuel channels. This system, however,
could not be made critical with the 7.5-in.-thick
berylliumreflector until the number of fuel channels
had been increased to 43 ond five channels had
been loaded with twice the uranium prescribed,
The critical loading was then 29.8 kg of U235 with
98.9 kg of steel. Flux and power distributions
have been determined throughout the core and re-
flector. The data will be reported in detail in a
topical report to be issved by ORNL when the
experiment is complete ond are discussed in
General Electric’'s Engineering Progress Reports.
 
 

 

   

 

 

 

 

 

 

 

 

   

 

 

 

-
. P
. - : ! -
; ; »
-
-
.
: ‘ x
i
; ¢
INTRODUCTION AND SUMMARY

The research on high-temperature liquids for
reactors has been primarily concerned with the
determination of phase diagrams of fluoride and
chloride systems with and without fissionable
material, although a substantial effort has been
devoted to the production and purification of
halides and hydroxides (sec. 5). Detailed study
of the NaF-ZrF4—UF system’ has continued be-
cause of its:intendej application in the Aircroft
Reactor Experiment, wherein Na,UF will be added
to Ndszs to produce a suitable fuel. Techniques
such as quenching, petrographic and x-ray analysis,
differential thermal analysis, and high-temperature
filtration, as well as the customary thermal
analyses, have been applied to the binary and
ternary systems NaF-ZrF4 and NaF-ZrF -UF .
Spectrophotometric measurements of fused salts
containing UF . show only two absorption-spectrum
patterns in the NaF-ZrF[UFé system.  Some
physical property data substantiate the concept
that the fused salts are essentially totally ionized.
A number of fluoride samples were purified in the
laboratory for various tests while the facilities for
the routine production of these fluoride mixtures
are being constructed. In an examination of various
phases of fluoride purification techniques, it was
found that thereductionrate of N02UF5 by hydrogen
increases with temperature. The reducing power
of NaH when added to NaZrF:5 is being studied,
as well as the kinetics of the H, reduction of
Nitt in NaZrF_. Chemical studies of hydroxides
were limited to the purification of NaOH ond
5r(OH)., and the determination of the carbonaceous
content of NaOH at 700°C.

The recent corrosion studies have been devoted
almost entirely to the effects of various parameters
on the corrosion of Inconel by fluorides, although
some work with hydroxides and liquid metals was
continued (sec. 6}. Studies of the corrosion of
inconel by the tluoride fuel NQFS-Zer-UFd (50-46-4
mole %) have substantiated earlier conclusions
that the total attack in 500 hr is independent of
temperature from 1500 to 1650°F and that the
attack rate is time dependent. The value of having
purified fluoride melts in clean container systems
has been investigated not only with regard to the
metallic fluoride impurities in the melt, the concen-
tration of which may be directly correlated with
corrosive attack, but also with regard to contami-

nation of the container surface, which has a less
deleterious effect. Pretreatments of the fuel with
chromium and Inconel; as well as the precirculdtion
of fuel, have each been effective in reducing
attack, and an observed correlation betweendepth
of attack and surface-to-volume ratio of the Inconel
and fluoride also indicates that the impurities in
the fluorides are the major factor in corrosion.
The attack by fluorides on Inconel at 1500°F
apparently increases in intensity, although not in
maximum depth, as the UF  content of the meltis
increased from O to 6.5 mole %. Static fests of
various combinations of metals in fiuorides ot
1500°F have shown that one of the metals is
attacked to a greater extent than the other, even
though the heavily attacked metal may be relatively
unattacked when alone in the fluoride. The fluoride
fuel NaF-ZrF4-UF4 (50-46-4 mole %) has also been
circulated in nickel and type 430 stainless steel
loops for 500 hr at 1500°F; no plugging occutred,
but some mass transfer was observed in the stain-
less steel loop. A number of metal specimens were
tested in circulating lead in quartz therma! con-
vection loops. Of the metals tested, oxidized
types 347 and 304 stainless steel and molybdenum-
nickel alloy produced the least mass transfer.
Other liquid metal studies included tests of
inconel, stainless steels, and potential bearing
materials in lithium, sodium, and lead. The more
fundamental  corrosion studies included ' the
determination of the pressure  of
and an

equilibrium
hydregen in hydroxide metal
attempt to determine the chemical equilibria for the
postulated  fluoride corrosion which results in
the selective leaching of chromium from Inconel.

systems

The fabrication of high-gerformance radidgtors
and high-conductivity radiator fins represents the
major undertaking of the metallurgical research
program, which also included stress-rupture tests
of Inconel, the forming of special alloys, the
fabrication of solid fuel elements, and the develop-
ment of ceramic materials (sec. 7).  Although
Inconel-clad copper fins are inadequate because
of diffusion of nickel into the copper, acopper-
aluminum bronze fin and types 310 or 346 stainless
steel clad copper fins have been fabricated .and
satisfactorily brazed to Inconel or stainless steel
tubing with a number of special brozing alloys.
The technique of “*backing-up’’ the heliarc-welded

47
ANP QUARTERLY PROGRESS REPORT

radiator joints with a brazing alloy produces
leak-tight radiators. The stress-rupture

various combinations of

sound,
life of Inconel
temperature, stress, and environment is being
measured; the tests at 3500 psi and 1500°F show
that, as far as rupture life is concerned, air is the
most beneficial environment and hydroger the most
deletericus.

unider

Stress-rupture data are also being
obtained for lnconel tubes and for fine- and coarse-
grained Incone! sheet. The limited effort on solid
fuel elements has been confined to the drawing of
tubular fuel elements and the plug welding of
steel-clad fuel plates, both of which
were effected with some success. Incone! tubing
for experimental equipment is being drawn satis-
factorily with reductions as high as 36%. The
rolling of various metals and ulloys was effected:
columbium sheet after an 87.5% reduction at 1100°F
showed complete recrystallization.

The physical properties of several fluorides and

stainless

metal alloys have been measured at temperatures of
up to 1000°C, and the heat transfer characteristics
of reactor coolants are being studied in various
systems (sec. 8). The viscosity and density of the
ARE fuel concentrate, Na UF , and the vapor
pressure of the fuel solvent, NchFs, have been
measured. All these values are in line with those
anticipated for these materials. Other heat
capacity, thermal conductivity, density, and vapor
pressure data for fluoride compositions of interest
have been obtained. Measurements of the velocity
in a thermal convection loop indicate that the radial
gradients cause turbulence in the
liquid flow at Reynolds numbers above 100; the
laminar tlow solutions for flow in this regime are
thus in serious error. Forced-convection heat
transfer data with the NafF-KF-LiF eutectic in
nickel and in lnconel tubes have been confirmed;
the heat transfer in Inconel is less than that in

temperature

nickel because of film formation. Preliminary data
are available on a forced-flow heat transfer experi-
ment in the laminar flow regime, and apparatus is

48

being prepared for measurements of fluid flow in an
annulus and for studies of surface-boiling phenomena.

The radiation damage program included studies
of the stability of fluoride mixtures in Inconel and
determinations of the creep of metals under irradi-
ation, as well as the construction of the in-pile
circulating loop (sec. 9). Although [nconel capsules
containing fluorides show more tendency toward
intergranular corrosion under irradiation than they
do in out+of-pile control tests, there is no definite
trend in corrosion behavior that can be correlated
with either irradiation time (from 53 to 810 hr) or
irradiation level (from 230 to 8000 watts/cm3).
There were no in-pile creep tests made in the LITR
or in the MTR during this quarter, because more
reliable apparatus for creep measurements in both
these irradiation facilities was being developed
and constiucted. The in-pile circulating loop,
types of which will be used for forced-convection
corrosion tests in the LITR aond in the MTR, is
being fabricated and assembled.

The analytical studies of reactor materials in-
cluded chemical, spectrometric, and petrographic
anclyses of fuel composition and corrosion products
(sec. 10). The chemical analyses were primarily
concerned with the oxidation states of the con-
stituents and the metallic corrosion products in the
fluoride mixtures. Techniques were
established for determining the amounts of FeF_,
Cer, and UF_ in fluoride fue! mixtures, as wefil
as the reducing power of NaZrF_. None of a
number of solvents tested were as effective for
dissolving the proposed fluoride fuels from the

various

walls of metal containers as the nitric and boric
acid mixtures whichare currently used. Petrographic
examination of a numbker of irradiated and wun-
irradiated fuel samples did not reveal any dif-
ferences in the two species. A technique has been
developed by which the mass spectrometer may
provide a measurement of the uranium burnup (due
to irradiation), as well as a quantitative analysis
of the uranium in the fuel.
PERIOD ENDING DECEMBER 10, 1953

5. CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS

W. R. Grimes
Materials Chemistry Division

Exploratory investigations of phase equilibria in
fused salt systems by the simple thermal onalysis
techniqgue have been extended to a number of
chloride systems containing UCI, and UCl, as the
fissionable components, as well as to examination
of several fluoride systems with ond without
uranium that are of possible interest as fuels,
fuel solvents, or coolants. In addition, the thermal
analysis technique has been applied to a number of
previously studied systems in attempts to verify
some doubtful values in various regions or to
amplify data for regions in which previous experi-
mentation was limited. '

During the past several months, other techniques
useful for more refined studies of systems pre-
viously shown to be valuable have been developed
and perfected.  These techniques, especially
quenching, with petrographic and x-ray examination
of the solids, differential thermal analysis, and
high-temperature filtration, have been applied,
in general, to the binary and ternary systems
NaF-ZrF, and NaF-ZrF -UF,. While it is not
yet possible to interpret even the binary system
uniquely, it appears that combining the data ob-
tained by these techniques and the thermal analysis
data will permit rapid progress to be made.,

Spectrophotometric measurements of the absorp-
tion spectra of UF, and UF, in fused salts were
made. Only two absorption spectrum patterns were
found for glasses from the NaF-ZrF -UF, system,
Glasses were obtained with compositions of
between 16.65 and 50 mole % UF, in the ZrF UF
binary system.

Additiondl electrolyses of various solutions of
metal chlorides in potassium chloride were made.
Erratic results obtained with chromium anodes
have been ascribed tentatively to oxidation of the
chromium. When a thin plate of nickel was used
to protect zirconium electrodes from oxidation,
reproducible emf’'s were observed.

Measurements of the density and electricadl
conductivity of the fused salt system LiCi-KCl are
indicative of considerable
persisting in these high-temperature fluids; these
and other measurements substantiate the concept
thot fused salts are essentially totally ionized.

shorf-range order

Facilities for the production of a variety of
fluoride compositions are being prepared and
laboratory~scale production is to be discontinued,
except for the preparation of small enriched-
uranium-bearing specimens for radiation stability
tests. A number of small fluoride samples were
purified for phase equilibria studies, plans were
made for producing the enriched material for the
in-pile loop experiments, ond various simple
structural metal fluorides were prepared for corro-
sion studies,

In a study of the reduction of Na, UF by hydrogen
it was found that the rate of reduction increased
as a function of temperature. A further check on
the possibility of retaining reducing species in
NaZrF. melts was attempted by treating NalZrF,
with NoH, and the kinetics of the reduction of
Nit* in NaZrF, by H, at 800°C was studied be-
couse of the |mportance of this reduction in fluoride
puritication procedures. Studies of the reaction of
sodium hydroxide with carbonaceous matter suggest
that it would not be possible to maintain sodium
hydroxide free from carbonate at high temperatures
in metals which contain oppreciable quontities of
carbon,

THERMAL ANALYSIS OF FLUORIDE SYSTEMS

C. J. Borton L. M. Bratcher
J. Truitt

Materials Chemistry Division

Additional data that are believed to be more
reliable than some obtained sarly in the phase
study program were obtained on a number of binary
and ternary fluoride systems. Revised diograms
will be published in topical reports when the
studies are completed. Although some data for
compositions in the NaF.-ZrF -UF, system were
obtained by the therma! analysis technique, it is
expected that the major emphasis will be placed
upon other techniques for study of this system.

MaF.ZrF -UF,

The probable existence of a ternary eutectic
NGF-ZrF4~UF4 (63.5-19.0-17.5 mole %} with a
melting point of 800°C was mentioned in the pre-
vious report. Closer study of this region has

49
ANP QUARTERLY PROGRESS REPORT

shown that this composition is probably near the
bottom of a very narrow valley extending down
toward the chszé-NaZrF eutectic. Additional
data between the 40 to 50 mole % NaF lines at
several uranium concentration levels essentially
confirmed the location of the isotherms in this
region that are shown in the published diagram for
the system.! Data for the 75% NaF mixtures are
shown in Fig. 5.1. These data indicate solid
solution formation between Na,ZrF, and Na,UF,
over a wide range of compositions, and they have
been confirmed by other methods (cf., subsection
on ‘‘Filtration Analysis of the NaF-ZrF -UF,
System”’).  Thermal data obtained from sealed
capsules of ternary mixtures containing more than
75% NaF have, in general, indicated melting points
that were lower than had been anticipated for this
region and lower than had been indicated by the
few runs made previously, Further study of this
system by other metheds is planned.

 

1. M. Bratcher, R. E. Traber, Jr., and C, J. Barton,
ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL-1375,
Fig. 41, p. 78.

RbF-LiF-UF,

Data for the system RbF-LiF-UF , were published
in a previous report.?2  Additional data obtained
during this quarter confirmed the existence of two
low-melting-point regions in the ternary system,
one close to the RbF-LiF binary eutectic and the
other in the vicinity of the LiF-UF, eutectic.
These data indicate that revision of the isotherms
shown in the published diogram is needed. [t
appears to be rather difficult to obtain reproducible
thermal data in this system, possibly because of
the oxygen introduced into the mixtures by the
hygroscopic alkali fluorides.

NaF-ThF

Thermal analysis of the system NaF-ThF,
supplemented by petrographic study of a number of
slowly cooled compesitions, continued during this
quarter, It seems to be a difficult matter to prepare

2J. P, Blakely et al,, ANP Quar. Prog. Rep. Sept. 10,
1951, ORNL.-1154, Fig. 13.3, p. 159.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DWG. 22251
‘ |
800 |- N SO D SN OUR e nPA7
//
)/ @
750 b L ey
g L-’/
w /
T
>
o
< 700 - )/ ,,,,,,,,,,,,,,,,,,
L !
o i
=

o

650 &

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B0 |reeeeerem L —
550 1 N [ L
NagUF, 10 20 30 40 50 60 70 80 90 NayZr F,

NazZr F; (mole %)

Fig. 5.1. The System N03UF7-N03ZrF7.

50
compounds in this system in the pure, well-crystal-
lized condition needed for petrogrophic identifi-
cation. Twocrystalline phases have been observed,
and one of them is believed to be Na,ThF , the
identity of the other is still uncertain, Further
study will be required before a satisfactory equi-
librium diagram for the system can be prepared.

LiF-RbF-BeF,

Thermal data were obtained in the system LiF-
RbF-BeF ., for 50 compositions with up to a BeF,
concentrotion of 60 mole %. The lowest melting
point observed with these mixtures was 345°C for
material containing (in mole %) 37% LiF, 8% RbF,
and 55% BeF,. The composition of this material
is close to that of the LiF-BeF, binary eutectic
reported to melt at 356°C.% It appears unlikely
that compositions with melting points substantially
lower than 350°C exist in this system, except
possibly in the high Bef, region which is not
amenable to thermal analysis; materials with such
high BeF, concentration would probably be of
fittle interest in the reactor program because of
their probable high viscosity, These results are
similarto those reported previously for the KF-LiF-
Be2 system.?

RbF-BeF 2" er'z‘1

The system RbF-BeF -ZrF, was investigated
with ZrF . concentrations of up to 50 mole % and
with Bef, concentrations of up to 60 mole %,
Halts in the cooling curves at temperatures near
300°C were observed for several compositions in
this system. |t is likely that the lowest thermal
effects observed can be attributed to a solid-phase
transition, since the RbF-BeF, and RbF-ZrF,
binary systems showed thermal effects in the 300
to 350°C temperature range which were almost
certainly due to such transition. No liguidus
temperatures significontly lower than 380°C (the
minimum melting peint in the RbF-BeF, system)
were observed with the compositions tested.

THERMAL ANALYSIS OF CHLORIDE SYSTEMS

C. J. Barton S. A. Boyer
Materials Chemistry Division

The UCI; used for most of the phase studies of
chloride systems made this quarter was prepared

 

3D. M. Roy, R. Roy, and E. F, Osborn, J. Am. Ceramic
Soc. 33, 85 (1950).

PERIOD ENDING DECEMBER 10, 1953

by the Y-12 Chemical Division by hydrogenation of
UCl,. The melting point (840°C) obtained for this
material agrees well with the melting point of
842 + 5°C reported by Kraus.® Uranivm and
chloride analyses of the as-received material
checked with the theoretical values very closely;
however, slight hydrolysis of the material occurs
on heating, even when considerable precaution is
taken to avoid exposure.

LiCl-UCI,

Thermal data obtdined with ¢ number of composi-
tions in the system LiCl-UCl; are shown in Fig.
5.2, The only eutectic in the system contains
approximately 25 mole % UCl, and melts at
495 £ 5°C, This is apparently the lowest melting
alkali chloride~UCI, binary eutectic, The thermal
data gave no indication of compound formation in
this system, but, since the solid phases have not
yet been studied, the absence of incongruently
melting compounds is not completely demonstrated.

NaCI-UCl,

An equilibrium diagram for the system NaCl-UCI,
was reported by Kraus,® who stated that there were
no compounds formed by these components. Thermal
data obtoined with nine mixtures containing 33 to
85 mole % UCl, are in reasonable agreement with
Kraus' data. However, the absence of the eutectic
break in the cooling curves for the mixtures con-
taining 65 mole % UCI, or more seems to indicate
the possibility of compound formation in the high
UCI; region. A thermal effect at 415 £ 5°C through-
out the region studied here, which was not reported
by Kraus, could be due to « solid transition of a
compound in the system.

KCl-ucl,

Krous also published an equilibrium diagram for
the system KC]-UCI3.5 Thermal analysis of the
system in this laboratory with mixtures containing
12 to 65 mole % UCI3 has essentially verified
Kraus' data. The maximum difference in liquidus
temperatures in the two studies was about 15°C,

 

4., M. Bratcher and C. 1. Barton, ANP Quar. Prog.
Rep. March 10, 1953, ORNL-1515, p. 113.

S¢. A Kraus, Phuse Dioyrams of S5oeme Complex Salts
of Uranium with Halides of the Alkali and Alkaline Earth
Metals, M-251 (July 1, 1943).

51
ANP QUARTERLY PROGRESS REPORT

 

900 (—- ‘

 

 

 

 

700 fromveeee

o

2.

PRI

o

o

E »
2 800
Ll

0.

=

wl

—

 

 

 

 

 

 

500

 

 

 

 

 

400 fo e e

 

 

 

 

 

 

300 term— e i ‘

 

 

 

 

 

 

 

 

 

LiCl 1O 20 30 40

50 60 70 80 90 Ucty

UCl3 (mole %)

Fig. 5.2. The System LiCl-UCi..

RLCI-UCI

Preliminary data indicate that there are two
eutectics in the system RbCI-UCl;, one ot about
15 mole % UCI,; that melts at 605 £ 5°C and
another lying between 40 and 50 mole % UCI; that
melts at 515 £ 5°C, The compound Rb,UCI, melts
congruently at about 730°C, while Rb,UCI appears
to melt incongruently at approximately 565°C,

CsCl-ucl,

Only a few thermal measurements have been made
on the system CsCl-UCI,, but there is indication
that there are probably only two eutectics in this
system. One eutectic has a high CsCl content
and melts at 580°C, and the other contuins approxi-

mately 47 mole % UCI; and melts at about 540°C.

KCi-LiCl-ucl,

The few studies that have been made on the
system KCI-LiCl-UCl, indicate that the addition
of about 4 mole % UCI; to the KCI-LiCl eutectic
; lowers the melting point from 355 to 345°C, but

52

further additions of UCI, result in a sharp increase
in melting point that is probably due to formation
of the high-melting-point K2UC|5 complex. Study
of this system will be resumed at a later date.

RbCI-UCI,,

Preliminary data for the system RbCI-UCI, indi-
cate that Rb,UCI, melts congruently at about
630°C and thot the eutectic in the 40 to 50 mole %
region melts at approximately 340°C.

CsCl-ucl,

Preliminary data for the system CsCl-UCI, were
presented in the previous report.® A tentative
equilibrium diagram for this system, based on the
available thermal data, is shown in Fig. 5.3, The
compound Cs,UCI,, with a melting point of
657 t 5°C, is believed to be the only congruently
melting compound in this system. There is some
evidence for a compound such as CsU,Cly or

 

®R. J. Sheil, 5. A. Boyer, and C. J. Barton, ANP
Quar, Prog. Rep. Sept. 10, 1953, ORNL.-1609, p. 59.
PERIOD ENDING DECEMBER 10, 1953

 

800

700

600 |—

 

500

TEMPERATURE {°C)

 

400

 

300 oo s o R

 

~.
L.SZUC!E

 

 

 

 

 

 

200

IR
DWG 22253

o
]

 

 

 

 

 

______ SRRSO SR S N U

Oy
T O — &

 

 

 

 

 

 

 

 

Cs{l 10 20 30 40

50 €0 70 80 90 UCi,

Ucl, {(mole %)

Fig. 5.3. The System CsCI-UCI .

CsU,Cl,, that would melt incongruently at about
382°C, which is only a few degrees above the
melting point of the eutectic containing approxi-

mately 58 mole % UCI .

KCl-Licl-ucl,

Preliminary data reported’ for the system KCI-
LiCI-UCI, indicated a eutectic of unknown compo-
sition that melts at approximately 275°C, Further
investigation of the system during this quarter
confirmed the existence of the low-melting-point
eutectic, which appears fo contain (in mole %)
approximately 45% KCl, 20% LiCl, and 35% UC|54.
However, a more complete study of the system will
be required for determining the eutectic composition
with more certainty.

 

7R, J. Sheil and C. J. Barton, ANP Quor, Prog. Rep.

June 10, 1953, ORNL-1556, p. 42.

NaCl-LiCl-UCH ,

The very complete investigation of the system
NaCl-LiCl-UCl, made during this quarter failed to
show a ternary eutectic with ¢ melting point
significantly lower than that of the lowest melting
NaCl-UCi, binary eutectic (375°C). Since this
investigofion covered the region where low melting
compositions are believed most likely to occur, no
further work with this system is contemplated ot
the present. [t is interesting to note that when
small amounts of UCI, (2 to 10 mole %) were added
to the NaCI-LiCl binary eutectic, the resulting
ternary mixtures had lower melting points than
those of the binary mixtures. The smallest addi-
tion of UCI, (1 mole %) to the KCI-LiCl eutectic
resulted in a higher melting point. This difference
in behavior probably can be atiributed to the higher
melting point of the K,UCI, complex as compared
with that of the Na,UCl, complex,

53
ANP QUARTERLY PROGRESS REPORT

QUENCHING EXPERIMENTS WITH
THE NoF-ZrF , SYSTEM

R. E. Thoma
J. Truitt
G. D. White

Materials Chemistry Division

Barton
Moore

C. J.
R. E.

Recent improvements in previously described®
techniques ond equipment for quenching small
samples of fused fluorides have greatly improved
The technique of
determining equilibrium conditions in solid phases
by quenching o composition to glass ond then
heating to the temperature of interest to grow
crystalline phases in the glass gives promise of
solving some of the more difficult problems in the
NaF-ZrF, system. Only preliminary results with
this technique have been obtained.

It is opparent that very occurate temperature
control is necessary to establish phase relation-
ships in the 40 to 60 mole % ZrF, region of the
NaF-ZrF , binary system in which most of the
quenching work has been done. The improved
equipment in use at present controls and measures
temperatures to t0.5°C. The two furnaces now
in operation make possible the quenching of about
40 samples a day.

the reproducibility of results.

Qil has proved to be as effective a quenching
medium as mercury and it is much more convenient
becouse it eliminates the necessity of having to
attach it to the capsules. The formation of crystals
during quenching has been kept to a minimum
through the use of small samples (about 3 mg) and
by severe pressing of the portion of the capsule
that contains the sample to decrease the sample
thickness. However, quench growth is still some-
times troublesome in the NaF-ZrF, system. An-
other problem that is not yet completely sclved is
the presence in most gquench samples of small
amounts of an unidentified phase believed to be a
complex sodium zirconium oxyfluoride. However,
the practice of hydrofluorination of small samples
(cf., subsection on ‘‘Production of Purified Fluo-
rides’’) and the handling of the finely ground
samples in vacuum dry boxes have been of value
in keeping these phases at a minimum.

Quenching of specimens containing 33 to 46
mole % ZrF, reveals a cubic crystalline phase,

Because the phase is isotropic, petrographic

 

8R. L. Moore and C, J. Barton, ANP Quar. Prog. Rep.
Sept. 10, 1953, ORNL-1609, p. 59.

54

identification is difficult; x-ray diffraction tech-
niques will be required to establish the liquidus
temperatures, The cubic matericl is believed to be
a form of Na,ZrF , but the possibility that it is
Na,Zr,F | has not yet been eliminated.

Material containing4émole % ZrF , shows NaZiF ¢
os the primory phase. The liquidus temperature is
512 10 515°C, and the sclidus temperature is 510°C,
Devitrification at 506°C of glass of this composi-
tion gave NaZrF . and a small amount of the cubic
material.

Material containing 48 mole % ZrF, shows as
its primary phase on easily crystollized speciss
believed to be o high-temperature form of Na,Zr F o
(designated R3).

In the mixture containing 30 mole % Zrf  the
primary phaose is R3; the liquidus temperature is
near 518°C; and the solidus temperature is slightly
above 510°C. Slow cooling of mixtures with this
composition invariably gives a phase referred to
as NoZrFS and No3Zr4F]9. However, devitrifi-
cation of the glass of 506°C provided only crystal-
line NaZrF, with some glass residue. It now
appears that solid solutions of Na,ZrF, in NaZrF
exist which show optical properties that differ
only slightly from those of NaZrF .. The devitrifi-
cation experiments seem to indicate that NaZrF,
is the stable solid phase at this concentration,

Material containing 53.8 mele % ZrF, shows R3
as the primary phase; the liquidus temperature has
not been established, but it is above 517°C; the
solidus temperature is about 511°C, BothNaZr F o
and NaZrF . ore present below the solidus tempera-
ture.

Mixtures contoining 57.2 mole % ZrF | correspond
to the formula Na,Zr F .. A glass devitrificotion
experiment produced an almost pure crystalline
phase with this composition. Although this single
experiment apparently confirms the formula, further
experimentation at this and neighboring composi-
tions will be required to establish beyond doubt
the composition of the compound.

DIFFEREMTIAL THERMOANALYSIS OF
THE NaF-Zr?4 SYSTEM

R. A. Bolomey
Materials Chemistry Division

The equipment used for differential thermo-
analysis was modified during this quarter by re-
placing the capsules that had thermocounles
welded to the side with small copsules that have
center thermocouple wells, With the new capsules,
much better agreement is obtained between heating
and cooling curves, and, for most samples, the
break for the liguidus transition is sharper and
better defined. Heating and cooling rates of 1°C
per minute can now be employed satisfactorily for
most samples. Date have been obtained with these
capsules for mixtures in the system NaF-ZrF ,
that were hydrofluorinated before use. A break at
about 512°C for samples containing 57 to 60 mole %
ZrF4 suggests that the composition of the phase
which precipitates from 50 mole % material is
probably above 57 mole %. It is possible, however,
that the existence of the 512°C break in thisregion
is due to o solid transition; this point will be
established by combining the data obtained from
differenticl analysis with that obtained by using
quenching techniques. Data are being collected in
the 40 to 50 mole % ZrF, range, and, when ex-
tended to higher ZrF concen_frafions, a complete
phase diagram will be presem‘ed,

FILTRATION ANALYSIS OF FLUORIDES
C. J. Barton R, J. Sheil

Materials Chemistry Division

The high-temperature filtration method of study-
ing phase equilibriac mentioned in the previous
was applied mainly to NaF-ZrF - 4~UF,
Two afiempfs to
mixfures were unsuccessful be-

repor‘t9
compositions during this quarter.
filyer UF“--,Z:’F4
cauyse the high vapor pressure of Z¢F , resulted in
separation by sublimation rather than liquid-solid
separation.  in the NGF-ZYF4-UF4 system, the
filtration method provided useful information
about an important fuel composition in addition to
data for fundamental studies of phase relation-
ships in this system.

Mixtures with 53 mole % NaF

The melting point of o mixture containing (in
mole %) 53.5% NaF, 40.0% ZrF4, and 6.5% UF‘1
has been reported as 545°C., However, in g
previous filtration at 560°C of a 7 mole % UF,
mixture, analysis of the residue and of the filtrate
some segregation of o high-uranium-
content material. This experiment has been re-
peated carefully with a mixture containing {in mole

indicated

 

9R. J. Sheil and C. J. Barton, ANP Quar. Prog. Rep.
Sept. 10, 1953, ORNL-~1609, p. 61,

PERIOD ENDING DECEMBER 10, 1953

%) 53.0% NaF, 40.0% ZrF,, and 7.0% UF,, which
was filtered at 604, 575, crnd 560°C. At the two
higher temperatures, there was no detectable dif-
ference between the filtrate and the residue, when
examined either petrographically or by x-ray dif-
fraction. At 560°C, more than 99% of the charge
material passed through the filter, but examination
of the residue indicated that it was a higher ura-
nivm solid solution than the filtrate. There was
not enough residue obtained in this experiment for
a chemical analysis, and therefore the significance
of the apparent phase separation is not clear. In
one filtration carried out at 525°C with the same
composition, 83 wt % of the material filtered. Both
filtrate and residue were found to be solid solu-
tions of the type No(U)ZrFs, with the residue hav-
ing the hlgher uranium content. It is considered
significant that no trace of free UF, was detected
in the solid phases.

Mixtures with 50 mole % NaF

Thermal data for NaF-ZrF -UF , mixtures'® con-
taining 50 mole % NaF seemed to indicate that
NaUF o end NaZrF, form solid solutions. Petro-
graphic and x-ray diffraction studies have shown
that solid solutions do occur in this system but
that UF, is probobly the primary phase that sepa-
rates from the mixture containing 50% NaF, 25%
UF,, and 25% ZrF . Filtration experiments ‘with
50% NaF mixtures contammg 10, 12.5, 25, 31, and
37.5 mole % UF, showed that UF, separated in
significant omounts from the 12,5, 25, and 31%
UF, compositions, with the maximum separation
occurting at 25%. Trace amounts of UF, were found
in several of the other solid phases, Because of
the UF, separation, chemical analysis of the
filtrate and residue did not give a relioble indica-
tion of the liquidus ond solidus temperatures for

most of the 50% NaF join.

Mixtures with 75 mole % NaF

The thermal data for the 75% NuaF compositions
in the NaF-ZrF «UF , system showed evidence of
more or less complete solid solution formation
between Na,UF, ond Na 3Z¢F .. Since Na UF
melts mcongruenfly, it was expected that the sohd
solutions would break down near the uranium end
of the series. However, prelimincry data indicate

 

 

]OL. M. Bratcher and C. J, Barton, ANP Quar. Prog.
Rep, Dec, 10, 1952, ORNL.-1439, Frg 10.2, p. 108.

55
ANP QUARTERLY PROGRESS REPORT

that o comparatively small amount of Zrf, (1.5
mole %) stabilizes the Na,UF_ structure,

Filtrations with mixtures containing 8%, 12%,
and ]62/3 mole % UF, (equivalent to 25, 50, and
75 mole % Na,UF,) gave evidence of complete
solid solution formation within this range, The
x-ray diffraction pattern for all six samples from
these filtrations showed only the NusZrF7 ond
Na,UF, lines with the maximums shifted to varying
extents. Petrographic examination, which is
capable of detecting trace aomounts of crystalline
phases not shown by x-ray diffraction, showed the
presence in some samples of other phases, such
das NGQZrFé and NaZrFs, that contained dissolved
uranium and, in one case, a small amount of UO,,.
An unidentified yellow-brown phase, which ap-
to be characteristic of NaF-ZrF -UF,
mixtures containing more than 75 mole % NalF, was
reported to be present in both the filtrate and the
residue from the filiration of the 8]/3 and 127/2 mole
% UF , mixtures. Since this phase did not appear
in hydrofluorinated mixtures, it seems probable
that it is an oxygen-containing complex,

The dato shown in Table 5.1 indicate that, within
the range of compositions and temperatures covered
in these experiments, the system NaF-ZrF -UF,
is a true binary system, since the phase separation
resulted in mixtures of very nearly the same NaF
content as the starting material. The liquidus
temperctures determined by the filiration method
agreed wel! with the thermal data shown in Fig.
5.1 (cf., subsection on ‘‘Thermal Analysis of
Fluoride Systems''). Since cooling curves for
these compositions failed to indicate solidus
temperatures, the data in Table 5.1 give the first
indication spread between
liquidus and solidus lines for this system. Further
filtrations with compositions approaching that of

pears

of the temperature

the pure compounds are planned for the near future
in an attempt to determine whether limits of solid
solution exist in this system,

FUNDAMENTAL CHEMISTRY
Spectrophotometry of Supercooled Fysed Salts
H. A. Friedman

Materials Chemistry Division

Measurements of the absarption spectra for UF
andUF ; in quenched fluoride melts with a Beckman
DU spectrophotometer and the procedure described
by Friedman and Hill'! have been continued. More
efficient quenching was obtained by flattening the
sealed samples in a hydraulic press to give larger
surface areas and thinner melts.

With glasses from the NaF-ZrF ,-UF, system,
only two absorption spectrum patterns have been
found. One of these superficially resembles that
of crystalline UF ; the other more closely resembles
crystalline Na,UF .. The Na,UF  type of pattern
was found only in alocalized region near the
Na,UF . composition. Glasses of the ZrF -UF
system were prepared with compositions in the
range from 30 to 70 mole % UF, in 10 mole %
intervals. Qutside this region, no glasses were
obtained; within this region, Beer's law was closely
followed by the UF , type of pattern. Samples along
the No ,UF .+Na,ZrF  join that contained less than
60% Na,UF . were not studied because of quench
growth; the Na,UF , type of pattern wos found in
samples that contained from 60 to 100 mole %
NGZUFG. Nine mixtures were prepared in which
the ZrF, content was held at 41 mole %, but the
NaF-to-UF , ratios were adjusted to cover the

114, A. Friedman ond D. G. Hill, ANP Quar, Prog.
Rep. Sept. 10, 1953, ORNL-1609, p. 62.

TABLE 5.1. LIQUIDUS-SOLIDUS EQUILIBRIA OBTAINED BY FILTRATION

 

 

INITIAL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FILTRATION FRACTION N ANALYSIS OF FRACTIONS (mole %)
COMPOSITION N OF CHARGE ) .
(mole %) TEMP(E‘Z‘;‘TU“E IN EILTRATE Filirate Residus
- o - T
NayUF, Na, ZeF, (wt %) NaF ZrF 4 UF, NaF ZE, UF,
75 25 695 63 74.9 5.1 20.0 74.9 8.4 16.7
50 50 742 29 75.0 10.5 14.5 75.1 13.8 11.1
25 75 786 60 74.9 17.4 7.8 75.6 20.4 3.8

56
region from 6 to 59 mole % UF,; all nine mixtures
formed glosses and exhibited the UF, type of
absorption in conformity with Beer's law. The
NaUF .-NaZrF . system is currently being investis
gated, :

Glosses were obtained with compositions of
between 16.67 and 50 mole % UF, in the 2.'1'!’74-UI:3
binary system. At higher concentrations, crystal-
lization of UF; could not be avoided. No shift in
pattern occurred in the glass region to indicate o
change in the nature of the UF, species in the
liquid melt ot concentrations below 50 mole %

PERIOD ENDING DECEMBER 10, 1953

EMF Measurements in Fused Salts
L. E. Topol

Materials Chemistry Division

The techniques described previously have. been
used for additional electrolyses of various  solu-
tions of metal chlorides in potassium chloride.
Containers of morganite, grophite anodes, nd
platinum or nickel cathodes are used, and the
electrolyses are carried out af a temperature of
850°C, Potassium chloride solutions containing
1 wt % of NiCl, (theoretical E% = 0.82 volt) gave
decomposition potentials at 1.20 to 1.26 volts,

 

UF,. The color of the glasses was rust-olive.. A while solutions containing 10 wt % NiCl, yielded
typical spectrum is shown in Fig. 5.4 values in the range 1.0 to 1.06 volts. Solutions
L
.400 [ OWG. 22254
0.300 e e e

2

0.200 |-

OPTiCAL DENSITY PER mg OF UF,

 

 

Q.00

 

 

 

 

 

 

 

 

500 500 - 700

800 200 1000

CWAVELENGTH (mu)

Fig. 5.4. Absorption Spectrum of the ZrF ~UF , System Containing 33.3 mole % UF ..

57
ANP QUARTERLY PROGRESS REPORT

containing 1 to 2 and 10 wt % of FeCl, (theoretical
£% = 1.11 volts) yielded decomposition potentials
at 1.50 to 1.55 and 1.20 volts, respectively, A
mixture containing 2 wt % of NiCI2 and 2 wt % of
FeCl, in KCI yielded two distinct brecks in the
I-£ curve at 1.18 and 1.56 volts, as shown in Fig.
5.5,

UNCLASSIFIED
DwG. 22255

0.50 [

0.40 e - e B— W

 

0.30 Los : - e e

1 (amp)

 

0.20 | ~ - e i

 

 

 

 

© INCREASING POTENTIALS

A DECREASING POTENTIALS
NICKEL CATHODE
GRAPHITE ANODE

2.0 30 4.0
E (volts)

 

 

 

Fig. 5.5. Electrolysis at 850°C of Mixture Con-
taining 2 wt % of NiCi, ond 2 wt % of FeCl, in
KCI,

in KCi in similar
experiments have yielded decomposition potentials
at about 1.15 to 1,30 and 1.50 to 1.75 volts, but
reproducibility of the results has been rather poor.
Electrolyses of KCl with chromium anodes have
shown erratic results that are tentatively ascribed

Attempts to electrolyze Crf

to oxidation of the chromium, even though attempts
Rather
erratic changes of slope at about 1.2 and 1.9 volts
are observed,

are made to maintain an inert atmosphere.

When a thin plate of nickel was used to protect
zirconium electrodes from oxidation, reproducible
emf's at 1.06 to 1.12 volts were observed in
electrolyses of KCI, Since nickel is not dissolved
in KCI ot potentials below 1.8 wvolts, the emf’s

58

obtained are considered to be characteristic of the
zirconium. Based on the value for zirconium and
the decomposition potential for KCi (3.1 volts), the

potential for the reaction

ZrCl“m-m‘%h Zr + 2Cl2

appears to be 2.0 volts, When a hydrogen atmos-
phere is used over the zirconium electrode system,
the decomposition potential rises to 1.12 to 1,18
volts. Whether this may be due to formation of
hydrides of zirconium is not certain,

Physical Chemistry of Fused Salts

E. R, Yan Artsdalen
Chemistry Division

insight into the fundamental physical
chemistry of fused salt systems may be realized
from the physical properties of these systems.
Accordingly, the depression of the freezing point of
sodivm nitrate by a wide variety of solutes was
determined, and measurements were made of the
density and electrical conductivity of the LiCI-KClI
system and the heat capacity of Cdl,. The molten
NaCl-KCI system is also being investigated in the
hope that it will yield not only practical data, but
will also contribute to the understanding of short-

Some

range order and the nature of electrical conductivity
of fused salt systems. This work is reviewed
brietly below ond is reported in detail in the

Chemistry Division semiannual report. ' 2

Detailed, highly reproducible measurements of the
density and the electrical conductivity of the fused
salt system LiCl-KCl were completed for the entire
composition field over wide temperature ranges. An
exact linewr relation between molar velume and
composition at any specified tempercture indicates
that no complex formation occurs in this molten
system. On a molar basis, lithium chloride is a
considerably better elecirical conducter than po-
It was therefore significant to
discover that substitution of lithium chloride for
some of the potassium chloride actually lowered
the equivalent conductance of the solution below
that of pure potassium chloride. This phenomenon
was interpreted to be indicative of considerable

tassium chloride.

short-range order persisting in these high-tempera-
ture fluids. A further tentative explanation of this

 

]2A., S. Dworkin, D. J. Sasmor, and £. R. Yan Artsdalen,
Chem, Div. Semionn. Prog. Rep, June 20, 1953, ORNL-
1587, p. 19-25,
behavior was based on the consideration of the
relative sizes of the ions. Substitution of small
Li* ions for some of the much larger K¥ ions will
shrink the short-range chloride semilattice of the
fluid ond make transport by K* appreciably more
difficult. As more and more Li% is substituted for
K*, the conductivity will pass through a minimum
and again increase until, finally, pure LiCl is
present. '

The depression of the freezing point of sodium
nitrate was investigated with a wide variety of
solutes. The results significantly substantiate the
concept that fused salts are essentially totally
ionized and behave as nearly ideal electrolytes,
even at relatively high concentrations of the
solute. Thus, calcium, strontium, barium, and lead
chlorides, for example, are completely dissociated,
as is LaCl;. FeCl; seems to form two particles
only, apparently Fe(:l; and Ci™. Chromate, di-
chromate, and bromate ions are stable in liquid
sodium nitrate, but the metaperiodate ion, 107,
decomposes to yield oxygen and the iodate ion.

The heat capacity of high-purity crystalline
cadmium iodide has been measured from 15°K to
room temperature. It remains necessary to de-
in the range well below 15°K because of
the observed high heat capacity at this tempera-
ture. The heat capacity shows approximately @
l.4-power temperature dependence at the lowest
measured range and, therefore, does not yet fit
either the usual Debye T3 law or the TZ taw ob-
served here for boron nitride. At 298.16°K, the
following thermodynamic values have been found
for cadmium iodide: ¢, =19,12cal/mole, 5°=38.5,,
(F° ~ H3)/T =23.19, (H® — HS)/T = 1531 The
$° value has an uncertainty of about 0.5 because
of the+high heat capacity value at 15°K and 'its
attendant inherent difficulty of extrapolation to
0°K. This uncertainty is, of course, reflected in
the two derived quantities,

fermine ¢

PRODUCTION OF PURlFI‘ED FLUORIDES
F. F. Blankenship G. J. Nessle
Materials Chemistry Division
L.aboratory-Scale Production of
Molten Fluorides

F. F. Blankenship F. P, Boody
C. M. Blood . M. Watson
Materials Chemistry Division

Twenty-eight batches fotaii’ng 55 kg of molten
fluorides and including one batch containing 15 g

PERIOD ENDING DECEMBER 10, 1953

of enriched uranium were prepared and.distributed
to various laboratories for testing, When the ex-
perimental production facilities have been expanded
(cf., subsection on “‘Experimental Production
Facilities’), production of materials in this lab-
oratory will be terminated, except for the prepara-
tion of small enriched-uranium-bearing specimens
for radiation stability tests.

Purification of Small Fluoride Samples
for Phese Studies

F. F. Blankenship C. M. Blood
F. P. Boody

Materials Chemistry Division

Studies of phase equilibria in the NaF-ZrF4 and
l’\lc!F-Zrl‘""l—UF4 systems by the quenching technique
have been hompered by contamination of the
materials with traces of water which caused hy-
drolysis of the ftetravalent halides during the
heating period ond consequent contamination of
the quenched melt with oxide or oxyfluoride phases.
Materials for this and similar phase equilibrium
studies that are essentially free from water and
oxygenated compounds are now routinely prepared
by amedification of the hydrofivorination-hydrogena-
tion pretreatment method used for larger specimens.

In this process a number (up to 28) of smali
{(~25 g) samples are placed in nickel or platinum
crucibles on a nickel rack inside a 4-in.-diameter
nickel reactor for processing. A typical processing
cycle of 2 hr of treatment with HF at 800°C fol-
lowed by stripping with H, at 600°C for 18 hr has
been shown to produce material satisfactory for
these purposes. To date, a total of six batches
comprising 67 samples has been prepared., It
appears, however, that handling of the specimens
in a vacuum dry box may be required if freedom
from oxygenated phases is to be maintained,

Production of Enriched Moterial for
In-Pile Loop Experiment

G. J. Nessle

 Materials Chemistry Division

Plans for preparation of the enriched fuel for o
miniature in-pile loop facility and for charging of
the material into it have been prepared in accordance
with the necessary technical and accountability
considerations. |t is planned that this loop will
be charged with a uranium-free mixture for prelimi-
nary testing before the enriched concentrate is
added for final testing.

59
ANP QUARTERLY PROGRESS REPORT

The processing equipment for this production
operation has been fobricated, When production
of the ARE enriched concentrate has been com-
pleted, the new process equipment will be installed
in Building 9212 in the area now occupied by the
ARE facilities. The enriched fluorides for the
in-pile loop will be transferred to the X-10 site
where the loop will be filled,

Experimental Production Facilities

J. P. Blakely C. R. Croft

Materials Chemistry Division

D, E. McCarty R. Reid
R, G. Wiley
ANP Division

A total of 329.2 kg of mixed fluorides was
processed ond dispensed to various requestors
during this quarter. Included in this total were
14 batches of approximately 2.5 kg each, 3 batches
of approximately 10 kg each, and 10 batches of
approximately 27 kg each.

Although the quontity of fluorides processed
decreased, the variety of demand increased sharply.
This tendency is becoming more pronounced, and it
appears that production of large quantities of a
few materials will be of considerably less im-
portance; versatility and adaptability of equipment
will be primary requirements, Consequently, an
installation is being planned for producing o variety
of fluoride compositions with a minimum loss of
time and equipment.

Reduction of Na UF , by Hydrogen
. F. Blankenship F. P. Boody

Materials Chemistry Division

The Na,UF, which had previously been hydro-
fluorinated and stripped with hydrogen so that
the structural content should have been
reduced to less than 30 ppm was bubbled with
hydrogen gas in the original nickel! reactor for

metal

considerable periods of time at 692, 785, and

882°C. It was found that the rate of reduction of
the uranium according to the equation
] —
UF, + 4H, = UF; + nF

increased with temperature, as shown in Table
5.2. When the data given in Table 5.2 are plotted
on a semilogarithmic scale as reaction rate vs.
(1/T) x 104 the straight line shown in Fig. 5.6 is
obtained, From the slope of this curve, the acti-
vation energy of the reduction can be calculated

to be 33,700 + 2600 cal/mole.

Treatment of Molten NuZa‘FS with
Strong Reducing Agents

. . Blankenship C. M. Biood
G. M. Watson
Materials Chemistry Division

It was previously reported!® that zirconium bars

suspended in molten NaZrF ; contained ot 800°C in
graphite-lined reactors and stirred by bubbling
argon undergo weight losses far in excess of the
amount required to reduce structural metal ions.
This observation was further confirmed by addi-
tional experiments thot involved the treatment of
NaZrF. in grophite-lined nickel ot 800°C with
zirconium chips; the unrecovered zirconium metal
amounted to ot least 20 g per kilogram of fused
salt, However, it has since been observed that
little if any loss of weight occurred when zirconium
metal chips were exposed in a parallel experiment
without the graphite liner. The difference might be
due to the formation of zirconium carbide, but the
available data are insufficient to show it. The
possibility that a portion of the zirconium enters
the melt as the trifluoride or as disscolved zirconium
metal has not been ruled out,

 

IsC, M. Bilood et al., ANP Quar, Prog. Rep. Sept. 10,
1953, ORNL-1609, p. 69.

TABLE 5.2 RATE OF REDUCTION OF Na,UF . WITH H, AS A FUNCTION OF TEMPERATURE

 

 

 

 

 

REACTION RATE AS MOLES OF HF
PER LITER OF H, BUBBLED

 

TEMPERATURE 1

% 104
oc °K TCK)
692 965 10.36
785 1058 9,45
882 1155 8.65

 

 

8.8 x 107°
3.7 x 107%
1.58 x 10~¢

 

 

60
~E—
—3 DWG. 22256

REACTION RATE as moles of HF per titer of H2

 

8.0 a5 9.0 g5 0.0 105 1.0

Fig. 5.6. Reaction Rate vs. Reciprocal Temper-
ature According to the Equation UF , + 3’2 H, = UF,
+ HF.

The effect of uranium metal on NaZrF, was
studied by adding 5.5 wt % of uranium chips to the
material in o graphite-lined reactor. Unlike the
previous experiments, the mixture was treated with
hydrogen for 2 hr at 800°C, and the material was
subsequently filtered ot 700°C.  The filtrate
started freezing at about 540°C and contained five
phases typically found in fuels which are reduced
to the extent that orange-red UF 4:2ZrF ; and purple
UF, oppear. The uranium content of the filtrate
was 5.00 wt %, and opparently about 90% of the
vranium reacted. [f it is considered that the dis-
solved uranium is present as UF; and that all
other reduced phases are insoluble, a reducing
power of 210 meq per kilogram of melt should have
resulted; if all reduced phases remained in solution,

PERIOD ENDING DECEMBER 10, 1953

the theoretical reducing power would be 840 meg/kgq.
Chemica! analysis by H, evolution showed 550
meq of reducing power per kilogram of filtrate,
which indicated the presence of soluble reduced
phases in addition to the UF,. Less than 20 ppm
of nickel and chromium ond 85 ppm of iron were
found, The unfiltered residue in the reactor con-
tained black opaque material which liberated gas
readily when treated with dilute acids. X-ray
analysis of the residue revealed the presence of
ZrH, UF,, and uranium metal, along with occluded
melt. Thus it appears that the uranium did not
react quite completely and that some zirconium
tetrafluoride was reduced to the metal by the urco-
nium. Since ZrH, is decomposed to the extent of
about 25% ot 800°C under 1 atm of Hz,]4 most of
the zirconium appears as ZrH.

A further check on the possibility of retaining
reducing species such as ZrS ZrF3, ZrC, ZrH,
and ZrH, in NaZrF. melts wos attempted by
treating NaZrF . with NaH (2 moles per kilogram of
melt). The mixture was sparged with hydrogen in
a graphite liner for 24 hr ot 800°C and then filtered.
The reactor heel contained a dark-black residue
which was tentatively identified os ZrH mixed
with the melt. Chemical analysis of the white
filtrate showed 106.5 meq of reducing power per
kilogram or about 5% of that expected if all re-
duced species were soluble. Efforts to learn more
about the nature of this apparent reducing power
will be continued, | |

Reduction of NiF, by Hydrogen

F. F. Blankenship C. M. Blood
G. M. Watson

Materials Chemistry Division

Considerable attention has been given to the
kinetics of the reduction of Ni*% in NaZrF . by H,
at 800°C because of the importance of this reduc-
tion in fluoride purification procedures. The use
of graphite linersin the nickel reactors for handling
molten salts is a great convenience in loading and
unloading and in avoiding excessive contamination
with NiF, during the HF treatment. Accordingly,
for most of the kinetic studies, graphite-iined
reactors have been used. Early suspicions that
graphite might reduce Ni*" were allayed by the
large, positive, standard free energy estimated for

 

14M. N. A. Hall, Trans. Foraday Soc., 41, 306 (1945).

61
ANP QUARTERLY PROGRESS REPORT

the reduction with carbon,

2NiF,(solid) + C(solid)

 

> 2Ni%(solid)

+ CF (gas) AF{ g0k = 118 keal ,

and by the failure to find o noticecble difference
in short experiments without graphite liners. |In
carrying out more detailed measurements over a
period of days, the graphite liners have been found
to play an important but incompletely understood
role in the reduction of Nit", As a result, the
apparent complexity of the reduction kinetics is
now explainable from o qualitative standpoint,
but 6 new set of experiments without the graphite
liner must supplant the many experiments made
with the use of the grophite liner before guanti-
tative data can be assumed,

In a series of triols which provide evidence for
the effect of grophite, a weighed portion of nickel
fluoride was mixed with previously purified NaZrf
which was then pretreated at 800°C in a graghite-
lined reactor by bubbling inert gas through it at an
approximate rate of 400 cm?/min ond, finally, treated
with hydrogen at the same approximate bubble rate.
The varying concentrations of hydrogen fluoride
in the effluent gases were continuously determined
by direct titration of standard potassium hydroxide
samples to the phenolphtalein end point as the
gases bubbled through.

In Table 5.3 are listed the initial concentro-
tions of nickel fluoride used, the volume of inert
gas passed during pretreatment, the volume of
hydrogen passed, and the percentage of hydrogen
fluoride recovered during the hydrogen treatment.
From the results given in Table 5.3, it appears
that the recovery of hydrogen fluoride decreases
as the length of pretreatment increases; that is,
it is apparent that a fraction of the nickel flueride
added is reduced by some agent other than hydrogen

and thot this fraction increases with increasing
tength of pretreatment.

Since it appeared that the graphite liner was in
some fashion effectively reducing the NiF,, one
experiment that essentially duplicated the lost
test shown in Table 53 was conducted without
the graphite liner in the nickel vessel. Copious
evolution of hydrogen fluoride occuired. The re-
covery measured amounted to 72.6% at the end of
passage of 60 liters of hydrogen. Kinetically, the
reaction appears complex in the presence of graphite,
as evidenced by the voriable over-all order with
respect to the apparent nickel fluoride concentro-
tions.

Preparation of Various Fluorides

8. J. Sturm
Materials Chemistry Division

The methods of preparation and some of the
physical properties of numercus complex fluorides
that resulted from the interaction of alkali with
structural metal ftluorides have been previously re-
ported. 19 During this quarter, the simple structural
metal fluorides have been used increasingly in
corrosion studies, and a major partion of the avail-
able time has been devoted to preparation of these
The compositions of these fluorides
were determined by chemical analyses, and some
supplementary x-ray and petrographic determinations
were made for further characterization.

Several batches of FeF, and FeF, have been
prepared by hydroflucrination of the anhydrous
chlorides. Emphasis was placed on the preparation
of relatively small batches of these two fluorides

materials.

 

ORML-1439, p. 123; G. J. Messle and H. W. Savsge,
ANP Quar. Prog. Rep. Morch 10, 1953, ORNL-1515, p.
115, 8. J. Sturm and L. G. Overholser, ANP Quoi. Prog.
Rep. June 10, 1953, ORNL-1556, p. 48,

TABLE 5.3. RESULTS OF TESTS OF THE REDUCTION OF Nin BY HYDROGEN

 

 

 

 

INITIAL VOLUME OF GAS PASSED HF RECOVERED
CONCENTRATICN ot P DURING HYDROGEN
OF Nif, elium .reweofment Hydrogen Pretreatment TREATMENT

(meg/kg) (liters) (liters) (%)
15.7 44 69 82.5
29.2 132 119 81.6

228.0 1285 60 9.0

 

 

 

 

62
of the highest possible purity. Several pounds of
anhydrous Nif, were prepared by hydrofluorination
of the hydrated N!F Small quantities of anhydrous
NiCl, were S|m|lurly prepared from hydrated N|C|
tor specnul uses.

Previous work had indicated that the precipitate
resulting from the addition of NH HF, toan aqueous
solution of CdCl, could not be quantitatively con-
vertedto CdF , by hydrofluorination. Further studies
have shown that the solids resulting from the addi-
tion of NHHF, to aqueous solutions of either
CdSO, or Cd(C H,0 )2 give the same x-ray pattern
as thclt found for the material prepared from CdCl,;
this pattern differs from the accepted pattern for
CdF, and apparently indicates complex or double
salt formation. |t was observed that the interaction
of HF with CdCO, in an aquecus medium gives a
precipitate of CdF, that has the x-ray properties
accepted for CdF ...

In an effort to prepare some of the flyorides of
molybdenum, the powdered metal wos brominated to
produce a material consisting. mainly of MoBr,. A
first attempt to hydrofluorinate this material was
incomplete; components of the mixture have not
yet been identified.

PURIFICATION AND PROPERTIES
OF HYDROXIDES

Purification of Hydroxides

E. E. Ketchen L. G. Overholser
Materials Chemistry Division

The work with hydroxides during this quarter has
been confined to the purification of NaOH and
Sr(OH)., according to the methods previously des-
<::r|bed:Zl Earlier studies indicated that Na,CO,
could be separated from NaOH by filtering the 50
wt % aqueous solution of NaOH through a fine
sintered-glass filter instead of decanting off the
NoOH solution, as had been done. Further work
has shown that the filtration method produces o
material with a slightly lower Na,CO, content and
without a significant increase m s:hcu content.

 

16D, R. Cuneo ef al., ANP Quar, Prog. Rep. Mareh 10,
1953, ORNL-1227, p. 104; E. E. Ketchen and L. G.
Overholser, ANP Quor. Prog. Rep. June 10, 1952,
ORNL-1294, p. 89. :

PERIOD ENDING DECEMBER 10, 1953

Seven batches of NaOH containing approximately
400 g each have been prepared by filtration and de-
hydration during this quarter. The material obtained
contained from 0.02 t0 0.10 wt % of Na,CO, andless
than 0.1 wt % of H,0.

The Sr(OH)2 purified during this quarter con-
tains considerably less SrCO, thon the material
previously. prepared; the improvement resulted from
the replacement of a faulty filter., For the 12
batches produced, the SarCO3 contents range from
0.02 to 0.06 wt %, and the total olkalinities, as
determined by titration, range from 99.8 to 100.0%.

Reaction of Sodium Hydroxide with
Carbonaceous Matter

E. E. Ketchen L. G. Overholser

Materials Chemistry Division

A few experiments were run to check observations
made elsewhere which suggested that the No,CO
content increased when NaOH was heated at 700°C
ina sealed nickel container. If the increase occurs,
it may be due to the presence in the hydroxide of
carbonaceous matter capable of oxidationto NaéCOs.
Samples of commercial NaOH {about 1.5 wt %
Na,CO;) showed no significant
heating ot 700°C for 24 hours.

analytical

change aofter
However, the
results were not sufficiently precise
to detect very small chonges in this range of
carbonate concentration. When samples of purified
NaOH (one purified by filtration and one by de-
contation) for which occurate COJ ~analyses could
be obtained were heated in nickel for 24 hr at
700°C, slight increases in carbonate content were
observed., Both purified NaOH samples contained
0.02 wt % Na,CO, prior to heating. After heating,
the NaOH purified by decantation contained 0.03
wt % Na,CO, and that purified by filtration con-
tained 0.06 wt %. The increose observed could be
due to oxidation of the carbon present in the nickel
or the NaOH. Additional tubes were filled with
NaOH to which had been added 0.2 wt % graphite.
After these tubes were heated at 700°C for 24 hr,
preliminary onalyses indicated a marked increase
in carbonate content, This suggests that carbon
is oxidized by NaOH ot 700°C; if this is the case,
NaOH could not be maintained carbonate free at
high temperatures in metals which contain appreci-
able percentages of carbon,

63
ANP QUARTERLY PROGRESS REPORT

6. CORROSION RESEARCH

W, D. Manly
Metallurgy Division

W. R, Grimes

F. Kertesz

Materials Chemistry Division

H. W. Savage
ANP Division

The stotic and the seesaw corrosion testing
facilities have been used to study the corrosion of
Inconel, cermets, and some spzacial alloys in the
fluoride fuel NaF-ZrF -UF, (50-46-4 mole %), The
effects of improper degreasing of metals and of
contaminants in the fluorides were studied. Several
tests of bearing materials and hard-facing alloys
have been completed in which the molten fluorides
were used as the corroding agent. The materials
tested are of considerable interest because of their
prospective uses as high-temperature bearings and
valve facings. Various cermets have also been
tested in the fluorides to determine their fitness for
these same applications. Static tests in graphite
capsules of Inconel specimens exposed to fluoride
fuel to which FeF3, FeF,, or Nil'”'2 additions were
made gave anomalous results that were presumably
due to the reduction of the added metal flucrides
by the graphite, Static tests of itwo dissimilor
metals indicated that one metal is usvally prefer-
entially attacked,

Fluoride corrosion tests were made in [nconel
thermal convection loops to supplement and extend
the data obtained in the above-mentionsd and
previous static tests. Although the thermal con-
vection loops normally operate with a hot-leg
temperature of 1500°F, the temperatures were
varied to determine the effects of both time and
temperature. Additional tests have substantiated
the earlier conclusion that attack by NaF-ZrF4-UF
(50-46-4 mole %)} on Inconel is independent 01
temperature, at least in the range of 1500 to
1650°F, and dependent on exposure time. The
purity of the fluoride mixture was shown to be
important when the attack depth was increased by
a factor of 3 in a loop operated with a low-purity
batch of fuel. Reductions in attack, in comparison
with that found in the standard loop, that is, a
loop operated for 500 hr at 1500°F, were achieved
by (1) precirculating the fluorides, (2) pretreating
the fluorides with chromium metal flakes, and (3)
holding the fluorides with

in contact Inconel

64

turnings before circulation in the loop. While
pretreatment of the fluorides with metallic chromium
was beneficial, a chromium plate on the pipe wall
was not, Increasing the UF, content in the
flucride mixture from 4 to 4.5 mole % caused an
incregse in mass transfer in the loops and possibly
a slight increase in corrosion. The maximum depth
of attack by the purified ARE fuel solvent, NaZrFS,
was comparable to that of the purified fuel mixture
(with 4 mole % UF ), although the intensity of the
NGZst attack was much less and the same mass
transfer wos noted.

Additional studies were performed in an attempt
to establish the chemical equilibria of the postulated
corrosion reaction which results in the selective
leaching of chromium from Inconel. Preliminary
values have been obtained for the equilibrium con-
stants involved, but the data are notyet sufficiently
reliable,

Nickel and type 430 stainless steel thermal
convection loops were also operated with the
fluoride NcJF«ZrF"'-UF4 {50-46-4 mole %). There
was no measurable mass transfer in the nicke!
loop, and the hot-leg surface had a polished finish.
The hot leg of the type 430 stainless steel loop
showed a small amount of smooth, even removal,
with no subsurface voids or intergranular attack;
metal crystals were deposited in the cold leg.

A number of tests, both static and dynamic,
were performed with the liquid metals lithium,
sodium, and lead. Tests have been completed on
beryllium oxide and a number of bearing materials
and cermets.

The investigation of the mass transfer and
corrosion of metals and alloys in quartz thermal
convection loops containing liquid lead have
continved. The metals tested during this quarter
included o molybdenum-nickel (25% Me--75% Ni)
alloy, a chromium-iron-silicon (14% Cr—-84% Fe-2%
Si) alloy, and types 304 and 347 stainless steel
specimens which were oxidized prior to contact

with the lead, The most significant result from
this study was the marked improvement in the
resistance to mass transfer of these metais be-
cause of the oxide film. Some static tests on the
corrosion and mass transfer characteristics of
various aolloys in contact with lithium were com-
pleted. ‘

Static tests on a number of solid fuel elements
were performed in sodium and sodium hydroxide.
While there was no evidence of attack in any of the
sodivm tests, all the fuel elements were severely
attacked in the hydroxide. In connection with a
study of hydroxide containment, the equilibrium
pressures of hydrogen over a number of hydroxide-
metal systems are being determined,

FLUORIDE CORROSION IN STATIC AND
SEESAW TESTS

E. E. Hoffman L. R. Trotter
J. E. Pope D. C. Vreeland
Metallurgy Division

H. J. Buftram R. E. Meadows
No V. Smifh

Materials Chemistry Division

Both the static capsule and the seesaw tests
provide relatively cheap and simple means of
investigating the many parameters which affect
fiuoride corrosion. The seesaw tests in which the
capsule contdining the fluoride is rocked in a
furnace were operated at 4 cps with the hot end of
the capsule at 815°C and the cold end at 700°C.
The static tests in which the fluoride mixture is
sealed in a capsule were vsually conducted for
100 hr at 816°C. In order to permit testing of
material which could not be readily fabricated into
capsules or to permit variation of the surface-to-
volume ratio of fluoride to metal, a number of tests
were performed in which the fluoride and metal
specimens were contained in graphite capsules.
All capsules are filled in a helium dry box, and
therefore oxidation or hydrolysis of the contents of
the capsules is improbable,

Inconel Corrosion by Fluorides with Metal
Fluoride Additives

In general, corrosion of Inconel specimens by
several fluoride preparations in graphite capsules
was similar to that obtained when inconel copsules
were used in static tests. Some carburization of
the specimens was obtained, but no other special
effects were observed. | :

PERIOD ENDING DECEMBER 10, 1953

However, when fluorides to which had been
added considerable quantities of FeFS, Fer', or
NiF , were tested in graphite capsules containing
inconel specimens, rather surprising results were
obtained, In each case, tests at 800°C showed
extensive deposition of iron or nickel, usually
in films on the graphite capsule and on the speci-
men. When Fef_ was the additive (in amounts up
to 5 wt %), the corrosion of the Inconei specimen,
as observed by metallographic examination and by
analysis of the final melt for chromium, was much
iess than expected. In similar tests with FeF_,
corrosion of the Inconel was virtually undetectdb?e
by either technique. |n this case, the precipitated
film of metal appeared to be protective.

When NiF_ was the added material, however, no
protection of the specimen by the film resulted,
in these experiments, exiremely heavy subsurface
void formation to a depth of 9 mils wes observed.
There is evidence that corrosion by fluoride mix-
tures with NiF added is more severe at 600 than
at 800°C.

These experiments suggest that graphite or some
unknown contaminant in the graphite is responsible
for reduction of the added metal fluoride. While
reduction of FeF_, for example, by graphite should
have an extremelzy iow equilibrium constant, it is
possible that the volatile reaction product (CF“)
escapes from the system and permits the reaction
to proceed, It is also possible that there is
sufficient oxygen or water adsorbed in or on the
graphite so that reduction of oxide by carbon (o
reaction with an appreciable equilibrium constant
at these temperatures) is responsible for these
effects. Further study of this and similar reactions
is in progress., ' '

Corrosion of Yarious Metal Combinntions

A series of experiments in which various metals,
either alone or in combination, were tested in
molten fluorides in graphite capsules have vielded
the following results. (1) Nickel alone shows no
attack. (2} Chromium-plated lncone! shows heavy
subsurface void formation ‘at all temperatures
above 600°C; the voids were 3 to 4 mils deep at
700°C and above. (3) Chromium metal yielded high
values for chromium content of the melt; it is
possible, however, that slight surface oxidation of
the chromium pellets used was responsible for this
dissolved material, (4} When Incornel and chromium
specimens were exposed fogether, the chromium

65
ANP QUARTERLY PROGRESS REPORT

test piece was nearly completely dissolved, while
the Incone! specimen, which was covered with a
thin metal layer, was almost completely unattacked.
(5) However, when Inconel and nickel were exposed
together, the nickel appecred unaffected, while
the Incone! was badly attacked; at 700°C, for
example, the heavy subsurface void formation
reached a depth of 3 to 4 mils.

Corrosion of Cermets

Several cermets which were considered for
bearings and hard-facing applications were tested
for resistance to fluoride corrosion. Three of these
ceramicemetal materials were subjected to static
tests. The materials consisted of heavy metal

carbides bonded by nickel and were cbtaired from
the Firth Sterling Company. Table 6.1 presents
results obtained when the cermets were exposed to
two fluoride mixtures in graphite capsules. The
results indicate fair resistance to the two fluoride
mixtures under the conditions of the $est, and
there wos no apparent change in the appecrance of
the specimens after testing.

Another set of specimens was subjected to see-
saw tests in Incone!l capsules for 100 hr, with the
specimens restricted to the hot zone of the Inconel
tube. The hot-zone temperature was 816°C, and
the cold-zone temperature was 730°C. Results of
these tests are presented in Table 6,2. Some of
these specimens contained chromium as a binding

TABLE 6.1. RESISTANCE OF CERMETS TO TWO FLUORIDE MIXTURES

Static tests at 800°C for 100 hr in graphite containers

 

 

 

CERMET FLUORIDE COMPOSITION WEIGHT CHANGE
(wt %) (mole %) (%)
TiC-CryCo-Ni (42.9-7.1-50.0)* NaF-ZrF, (50-50) .12
NaF-ZrF ~UF, (53-43-4) +1.27
TiC-Mo,C-Ni (72-5-23) NaF-Z:F ; (50-50) +0.047
NaF-ZrF ,-UF, (53-43-4) ~0.005
CryC-Ni (89-11) NaF-ZrF  (50-50) 40.152
NaF-ZeF UF, (53-43-4) *0.153

 

 

 

*

Firth Sterling 27.

TABLE 6.2. RESULTS OF SEESAW TESTS OF VARIOUS CERMET MATERIALS TESTED IN
N(fl"’-ZrF“--UF4 (50-46-4 mole %)

 

 

MATERIAL

COMPOSITION (wt %)

METALLOGRAPHIC NOTES

 

Metamic L. T-1 (not heat treated)

Metamic LT-1 (heat treated)

Firth Sterling 27

Kenametal 151A

 

CrAl, 0, (77-23)
Cr-A1,0, (7723)

TiC2 with Mi-Cr binder*

TiC2 with 20% Ni

Specimen completely penetrated

Specimen complztely penstrated

Attack zone 2 to 5 mils deep, 3

mils in most ploces

Slightly affected zone 1 tc 2
mils deep

 

 

*C#., Table 6.1.

66
agent, and, os expected, these materials were
attacked by the fluorides. Metamic L.T-1, in hoth
the heat-treated and unheat-treated conditions,
was attacked severely. While the attack on the
Firth Sterling 27 and on the Kenometa! 151A in the
seesaw tesfs was only modérofe, the increased
severity of attock in the seesaw tests compared
with that in the static tests is evidenced by the
test data on Firth Sterling 27.

Inconel with Dil and Trichloroethylene Additives

Several thermal convection loops had more than
average depths of attack which were thought to be
due to improper degreasing of the loops prior to
filling. Therefore, a static test was run in which
six drops of oil and six drops of trichlorcethylene
were added to an Inconel capsule containing an
inconel specimen and the solution was allowed to
evaporate in an attempt to duplicate improper de-
greasing. This capsule was then loaded with the
fluoride mixture NoF<ZrF -UF , (50-46-4 mole %)
and tested for 100 hr at %16"& under static cons
ditions, along with a control capsule for com-
parison. The specimen in the control capsule had
greater weight ioss and depth of attack than the
specimen contaminated by the evaporation of the
trichlorcethylene-cil solution. |t appears that
under the conditions of this test, the contamination
was not detrimental and might even have been
beneficial.

Inconel Corrosion by Fluorides with
MoS$, Additive

Molybdenum disulfide is being used as a Swage-
lok lubricant in fluoride production systems, and
therefore it was pertinent to determine whether
increased corrosion of the Inconel tubing would
result if some of the MoS_ were to contaminate the
fluoride. Accordingly, two seesaw tests of the
fluoride mixture NuF-ZrF[UFA (50-46-4 mole %) in
Inconel tubing were run for 100 hr. In one of these
tests, an addition of 3 wt % M052 was made to the
fluoride; the other test was run with no addition to
fluoride.

Metallographic examination of sections of tubing
from hot and cold zones of both tests showed that
more severe attack had taken place in the test with
the MoS,, addition. The cause of this increased
corrosion is not known. The type and extent of
attack both with and without Mr.aS2 additions .is
shown in Fig. 6.1. ‘

PERIOD ENDING DECEMBER 10, 1953

Screening Tests of Metallic Bearing Materials

Several tests have been completed in the screen-
ing program on materials for potential use in bearing
and hard-facing applications. Materials for these
uses must be resistant to corrosion by molten
fluorides. One set of specimens consisted of
Hastelloys B, C, and D, and Stellites 1, 6, 12, 21,
25, 40, and 41 deposited on pieces of type 347
stainless steel. These composite specimens were
tested in the fluoride mixture NaF-ZrF -UF,
(50-46-4 mole %) for 100 hr at 816°C under static
conditions. In these tests, it appeared that
Hastelloys B, C, and D and Stellites 21, 25, and
40 satisfactorily resisted attack by this fluoride

mixture. Stellites 1, 6, 12, and 41 appeared to be

attacked severely. The appearance of Stellite 40
after the test is shown in Fig. 6.2; the attack is
limited to a depth of 4 or § mils. As reported
previously, it is apparently the carbides in the
Stellites which are preferentially attacked by the
fused fluoride salts, and it appears that the
Stellites of low carbon content are not so severely
attacked. The results of these tests are presented

in Table 6.3.

FLUORIDE CORROSION OF INCONEL IN
THERMAL CONVECTION LOOPS

G. M. Adamson
Metallurgy Division

The use of thermal convection loops for determi-
ing dynamic corrosion by liquids has been previously
described.! Unless otherwise stated for the tests
described in the following sections, the tempera-

ture of the hot leg of the loop was maintained at

1500°F and the temperature of the uninsulated
cold leg was approximately 1300°F. With the
fluoride salts, this temperature difference results
in a fluid velocity of about & to 8 fpm. The usual
testing period is 500 hours.

Effect of Fluoride Batch Purity

Pilot-plant batches of the fluoride fuel mixture
NaF-ZrF4-UF4 {50-46-4 mole %) which receivgd a
24-hr gas purification instead of the customary
8 hr or less are now available for thermal ‘con-
vection loop work. The longer gas purification
results in a reduction of the hydrogen fluoride
content of the melt. Corrosion results from three

 

'D. C. Vreeland et al., ANP Quar. Prog. Rep. March
i0, 1953, ORNL-1515, p. 121. -

67
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

 

 

 

Fig. 8.1. Corrosion of Inconel after a 100-hr Seesaw Test ot 815°C in the Fluoride NaF-ZrF ,-UF,
(50-46-4 mole %) with and Without3 wt% Mo5, Additive. a. Tested with no additive. 250X. b. Tested
with 3 wt % Mo5, added. 250X,

68
PERIOD ENDING DECEMBER 19, 1953

TABLE 6.3. CORROSION OF HARD-FACED TYPE 347 STAINLESS STEEL SPECIMENS EXPOSED TO

Na F- ZrF -UF, FOR 100 hr AT 816°C UNDER STATIC CONDITIONS

 

HARD-FACING
MATERIAL

COMPOSITION (wt %)

METALLOGRAPHIC NOTES

 

Hastelloy
B

Stellite
1

12

21

 

Mo, 26 to 30
Fe, 416 7
C, 0.12 mox

Mi, balance

Mo, 16 to 18
Fe, 4.5t0 7

C, 0.15 max
Cr, 15,510 17.5
W, 3.75 to 5.25
N

Cu, 3
Al, 1
Si, 10
Mn, 1
Ni, balance

C, ?

, balance

Cr, 30.5
W, 12.5
Fe, 3 max
C, 2.5
Ni, 1

- Co, balance

Cr, 27.5

W, 4

Fe, 2 to 3 max
G 1

Co, balance

: Cr; 29'5

W, 8.25

© Fe, 2 to 3 max

: cl 104

Ni, 1
Co, balance

Cr, 25 to 30
Ni, 1,5 tc 3.5

_MOJ 4.5 10 6.5

- Fe, 2 max

Co, balance

C‘ 0-2 to 0.35

 

A few voids to a depth of (.5 mil
A little spalling and a few voids to a

maximum depth of 1 mil

Subsurface voids to a depth of 1,5 mils

Many voids to o depth of 5 to 10 mils

. Erratic attack up to 15 mils deep along

carbides in grain boundaries
Erratic intergranular attack 0 to 13

mils deep

A few voids to u depth of 0.5 mil

69
ANP QUARTERLY PROGRESS REPORT

 

 

TABLE 6.3, (continued)
ARD-FACING
HAR COMPOSITION (wt %) METALLOGRAPHIC NOTES
MATERIAL
Stellite
25 Cr, 19 to 21 A few voids to a depth of 0.5 mil
C, 0.15
Ni, 7 to 11
W, 14 tc 16
Si, 1
Mn, T to 2
Fe, 2
Co, balance
40 Composition not available Voids to a depth of 3 to 5 mils
41 Composition not available Heavy, erratic intergranular attack
4 tc 23 mils deep

 

 

 

 

 

 

 

 

Fig. 6.2, Stellite 40 After Static Testing in NaF-ZrF ,-UF , (50-46-4 mole %) for 100 hr ot 816°C,

incone! loops operated with this material showed
general attack to a depth of 3 mils and some
intergranular penetration to a depth of 6 mils. The
importance of purity of the materiols is evident
from these tests, since the average maximum pene-
tration in standard production batches has been

70

9 mils. The reduction in attack, shown on Fig.
6.3, was also reflected in the chromium pickup,
which was down to 450 to 600 ppm, that is, less
than one half the previous contamination. Some
reduction was also obtained in the amount of iron

present.
 

Unclassified
T~3480

 

 

 

E Unclassiiied 3
 rass §

 

 

 

   

Fig. 6.3. EEffect of Fluoride Batch Purity on the
Corrosion of Inconel Exposed in Thermal Convection
Loops to MaF-ZrF -UF, (50-46-4 mole %) for 500
hr ot 1500°F. o. Exposed to standard production
batch, 250X. b. Exposed to high-purity batch.
250X, Reduced 17%.

Fluoride fuel mixtures with high hydrogen fluoride
content were circulated in a number of loops.
Increased corrosion, as indicated by heavy attack
and maximum penetrations of 14 and 15 mils, was
found under standard operating conditions.

Effect of Chromium Additives

Data were presented previously? which indicated
doubtful improvement in corrosion characteristics
when chromium powder was added to the flueride

PERIOD ENDING DECEMBER 10, 1953

mixtures. However, in 0 new fest, in which the
chromium was added in the form of coarse metallic
flakes, a reduction in the depth of attack was
found. The fluoride mixture was held in contact
with the chromium flakes overnight at a temperature
of 1300°F before being charged into the loop.
Examination subsequent to operation of the loop
showed a maximum penetration of only 4 mils, as
compared with penetration of the control loop 1o a
depth of 15 mils. As in previous tests with
chromium additions, a thin, continuous, metallic
deposit was found in the cold leg of the loop.

One thermal convection test was run in which
the chromium was added in the form of an electro-
plated layer on the inner surface of a G-in. length
of Inconel pipe. This pipe was welded into the
top of the hot leg of a standard loop. After oper-
ation of the loop, no trace of the chromium could
be found .in the hot leg by metallegraphic exami-
nation. Typical subsurface void attack was found
in the Inconel, with a maximum penetratfon of
10 mils. No data are available on the hydrogen
fluoride content of the batch circulated in this
loop. The 10 mils of attack is slightly higher
than the average, but no change in attack could be
seen on opposite sides of the hot leg where the
section was welded into the loop. Scattered
metallic particles were found on the cold-leg wall.
This test confirms the earlier finding that, while a
direct chromium addition to the fuel may result in a
beneficial effect, a chromivm plate on the pipe
wall does not offter any protection in o dynamic
system. :

Although large additions of chromium were made
to the fluoride mixiures in both these loops, the
chromium - concentration in the fluoride mixtures
after circulation in both loops was uniform and was
only 1600 ppm. A study of the chemical analyses
of fluoride mixtures from other loops showedthis
concentration to be the maximum reported. Attempts
have been made by the Materials Chemistry Di-
vision to obtain apparent sclubilities of chromium
of over 1% ot 800°C in this fluoride mixture, but it
is speculated that 1600 ppm is an equilibrium
concenirafion.  Attempts fo obtain this concen-
tration in the laboratory have not yet been success-

ful.

 

2G.. M. Adamson, ANP Quar. Prog. Rep, June 10, 7953,
ORNL‘]SSér P- 60- :

71
ANP QUARTERLY PROGRESS REPORT

Pretreatment of Fluoride with lnconel

Some reduction in corrosive attack was obtained
by first holding the fluorides in contact with
Inconel turnings for 100 hr at 1300°F. After
subsequent operation of the convection loop,
examination revealed the maximum depth of attack
of the Inconel to be 11 mils, as compared with a
maximum depth of 14 mils in the control tests.
This pretreatment, although it apparently does not
have as beneficial an effect on the corrosion
properties as does the chromium addition, has the
advantage that it prevents the formation of o layer
on the cold-leg wall. It is likely that larger
reductions in attack would be obtained with higher
treatment temperatures, although it is not come
venient to do this with existing equipment.

Effect of Exposure Time

Twe series of Incone! loops have been filled
and are now being operated to determine the
effect of time on corrosion and mass transfer.
One series is circulating the fluoride fuel NaF-
ZrFA-UF4 (50-46-4 mole %), and the other is
circulating the fuel solvent NoZrF_; periods of
operction have ranged from 500 to 5000 hours.
Both series of loops were filled with especially
purified fluorides. Metallographic results are now
available for the loops which circulated fuel for
500 and 1000 hours.

Moderate to heavy subsurface void formation
with @ maximum penetration of 8 mils was found in
the hot leg of a loop that operated for 1000 hours.
The voids tended to concentrate more in the grain
boundaries than they do in standard loops. An
intermittent very thin layer that appeared to be
metallic was noted in the cold leg, and a thin,
metallic ring of chromium metal was visible in the
cold trap. This is the first Inconel loop circulating
the NaF-ZrF4-UF4 fuel in which a definite cold-leg
layer has been found.

The standard loop that opesrated for 500 hr
showed moderate to heavy attack with @ maximum

penetration of 6 mils. No metallic deposit was

visible in the cold leg or in the trap of this loop.

A dark layer was found in the trap, but it was not
metallic. Comparison of the results from these
two loops helps to confirm the previous observation
that a decrease in rate of attack is apparent with

longer operating times; however, the attack does

not stop completely. The increase in the metallic
deposit on the cold leg and in the trap points to

72

mass transfer as the cause of the continuing

attack.

Effect of Temperature

Additional confirmation of the conclusion pre-
sented in the previous report® that attack is
independent of hot-leg temperature from 1400 to
1650°F has been obtained. A loop operated for
1000 hr with a hot-leg temperature of 1650°F
showed attack comparable to that of another loop
which operated for 1000 hr at 1500°F. Both loops
were filled with fluoride fue! of comparable purity,
and the attack in both was moderate, with a maxi-
mum penetration of 8 mils. Furthermore, the voids
in both loops were similar, although they were
larger and more concentrated at the grain boundaries
than they were in similar loops operated for only

500 hours.

Effect of Surface-to-Yolume Ratio

LLarge differences have always been observed
between the corrosion results obtained in seesaw
tests and those obtained in convection loops. One
of the main variations is in the surface-to-volume
To determine the effect of this variable, a
loop was constructed from 1-in. tubing instead of
from the standard ]/2-in.-IPS pipe. The lecop with
the l-in. tubing has a surface-to-volume rotio of
4.5 in.2/in.3, while that of a standard loop is
6.5 in.2/in.3. The attack in the hot-leg of the
loop with 1«in. tubing was light to moderate, with
a maximum penetration of 9 mils. A standard loop
filled from the same fuel batch showed similar
attack, with a maximum penetration of 5.5 mils.
The percentage in ottack depth was
about the same as that for the decrease in the
surface-to-volume ratios. This increase in attack
depth with a decrease in surface-to-volume ratio
again points to impurities in the fluorides as the
major factor in corrosion.

ratios.

increase

Fluoride with 6.5 mole % UF4

Tests have been completed on two Inconel loops
circulating the fluoride fuel mixture with the
highest UF, concentration contemplated for use in
the ARE, that is, MNoF-ZrF -UF, (53,3-40.0-6.5
mole %). For the first test, which was run to
provide preliminary data, the loop was not thoroughly

 

3G. M. Adamson, ANP Quor. Prog. Rep. Sept. 10, 1953,
ORNL'1609, p- 77.
cleaned, and the fuel preparation was not of
optimum purity. The preliminary test loop was
operated for 350 hr at 1500°F, and, upon exami-
nation, was found to have moderate to heavy attack
in the form of subsurface voids to a maximum depth
of 12.5 mils. A very thin metallic deposit was
found in the cold leg. The second test loop was
cleaned by circulating NaF-ZrF,, and it was
charged with high-purity fuel. After 500 hr of
operation, this loop showed heavy attack to «
maximum penetration of 7 mils, as shown in Fig.
6.4. Again, a thin, metallic layer was found in
the cold leg. Two more standard Inconel loops are
now being operated with this high-purity fuel
mixture. The preliminary tests indicate that the
higher uranium content of this fuel may cause a
small increase in the depth of attack, but additional
tests are needed for confirmation. The higher
vranium content also causes an increased amount
of mass transfer, as shown by the thin, metallic
layer in the cold leg.

The Fluoride NaZrFS
One batch of NquFs, the ARE fuel solvent, was

tested in a series of six Inconel loops. Upon
examination, the loops showed light to moderate
attack, with maximum penetration ranging from 4 to
8 mils, and the fluoride was found to contain
average impurities, that is, 230 ppm of Cr, 55 ppm

 

E Unclossified
i 4229

 

 

 

 

 

 

   

Fig. 6.4. Corrosion of Inconel Thermal Con-
vection Loop by High-Purity NaF-Z¢F -UF, (53.5-
40.0-6,5 mole %) After 500 hr ot 1500°F. 250X,
Reduced 15%.

PERIOD ENDING DECEMBER 10, 1953

of Fe, and less than 20 ppm of Ni. A typical
corrosion specimen taken from the hot leg of one
of these loops is shown in Fig. 6.5. The voids
are larger, and they are more definitely concen-
trated in the grain boundaries than they were in the
tests with the uranium-bearing fluoride mixture.
The low chromium content of the fluoride after
circulation in these loops is not understood be-
cause, even though the attack was less, it was not
decreased so much as was the chromium content.
There was a gray layer above the freezing line in
the trap of each of these loops that was about: | to
Y in. in thickness. This material was identified
petrographically as fine, needle-like, metallic
crystals of chromium metal dispersed in the fluoride.
Some mass transfer had, therefore, taken place even
with the low chromium concentrations found in
these loops.

FLUORIDE CORROSION OF NICKEL AND
STAINLESS STEEL LOOPS ‘
G. M. Adamson
Metallurgy Division
A small amount of mass transfer was previously
observed® in a nickel loop in which the fuel NaF-

ZrF -UF, (50-46-4 mole %) was circuloted. To

4
confirm tAhiS single test, another nickel loop was

 

4G, M. Adamson, ANP Quar. Prog. Rep. June 10, 1953,
ORNL 1556, p. 63-64.

 

Unclassified
T 4075

MILS

 

 

 

Fig. 6.5. Corrosion of Inconel Thermal Convection
Loop by NaZrF, After 500 hr ot 1500°F. 250X.
Reduced 12%. : :

73
ANP QUARTERLY PROGRESS REPORT

cleaned with NaZrF_ and operated with high-purity
fuel. After circulation of the fuel, no metallic
crystals were found in the flucrides in the cold leg
or in the trap nor could any layer be identified on
the cold-leg surface. The surface of the hot leg
was bright and pelished, and there was no evidence
of subsurface voids or intergranular attack. The
wall thickness was within commercial tolerances,
but it was again on the low side. lt seems likely
that a small amount of even removal had taken
place, since the surface was polished; however,
this seems to be the smallest amount of mass
transfer yet found.

One type 430 stainless steel loop operated with
the fuel NaF-ZrF -UF , (50-46-4 mole %) for 500 hr
at 1500°F with no sign of plugging. However, a
preliminary metallographic examination revealed
unexpected even surface removal in the hot leg
rather than the subsurface voids in the grain
boundaries that were found in the type 316 stainless
steel loop. The hot-leg surface was smooth, with
no subsurface voids or intergranular type of attack.
(Carbides appear to have been leached (or possibly
to have gone back into solution) from a zore about
2 mils wide on the surface, and considerable grain
growth had taken place in this zone. The outer
surface of the pipe was quite rough, and therefore
the wall thickness measurements are questionable,
Variations in thickness of from 0,106 to 0,109 in.
were found. Since the specified thickness of the
wall was 0.109 in., it is evident, at least, that no
large omount of even removal occurred. However,
some removal had taken ploce, as shown by the
scattered metallic erystals adhering to the cold-leg
surface.

LIQUID METAL CORROSION

W. H. Bridges J. E. Pope
J. V. Cathcart G. P. Smith
E. E. Hoffman L. R. Trotter

D. C. Vreeland
Metallurgy Division

Mass Transfer in Liquid Lead

Investigations of mass transfer and corrosion of
various metal specimens in small, quartz, thermal
convection loops containing liquid lead have been
tontinued. Details of the construction and operation
of the loops and the results obtained for Inconel,
columbivm, molybdenum, types 304, 347, 410, and
446 stainless steel, Armco iron, nickel, chromium,

74

Nichrome V, and a nickel-iron (30% Ni-70% Fe)
alloy were reported previously.® The metals tested
during this quarter included a molybdenum-~nicke!
(25% Mo~-75% Ni) alloy, a chromium-iron-silicon
(14% Cr—84% Fe~27% Si) alloy, and types 304 and
347 stainless steel specimens which were oxidized
prior to contact with the lead.

The molybdenum-nickel alloy loop was operated
for 672 hr with hot- and cold-leg temperatures of
820 and 480°C, respectively. Plugging did not
but o small omount of mass-transferred
material was found in the bottom of the cold leg.
The extent of the corrosive attack may be seen
in Fig. 6.6. A spectroscopic analysis showed
that the mass-transferred material consisted al-

occur,

most entirely of nicke!l, the concentration of mo-
lybdenum being less than 0.01%. The results for
this loop were comparable to those previously
obtained for the 400 series stainless steels.

The loop containing chromium-iron-silicon alloy
specimens was operated for 550 hi with hot- and
cold-leg temperatures of 805 and 490°C, respec-
tively. Again, the mass transfer which occurred
was not sufficient to stop circulation, but a plug
was starting to form in the cold leg. The amount
of mass-transferred material was slightly greater
than that found in loops containing the 400 series
stainless steels. Rather severe intergranular at-
tack to a depth of 12 mils was found.

The effect of an oxide coating on the metal
specimens was also studied. Sections of types
347 and 304 stainless steel tubing were oxidized
until they showed interference colors corresponding
to oxide thicknesses of 500 to 1000 A. These test
specimens were then tested in quartz loops, and
marked improvement in resistance to mass transfer
was noted with both metals. A loop containing
oxidized, tubular, type 347 stainless steel speci-
mens was operated for 550 hr with hot- and cold-
leg temperatures of 815 and 510°C, respectively.
The loop was terminated on schedule and only a
small amount of mass-transferred material was
found in the cold leg. When the type 347 stainless
steel specimens were used in the as-received
condition, a plug formed in the loop within 140
Although uniform intergranular attack was
surface of the oxidized

hours.

observed on the inner

SANP Quar. Prog. Rep., ORNL-1515, p. 128; op. cit.,
ORNL-1556, p. 64; op. cit., ORNL-1609, p. 80; Met. Div.
Semiann. Apr. 10, 1953, ORNL-1551, p. 17.
 

PERIOD ENDING DECEMBER 10, 1953

 

| Unclassified
| Y.10393
S
[~ Q
@
%
o
O
S
0
e
O
L0
X
. &
=
[y
<
O
| <

 

 

Fig. 6.6, Corrosion of a Molybdenum-Nickel {25% Mo~75% Ni) Afloy Specimen in a Quartz Thermal Con-
vection Loop After Circulation of Lead for 672 hr at 820°C. 500X,

tubular specimen, much less corrosion was ob-
served on the outer surface. The attack appeared
to have started on the outer surface of the specimen
at isolated points along the surface, perhaps cor-
responding fo breaks in the oxide film, and to have
spread out in fan-like fashion (Fig. 6.7). Cor-
rosion of the specimen in the cold leg of the loop
was much less pronounced.

A type 304 stainless steel loop containing previ-
ously oxidized specimens was also operated for
550 hr without the formation of a plug large enough
to stop circulation. The hot- and cold-leg temper-
atures were 800 and 550°C, respectively. Metallo-
graphic examination of this loop is incomplete.
In previous tests with type 304 stainless steel
specimens in the as-received condition, the loops
had to be terminated within 100 hr because of plug
formation. :

Static Tests of Beryllium Oxide in Sodium,
Lithium, and Lead

Specimens of beryllium oxide, approximately
by 14 by 1{‘ in., cut from 2-in. hexagonal blocks
hot pressed by the Norton Compony were tested
in sodium, lithium, and lead. One haif the speci-
mens were tested as cut from the original block,
and one half were tested after having been refired
for 2 hr at 1800°C. The refired specimens had
higher density ond lower apparent porosity than

those specimens which were not refired. ' The
sodium was contained in inconel tubes, cmd the
lithium and the lead were contained in Armco iron
tubes. The test time was 100 hr, and the temper-
ature was 538°C. The weight ond dimension
changes after test were less than 1.5%. ' The
weight gains in sodium and lithium were presumed
to be due to absorption of these metals. = The
specimens tested in lead appeared to be affected
the least. Edges of the specimens appeared to
be just as sharp after test as before. The refired
beryllium oxide specimens apparently had greater
resistance to aftack than did the specimens that
were not refired. The one exception appeared in
the sodium test, and, even in this fest, the refired
specimen was more dimensionally stable.

Spinner Tests of Inconel and Type 405
Stainless Steel in Sodium

A spinner test was run with Inconel-/ in sodium,
but the apparatus was not rotated. This test was
run to check the possibility that attack noted in
a previous spinner test might be due to impurities
in the sodium or in the atmosphere rather than to
the velocity of test. However, much lighter attack
and less weight change was noted on these speci-
mens than on the Inconel-X specimens previously
tested at fluid velocities of 405 fpm; therefore it
is evident that most of the attack in spinner tests
is due to velocity effects.

75
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

Fig. 6.7. Corrosion on Quter Surface of a Tubular Type 347 Stainless Steel Specimen Exposed to Cir-
culating Lead in a Quartz Convection Loop for 550 hr ot 815°C, 500X,

A spinner test with type 405 stainless steel
(12% Cr) has also been run at a velocity of 405
tpm. These tests were conducted at a temperature
of 816°C for 100 hr in sodium. The weight change
was of the order of +0.5%, and the attack was
intergranular to a maximum depth of 3 mils, with
surface roughness to approximately 1 mil,

Static Tests of Bearing Materials in Sodium,

Lithium, and Leod

A number of metal alloys of potential value in
bearing and hard-facing applications in liquid
metals have been tested in sodium, lead, and
lithium under static conditions for 100 hr at 816°C.
The results of these screening tests are given in

Tabie 6.4,

Static Tests of Stainless Steels in Lithium

A series of static tests of various stainless
steels in two types of lithium (helium packed and
oil packed) have been run for 100 hr at 816°C, In
this group of tests, types 310 and 446 stainless
steel showed the least resistance to attack by
fithium, while type 347 stainless steel showed the
best resistance. Types 304, 309, 316, and 347
stainless steel showed very little or no subsurface
attack. Crystals of what were identified by x-ray
diffraction as a chromium carbide phase (DPH
1600 to 1100) were detected on the exposed sur-

76

face of all the austenitic stainless steel speci-
mens, as shown in FFig. 6.8. These crystals were
more numerous and slightly larger on the speci-
mens tested in the oil-packed lithium, as would
be expected because of the higher carbon content
of this material. Chemical analyses of the lithium
baths after testing revealed no significant changes
in the iron, nickel, and chromium contents of the
oil-packed and helium-packed lithium. In the tests
in which weight changes were noted, the weight
changes of the specimens in helium-packed lithium
were less than those of the specimens in oil-
packed lithium, possibly because of the higher
purity of the gos-packed lithium. Results of these
tests are summarized in Table 6.5,

Static Tests of Solid Fuel Elements in Sodium
and Sodium Hydroxide

Core materials (UOz) that had been clad with
nickel, copper, and types 304, 316, and 347 stain-
less steel were tested in sodium and sodium hy-
droxide for 100 hr at 816°C under static conditions.
These specimens were prepared to the specifi-
cations given in Table 6.6.

Specimens clad with nickel and types 304, 316,
and 347 stainless steel were annealed for 15 min
at 1000°C in hydrogen; copper-clad specimens
were annealed for 15 min at 450°C in hydrogen.
These specimens were approximately 1 by 0.35
PERIOD ENDING DECEMBER 10, 1953

TABLE 6.4. STATIC TESTS OF VARIOUS ALLOYS IN SODIUM,
~ LITHIUM, AND LEAD FOR 100 hr AT 816°C

 

 

CORRODANT

 

MATERIAL COMPOSITION METALLOGRAPHIC NOTES
Hastelloy '
B Mo, 26 to 30 Sodium Some voids to 0.5 mil in a few places
Fe, 4.7 Lead 0 to § mils of erratic alloying type of attack
C, 0.12 mux :
. Lithium 5 mils of subsurface voids
Ni, balance .
C Mo, 16 to 18 Sodium Some voids in a few places to 0.5 mi!
Fe, 4.510 7 Lead Erratic attack; practicolly complete penetration
C, 0.15 max. in some places, very little in others
Cr, 15.5 t0 17.5
W. 3.75 to 5.25 Lithium 4 to 8 mils of subsurface voids
Ni, balance
D Cu, 3 Sodium Layer of subsurface voids 1 to Z mils deep on
Al 1 some portions of specimen
Si, 10 Lead Specimen appearedto be completely penetrated
Mn, 1 '
Ni ’ balance Lithium 4 to & mils of subsurface voids
c?
Stellite
1 Cr, 30.5 Sodium Apparestly no attack
W, 12.5 Lead 2 mils of voids and intergranular attack
Fe, 3 max
C, 2.5 Lithiom 0.5 to 1 mil subsurface voids where carbides
Ni 1 were attacked
Co, balance
6 Cr, 27.5 Sodium Yery slight intergranular penetration, less than
W, 4 0.5 mil
Fe, 2 1o 3 max Lead Intergranular attack 1 to 2 mils deep
c, 1 "
Co. balance Lithium 4 to 6 mils of subsurface voids where carbides
. :
were attacked *
12 Cr, 29.5 Sodium Very slight intergranular penetration, less than
W, 8&25 0.5 mit
231124"‘ 3 max Lead Apparently no attack
N'i ; Lithium Less than 0.5 mil of attack on carbide phase
Ceo, balance
21 Cr, 25 10 30 Sodium Apparently no attack
Ni, 1.5 to 3'5 Lead A slight roughening and intergranular pene-~
Fe, 2 max : tration to about 0.5 mil
Mo, 4.5 to 6.5
l.ithium 0.5 mil of small subsurface voids

 

C, 0.2 10 0.35

Co, balance

 

 

77
ANP QUARTERLY PRGOGRESS REPORT

TABLE 6.4 (continued)

 

 

 

MATER!AL COMPOSITION CORRODANT METALLOGRAPHIC NOTES
Stezlsl’”e Cr, 19 to 21 Sodium Some veoids and intergronular attack to 1 mil
C, 0.15 maximum in a few places

Ni, 910 T Lead Yoids to about 0.5 mil
W, 1410 16
$i 1 Lithium 2 mils of subsurface voids in most places,
¢
Mn, 1 to 2 10 to 12 mils in several places
Fe, 2
Co, balance
40 Compesition not Sodium
avoilable Lead Yery heavy penetration to about 50 mils
l.ithium Subsurface voids to a depth of 10 mils
41 Compeoesition not Sodium Apparently no attack
available lead Apparently complete penetration
Lithium Subsurface voids to a depth of 10 to 15 mils

 

 

 

 

 

 

 

 

 

Fig. 6.8. Type 316 Stainless Steel After Static Testing in Lithium for 100 hr at 816°C. Note crystals
attached to surface. 1000X. Specimen copper plated to preserve edge during polishing.

78

 
by 0.050 inch. The thickness of the core and of
each of the clodding layers was approximately
16 mils. | '

in sodium, the specimens clad with nickel and
types 304, 347, and 316 stainless steel showed
no evidence of attack. The sample clad with
copper had a slightly irregular edge and showed
weight and dimensional increases after testing.
The many, copper, mass-transferred crystals at-

PERIOD ENDING DECEMBER 10, 1953

tached to this specimen were probably caused by
temperature gradients during the test. In the tests
in sodium hydroxide, the copper- and nickel-clad
specimens, were unattacked, and the types 304,
316, and 347 stainless steel-clad materials were
attacked more or less severely, that is, from 5 to
18 mils. The type 347 stainless steel-ciad speci-
men tested in sodium hydroxide is shown in Fig.

6.9.

TABLE 6.5 STATIC CORROSION OF STAINLESS STEELS TESTED IN TWO TYPES
OF LITHIUM FOR 100 hr AT 816°C

 

 

 

 

 

 

TYPE OF WEIG_HT.CZIANGE ANALYSIS METALLOGRAPHIC NOTES
STAINLESS STEEL {g/in.%) {wt %) .
304 {a)* 0 Fe, 0.024 Small erystals (0.5 mil thick) attached to
Ni, 0.017 surface; some of this same phase concen-
Cr, <0.001 trated in grain boundaries near surface
{b)** 0 Fe, 0.026 Same as obove, with surface of specimen
Ni, 0.021 slightly irregular, 0.5 mil of attack
Cr, 0.002 '
309 (o) ~0.0010 Fe, 0.017 Specimen attacked to 4 mils in one area; sev-
Ni, 0.034 eral small crystals attached to surface
Cr, 0.018 '
{b) -0.0013 Fe, 0.017 No attack on specimen; crystals (0.25 mil
Ni, 0.040 thick) attached to surfcjxce in some dgreas
Ce, 0.011
310 (o) ~0.0007 Fe, 0.012 - Grain boundaries of specimen exuded moisture
Ni, 0.032 to a depth of & to 7 mils, which indicated
Cr, 0.028 penetration by lithium to this depth; crystals
attached to surface in some areas
{b) ~0.0007 Fe, 0,018 Grain boundaries of specimen exuded moisture
Ni, 0.039 to a depth of 6 to 8 mils, as above; surface,
Cr, 0.046 layer on port of specimen approximately 0.25
mil in thickness
316 (a). 0 Fe, 0.007 Grain boundaries exuded moisture to a depth
Nj, 0.037 of 5 mils in some areas; specimen heavily .
Cr, 0.01 attacked to 0.5 mil in one area; crystals
attached to surface
(b) +0.0009 Fe, 0.006 Approximately 0.5 mil of grain boundary
' Ni, 0.038 attack; 1-mil-thick crystols attached to
Cr, 0.01 surface

*Tested in helium-packed lithium,

**Tested in oil-packed lithium.

79
ANP QUARTERLY PROGRESS REPORT

TABLE 6.5 {(continued)

 

 

 

 

 

 

 

 

 

 

TYPE OF WEIGHT CHANGE ANALYSIS METALLOGRAPHIC NOTES
STAINLESS STEEL (¢/in.2) (wt %)
317 (a) +0.0007 Fe, 0.009 No attack except for surface layer 0.1 to 0.2
Ni, 0.028 mil thick
Cr, 0.008
(b) +0.0009 Fe, 0.004 Some as above, with crystals to a maeximum
Ni, 0.009 thickness of 0.7 mil
Cr, 0.035
347 (a) —~0.0003 Fe, 0.11 No attack except for 0.25-mil-thick crystals
Ni, 0.038 attached to surface in some areas
Ce, 0.003
(b} +0.0003 Fe, 0.015 Same as above
Ni, 0.039
Cr, 0.002
446 (b) +0.0026 Grain boundary voids to a depth of 10 mils
TABLE 6.6, SPECIFICATIONS OF SOLID FUEL ELEMENTS
ROLLING HOT
CLADDING CORE TEMPERATURE REDUCTION
(°Q) (%)
Type 304 stainless steel Type 302 stainless sfeeI—U02 1225 60
Type 316 stainless stee! Type 302 stainless steel-UQ, 1225 60
Type 347 stainless steel Type 302 stainless sfeel-~UO2 1225 60
Nickel Nicke|~UQ2 1100 60
Copper Copper-U0, 1000 60

 

 

 

 

FUNDAMENTAL CORROSION RESEARCH

J. V. Cathcart G. P. Smith
Metallurgy Division

Equilibrium Pressure of Hydrogen in
Hydroxide-Meta! Systems

F. Kertesz F. A. Knox

Oxidizing Power of Hydroxide Corrosion Products Materiats Chemistry Division

During this quarter, equilibrium pressures of
hydrogen over the KOH-Ni system were partially
determined, and values previously reported for
NaOH-Cu and NaOH-Au systems were redetermined
by using a modified form of the apparaius previ-
ously described.®

The work on trivalent nickel compounds formed
in hydroxides was continued. The compound
NaNiO, was found to undergo a change in crystal
structure at about 225°C, and the structure of the
high-temperature form has been determined. The

existing evidence indicotes that the transformation
may be of the martensitic type.

 

6éANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609,
p. 84.

80

The modified apparatus includes a long nickel
tube which serves as a reaction chamber in place
of the metal crucible previously uvsed. This tube
is surrounded by a quartz tube to which it is
 

PERIOD ENDING DECEMBER 10, 1953

Unclosstfied
Y.7256

Fig. 6.9. Static Test of o Type 347 Stainless Steel-Clad Fuei Plote After 100 hr at 816°C in Sodium

Hydroxide. 80X.

sealed by means of a Kovar seal, The pressures
established inside the nickel tube and in its
quartz jacket (nickel being permeable to hydrogen
at the temperatures utilized in these experiments)
gre measured by two mercury monometers con-
nected to the sysfem by seals; therefore the pres-
sure built up by the hydregen that diffuses through
the nickel can be determined independently from
the pressure inside the tube. This apparatus
utilizes both the passivity of nickel to sodium
hydroxide and the nonpermeability of quortz to
hydrogen. Equilibrium pressures may be estab-
lished more rapidly by adding hydrogen to the
quartz-nickel cavity to bring its pressure close
to that inside the nickel tube. :
The hydrogen pressure of the KOH-Ni system,
as determined in the above-described apparatus,

was found to be less than 1T mm at 600°C, while
it is in the range of 8 to 12 mm at 700°C. Measure-
ments gt higher temperatures are being made. -

The NaOH-Cu and NaOH-Au hydrogen equilibrium
pressures were redetermined, and the results
agreed closely with those previocusly reported:®
for the NaOH-Au system at 900°C, 3.5 mm com-
pared with 3.0; ot 1000°C, 7.5 compared with
6.0 mm. The results for the NaOH-Cu system
show reasonable agreement ot 700°C (2.2 mm com-
pared with 3.0 mm); however, the more recent
determination at BOO°C shows a thus far unex-
plained time dependence of the hydrogen pressure.
An initial pressure of 14 mm gradually fell to
about 10 mm after a few hours; the 10-mm pressure
then remained constant.

81
ANP QUARTERLY PROGRESS REPORT

Mass Transfer of Chromium in Incone!-Fluoride
Systems

L. G. Overholser J. D. Redman
C. F. Weaver

Materials Chemistry Division

When
at elevated temperatures, chromium is selectively
leached from the alloy. While a considerable frac-
tion of this corrosion may be ascribed to oxidizing
impurities in the molten material or on the capsule
walls, reactions of UF , with the container metals
to yield UF, and complex metal fluorides could
well be responsible. The reaction

Cr + 2UF4'-=:'“"'—""' 2UF3 + CrF2

lnconel is exposed to molten fluorides

has at 600°C a stondard free energy change of
~7 kecal and should, if the ionic species were
soluble and showed activities proportional to their
concentrations, proceed nearly to completion. That
such is not the case, or at least that the reaction
does not equilibrate ot a reasonable rate, is well
demonstrated in a large number of corrosion tests.

if the assumption that the physical solubility
of chromium as metal in fused Nc::ZrF5 is quite
low is permissible, then equilibration of UF , dis-
solved in NaZrF_ with chromium metal followed
by separation of the phases and analysis of the
filtrate would permit evaluation of the equilibrium
constant of the reaction. For these studies, the
purest NaZrF, available was used, along with
chromium metal hydrogen fired ot 2000°F in nickel
equipment of the type previously described. All

The fil-
trates obtained were submitted for chemical onal-
ysis, and the results shown in Table 6.7 were
obtained.

formed in on inert-otmosphere dry box,

These dota indicate that equilibration is rapid
because the values obtained after 5 hr at temper-
ature are not distinguishable from those obtained
for a shorter exposure. It is obvious from the ex-
posure of the nonuranium-bearing material that the
chromium contains some oxide and that hydrogen
tiring of the nickel equipment is not sufficiently
complete, (Subsequent experiments, not suffi-
ciently complete to report, in which the chromium
was fired under dry hydrogen directly in the nickel
reaction vessel demonstrate that it is possible
to have less than 200 ppm of Cr in the nonuranium-
bearing material.) The chromium concentration in-
crease resulting from the reduction of UF, to UF,
was obtained by subtracting the chromium content
of the monuranium-bearing material from the total
chromium found, ond this value was used for ob-
taining the UF, concentration, since the methods
for determining UF3 in the presence of divalent
chromium are not sufficiently accurate to permit
a direct determination of UF,.

If unit activity is assumed for the chromium and
activities equal to concentrations (in moles per
mole of solvent) are assumed for the uranium and
chromium species, these data indicate that the
equilibrium constant at 600°C is 1,2 by 1073,
and, at 800°C, it is 1.7 by 1073, The data ore

certainly not sufficient in amount or in precision

 

 

 

 

handling of the materials before testing was per-  for an evaluation of this small difference. It does
TABLE 6.7. ANALYSES OF FILTRATES FROM EQINLIBRATION
OF CHROMIUM WITH NaF-—ZrF4-UF4 MELTS
CONDITIONS OF EQUILIBRATION FILTRATE AMALYSIS

Time Temperature Total Uranium Chromium Mickel lron
{hr) (°C) (%) {ppm) (ppm) (ppm)

5 800 0 935 355 75

0 930 200 40

3 800 8.47 3420 30 65

8.80 3330 40 60

3 600 8.80 3200 130 5¢

8.65 3100 125 50

5 600 8.65 3100 130 70

8.88 3200 130 60

 

 

 

 

 

 

82
appear, however, that the standard free energy
change for the reaction at 600°C is about +11.5
kcal. The discrepancy between this measured
value and the value of ~7 kecal calculated for the
reaction by assuming no complexing is probably
to be ascribed to solvation of the UF,.

PERICD ENDING DECEMBER 10, 1953

These experiments are being refined, insofar as
purity of the materials and improvement of tech-
niques are concerned, and additional experiments
will be conducted at considerably lower reagent
concentrations so that the assumptions regarding
activities will have more validity.

 

7. METALLURGY AND CERAMICS

w. Di Mflnly

Jo Ma Warde

Meffillurgy Division

A sodium-to-air radigtor was fabricated by brazing
the tube-to-fin joints and backing up the previously
heliarc-welded tube-to-header ond manifold joints
with brazing alloy G-E No. 62, a nickel-chromium-
silicon alloy. The radiator assembly was welded
and then dry-hydrogen brazed ot 2100°F for ap-
proximately 30 minutes. Subsequent testing after
slow cooling showed all the welded and brazed
joints to be satisfactorily leak and pressure tight.
The duplex fabrication technique, that is, the
combination of heliarec welding 'and_high-'rempem’rure
brazing, apparently produces satisfactory, sound,
leak-tight radiators and minimizes many of the
problems encountered in brazing operations. The
use of the G-E No. 62 brazing clloy eliminates
many of the deleterious effects of the Nicrobraz
alloy that was formerly used.

The investigation of the fabrication of high-
conductivity fin materials included o study of
diffusion barriers for use between lnconel and
copper, electroplating copper to obtain oxidation
resistance, and the fabrication of clad-copper fin
material in sufficient quontities for experimental
radiators. Furthermore, several new high-conduc-
tivity fin materials were studied because it was
found that diffusion of nickel and copper made the
brazing of Inconel-clad copper fins to Inconel
tubing less sdatisfoctory than it was previously
thought to be., Each clad-copper and copper-alumi-
num bronze fin material was tested by brazing to
either Inconel or stainless steel tubing with each
of the following brazing alloys: low-meiting-point
Nicrobraz (LMNB), "‘electroless’’ nickel-phospho-
rus, LMNB plus 10 wt % ‘“‘electroless’ nickel-
phosphorus, and nickel-phosphorus-chromium alloys.
A summary of the results of this study is given. A

series of one hundred 3 x 3 in. nickel fins is being
“‘electroless’’ preplated with nickel-phosphorus
alloy for brazing into a rodiator assembly.
preliminary test, good-quality brazes were obtained
with this material at-tube-to-fin joints,

The assembly of stainless steel-clad fuel plates
into fuel elements is being studied, and it appears
that inert~-arc plug welding may be a feasible
method of fabrication. The configuration investi-
gated consisted basically of « 0.02é-in.-thick type
304 stainless steel-clad type 304 stainless steel
sheet 24 in. long that was to be joined on the edge
at 2-in. intervals to a 0.041-in.-thick 'rype 304
stainless steel side plate.

Stress-rupture tests of coarse- and fine—gmined
Inconel in argon and in fluoride fuel NaF-ZrF - +UF,
at temperatures of 1300, 1500, and ]650°F: are
being made. The tests completed at 1500°F show
that hydrogen is the most detrimental environment
and that air is the most beneticial, as far as rupture
life is concerned, and that increased ductility in
air and in sodium is probably the result of decarbu-
ization.  Tube-burst tests of triaxially stressed
Inconel tubes are being made with fluoride fuel
NaF-ZrF -UF , inside the tube and purified argon
outside, Comparison of the data obtained with
creep data shows that the tube-burst time is about
the same as the time for 2% elongation in the
tensile creep tests.

In a

Several special alloys for the corrosion testing
program were melted in the high-vacuum melting
furace and extruded into o drow biank. The draw
blank was drawn into tubing by the Superior Tube
Company. Columbium cold worked to a reduction
of 87.5% has been shown to have a recrystallize
tion temperature of between 1000 and 1100°C.

83
ANP QUARTERLY PROGRESS REPORT

Tubular fuel elements are being produced by
using plug drawing techniques. The drawing
schedule was found to be too severs in that it
caused premature failure during drawing. However,
even this severe working schedule produced tubing
that had excellent inside and outside finishes and
there was no evidence of rippling or folding in the
core region.

Ten test rings of a beryllium fluoride glass were
prepared for determining their suitability as high-
temperature pump seals. A ceramic container for
use in the investigation of the electrical resistivity
of fluoride fuels is being prepared by isostatically
pressing and sintering beryllium oxide. High-
density graphite (Graph-i-tite) is being tested for
compaoatibility with fluorides and other molten salts
and molten metals.

Studies of the flammability of jets of sodium
alloyed with numerous metals have been made.
Sodivm-mercury alloys containing less than 34
mole % sodium and sodium-bismuth alloys contain-

 

Fig. 7.1,
Brazing Alloy and Exposed te Static Air at 1500°F for 1200 Hours, Etchant: glyceria regia. 75X

84

ing less than 40 mole % sodium did not burn; all
other alloys tested burned,

WELDING AND BRAZING RESEARCH

P. Patriarca G. M. Slaughter
Metallurgy Division

J. M. Cisar, ANP Division

Brazing of Radiator Assemblies

It was stated in the previous report that the G-E
No. 62 brazing alloy (69% Ni~2% Cr-11% Si)
might be svitable for use on ANP-type sodium-to-air
radiators. Stainless steel joints brazed with this
alloy were considered to possess adequate resist-
ance to cracking. This alloy was also shown to
exhibit excellent resistance to high-temperature
oxidation. A metallogrophic section of a joint of
an Inconel tube brazed to a type 304 stainless
steel fin and exposed in static air for 1200 hr ot
1500°F is shown in Fig. 7.1. It can be seen that

 

 

 

 

 

 

Section of Joint of Inconel Tube Brozed to a Type 304 Stainless Steel Fin with G-E No. 62
the extent of attack is minor. Previous tests had
also shown that this alloy was attacked by molten
sodium only slightly a 1500°F in a 100-hr test.

Therefore the sodium-to-air radiator shown in
Fig. 7.2 was fabricated by brazing the tube-to-fin
joints and backing up the previously heliarc-welded
tube-to-header and manifold joints with brazing
alloy. This method of fabrication is thought to be
the most reliable for obtaining a pressure-tight
system with one brazing operation.

The tube-to~header welds were manually inert-
gas tungsten-arc welded with argon as the shielding
gas and helium as the back-up gos. The header
caps and the manifold were also heliarc welded by
using suitable joint designs to permit complete
penetration. e

The sequence of manifold and header cap welding
was such that the nickel-chromium-silicon alloy
could olwoys be opplied to the underside of the
jioint after welding. Each tube-to-header weld was
alse backed up with this brazing alloy. A typical

 

Fig. 7.2.

PERICD ENDING DECEMBER 10,:1953

heliarc-welded joint backed up with the nickel-
chromiym-silicon braze is shown in Fig. 7.3. [t
can be seen that the probability of obtmining a
[eak-tight joint is very good.

The completed assembly was then dry- hydroqen
brazed by using the thermal cycle shown in Fig.
7.4. A preheat at 1600°F was utilized to permit
an equilization of temperatyres over the assembly.
The thermocouple, which was ploced directly on
the radiator, also indicated that a maximum temper-
ature of 2150°F was attained and that the unit was
held above 2100°F for approximately 3 minutes.
The brazing technique used for this fabrication
ensures a relatively slow cooling rate so that
cracking upon cooling as d result of uneven thermal
contractions should not be a major problem.

Helium leak testing at fess than 0,034 vacuum
revealed that all welded and brazed joints in the

radiator were leak tight. The assembly was also

pressure tight to a 20-psi helium pressure when it

E UNCLASSIFIED
[ PHOTO 20696

was immersed in water,

Sodium-to-Air Radiator Fabricated by the Combined Welding and Brazing Technique,

85
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

Fig. 7.3.

It is thus believed that a combination heliare-
welding and high-temperature-brazing procedure is
satisfactory for fabricoting sound leak-tight radi-
ators. This duplex technique has been shown to
minimize many of the problems and difficulties
vsually encountered in brozing operations. The
use of the G-E No. 62 brazing alloy alsc seems to
minimize or completely eliminate many of the
deleterious effects of the Nicrobraz alloy that was
formerly used,

Brazing of High-Conductivity Radiator Fins

Although the problem of brezing lnconel-clod
copper high-conductivity fin material to lnconel
tubing was satisfactorily resolved, as described in
the previous report,| the subsequent determination
of the detrimental effects of diffusion of nickel
and copper has proved to be o limitation of the

e s, Bomar et al.,, ANP Quar. Prog. Rep. Sepi. 10,
1953, ORNL-1609, p. 95.

 

86

 

 

 

 

 

 

 

 

Inconel Tube-to-Header Heliorc Welded Joint Backed Up with G-E No. 62 Brazing Alloy.
Etchant: glyceria regia. 50X

method. Accordingly, a group of other high-conduc-
tivity fin materials was fabricated for brazing tests,

The high-conductivity fin materials studied
during the current investigation were types 446
and 310 stainless steel-clad copper and a copper-
aluminum (94% Cu—-6% Al) bronze. The fabrication
of these moaterials is described below under *“High-
Conductivity Metals for Radiator Fins.,'”' The
brazing alloys investigated were low-melting-point
Nicrobraz (LMNB), “‘electroless’’ nickel-phospho-
rus (90% Ni—~10% P) dlloy, low-melting-point Nicro~
braz plus 10 wt % *‘electrocless’’ nickel-phosphorus
alloy, and nickel-phosphorus-chromium (8% Ni—10%
P—-10% Cr) olloy. Each of the brazing alloys was
used in conjunction with each of the high-cenduc-
tivity fin materials against Inconel tubing, os wel!
as against tubing of the same nominal composition
as that of the stainless steel component of the fin
material,

The test specimens were designed to simulate
T flowability specimens. A sheet and a tube were
PERIOD ENDING DECEMBER 10, 1953

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DWG, 22215
2400 1 | [ : ‘
| i | i
| ; 3 1 1 !
_________________ ? — BRAZING FERIOD COOLING PERIOD
| PREHEAT _ i | }
2000 FERIOD T tmmmees ‘ \ -
INITIAL l/,:,__o/a" | \\1
HEATING| | | P o . b
PERIOD " q\
1600 | N I N T N I A
I t
o et Ll I
(KN
% |
E 1200 pfdd e e R —
t ;
& |
e L4 bl
i |
800 - bed e
|
|
400 L
o1
0 20 40 &0 80 - 100 120 140 160 180 200
TIME (min)
Fig. 7.4. Thermal Cycie Employed ir Furnace Brazing a Heat Exchanger with G-E No. 62 Alloy.

Temperatures were obtained with a chromel-alumel thermocouple secured to the heat exchanger.

placed side by side, and the brazing alloy was
applied initially to one end of the specimen. All
test specimens were brazed in -70°F dew-point
hydrogen for 10 minutes. A summary of the data
and the results of this study are presented in Table
7.1, along with an evaluation of the flowability of
of the braze material based on visual observations,

The foliowing conclusions have been drawn from
these tests:

1. Ali the brazing alloys investigated, which
were chosen because of oxidation resistance and
favorable melting temperature, exhibited excellent
flow against types 310 and 444 stainiess steel-clad
copper when used against Inconel tubing. It
appears that the Inconel behaved as a '‘carrier”
for the ailoys.

2. The oxide films present on the as-received
type 310 5ffiiniess steel tubing used in this in-
vestigation inhibited in vorying degrees the flow
of all the brazing alloys tested. The removal of

this oxide by prebright annealing at 1150°C or by
electropolishing improved the observed flow., The
most marked effect was the improved flow of
LLMNB.

3. Removal of oxide films by prebright annealing
or by electropolishing the type 4456 stainless steel
tubing did not markedly improve the flow of the
brazing alloys and was particularly ineffective for
the LMNB alloy.

4. The addition of 10 wt % nickel-phosphorus
{90% Ni-10% P) alloy to the LMNB alioy markedly
improved flow against type 446 stainless steel
tubing, although this addition was not beneficial
against type 310 stainless steel tybing.

5. The oxide film which formed on the copper-
aluminum (M% Cu-6% Al) fin material during the
brazing cycle impeded flow of all the brazing al-
loys investigated, Prevention of this oxide forma-
tion by preplating with electrolytic nickel permitted
excellent wetting and flow of all the brazing
alloys investigated.

87
ANP QUARTERLY PRCGRESS REPCRT

TABLE 7.1, SUMMARY OF B

 

Fin Matarial

_ TSPECIMEN .

Tubing

RAZING STUDIES OF HIGH-CONDUCTIVITY FIN MATERIALS

 

BRAZING ALLOY

BRAZING
TEMPERATURE
(°C)

 

Type 446 stainless steel-clad
cepper {0.004 in. of coppar
cled en both sides with
0.002 in. of the steel)

Type 310 stainless steel-clod
copper (0.004 in. of copper
clad on both sides with
0.002 in. of the stesl)

Copper-oluminum
{94% Cu--6% Al) alloy
electraplatad witk 0.0002

in, of nickel**

 

Inconel, as-received

Type 446 sioinless steal, |

as-received

Type 446 stainless steel,
prebright annealed in
dry H,, ot 150°C for
15 min

Type 446 =tainless steel,
electropolished

Inconel, as~received

Type 310 stainless stesl,

as-recaived

 

Type 310 stainie=s stsel,
prebright annealed in
dry H, at 1150°C for
15 min

Type 310 stainless steel,

elactropolished

Inconat, as-received

Type 446 stainlass steel,
prabright annealed

Type 310 stoinless st=el,
prebright annealed

 

LMNB

90% Ni--10% P

LMMEB + 10% Mi-P*
80% Ni~10% P-10% Cr

LMNE

90% Mi-10% P

LMME -+ 10% Ni-P

80% Ni—~10% P-10% Cr

LMNBR

90% Ni-10% P

LMMB + 10% Ni-P

80% Ni--10% P-10% Cr

LMNB

20% Ni-10% P

LMNB + 10% Ni-P

80% Mi-10% P-10% Cr

LMNB

90% Mi-10% P

LMNB + 10% Mi-P

80% Mi-10% P-10% C»

LMNB

0% Ni—-10% P

LMNB + 10% Ni.P

80% Ni~10% P-10% Cr

L.MNB

90% Ni-10% P

LMNS + 10% Ni-P

80% Ni-10% F..10% Cr

LMNB

20% Ni-10% P

LMNB + 10% Ni-P

80% Ni-10% P-10% Cr

LMMB

20% Ni-10% P

LMNB + 10% Ni-P

80% Ni—-10% P~10% Cr

LMNE

90% Wi—-10% P

LMNS + 10% Ni.P

80% Ni—-10% P~10% Cr

LMNB
90% Ni-10% P
LMMNB + 10% Ni-P

B0% Ni-16% P-10% Cr

 

1050
1000
1630
1050

1050
100
105¢
1030

1058
1000
1050
1050

1050
1000
1050
1050

1050
1000
1650
1050

1050
1000
1050
1050

1030
1000
105¢
1050

1050
1000
1050
1050

1030
1000
1030
1030

1030
1000
1030
1030

1030
1060
1030
1030

 

FLOWABILITY

(visuval chsarvation)

Good
Excellent
Excallent

Excellent

Mane
Excallent
Good

Excallent

None
Exc=!lent
Good

Excellent

None
Excellent
Excellant

Excsllant

Exceliant
Excelliant
Excellent

Excellent

None
Good
None
Fair
Goed
Excellent
Excellont
E xcellent

Good
Excellent
Exceltant

Excellent

Excellant
Excellent
Excellent

Excclisnt

Excellent
Excelient
Excellent

Excellent

Exceilent
Excellent
Exceoliont

Excellent

 

 

**Without the electroplate, this alloy wos net wet by ony of the brazing olloys,

88
“Electroless’ Preplating of Brazing Alloys.

Work on the ‘“‘electroless’’ preplating of nickel-

phosphorus brazing alloy, as described pre:vicsus-;l',v,2
has continued, and the method is being applied to
practical brazing problems., A commercially availe-
ble system of plating solutions is now being
successfully used to plate the nickel-phosphorus
alloy on to-be-brazed specimens, The system,

Lustralloy, manufactured by Metal Processing

Company, lnc., is cheaper and as satisfactory as
the procedure developed by the ORNL Research
Shops. A series of approximately one hundred
3 x 3 in. nickel fins is currently being preplated
with nickel-phosphorus alloy for brazing into a
radiator assembly for the Griscom-Russell Company.
As a preliminary test, two of these fins were
sheared into 3/“-in. sections; nickel tubes were

 

2p, Patriarca and G. M. Slaughter, ANP Quar, Prog.
Rep, Sept. 10, 1953, ORNL-1609, p. 91.

 

PERIOD ENDING DECEMBER 10, 1953

inserted into holes in these fins, and the mock
assembly was placed into the hydrogen furnace at
1020°C for approximately 10 minutes, The speci-
men was then cooled slowly under a hydrogen
atmosphere. The finished specimen was bright,
and the joints were well brazed. A similarly
brazed assembly is shown in Fig., 7.5. As can be
seen, the brazing alloy, originally uniformly distri-
buted on the fins and obsent on the tubes, was
drawn to the capiliary. Figure 7.6, a photomicro-
graph of the tube-to-fin joints, shows the good
quality of the broze. : '

inert-Arc Plug Welding of Solid Fuel

Elements

The assembly of stainless steel-clad fuel plates
into fuel elements of the size and configuration of
interest presents a problem for which there are
several possible solutions. One method that may
prove useful in certain applications utilizes inert-

UNCLASSIFIED
PHOTO Y 10200

Fig. 7.5. Tube-to-Fin Assembly Brazed with Prepioted ‘'Electroless’ Nickel-Phosphorus Brazing

Alloy.

89
ANP QUARTERLY PROGRESS REPORT

 

 

Fig. 7.6. Tube-to-Fin Joints Brazed with Preplated ‘'Electroless’ Nickel-Phosphorus Brazing Alloy.

arc plug welding techniques. The configuration
investigated in these preliminary experiments
consisted basically of a 0.026-in.-thick type 304
stainless steel-clad type 304 stainless steel sheet
24 in, long that was to be joined on the edge ot
2-in, intervals to a 0,041-in,-thick type 304 stain-
less steel side piate,

The joint design considered to be optimum for
welding is shown in Fig. 7.7. A hole 0,067 in, in
diameter is drilled in a 0.041-in.-thick header
sheet to receive a projection 0.062 x 0.026 x 0,125
in. obtained by edge milling the 0.026-in. type 04
stainless steel-clad sheet, The 0.84-in. extension
of the projection above the header sheet was pre-
determined from the following volume considera-
tions: (1) the volume of the 0.067-in.-dia, 0.041-
in.-deep header hole was 0,00015 in.3; (2) the
volume of the 0,062 x 0,026 x 0,125in, projection
was 0.0002 in.3; (3) the excess of 0.00005 in.>
would ensure adequate weld buildup and could be
removed by finish machining.

90

A series of specimens of the design shown in
Fig. 7.7 were welded with variations in arc current,
arc time, arc distance, and welding sequence,
Since precise control of these variables was re-
quired, cone-arc equipment was used throughout
the study. The welds were examined metallograph-
ically for penetration and flaws.

E xamination of test specimens indicated that the
most favorable weld geometry could be achieved
by two separate inert-arc welding operations, The
projection was melted info a ‘‘rivet’’ by an initial
choice of arc current, orc time, and arc distance.
The arc distance was then reduced, and on addi.
tional application of current was used to form the
the finished weld, the filler metal being provided
by the rivet.

The rivet formed during the first operation by
using the optimum welding conditions is shown in
Fig. 7.8, which is a section through the weld joint
corresponding to section A-A in Fig. 7.7. The plug
weld formed in the second operation is shown in
Fig. 7.9. As may be noted, the penetration of the
type 304 stainless steel-clad sheet was complete
and reasonably uniform, which indicates that an
apparently sound weld joint was obtained. [t thus
appears that inert-arc plug welding may be a feas-
ible method of fabrication of these assemblies.

UNCLASSIFIED
DWG 22216

DRILLED HOLE

// z"

- -

Lo
e
g S

e Aoy~ 0.067=in~dia
i | H 7 "
1] 7
11

 

“\

N

 

0.062 in,

 

 

 

 

 

 

 

~~-HEADER SHEET
s
SECTION B-8

E T i

1 0.084 1
A

: N ‘*\.‘\\

N

 

 

 

[ TYPE 304 STAINLESS
STEEL-CLAD SHEET

 

/% : SO ONN,
r.

D.026 inqad-t=-—

SECTION A~A

Fig. 7.7. Assembly of Type 304 Stainless Steel
Components Prior to Inert-Arc Plug Welding,

B UNCLASSIFIED
[ eHOTO Y 10336

 

Fig. 7.8. Section of ‘‘Rivet'' Formed by Initial
Application of a 30-amp (DCSP) Arc Current for an
Arc Time of 2 sec at an Arc Length of 0,140 in.
Etchant: aqua regia. 21X '

PERIOD ENDING DECEMBER 10, 1953

However, ‘ the application of these techniques to
the fabrication of a multiweld unit should be
investigated further, '

MECHANICAL PROPERTIES OF INCONEL

R. B. Oliver D. A. Douglas
K. W. Reber J. M. Woods
Metallurgy Division

Stress-Rupture of Inconel in Fluoride Fuel

Stress-rupture tests of coarse- and fine-grained
Inconel in argon and in fluoride fuel NaF-ZrF -UF
at temperatures of 1300, 1500, and 1650°F are be-
ing made, Data for the higher stress ranges are
complete, and the tests being made currently are
expected to run from three to eight months, Test.
ing in the stress range from 500 to 1500 psi will
commence in about six months when the six new
units designed for low loads are completed.. The
results obtained to date for fine-grained Inconel
are summarized in Table 7.2.

Tube-Burst Tests of Triaxially Stressed Tubes

In the past, data obtained in the tube-burst tesis
of triaxially stressed Inconel tubes have shown
considerable scatter and interpretation has been
difficult. However, data obtained in recent tests
have shown reasonable reproducibility, and it is
believed that most of the operating difficulties
have been overcome. The present procedure is to

UHCLASSIFIED |
PHMOTO Y 10335

 

Fig. 7.9. Section of Completed Weld Formed by
Application of a 30-amp Arc Current for 4 sec at an
Arc Length of 0.060 in. to the ‘““Rivet’’ Shown in
Fig. 7.8. Etchant: oqua regia. 21X

N
ANP QUARTERLY PROGRESS REPORT

TABLE 7.2, RESULTS OF STRESS-RUPTURE TESTS OF FINE-GRAINED INCONEL IN NaF-ZrF4-UF4

 

 

 

 

 

RUPTURE TIME RUPTURE TIME IN
TEMPERATURE STRESS IN ARGON L UORIDES

(°F) (psi) (hr) (hr)
1300 10,000 20600

1500 10,000 50 30
1300 5,000 7000

1500 5,000 540 260
1650 5,000 25 25
1500 2,500 4000 1200
1650 2,500 450 175

 

 

test the tube with the fluoride fuel NaFoZrF4-UF4
inside and purified argon outside. The stress
range from 500 to 3500 psi is currently being in-
vestigated ond tests have been completed in the
vpper half of the range. In comparing the data with
creep data at the same stresses in the same fuel,
it is found that the tube-burst time is about the
some as the time for 2% elongation in the tensile
creep tests, T he tube specimens looded by internal
pressures give an axial-to-tangential stress ratio
of ]/2, which results in minimum ductility; thus the
results probably represent the minimum life ex-
pectancy, other factors being equal. In the near
future, these tube specimens will be tested with
the fuel inside the tube and sodium, instead of
argon, outside.

Environmenta! Effects on Creep of Inconel

Creep tests of Inconel specimens in nitrogen and
in sodium environments at 1500°F have recently
been completed. Thus it is now possible to com-
pare the effects of six types of environments that
Inconel structures might encounter. The results
of these creep tests are summarized in Table 7.3.

Table 7.3 shows that hydrogen is the most detri-
mental environment and that acir is the most bene-
ficial, as far os rupture life is concerned, It is
interesting to note that, with the exception of the
tests in air and sodium, all the specimens reached
about the same total elongation regardless of the
rupture time, The increased ductility in air and in
sodium is probably the result of decarburization.
Chemical analyses are being mode to test this
hypothesis., The results presented in Table 7.3
are for Inconel specimens that were annealed at

92

TABLE 7.3. RESULTS OF CREEP TESTS OF
INCONEL. ANNEALED AT 1650°F AND TESTED
UNDER A STRESS OF 3500 psi ot 1500°F
IN VARIOUS ENVIRONMENTS

 

 

 

LIFE FINAL

ENVIRONMENT | RUPTURE | ELONGATION

() (%)

\
Hydrogen 446 13.0
NaF~ZrF4-UF4 550 12.0
Sodium 1333 35.0
Argon 1467 12.0
Nitrogen 1770 11.6
Air 2567 50.0

 

 

1650°F hefore being tested. Specimens annealed
at 2050°F and then stressed in hydrogen or Naf-
ZrF4-UF4 have shown more resistance to the
environmental effect. Tests of this type of speci-
men in the other environments are now being made.

HIGH-CONDUCTIVITY METALS
FOR RADIATOR FINS

E. S. Bomar J. H. Coobs
H. Inouye, Metallurgy Division

The investigation of the fobrication of high-
thermal-conductivity fins was confined during this
period to a study of diffusion barriers for use be-
tween Inconel and copper, the fabrication of clad-
copper fin material in sufficient quantities for the
assembly of experimental radiators, and the
investigation of special materials,
Diffusion Barriers

Barrier material for use between Inconel and
copper was selected on the bosis of low solubility
of the element or alloy in copper or Inconel. In
addition, it was desired to fabricate thin composite
clads to determine their rolling characteristics.
The barrier materials used were tungsten, molybde-
num, titanium, zirconium, tantalum, vanadium,
types 310 and 446 stainless steel, and iron plus
silver, ‘

Several composites were hot rolled approxi-
mately 50% ot 1800°F and then finished cold to
0.008 in, total thickness. Bonding was not achieved
between molybdenum and Inconel or between va-
nadium and Inconel, and the following composites
could not be successfully rolled: (1) Inconel-Ag-
Fe-Cu-Fe-Ag-Inconel and (2) Inconel-W-Cu-W-
inconel, The first of these composites was un-
successful because of improper design, and the
‘second showed severe stretcher marks, Oxidotion
and diffusion fests of the other composites are
being made. . '

Some of the earlier tests of Inconel-clad copper
showed the absence of diffusion voids, and there-
fore a few tests are now being made to determine
the effect of surface preparation on void formation
and diffusion. The composites for this study have
been fabricated; they consist of a layer of oxide

on Inconel formed by heating in wet hydrogen and
in air at 1300°F,

Clad Copper

As reported previously,? types 310 and 446 stain-
less steel are suitable cladding material for copper
for use at 1500°F, while Inconel cladding couses
severe diffusion. In order to establish which of
the claddings is most suitable on copper and also
to determine the effect of diffusion on thermal con-
ductivity, opproximately 15-ft2 samples of copper
clod with Inconel, type 310 stainless steel, and
type 446 stainless steel have been or are being
made. The fin material is to be strips 4 in, wide
and 0.010 in, total thickness; the copper core will
be 0.006 in. thick and the cladding material will
be 0.002 in. thick. This material is to be used for
the fabrication of experimental heat exchangers.

Electroplated Copper

Plates of chromium plus iron and of iron plus
nickel plus chromium to give compositions corre-
sponding to types 446 and 310 stainless steel,

PERIOD ENDING DECEMBER 10, 1953

respectively, did not protect copper from oxidizing
at 1500°F. A composite made entirely by electro-
forming showed the same results. The inability of
these plates to protect copper ot 1500°F is ascribed
to the lack of a metallurgical bond between the
copper and the plate and to the brittleness of the
deposits. Although the ductility can be restored by
heating to 1600°F, the domaging cracks are
present in the as-received or annealed material,

Solid Phase Bonding

An calternate investigation for obtaining uniform
thin claddings is being made., A bond between the
copper and the cladding can be obtained by seam
welding maoterial of the final desired thickness.
The Glen [, Martin Company proposed this method
and has supplied samples up to 4 in. long which
appear satisfactory. Joint efforts have, to date,
produced claddings 0.002 and 0,001 in. thick.

The thermal properties of the aiuminum bronzes
at low temperatures indicate that the conductivity
of alloys containing about 6% aluminum is greater
than that of the stainless steels or Inconel by a
factor of about 6 at 1500°F. To verify this, meas-
urements are now being made on & and 8% aluminum
bronzes. These alloys are oxidation resistant,
and if the expected thermal properties are correct,
the need to fabricate composite bodies will be
eliminated because these alloys may be suitable
for use as radiator fins. This is especially im-
portant because difficulties have been experienced
in obtaining clad material from commercial suppliers,

FABRICATION OF SPECIAL MATERIALS

E. S, Bomar J. H. Coobs
H. Inouye, Metallurgy Division

Extrusion of Inconel-Type Alloys

Extruded Inconel tubing was redrawn from 1 in.
in outside diameter with a 0.125-in. wall to 0.500
in. in outside diameter with o 0.035-in. wall by the
Superior Tube Company., These moterials were
drawn satisfactorily at reductions as high as 36%
per pass with annealing after each stage at 1900°F,
The properties of the tubing are listed in Table 7.4,

The extrusion of near-lnconel alioys from seven
other vacuum meits was dttempted with varying
degrees of success., Temperatures from 2150 to

 

3E. S. Bomar et al.,, ANP Quar, Prog. Rep. Sept, 10,
1953, ORNL.-1609, p. 96. '

93
ANP QUARTERLY PROGRESS REPORT

TABLE 7.4. PROPERTIES OF ANNEALED HIGH-PURITY INCONEL TUBING 0.505 in.

IN OUTSIDE DIAMETER WITH A 0.035-in. WALL

 

 

 

 

 

GRAIN SIZE TENSILE STRENGTH YIELD STRENGTH ELONGATION
MELT NO.* . .
(mm) (psi) (psi) (%)
DPL.7 0.090/0.130 £1,000 35,000 52
DPI-8 0.075 83,400 36,000 50
DP9 0.090/0.110 83,500 40,000 48
DPIi-10 0.090/0.130 80,400 40,800 46

 

 

 

 

 

 

*Analyses of melts 8, 9, and 10 were given in the Metallurgy Division Progress Report for the Period Ending April

10, 1953, ORNL-1551, Table 22, p. 54.

235)°F were employed to produce the tubing (or
rod), ond, in some instances, extrusion rotios as
low as 10:1 were required, All the extruded tubing
has been sent to the Superior Tube Company for
further reduction., A molybdenum-nickel alloy (20%
Mo-~80% Ni, nominal composition) did not extrude
at a reduction of 13:1 at 2350°F. An attempt will
be made to extrude this alloy at a lower reduction
ratio,

Rolling of Chromium-Cobalt Alloy

A chromium-cobalt alloy ingot 6 in. long and 1
in. in diameter (45% Cr—55% Co, nominal composi-
tion) was cost, and attempts to roll it at 2350°F
showed that this alloy was hot-short, Additions of
0.5% manganese and 0.5% aluminum did not im-
prove the hot-rolling properties. A solid rod of this
alloy which was precision cast was readily machin~
able, but it cracked severely, The present plans
call for precision casting the alloy tubing. The
hardness of the cast alloy is Rockwell C-50,

Rolling of Cebalt

A ]l/z-in.-dics, 6-in.-long cobalt ingot containing
0.5% moanganese and 0.5% aluminum was success-
fully rolled to Y -in.-dia rod at 2300°F, Previously,
it was found that cobalt without the manganese
and aluminum additions was hot-short, The alloy
contgining aluminum and manganese has anas-cast

hordness of Rockwell A-49,

Rolling of lron-Chremiuym-Nicke! Aloy

A 'Il/-in.»dm, 6-in.-long ingot of iron-chromium-
nickel alloy (45% Fe-40% Cr-15% Ni) was rclled
o yz-in.-dia rod at 2350°F, This alloy was found
to be hot-short at 1850°F, even though it contained

94

1% of a nickel-manganese-titanium master alloy,
Previously, attempts to hot roll this alloy at 1850°F
without the alloy addition were unsuccessful be-
couse of hot-shortness. The as-cast alloy has @
hordness of Rockwell A-62,

Cold-Ralled Columbiem

Columbium sheet 0,20 in. thick obtained from
the Fansteel Corporation was annealed to a grain
size of ASTM 7 at 1100°C in a vacuum. The sheet
was cold rolled to 0,025 in., which represented a
reduction of 87.5% i thickness. The room-temper-
ature mechanical properties were determined on
specimens vacuum annealed for ]/2 hr at various
temperaiures and are summarized in Table 7.5,

The microstructure of columbium under these con-
ditions shows partial recrystallization at 1000°C,
and, at 1100°C, recrystallization is complete.
Vacuums of the order of 107> mm Hg were obtained
in the annealing furmnace. The analysis of the
sheet shows 0.75 + 0,3% tantclum, and traces of
copper, iron, and titonium, Analysis of the cxide
content of the sheet is not complete, but it is
expected to be greater than 0.3%.

TUBULAR FUEL ELEMENTS
J. H. Coobs, Metallurgy Division

Drawing of tubular fuel elements has been con-
tinved, Plans called for the reduction of six tubes
on each of two schedules in steps of 15% and 2%
reduction per pass, respectively, All tubes were
to be reduced from 0.750 in. in diometer with o
0.042-in. wall to a final size of 0.250 in. in dicmeter
with a 0.015-in. wall, or o total reduction of 87%.
Plug drawing was used in preference to drawing on a
mandrel, as explained previously.? However, be-
cause of failure of the tubes at intermediate steps,
the drawing schedules could not be carried to com-
pletion. Reduction of the tubes was continued as
long as was considered practical, that is, until
each tube had failed within the core region or near
the starting end in such a way as to make further
drawing difficult or impossible. The final sizes
of the tubes processed on the 15% schedule are
given in Table 7.6, along with core composition
and total reduction. : |

Because the results with the 20% schedule were
discouraging, only three tubes were processed; dll

 

41bid., p. 99.

PERIOD ENDING DECEMBER 10, 1953

tubes failed on the second pass. Total reduction

of these tubes was 36%.

In general, the tubes reduced by both schedules
developed excellent finishes on both inner and
outer diameters, with no evidence of rippling or
folding in the core region. In most cases, failure
occurred near the frailing end of the core. The
failures were atiributed mainly to inability to keep
the plugs positioned in the die properly, with the
result that the reductions did not conform strictly
to the schedule. The reductions were somewhat
higher than scheduled on severcl passes and
caused galling in the dies and foilure of the tubes
because of severe tensile stresses, Samples have
been cut from the nine tubes drawn and are now
being examined metallographicalily.

TABLE 7.5. MECHANICAL PROPERTIES OF COLUMBIUM COLD ROLLED TO A REDUCTION OF 87.5%

 

 

 

 

 

 

 

 

 

 

 

 

ANNEALING YIELD POINT AT | TENSILE | ELONGATION | MODULUS OF
SPECIMEN HARDNESS
NO. TEMPERATURE | = 0.1% OFFSET | STRENGTH | IN1in. | ELASTICITY
’ (°C) (psi) (psi) (%) (< 10~¢ psi)
1 As rolled 145 76,500 4.0
2 300 154 69,500 80,700 2.6 17.66
3 600 _ 147 62,300 73,400 8.0 16.60
4 900 136 57,300 63,100 8.3 17.30
5 | 1000 | 100 37,500 50,800 12,6 17.88
6 1100 85 23,400 42,200 26.6 19.30
7 ; 1200 86 24,000 45,000 18.6 17.20
TABLE 7.6. FINAL SIZES OF TUBES PROCESSED ON THE 15% SCHEDULE
TUBE fi NO. OF | FINAL SIZE TOTAL REDUCTION
COMPOSITION (BY VOLUME
NO. _ ( ) PASSES (in.) (%)
1 80% type 302 stainless steel-20% U02 5 ’ 0.562 56
2 70% Ni-30% UO, 9 0.396 77
3 80% Ni-20% UO, 9 0.396 77
4 70% type 302 stainless steei-30% UO, 10 0.355 80
5 80% Fe—20% UO, 3 0.625 ' 39
6 70% Fe~30% UO, 5 0.562 56

 

 

 

 

 

 

95
ANP QUARTERLY PROGRESS REPORT

CERAMIC RESEARCH

L. M. Doney J. A, Griffin
J. R, Johnson, Metallurgy Division

Glass«Type Pump Seals

Work continued on the development of a suitable,
viscous, high-temperature pump seal, Ten test
rings, 2) in. ID, 3% in, OD, % in, thick, of a
beryllium fluoride glass® were prepared by melting
the glass batch in a platinum crucible and casting
the glass in grophite dies (Fig. 7.10). The rings
were cooled slowly in o blanket of glass wool to
minimize cracking, The finished rings were sent
to the Experimental Engineering Group for testing.

Ceramic Container for Fuel

Further work is being done on the fabrication of
a ceramic container for yse in the investigation of
the electrical resistivity of the fluoride fuels (cf.,
sec, 8, ‘‘Heat Transfer and Physical Properties'’).
Three specimens, 5]/2 x 1 x 1 in, were cut from
hot-pressed beryllium oxide blocks (ARE blocks),
A hole z’] in. in diameter was drilled 4‘/2 in. deep
along the fong axis of the specimen and then tapered
to 0.158 in. in diameter for the remaining length of
the specimen. These specimens were tested and
found to be resistant to attack by the fuel; how-
ever, they were too short to satisfy the test condi-
tions. Work has commenced on the fabrication of a

 

SL. M. Doney, J. A. Griffin, and J. R. Johnson, ANP
Quar, Prog, Rep. Sept, 10, 1953, ORNL-1609, p. 98.

~< I
PHOTO 20810

  

Fig. 7.10. Giass-Type Pump Seal. Composition

(by weight): BeF,, 50%; KF, 25%; MgF,, 16%;
AlF 5, 9%.

96

suitably sized beryllium oxide shape by isostati-
cally pressing and sintering.
High-Density Graphite

Specimens of high-density graphite (Graph-i-tite)
supplied by the Graphite Specialties Corporation
were obtained and sent to the Metallurgy Division’s
Corrosion Group for compoatibility tests with the
fluorides and other molten salts and molten metals.
Typical properties of this graphite are given by the
manufacturer as:

Apparent density,

g/cm® 1.85 to 1.92
Tranverse breaking
strength, psi 4000 to 4500

Modulus of elasticity,

Ik/in.2/in./in.

Electrical resistance,
ohm-in.

20 x 10° to 25 x 10°

0.00032 to 0.00040

Permeability
1.85-g/cm® density Impermeable to water
at 40 psi and room
temperature for 10
min minimum

1.92-9/cm® density Impermeable to air
at 40 psi and room
temperature for 10

min minimum
COMBUSTION OF SODIUM ALLOYS

G. P. Smith M. E, Steidlitz
Metallurgy Division

Studies of the flammability of jets of sodium
alloys in wet and dry air at temperatures of up to
800°C have been continued. Sodium-mercury alloys
containing less than 34 mole % sodium ond sodium-
bismuth alloys containing less than 40 mole %
sodium did not burn. All other clloys tested burned
in degrees ranging from slight to violent, The com-
bustible alloys tested were: binary alloys contain-
ing, in addition to sodium, 90% aluminum, 50 to 60%
bismuth, 90% indium, 90% lead, 0.6 to 66% mercury,
90% silver, or 90% zinc; ternary alloys having a
45% sodium—50% bismuth base and containing 5%
each of calcium, copper, magnesium, mercury,
potassium, or silver; and ternary alloys having a
0% sodium—65% mercury base with 5% each of
aluminum, bismuth, calcium, copper, magnesium,
potassium, or silver. The humidity of the air was
noted to have a very slight effect on the rate of
combustion.
PERIOD ENDING DECEMBER 10, 1953

8. HEAT TRANSFER AND PHYSICAL PROPERTIES RESEARCH
H. F. Poppendiek

Reactor Experimental Engineering Division

The physical properties of o number of fluorides
and other moterials of interest to aircraft reactor
technology have been measured.  Preliminary
density and viscosity measurements have been
obtained for the ARE fuel concentrate over the
temperoture range of &60 to 1000°C; the viscosity
varied from about 17.5 cp at 725°C to 9.9 cp at
975°C. The vapor pressure equations for two
NaF-ZrF, mixtures were determined. The thermal
conductivity of the solid fluoride mixture NaF-KF-
LiF-UF, (10.9-43.5-44.5-1.1 mole %) was deter-
mined to be 1.7 Btu/hrft°F, while that of the

solid heat transfer salt NaNO «NaNO ,-KNO, -

(40-7-53 wt %) was only 0.6 Btu/ilr'ff’oF. The
heat capacity of NaF-ZrF -UF, (50-25-25 mole
%) was 0.17 cal/g-°C in the solid state and 0.27
cal/g°C in the liquid state over the temperature
range of 610 to 930°C, with a heat of fusion of
42 cal/g. The heat copacities of two special
samples of type 310 stainless steel and of G-E
No. 62 brazing alloy have also been measured.
Additional  forced-convection heat transfer
experiments with NaF-KF-LiF eutectic in a nickel
tube were made, and the data obtained were found
to be in agreement with data from the former experi-
ments. No additional thermal resistance was
present and no film deposits on the inner surface
of the nickel tube were cobserved, in contrast to
the conditions found with this fluoride mixture in
Inconel. _
Temperature and velocity measurements were
obtained in o glass thermal convection loop over
a Grashof modulus range from 2 x 10% to 67 x 104,
The experimental velocity data are compared with
predicted values which were obtainedby the numeri-
col solution of the laminar-flow heat-conduction
equation. The measured values diverge radically
from the theoretical values, undoubtedly because
of the turbulence which was introduced by the

radial temperature gradient and which was observed

at Reynolds numbers as low as 100.
Forced-laminar-flow volume-heat-source velocity
and temperature solutions have been derived for
the case in which the fluid viscosity is temperature
dependent. These solutions are functions of a
new dimensionless modulus which is a measure of

the importance of the viscosity variation of the
system.

Experimental apparatus is being assembled for
measurements of fluid flow in an annulus ond for
studies of transient surface-boiling phenomena.
The annulus flow will be determined by photo-
graphing dust particles; the essential photographic
technique is being developed. The boiling studies
will be conducted on a flat metal filament sus-
pended in water.

HEAT CAPACITY
W. D. Powers G. C. Blalock

Reactor Experimental Engineering Division

The enthalpy and the heat capacity of NaF-ZrF -
UF, (50-25-25 mole %) were determined' in the
liquid and the solid state with a Bunsen ice calo-
rimeter. The data can be represented by the fol-
fowing equations: '

H (solid) — H o .(solid) ~18 + 0.177,

c, = 0.17 + 0.02,
at 280 to 610°C,
HT(quuid) - Hfloc(soiid) -39 + 0.27T1,
c, = 0.27 +0.02,
at 610 to 930°C,

where H is the enthalpy in cal/g, T is the tempera-
ture °C, ¢, is the heat capacity in cal/g'°C. The

it

0°cC

-heat of fusion is 42 cal/g at 610°C,

The enthalpy and the heat capacity of two
special samples of type 310 stainless steel and
of brazing compound G-E No. 62 (69% Ni-20%
Cr-11% Si) were determined for the General
Electric (.:<;'om;:u:m),v:2

1. for type 310 stainless steel (heat 64177) from
238 to 858°C,

f{T — HOOC

f

~7.2 + 0.143T,
0.143 + 0.007;

“p

it

 

'W. D. Powers and G. C. Blalock, Heat Capacity of
Fuel Composition No. 33, ORNL CF 53.11-128 {Nov.
23, 1953),

2W. D. Powers and G. C. Blolock, Heat Copocity of
Two Samples of 310 Stainless Stee! ond of a Brazing
Compound, ORNL CF 53-9-98 (Sept. 18, 1953).

97
ANP QUARTERLY PROGRESS REFPORT

2. for type 310 stainless steel (heat 64270) from
240 to 834°C,

f

T

"HOOC

c
P

~5.5 + 0.1397,
0.139 + (0.005;

3. for G-E No. 62 brazing alloy from 211 to 840°C,

}

il

-4.3 + 0.142T,
0.142 + 0.012.

THERMAL CONDUCTIVITY OF
SOLIDIFIED SALTS

W. D. Powers R. M. Burnett
S. J. Claiborne
Reactor Experimental Engineering Division

The thermal conductivities of solid flucrides have
been measured by a transient cooling methed, A
sphere of the material being studied is allowed to
until a uniform initial
The sphere is then
rapidly transferred to a large circulating bath at a
lower temperature, A fine-gage thermocouple at
the center of the sphere is used to determine the
time-temperature curve as the sphere cools. Clas-

sical transient temperature solutions in terms of

remain in o calorimeter

temperature is established.

the physical properties of the sphere are available
and may be usedto extract the thermal conductivity
of the sphere from the experimental time-tempera-

ture dota. In this way the thermal conductivity of
the solid fluoride mixture NaF-KF-LiF-UF, (10.9-
43.5-44.5-1. 1 mole %) was foundto be 1.7 Btu/hrft?
(°F /1), as compared with a value of 2 Biu/hrft?
(°F/ft) for the same fluoride mixture in the liguid
state. The thermal conductivity of a solid heat
transfer salt (NaNO,-NaNO,-KNO,; 40-7-53 wt %)
is also being determined by the transient cooling
method.,  Preliminary measurements indicate a
value of 0.6 Btu/hr ft2(°F/ft) over the temperature
range 80 to 170°F.

DEMSITY AMND VISCOSITY OF FLUORIDES
S. |. Cohen T. N, Jones

Reactor Experimental Engineering Division

Preliminary viscosity and density measurements
on Na,UF ( were made® on a Brookfield viscometer
and displacement apparatus, respectively, The
density, as plotted in Fig. 8.1, is given by the

equation
p = 5598 - 0.00119T,

where p is in g/cm3 and $60°C < T < 1000°C.
The viscosity varies from about 17.5 cp at 725°C

3& l. Cohen, Preliminary Measuremant of the Density
and Viscosity of Composition 43 (N02UF6), ORNL
CF 53-10-86 (Oect. 14, 1953).

S
DWG. 22217

 

50 ¢

 

4.9 -

4.8

4.7

 

plg7em3)=5598-0.001197,

NSITY (g/em3}

2
B
\
|
i
1

T E60°C < 7T < 1000°C

D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

650

800 900 1000

TEMPERATURE (°C)

Fig. 8.1, Density of Ma UF  as o Fuactior of Temperature,

98
to about 9.9 cp at 975°C (Fig. 8.2), and the experi-

mental data are represented by the equation
p o= 0.92 e2936/T ,

where yiis in centipoises and T is in °K.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

pr—
50 DWG. 22218
1 . e dameiaeaaan e e becrmebeee s
2O b
-
8
> 40
5 B
& _
3 -
@
>y 1
i B -
1 (cp)=0.92 £2938/7 ook
2 . T e e e e
E
!
|
f ! ‘
300 560 000 1100 1200 1300 1400

TEMPERATURE {°K}

Fig. 8.2. Viscosity of No,UF , as a Function of
Temperature,

The following solid density measurements at
room temperature were made on NaF-ZrF, mixtures
in support of phase studies of these materials:

DENSITY (g/cm?)

Batch B-62 (53-47 mole %) 4.00
Batch R-132 (52-48 mole %) 4,103
Batch R-133 4.124

A method for predicting densities of fluoride
mixtures have been developed and will be described
in a forthcoming report. Plots of liquid densities
at any temperature vs, the calculated room-tempera-
ture densities have been developed that correlate
all the experimental data available to within 2%.
Similar plots have been devised for viscosity datq,
but the correlation is not dependable.

VAPOR PRESSURES OF FLUORIDES

R. E. Traber, Jr. R. E. Moore
C. J. Barton
Materials Chemistry Division
Additional measurements of vapor pressures by
the method of Rodebush and Dixon? were madg.

PERIOD ENDING DECEMBER 10, 1953

Chemical analysis has shown that the material
used for the previously reported® measurement of
the vapor pressure of NoF-ZrF,, nominally 50-50
mole %, actually contained 47 mole % Z¢F,. In-
stead of being the pure compound NaZrF,, it was
probably a solid solution of NaZrF, and Na ZrF
The vapor pressure equation for fl'us 47 mole %
ZrF (~53 mole % NaF mixture is

7213

log P = — + 7.635,

 

where P is in mm Hg and T is in °K. The vapor
pressure equation for the mixture NaF-ZrF , (50-50
mole %) is

6827.7

log P = ~ + 7.503

over the temperature range 812 to 942°C.

ELECTRICAL CONDUCTIVITY OF FLUORIDES
N. D, Greene

Reactor Experimental Engineering Division

Preliminary electrical conductivity measurements
of NaF-ZrF , (50-50 mole %) and NoF-ZrF, (57-43
mole %) hqve been obtained for the tempera'rure
range 1000 to 1800°F, For these measurements,
a thick-walled cylindrical tube of beryllium oxide
is mounted vertically in an open Inconel crucible
contdaining the molten fluoride. An electrode is
inserted through the tube, and the electrical path
is defined from the electrode just under the fiquid
surface to the Inconel at the bottom of the crucible.
However, the beryllia tube was not sufficiently
long to prevent noniscthermal regions in the cell,
and therefore accurate measurement of the salt
temperatures were difficult. A longer tube has
been ordered.

An attempt was also made to measure the con-
ductivity of NaF-KF-LiF by using the apparatus
described above, but this salt penetrated ond
eroded the tube so badly that the data obtained
were deemed unrelioble. A new cell in which
only platinum is in direct contact with the salt
haos been designed and is now being constructed.
This cell will be used to measure the conducti»
vities of salts which are not compatible with

 

4W. H. Rodebush and A. L. Dixon, Phys. Rev. 28,
851 {1925).

R. E. Moore and R. E. Traber, ANP Quar. Prog. Rep.
Sept. 10, 1953, ORNL-1609, p. 106.

99
ANP QUARTERLY PROGRESS REPORT

beryllia. Since slight contamination of the melts
occurred as a result of the constituent materials
of the crucible going into solution, a more noble
material than either Inconel or stainless steel is
indicated. Accordingly, a platinum crucible has
been designed and constructed for this purpose.

FORCED-CONVECTION HEAT TRANSFER
WITH NoF-KF-LiF EUTECTIC

H. W. Hoffman J. Lones
Reactor Experimental Engineering Division

Additional data have been obtained for NaF-KF-
LiF eutectic (11.5-42.0-46.5mole %)} flowing through
a heated nicke! tube. The experimental results
are presented in Fig. 8.3 in terms of the Colburr
j-function. The previously reported results for
NaF-KF-LiF in both nickel and Inconel tubes are

also shown.

In review, NaF-KF-Lif heat transfer data were
first obtained by using a nickel tube, The results
(shown by the inverted triangles in Fig. 8.3) indi-
cated that the anticipated heat transfer correlation
was followed. The experiment was then repeated
with the eutectic flowing through an Inconel tube,

and the data {circles in Fig. 8.3) were foundtolie
approximately 50% below the expected correlation.
The discrepancy was believed to be due to the
formation of a film of corrosion products at the
fluid-metal interface. The heat transfer results
and the existence of the film were both definitely
verified for NaF-KF-LiF in Inconel tubes in a
subsequent experiment (squares in Fig. 8.3). The
film was analyzed and found to be K,CrF [ plus a
small amount of Li CrF ..

The data reported ot this time were obtained to
establish the validity of the original results for
NaF-KF-LiF in nickel tubes. The results (normal
triangles in Fig. 8.3) are essenticlly in agreement
with those for nicke! tube No. 1;visual observation
of the tube showed no film. The experiment was
terminated because of fatigue failure of the nicke!
tube at the high operating temperatures used.

Experimental apporatus is being constructed for
studying fluoride films formed on Inconel as a
function of time and temperoture, Because of the
unavailability of pure K CrF, it has not yet been
possible to measure the thermal conductivity of
this material. Several other fluoride mixtures are
being studied to determine the thermal conductivity
of such films because they may form on Inconel tubes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UNCILLASSIFIED
DWG. 22219
0.010 e ey . - e -
- 0 B ¥ NICKEL TUBE NO I | 3
| A NICKEL TUBE NO.2
T T " B 1}* 1 - " ® INCONEL TUBE NO.1
R —1- -— ® INCONEL TUBE NG 2 g
| = | \ ‘
_ b L ~ bl S ,,,,,,] Lgl i
0.005 — ‘ ‘
_ I lL { [ ‘ | : I‘
5 v i
!G """" I ¥ 1 /.«—-‘t“‘*'::;;_l____\i """"" o ' ; | - ; 4-—
5 ‘ /,7" A b‘“"‘*--._._\\ ; ‘ | !
w i v A ‘ | ' /=0023 Re™9? | l
'\‘ — —_— e ——— ——— —;}J EE -‘ ‘ ,,,,,,,, 5 i, s
% 1 // ‘ ! ‘ ‘ { i
% v'/ | ! i ! ‘
E)J /v | | [ ! t Co
S 0002 L 7+7a oL Lol e foo b
| e o
o e -
| 5 | | ‘
el - 0
| | | | ||
Lo ‘ | |
\ | 1 ‘ | o !
0.001 ‘ L L I L
103 2 5 10% 5 103
REYNOLDS MODULUS
Fig. 8.3. The j-Modulus vs. Reynolds Modulus for NaF-KF-LiF {i1.5-42,0-46.5 mole %}.

100
FLOW IN THERMAL-CONVYECTION LOOPS

D. C. Hamilton F. E. Lynch
L. D. Palmer

Reactor Experimental Engineering Division

Thermal-convection loops have been used for
observing the nature of flow of ordinary fluids and
to determine the accuracy and range of usefulness
of an analytical method for predicting the mean
velocity, or Reynolds modulus, from wail tempera-
ture measurements, A discussion of the velocity
measurements ond flow observations was given in
a previous report.® Additional velocity and wall
temperature data have been taken, fluid temperature
traverses have been made, the analytical analyses
have been completed, and ¢ report of the results
is being prepared.” '

Measurements of temperature and velocity were
taken for five specific conditions over the range
of Grashof modulus from 2 x 104 to 67 x 104, The
corresponding variation of Reynolds modulus was
from 50 to 270. The Grashof and Reynolds moduli
are, respectively,

a’?’figfiw(max)
Gr, = ————————

and

 

where d is the tube diameter, v is the kinetic
viscosity of the liquid, 3 is the thermal coefficient
of expansion of the liquid, g is the acceleration
of gravity, ! -
6 (max) is the maximum variation in wall tempera-
ture,

is the mean fluid velocity, and

The flow ond heat transfer observed in these
experiments were characteristic of free convection
rather than forced convection, and thus the signifi-
cant parameter which dictates the character of the
flowis the Grashof modulus rather than the Reynolds
modulus. |

At Reynolds moduli of up to about 100, the flow
was essentially laminar. Above this value, the
Grashof modulus was superceritical, ond the flow
was laminar over part of the circuit ond turbulent

 

8p. €. Hamilton, F. E. Lynch, and L. D. Palmer, ANP
Quar. Prog. Rep. June 10, 1953, ORNL-1556, p. B9.

p. Hamilton, F, E, Lynch, end L. D. Palmer,
The Nature of the Flow of Ordinary Fluids in « Thermal
Convection Harp, ORNL-1624 (to be published),

PERICD ENDING DECEMBER 10, 1953

in other parts. As the Grashof modulus {and thys
Reynolds modulus) was increased, the length of the
circuit in laminar flow decreased. In addition,
when the Groshof modulus was high, the radial
temperature gradient was sufficient fo cause vigor-
ous secondary flow cells in the laminar part of the
circuit,

The method for predicting the Reynolds modulus
from wall temperature measurements employed the
numerical solution of the laminar flow equation for
heat conduction. The problem waos solved on the

ORNL high-speed computer, the ORACLE.

In Fig. 8.4, the predicted and measured values
of Reynolds moduli are compared. For the first
two sets of datq, for which the flow was essentially
l[aminar, the correlation is no worse than 30%; for
Reynolds moduli grzater than 100, the flow was
observed to be partly turbulent, and o laminar flow
solution is obviously not applicable. The predic-
tion is satisfactory for the laminar flow regime, but
a similar prediction for the turbulent flow regime
cannot be made because the turbulent diffusivity is
not known for this type of turbulent flow.

UNCLASSIFIED
DWG 22220

 

 

 

 

 

 

 

 

 

1000 i l J ,

800 R - l ,,,,,,,,,,, ...._A_ .......
................. ‘ POSTULATED VELOf‘ITY e 4o
600 | ...l DISTRIBUTION: 4 elbo
................. o -U_m_ =4 |

400 |m—ed A @--efi-u—m s e //
30O o e fi—r—- seesquesqenpeepasgeen e Dt

3 o ] ° A

@ 200 w#ul - // /.

o

= 1 /

2 | 5

9 | +30 % //

O \ ]

£ 00 e YR

[ '1' /-(

a ‘

',_

o

o

Ll

1

.

 

 

 

 

 

 

 

 

 

 

 

. 'z J ________________

 

20 40 80 BO GO 200
MEASURED REYNOLDS MODULUS

10 20 300 40@

Fig. 8.4. Comparison of Predicted and Measured
Reynolds Moduli for a Therma! Convection Loop.

107
ANP QUARTERLY PROGRESS REPORT

The method used in the previcus work for measur-
ing velocity is not applicable to higher flows
Therefore a velocity measuring device is being
developed which will be applicable at higher
velocities and which will not offer resistance to
the flow.

FLUID FLOW IN AN ANMULULIS
J. O, Bradfute

Reactor Experimental Engineering Division

The experiment to be used for measuring fluid
velocity profiles in annulus systems was described
previously.? In essence, the method consists of
photographing dust particles which are entrained
in air flowing in an annulus. Most of the recent
effort on this research was devoted to phatographic
experimentation to develop a procedure for photo-
graphing the grid system and the dust particles
with the light available from a single flash of a
Strobolume. This has been accomplished.

The problem stemmed from the fact that colimating
a beam of light into a sharply defined region in
space, G necessary
inefficient,

is extremely
A pair of parallel slits, each about
2 mm wide and opproximately 25 cm apart, with the

requirement,

Strobolume placed immediately behind the last one,
produced what seemed to be the optimum combina-
tion of intensity and definition in the usable por-
tion of the beam. Although about 700 lumen-seconds
per flash of the light energy was dissipated by the
light source, the developed image of the grid sys-
tem obtained by using Kodak Tri-X panchromatic
film ond o stop opening of f:4.7 was so under-
exposed that even the doubly accented centimeter
lines were barely visible upon careful examination.
Several film sensitization and latensification
materials, as well as high-speed developers, were
tried with varying degrees of success and failure,
A :1.8 lens was obtained, and, by employing this
lens and following a complex developing process,
negatives were produced which have sufficient
contrast for satisfactory printing.  Chalk-dust
particles were also photographed by applying this
procedure,

TRANSIENT SURFACE-BOIJLING STUDIES
M. W. Rosenthal

Reactor Experimental Engineering Division

In the event o water-cooled reactor suddenly
become supercritical, the time required for produc-

102

tion of a certain volume of steam would be of im-
portance, Of primary interest would be the effect
of generation of heat ot rates increasing exponen-
tially with time in vertical plates (such as the fuel
plates of an MTR-iype reactor) submerged in water
that was initially at below the boiling point.

To obtain information on this phenomenon, an
experimental study of heat transfer to water under
transient conditions has been undertoken. The
basic experimental system is a flat metal filament
positioned vertically in o tank of water with heat
generated electrically in the filament. The resecrch
program consists of twe phases: a study involving
exponential generation of heat, which will be
started when the necessary electrical equipment
has beern developed, and a preliminary general
study of boiling from a vertical filament, which
will begin immediately.

Arrangements have been made for the Instrumenta-
tion and Controls Division to develop and construct
equipment to regulate the flow of current in the
filament and to produce an exponential heating
curve. Thisequipment will also measure and record
instantaneous values of the variables of interest.
The filament and the adjacent water will be photo-
graphed by using a high-speed
camerg

imotion-picture
syiichronized with the heat generation
system.

CIRCULATIMG-FUEL HEAT TRANSFER

H. F. Poppendiek .. D. Palmer
G. M. Winn

Reactor Experimental Engineering Division

Some preliminary experimental temperature meas-
urements have been obtained in a forced-flow
volume-heot-source system under !aminar flow
conditions. The volume-hect-source is generated
within the flowingelectrolyte by passing an electric
current through it. In general, the experimental
results are in ugreement with the previcusly de.
veloped laminar flow theecry., At low flow rates
(Reynolds number of 600 or less), the effect of the
free convection which is superimposed on the
forced convection was observed. The variable
ronges within which theoretical and experimental

forced-flow volume-heat-source behavior have been

SJ, 0. Brodfute and J. I. Lang, ANP Quor, Preg. Rep.
March 10, 1953, ORNL.-1515, p. 160.
compared to date are the following:

600 < Re < 14,000,

4.6 < Pr < 8.7,

to~— 1§

3 x 10—+ <6 x 10-2 ,

m
2
Yo

k

The next set of experiments is to be made with the
test section in a vertical rather than a horizontal
position; also, concentrated sulfuric acid is to. be
used as the electrolyte so that a new Prandtl
modulus range con be studied. '

Currently, the forced-turbulent-flow temperature
solution for the case of fluid flow between parallel
plates and annular spaces with volume heat sources
in the fluids is being evaluated, This solution
will be useful in estimating fluid temperature
distributions in reactor systems such as the
reflector-moderated reactor.

PERIOD ENDING DECEMBER 10, 1953

Forced-laminar-flow volume-heat-source velocity
and temperature solutions have been derived for
the case in which the fluid viscosity is temperature
dependent. A simple, algebraic viscosity-tem-
perature expression which closely opproximates
experimental viscosity data was substituted into
the hydrodynamic equation. A power series with
unknown coefficients was substituted into the
heat transfer equation, and two integrations were
carried out. The resulting temperature solution
was then substituted into the hydrodynamic
equation, and an integration: was carried out.  The
coefficients of the resulting power series velocity
solution were determined by relating them to the
initially proposed velocity power series and the
integral equation describing the mean fluid velocity.
The resulting velocity and temperature solutions
are functions of a new dimensionless modulus, as
well as radial position. = This dimensioniess
modulus is a measure of the importance of the
viscosity variation in the flow system.

 

9. RADIATION DAMAGE

J. B, Trice, Solid State Division
A. J. Miller, ANP Division

Inconel capsules containing the fluoride fuel
NoF-ZsF ,-UF , were irradiated in the LITR and in
the MTR. Examination of the capsule walls showed
a tendency toward intergranulor corrosion that does
not occur in unirradiated static tests. The dato on
depth of penetration showed considerable scatter,
and therefore no correlation with power production
rate and exposure time could be made. Design and
construction ‘continued on the inepile circulating-
fuel loops and the MTR creep test equipment.

IRRADIATION OF FUEL CAPSULES
G. W. Keilholtz C. C. Webster
J. G. Morgan M. T. Robinson
H. E. Robertson W. R. Willis

W. E. Browning
Solid State Division

Inconel capsules containing various compositions

‘of the fluoride fuel system, NaF-ZrF 4-UF,, have

been irradiated in both the Lowwlntensny Test
Reactor (LITR) and in the Moterials Testing Re-

actor (MTR). These tests have included irradiation
times of from 53 to 800 hr with power production
from the 230 watts/cm3 with a 4 mole % UF , fuel
in the LITR to the 8000 watts/cm® with a 12 mole
% UF , fuel in the MTR. In general, Inconel sur-
faces show more tendency toward intergranular
corrosion when in contact with irradiated fuel than
they do in out-ofepile tests. However, the data so
far available indicate no definite trend in corrosion
penefration either as a function of rate of power
production in the fuel or of irradiation time.

Six Inconel specimens containing the fluoride
mixture NoF-ZrF -UF  (50-46-4 mole %) which
were irradiated in the LITR for 53, 140, 270, 565,
685, and 810 hr each are shown in Fig. 9,1. In
each case, the power dissipation in the fluoride
was 230 watts/em3. Out-of-pile control capsules
for this series of tests are shown in Fig. 9.2.

Inconel capsules which were irradiated in the
MTR with power production in the fluoride mixture
of 2500, 3900, and 8000 watts/cm? are shown in
Figs. 9.3, 9.4, and 9.5, respectively. The Naf-

103
ANP QUARTERLY PROGRESS REPORT

 

 

Fig. 9.1. Static Corrosion of Inconel by NuF-ZrF4-UF4 (50-46-4 mole %) After Exposure at 1500°F in the
LITR to 230 watts/cm> for Various Irradiation Times.

ZrF -UF , system wos used in each of these tests,
although the composition was varied and the UF
content was 4, 6.5, oand 12 mole %, respectively,
in the three tests. The sample shown in Fig. 9.5,
which was contacted for 575 hr in the MTR with
fuel producing power at a rate of the order 8000
watts/em3, has only 3 or 4 mils of corrosion pene-
tration and is similar to the sample shown in
Fig. 9.1, which was exposed at only 230 watts/cm3.
In a few samples in which the Inconel had a very
large grain size, occasional penetration to a depth
of 12 mils has been observed. This phenomenon
appears to be fairly independent of power density
or time of irradiation, and, as with the other effects
observed, there is no certainty as to how much of
the effect is due to radiation,

104

In confirmation of the analytical werk reported
previously, petrographic examinations of irradiated
fuels showed no decompasition or segregotion of
the fuel as a result of irradiation.

CREEP UNDER IRRADIATION

W. W. Davis J. C. Wilson
N. E. Hinkle J. C. Zukas
Solid State Division

No new creep tests were mode during this quarter,
The recent increase in power of the LITR causes
excessive microformer operating temperotures,
and the increasing incidence of microformer foilures
(which occurred even before the power was in-
creased) prevents relicble strain measurements.
| uncLAssiFieD |
| PHOTO 11985

Fig. 9.2. Static Cotrosion of Inconel Exposed to
NoF-ZrF -UFJ {(50-46-4 mole %) ot 1500°F for

Various Times.

Accordingly, no more creep tests will be operated
in the LITR until a svitable strain transducer is
found. Both the Bourdon-tube extensometer and a
bimetal expansion extensometer now being de-
veloped show promise of becoming reliable meas-
uring instruments for use in the LITR ond in the
MTR. |
Several models of an in-pile stress-corrosion
apparatus are being built to determine the best
system for conducting stress-corrosion tests. The
discovery of an effect of straining on the corresion
of Inconel! and the recent observation of inter-
granular corrosion in static, inspile fuel irradiations

 

'R, B. Oliver et al., Met. Div. Semiann. Prog. Rep.
April 10, 1953, ORNL.1551, Figs. 43 and 44, p. 52, 53.

 

PERIOD ENDING DECEMBER 10, 1953

| UNCL ASSIEIED
E PHOTO 11961

 

Fig. 9.3. Static Corrosion of Inconel by NaF-
Zrl=4-l.ll"'4 (50-46-4 mole %) After Irradiation in the
MTR at 2500 wotts/cm3 for 575 hr ot 1500°F.

UNCLASSIFIED
PHOTO 11962

OUT-0OF~PILE
SPECIMEN

IRRADIATED
SPECIMEN

Fig. 9.4. Static Corrosion of Inconel by NaF-
ZrFA-UF - {53.5-40-6.5 mole %) After 419 hr at
1500°F \éith and Without lrradiation in the MTR at
3900 watts /cm3.

105
ANP QUARTERLY PROGRESS REPORT

 

Fig. 9.5. Static Corrosion of Inconel by NaF.
Zer'UF4 (63-25-12 mole %) at 1500°F With and
Without Irradiation in the MTR at 8000 watts/cm3,

have pointed to the need for data on the effects of
in-pile stress corrosion.

The MTR creep test rig has been welded shut,
but considerable difficulty has been experienced
with legks in the helium supply system and in the
seals where the lead wires exit from the pressurized
portion of the experimental plug. The complete

106

instrument assembly has been tested ond found
satisfactory.

IN-PILE CIRCULATING LOOPS

0. Sisman C. Ellis
W. E. Brundage M. T. Morgon
F. M. Blacksher A. S, Olson

W, W. Parkinson
Solid State Division

In-pile circulating loops for studying the corrosion
of Inconel by fluoride fuels under dynamic con-
ditions in a neutron flux are being built for insertion
in hale HB-2 of the LITR. It is expected that
similar loops con be used in the MTR. A loop
consists essentially of U-shaped sections of
0.225-in.-0D tubing, which will be in the neutron
flux, thick-walled 1/2-in. tubing thot will connect
the irradiated section with a Venturi flowmeter and
an air-cooled heat exchanger just inside the reactor
shield, and a centrifugal sump pump outside the
shield, All these components, except the shaft
and seal portion of the pump, will be enclosed in a
helium atmosphere. All in-pile components, other
than the pump, will be surrounded by a water-
cooled jacket. _

The fabrication of components for the first loop
is 80% complete and most of the work remaining to
be done on this loop consists of welding or brazing
together the finished parts. The components for a
second loop are 40% complete. The withdrawal
shield is being fobricated, but work on the shield
for the pump has not started. The design of the
instrumentation is 90% complete, and about 85% of
the instruments and controls are on hand.
PERIOD ENDING DECEMBER 10, 1953

10. ANALYTICAL STUDIES OF REACTOR MATERIALS

C. D. Susano, Analytical Chemistry Division

J. M. Warde, Metallurgy Division

R. Baldock, Stable lsotope Research and Production Division

Developmental work was mainly concerned with
the investigdtion of methods for the determination
of the oxidation states of the constituents and the
metallic corrosion products in fluoride fuel mix-
tures and fuel solvents. Divalent iron fluoride was
ieached from samples of fluoride mixtures with 0.1
M H,50,, and the iron was determined as the o-
phenanthroline complex. Divalent chromium fluoride
was found to be complexed by 0,17 M (NH )2C204
and could be distinguished from trivalent chromium
fluoride, which is readily complexed with a dilute
solution of disodium dihydrogen ethylenediamine-
tetraacetate (EDTA) at pH 4.0.

Studies were continued on the use of ammonium
oxalate for the determination of traces of UF ., in the
presence of UF . A separation based on solubility
difference is possible but not for low concen-
trations of UF,. The absorption spectra of so-
lutions of UF, in EDTA and HCl were determined.

The effectiveness of solutions of ammonium
oxalate, oxalic acid, and acetic acid as solvents for
removing fluoride fuels, which contain zirconium,
from metallic thermal convection loops was in-
vestigated. Ammonium oxalate proved the most
effective solvent tested; however, it was not con-
sidered so satisfactory as the nitric and boric acid
mixtures which are currently used.

Petrographic examinations were made of irradiated
samples of fluoride fuel mixtures. No difference
was noted under the petrographic microscope be-
tween the irradiated specimens and the comparable
control samples.

The mass spectrometer method for quantitative
analysis of the uranium content of irradiated fluoride
fuel mixtures has recently been extended for use as
a measure of uranium burnup during irradiation. A
procedure has been developed for calculating the
percentage of uranium by weight in both the un-
irradiated and the irradiated conditions. Hence, the
loss in uranium content can be determined to the
accuracy that the isotope abundances are known in
the originol material. The portion of this total
uranium loss that is due to burnup is then determined
by mathematical analysis.,

The activities of the Analytical Service Laboratory
have increased markedly during the past quarter.
A total of 1,378 samples was received as compared
with 878 for the previous quarter, and 1,201 samplies
invelving ‘o total of 11,034 determinations were
reported. i

ANALYTICAL CHEMISTRY OF REACTOR
MATERIALS

J. C. White
Analytical Chemistry Division

The developmental work during this quarter con-
sisted almost entirely of a search for methods for
determining the oxidation states of the constituents
and the metallic corrosion products in ARE fuels
and fue! solvents. Three phases of the problem are
currently being studied: the determination of the
oxidation states of iron in fuel solvent and in the
fuel; the deterinination of the oxidation states of
chromium in similar media; and the determination
of the oxidation states of uranium in the fuel. The
ultimate aim of these investigations is the determi-
nation of each specie in the presence of the other
species. The reducing power of specially prepared
NaZrF_ with zirconium additive was also de-
termined. The effectiveness of certain solvents
for disselving proposed fuels from the walls of
metal contciners was investigoted. No additional
work was conducted on the determination of metallic
oxides in reactor fuels because of the urgency of
other problems. '

Oxidation States of {ron

P. V. Hottman
Analytical Chemistry Divisien

A number of samples of the fuel mixture NaF-
ZrF ~UF | (50-46-4 mole %) and iron (5 wt %) were
heated and filtered. Divalent, trivalent, and total
iron contents were determined, in addition fo
vranium frifluoride. The finely divided samples
were leached with 0.1 M H2504 on the steam bath
for 1 hr to remove iron. Ferrous iron in the filtrate

107
ANP QUARTERLY PROGRESS REPORT

from the leach was determined with o-phenanthro-
linz;! iron was determined in the leach filtrate by
first reducing the iron present with hydrogquinone
and then determining it as the o-phenanthroline
complex. In order to test the efficiency of the
leaching process, a separate sample was dissolved,
and the iron was then determined by the o-phen-
anthroline method.

The results, with a relative standard error of 10%,
show that a single acid leach is sufficient to leach
iron quantitatively from the fuel mixture, Further
evaluation of the data shows that only o small
fraction of the added iron reacts with the fuel and
passes in solution through the filter. The oxidation
state of this iron is predominantly, if not com-
pletely, bivalent. {t should be recognized, however,
that if FeF_ and UF, exist as such in the mixture,
reduction of trivalent iron to the bivalent state
vndoubtedly occurs on contoct with sulfuric acid.
The absence of UF_ was indicated by the hydrogen
evolution method.2” However, this method is not
sufficiently sensitive to determine the probable
small concentration of UF, which may be present
in these samples. The need for 0 sensitive method
for determining UF, in reactor fuels is evident.

Oxidation States of Chromium

D. L. Manning
Analytical Chemistry Division

Work has continued on the determinotion of the
oxidation states of chromium following its reaction
with UF& in fuel solvent NcquFs. The tormula for
the reaction is

NaZrF
=== Cr® + UF, .

800°C

It has besn shown that chromium trifluoride can
be separated from chromium meta! by leaching with
an acetate-buffered solution (pH, 4,0) of disodium
ditydrogen ethylensdiaminetetraacetate (EDTA).
Chromium metal is insoluble in this medium. Several
samples were leached with EDTA at 60 to 70°C for
3 hr to determine the extent of reduction of CrF, to
metallic chromium during the heating peried. No
residues

 

(1) CrF, *+ 3UF,

indicotive of metallic chromium were

 

15, C. White et al., AMP Quor, Prog, Rep. June 10,
1952, ORNL-1294, p. 177.

2D, L. Manning, W. K. Miller, and ®&. Rowan, Js.,
Methods of Determingtion of Urenivm Trifivoride, ORNL.
1279 (Apr. 25, 1952).

108

obtained. This suggests that chromium trifluoride
is either not reduced to metallic chromium or that
the metal obtained is in such a finely divided and
reactive state that it dissolves in the leach so-
lution, However, previous tests weuld indicate
incomplete reduction, probably to divalent chromium
fluoride.

Since EDTA solutions complex both di- and tri-
valent chromium, other complexing agents were
studied to distinguish between CrfF, and CrF .
Preliminary experimenis have indicqteé that a 0.17
M (2.5%) solution of ammonium oxalate will readily
dissclve CrF_ at the refluxing temperature, and a
purple complex will form.  Metallic chromium,
CrF,, and Cr 03 appear to be insoluble, and they
show no te«néency to form complexes. The ab-
sorption spectrum of the complex formed by divalent
chromium fluoride and ammonium oxalate is identical
with that of the complex formed by the recction
between trivalent chromium chloride and ammaonium
oxalate. Further investigation of the nature of this
complex is ynder way.

The work on the determination of the oxidation
states of chromium in reactor fuels has shown that
(1) divalent chromium fluoride can apparently be
selectively leached with 0,17 M ammonium oxalate,
(2) divalent and trivalent chromium fluoride can be
leached simultanecusly with EDTA, and (3) the
metal and the trivalent oxide are not attacked by
either of these solvents.

Determination of UF3 and UF’4

W. J. Ross
Analytical Chemistry Division

Work on the development of a msthod for the de-
termination of UF, and UF_ in NeZrF, hos been
continued, and the more complete studies show, as
reporfed previously, that EDTA dissolves UF, but
apparently hes little effect on UF,. The cftects
of temperature and the concentration and acidity of
EDTA on the recction between UF, and EDTA
were investigated. Concentration o? the reagent
was not critical over the range 0.08 to 0,15 M; a
concentration of 0.1 M was used in most of the
tests. Solutions of EDTA with acid concentrations
of pH 4 were not appreciably more effective in
dissolving UF, than thoss with acid concentrations
approaching a pH of 7. Use of solutions less acidic
tharn pH 7resulted inhydrclysis of uranium. Temper-
ature was the most critical condition. The rate of
reaction at 100°C was approximately ten times that
at 25°C; however, the data obtained at 25°C were
not reproducible because of variations in particle
size. ‘ -

Under various conditions of temperature uand
acidity, the absorption spectra of UF -EDTA are
essentially identical with those for UF, dissolved
in HCl, except that the spectra for the 4UF4 EDTA
systems are shifted by approximately 20 my toward
the infrared by the presence of EDTA. The maxi-
mum molar extinction coefficients for the four
spectra shown are nearly equivalent; they range
from 35.7 for UF -HC| to 25.0 for UF -EDTA at a
pH of 6.8 and a tempemture of 100°C.

Petrographic and chemical analyses of the three
batches of UF:3 used for the study of solubility in
0.1 M EDTA showed UF, contents of 92.4, 90.4,
and 90.7%, respectively. %ampies of these batches
were treated with 0.2 M ammonium oxalate at reflux
temperature and filtered, and petrographic exami-
nation revealed that the residues were essentially
free from UF,. The data indicate that during the
first hour of reflux the greater portion of UF , present
is removed but that quantitative separation is
questionable. The solubility of UF,
oxalate is low; however, UF, may oxidize slowly
upon contact with this solvent, as indicated by the
appreciable concentration of uranium in the solution
when allowed to stand overnight. The UF content
of the sample which had been treated wn‘h oxala’re
is being determined and is expected to be near
100%.

On the basis of the data obtained thus far, UF
and UF, can be separated by ammonium oxuiate.
However, the method is not sufficiently precise for
the separation of traces of UF, from UF,. Similar
tests are under way to determine the feasibility of

using 0.1 M EDTA for separating UF .

in ammonium

Reducing Power of NaZ(F . with Zirconium Addition

D. L. Manning
Analytical Chemistry Division

The ftotal reducing power of NaZrF, specially
prepared by treatment with metallic zlrconlum was
determined. In the hydrogen evolution method?® ysed
for this determination, the sample is treated with
0.2 M HF to liberate the hydrogen which is formed

by the reaction between zirconium and hydrofluoric

 

3. C. White and W. J. Ross, ANP Quar. Prog. Rep.
March 10, 1953, ORNL-1515, p. 173.

PERICD ENDING DECEMBER 10, 1953

the gas is collected and measured; and the
reducing power is calculated. Essentially no re-
ducing power was observed in these somples. A
second method which involved complete dissolution
of the sample in standard ceric suifate solution in
9 M H2504 and back-titration with standord ferrous
sulfate solution was used to confirm the results.
Further tests by the hydrogen evelution method on
synthetic samples of NaZrF_ and zirconium demonw
strated that the method is quantitative,

acid;

Dissolution of Fluoride Mixtures
Containing Zirconium

P. V. Hoffman
Analytical Chemistry Division

Nitric and boric acid mixtures have been used
successfully as solvents for fluoride mixtures, but,
in an effort to find a less-corrosive solvent, small-
sized lumps of NaF-ZrF --Ut':4 (50-46-4 mole %)
were contacted with varlous concentrations of
ammonium oxalate, oxalic acid, and acetic acid at
70 to 80°C for various periods of time. The optimum
concentration of ammonium oxalate was 1 M, which
is approximately o saturated solution at 50°C.
[(One hundred ml of T M (NH ) C O M O will
dissolve 1 g of small lumps of Fuel un upproxnmcately
2 hr.] A 2 M solution was the most effective
oxalic acid concentration and was approximately
equivalent to 1 M ammonium oxalate in dissolution
effectiveness. The optimum acetic acid concen-
tration was 0.2 M, which yielded only 50% of the
solubility strength of the oxalates. A comparison
of the effectiveness of the three reagents tested is
shown in Fig. 10.1. The rate of dissolution with
1 M ammonium oxalate is greater than the rate with
2° M oxalic acid, although their effectiveness is
nearly equivalent. Ammonium oxalate (1.0 M),
because of its higher pH {6.5), is to be preferred
over 2 M oxalic acid (pH (.7), however, with respect
fo possible corrosiveness. None of these solvents
are svitable, however, for dissolving large-sized

lumps of Nan-ZrFA-UF4

from metallic containers.

PETROGRAPHIC EXAMINATION OF FLUOR?DES

G. D. White
Metallurgy Division

T. N. McVay
Consultant, Metaliurgy Division

Petrographic examinations were made of fluoride

fuel mixtures NaF-ZrFé-UFd (50-46-4 mole %) and

109
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i
100 OWG. 22221
i | T
14 (I“JH4)2C204""*"‘-\\‘>//
ap __..__‘;_________i // // ________ ]
|
| / //2 M 1,0,0,
80 7 —
7o // R
.. 60 // : S SR
% ‘
£ 0.2 M HC,H30, ‘
2 '3¥e |
a |
S 50 — e
w
a
n
3 /
3 40 HH—-- / ______
30 // ————————— -
20 R —_
10 H— e
o ‘
0 0.50 100 1.50 2.00

TIME (fr)

Fig. 10.). Comparisen of Effectiveness of Solu-
tions of Ammonium Oxclate, Oxalic Acid, ond
Acetic Acid as Solvents for Nr.aF-Z’.rFA-UF4 (50-46-4
mole %).

(53.5-40-6.5 mole %) which had been irradiated a
maximum of 492 hr in the MTR. Unirradiated conirol
samples were examined at the same time. These
fuels are normally composed of Nqu(U)Fs, which
is green, and perhaps a very smallamount ofa lower
refractive index complex, which is designated E-2.
The irradiated samples were normal, except that

 

the NaZr(U)F_ had a slightly low refractive index
and was slightly reduced. There were no differences
between the control samples and the irradiated
samples which could be observed with a petrographic
microscope.

MASS SPECTROMETER IMYESTIGATIONS OF
IRRADIATED FLUORIDE FUELS

C. R. Baldock

Stable |sctope Research and Production Division

The mass spectrometer has been successfully
employed in the quantitative anclysis of the uranium
content of irradiated fluoride fuel mixtures. Initially,
the mass spectrometer method was used to test the
validity of chemical analyses, but it was recognized
that the method could be extended to obtain an
accurote measure of fuel burnup. In addition, by
using accepted values of cross sections,an average
valve of fthe flux in which the fuel has been ir-
radiated can be determined. This gives an in-
dependent check of values predicted from other
types of measurements and of the internal con-
sistency of the celculations. The isotope dilution
method as applied to the analysis of irradiated fuel
has been used extensively with a routine accuracy
of £1%, which is not the ultimate limit of accuracy.

Calcuvlation of UF4 in Unirradioted Fuels

For the calculation of the percentage of UF,
(by weight} in unirradiated fluoride fue! samples of
known isotopic composition, the exact values for
the isotopic masses are used to compute the atomic
weight of the uranium in the fuel:

(1) At.wt = 235124, + 234124,

where A_ denotes abundance of uranium isotope of
mass number m. The ratio, R, of U235 to U?38 js
obtained from the mass spectrometer measurement,
and since the atoms of U235 gnd U238 gre obtained
from both the vranium in the fuel and the natural
U3OB spike materia! added,

(atoms of U235 from fuel) + (atoms of U235 from spike)

(2) R =

 

(atoms of U238 from fuel) + (atoms of U238 from spike)

110
The atomic weight of uranium in the spike is 238.07,
and therefore

 

 

W1F1YA235 W2F2A235
@) 2 at. wt 238.07
= ?
W F YA, . WoFaAaag
af. wt 238.07

where
W, = weight of uranium-bearing fuel,
W, = weight of U308 spike material,
F, = U/UI"*"4 = gravimetric factor for uranium
in fuel,
F, = 3U/U,0; = gravimetric factor for urani-
um in spike,
Y » 100 = percentage of UF, (by weight) in fuel.
if Eq. 3 is solved for Y,

 

 

WoFdsss WoF 4,38

) v 238.07 238.07
W1F1A233 W1F1A235

at. wi at. wt

 

(szz) (at. wt) (Azas - RA238>
Calculation of U235 Lost from Irradiated Fuel

if C is defined as the number of gram atomic
weights of U235 in W, grams of fuel before ir-
radiation, then

(5) c WIF!YA235

at. wt
and, for the irradiated spiked fuel sample, the
following equation must hold:

W,F A

 

 

2235
+ — s
@ R = 238.07
Wi F\ Y435 Waladysg
at. wt 238.07

where X is the gram atomic weight of the U235
present in W, grams of fuel after irradiafion. Solving
Eq. 6 for X gives

PERIOD ENDING DECEMBER 10, 1953

WrFIYAzaa WzeAzas
(7) X = R ————~+ R ———me
at. wt 238.07
wzeAzss
238.07

Then, the gram atomic weight of the U233 lost from
W, grams of fuel, from Eqs. 5and 7, is
]
W1 F1 Y

RA

@) C - X = (Ayg5 — RAygy)

af. wt
W2F2

238.07

 

and the percenfage of U235 lost from the fuel
follows as

C - X
—x 100 .
C

Determination of U235 Burnup

Equation  givesonly the total percentage of U235
lost from the fuel. However, this loss may be due
to several factors, including burnup, radiation
damage, and replacement of chromium leached from
the Inconel walls, A graphical method has been
developed to test this loss to determine whether it
can be completely ascribed to the capture of neutrons
to produce fission preducts and U236, that is, to
burnup.

The solution of the differential equation for the
first-order rate expression for burnup is found to be

ug?s (a, + a) et

9 IO = 7
® ¥ 235 2.303
i

 

where
a, = fission cross section =580 barns,
a, = capture cross section = 105 barns,

N e

= neutron fluxdensity=2.17 x10"¥n/cm?.sec,

- G-

H

irradiation time, sec,

and the slope is
(a, + az) b

10 e = 6.4544 x 1078
(10) 2.303

111
Therefore
U(2)35
(1) log =

U235
!

6.4544 x 10~8; .

 

If a plot is made of Eq. 11, which might be called
the theoretical curve, it is possible to compare the
experimental results on burnup with the results
predicted by theory, In addition, Eq. 11 can be
used to check the product of the three constants
Qyp Oy and ¢. Or, if the (a, * a,) values are
assumed to be correct, Eq. 11 can be used as a
check on ¢, the averoge flux density.

The values of a, = 580 barns and a, = 105 barns
are rather well established. The value of ¢ =
2.17 x 1014 n/cm2.sec has been tentatively chosen
because it gives the best fit to the data based on
samples from four test capsules and it is consistent
with values determined by other means,

SUMMARY OF SERVICE CHEMICAL ANALYSES

J. C. White A. F. Roemer
C. R. Williams
Anclytical Chemistry Division

Sixty-four samples of UF, and ten samples of
NGQUF6 (the ARE fuel concentrate) have bheen
analyzed to date. The UF | samples were analyzed
for uranium only, but comp?ate analyses were made
of the fuel concentrates, including determinations

TABLE 10.1.

of iron, nickel, and chromium. The cooperative
program with the {_aboratory Division of Y-12 was
continued on the determination of uranium, as
described in the previous report.? The relative
difference between loboratories on the determi-
nation of uranium in the fue!l concentrate was
0.04%, ond the difference on the UF4 analyses was
0.18%. The larger difference was ottributed mainly
to the heterogeneity of the samples of UF . which
contained black particles and comparatively large
hardpieces ofa lighter hue than that of the powdered
salt; the large pieces were probably UO,F,,
Uranium metal that was considered to be 99.94%
pure was used as the standard. The relative
standard error of 12 samples at a 95% confidence
level was 0.02%.

Some specimens of Inconel metal used in corrosion
tests were analyzed for uranium, zirconivm, sodium,
and fluoride and the major constituents nickel, iron,
and chromium. Traces of sodium carbonate in the
range of 0,05% and less in sodium hydroxide were
determined.

The major portion of the work wos concerned with
the anolysis of fluoride fue! mixtures. A decided
increase occurred in the number of samples received
during the quarter. A total of 1,378 somples waos
received, 1,201 were reported, and 11,034 determi-
nations were made {Table 10.1).

 

4. C. White er ol., ANP Quar. Prog. Rep. June 10,
1953, ORNL«1609, p- 121.

SUMMARY OF SERVICE ANALYSES REFPORTED

 

 

 

 

NUMBER OF NUMBER OF
SAMPLES DETERMINATIONS
Experimental Engineering 595 6,411
Corrosion Studies 271 3,247
Reactor Chemistry 256 1,000
ARE Fluid Cirguit 74 328
Heat Transfer and Physical Properties Studies 4 48
1,201 11,034

 

 

 

112
 
INTRODUCTION AND SUMMARY

E. P. Blizard, Physics Division

The spectra of gamma rays and of fast neutrons
from the Bulk Shielding Reactor were measured
(sec. 11). The neutron spectrometer was altered
for this work to improve the sensitivity at some
sacrifice
water shields can now be studied. The gamma-ray
spectrum of the Bulk Shielding Reactor without a
shield showed the usual broad peaks at 2.2 Mev
that are due to water capture and those at 7 to 8
Mev that are due to capture in aluminum and possi-
bly other materials., The fast-neutron leakage
spectra of the BSR were obtained from profon-recoil
measurements for the water-reflected reactor {oad-
ing No. 22, Data obtained with the nuclear plate
camera have been used to extend the spectrometer
data from 1.3 Mev down to 0.5 Mev. A measure-
ment of the angular distribution of fast neutrons in
water was also made. At neutron energies above
8 Mev, the intensities at various angles converge.

The Lid Tank Facility was used for studying the
slant penetration of fission neutrons in a hydro-
genous shield (sec. 12). This problem is encoun-
tered in specifying a crew-shield side-wall thick-
ness. 1he uncertainty lay in the importance of the
short circuiting of the shield by scattering into
paths more nearly normal to the shield surfaces.
For the shield thicknesses which have been meas-
ured, it appears thot this short circuiting is un-
important, and the attenuation is characteristic of
the slant thickness. This is not unexpected for
the thin shields (< 7.5 ¢m) which have been meas-
ured to date. There is as yet nothing conclusive
about the behavior of very thick shields with
slant-incidence radiation, but it will be advan-
tageous if the slant thickness gives the ottenuation
for the thicker shields,

The Lid Tank Facility was also used for air duct
experiments designed to give some insight into the
interaction between paralle! ducts (sec. 12). Data
from three sources, (1) a duct consisting of three
sections, 22 in. long and 3,5/] in. in diameter,
joined at angles of 45 deg, (2) a second duct con-
sisting of two such sections, and (3) the siant
penetration experiments, were put into the Simon-
Clifford formula, and the formula was solved for
the reflection factor. The values obtained were
in agreement with those obtained from previous
measurements and thus constituted an

in energy resolution; however, thicker

thermal
important confirmation of those measurements. It
is now felt that engineering designs based on the

Simon-Clifford theory are on a sound basis when
applied to-a single duct, especiaily when the di-
mensions are similar fo those studied. Since it
was observed that the peak values of dose meas-
vred at the end of the 19-duct array were a factor
of approximately 10 higher than those observed
with the single duct, o study of interference be-
tween adjacent ducts has been started. Tests are
planned to give more quantitative information on
the interaction between neighboring ducts. and
possibly to improve the formula for streaming
through ducts.

The effective removal cross section work for
which the Lid Tank Facility is used was con-
tinved, and old experiments have been reviewed.
A new set of values is reported (sec. 12).

A summary of the first 30 L.id Tank experiments
is being prepared (sec. 12). The summary will be
useful because of the large role the Lid Tank
Facility has had in the development of various
reactor shield designs and because of the constant
need for reference to the data.

The stee! structure of the Tower Shielding Fa-
cility has been erected, and the ground structures
are approximately 75% complete (sec. 13). The
optimistic date for completion of the facility is
Jonuary 15.

Some basic shielding research which is not re-
ported here because it is supported by another
activity is, nevertheless, of vital interest to ANP.
The development of a scintillating fast-neutron
spectrometer has received considerable impetus
recently with the growth of o few clear europium-
activated lithium iodide crystals. These crystals
have been tested on the ORNL Van de Graaff
generator and show encouraging energy sensitivity,
but an anomalous lack of energy resolution indi-
cates that the development is far from complete.
Most of the theoretical shielding work has been
reported in the Physics Division Semiannual Re-
port with the exception of o few problems which
originated in the experimental program (sec. 14).
An unusual calculation has been made of the vis-
ible light to be expected from divided-shield
aircraft.  An estimate of the neutron reflection
coefficient for water was obtained by using @
simple model with a coincident source and a
nondirectional receiver located above the water
surface.

115
ANP QUARTERLY PROGRESS REPORT

11. BULK SHIELDING REACTOR

R. G. Cochran
F. C. Maienschein

H. E. Hungerford
E. B. Johnson
T. A. Love

G. M. McCammon

Physics Division

During this quarter, the main effort at the Bulk
Shielding Facility was directed toward obtaining
neutron and gomma-ray spectroscopic data. The
gamma-ray leakage spectrum of the Bulk Shielding
Reactor has been meosured becouse it is typical
of the spectrum of gamma radiation incident upon
a shield. |

In addition, considerable fast-neutron spectral
data were obtained on the present loading for the
water-moderated reactor. The measurement of the
energy spectrum of neutrons attenuated by various
thicknesses of water was repeated and extended
to a 50-cm distance from the face of the reactor.
A meosurement of the angular distribution of fast
neutrons in water has been made, and the attenuated
neutron spectrum through o shield mockup was
measwed,

SPECTRUM OF GAMMA RAYS
EMITTED BY THE BSR!

Results of measurements of the gamma-ray energy
spectra and angular distributions through lead and
water were published ;::ria‘vic.n,:sly.2 in the following,
the results of measurements of the gamma-ray
spectrum of the Bulk Shielding Reactor without a
shield are presented. These measurements may
be considered as an extension of the earlier work.

The measurement of the spectrum from the un-
shielded reactor was necessarily delayed until the
BSR was loaded with “‘cold’’ fue! elements, that
is, elements which had not been used when the
reactor power was higher than 3 watts. The low
power level of the reactor was dictated by the
counting-rate limitation of the spectrometer. With

 

1F. Maienschein and T. A. Love, Spectrum of Gamma
Rays Emitted by the Bulk Shielding Recctor, ORNL
CF 53-10-16 (to be published).

2r, Maienschein, Gommo-Ray Spectral Measure-
ments with the Divided Shield Mockup, Part {, ORNL
CF 52-3-1 (Mar. 3, 1952); Part 1l, ORNL CF 52-7-71
(July 8, 1952); Part |1}, ORNL CF 52-8-38 (Aug. 8,
1952); and T. A. Love, Part IV, ORNL CF 52-11-124
{(Nav. 17, 1952).

116

the reactor ‘‘hot,’’ the power level from fission-
product activity alone would have been too great.

For this particular experiment, a 6 by 5 fuel
element arrangement was used in on unreflected
reactor. The total power and the power distribution
were determined with gold foils in the usudl
manner.

Spectrometer Arrangement. One end of on air-
filled oluminum tube was placed ogainst the face
of the reactor, and the tube extended to the three-
crystal gamma-ray spectrometer® (Fig. 11.1). For
the first run, the air tube was only 5 in. long ond
was surrounded by a 5-in. thickness of lead which
shielded the spectrometer proper from gamma radia-
tion from other parts of the reoctor. The results
of this run showed that the neutron-induced back-
ground was so high os to moke a meaningful in-
terpretation of the dota impractical,

In o second run, o 40-in. air column surrounded
by water, except for the 5-in. thickness of lead at
the end next to the spectrometer, served as a
collimator. The additional water shielding reduced
the neutron level sufficiently for the nevtron-induced
background to be negligibly small. The total back-
ground was measured by plugging the collimator
within the spectrometer with lead. It was also
found that the background from reactor fission
products was small,

Results, The absolute photon (gamma-ray) flux
obtained is plotted as a function of photon energy
in Fig. 11.2.  The flux volues were determined
from the known sensitivity of the spectrometer®
and from the solid angle defined by the collimator
shown in Fig. 11.1. Since the water surrounding
the aluminum tube was not very effective in pre-
venting gomma rays from entering the sides of the
air column, the spectrum shown in Fig. 11.2 is too
large by a small fraction, perhaps 20%.

 

3F. C. Maienschein, Multiple-Crystal Gamma-Roy
Spectrometer, ORNL-1142 (July 3, 1952).
 

 

PERIOD ENDING DECEMBER 10, 1953

UNCLASSIFIED
DWG. 24656

THREE-CRYSTAL GAMMA-RAY

SPECTROMETER-
7 '\\. /
; /7 \\-
///’ S
.

SIS
W,
. ! /{

i

 
 
       
  

   

 

 

 

 

t5 in. BSR T T I ””{_

 

 

 

 

 

 

 

 

 

 

l ZiALUMINUM TUBE

FHLED WITH AR

 

   

 

 

MREMOVAE!&

LEAD PLUG

    

Fig. 11.1. Experimental Arrangement for Measurement of Bulk Shielding Reactor Gamma-Ray Spectrum,

UNCLASSIFED
OWG. 2657

 

,,,,,,,,,

o

n

r {photons femEssec/ Mev wott Ssteradian)
o

0 1 2 3 4 5 6 7 8
GAMMA-RAY ENERGY (Mev)

Fig. 11.2, Spectrum of Gamma Rays from the
Bulk Shielding Reactor.

The spectrum shape shows the usual broad peaks
at 2.2 Mev that are due to water capture and those
at 7 to 8 Mev that are due to capture in aluminum
and possibly other materiadls. The couse of the
0.4-Mev pedak, which appears to be redl, has not
been found. The 0.4%-Mev pedk, however, may
well be due to gamma rays from the 0.478-Mev level
in the Li’ formed by neutron capture in the boron
liner of the spectrometer, ‘

The smooth curve in Fig. 11.2 represents the
spectral shape (not intensity} obtained by Mot z4

4). W. Motz, Phys. Rev. 86, 753 (1952).

 

with a Compton-recoil magnetic spectrometer look-
ing at a U233 slug in the center of the Los Alamos
water boiler and one ot the active core of the fast
reactor. The curves cannot be expected to corre-
spond closely because of the differences in neutron-
capture gamma-ray sources and because the low
resolution of the spectrometer used by Motz would
smooth out any structure present. The generdl
similarity in slopes, however, is interesting.

FAST-NEUTRON LEAKAGE SPECTRA
OF THE BSR

The neutron spectrum of the BSR obtained with
the reactor surrounded by cans of beryllium oxide
as a reflector® was reported previously. For the
data reported here, the previous spectral measure-
ments were repeafed and additional measurements
were made; however, a waterweflected reactor
(loading Né. 22) was used.

The measurements were made with the BSF
proton-recoil fast-neutron spectrometer,’ Except
for the spectrum at the face of the reactor, the

 spectral measurements in water out to 30 cm from

the fuel® are in agreement with the previous meas-
urements. The spectrum at the face of the reactor

 

SR. G. Cochran aond K. M. Henry, Fast Neutron Spec-
trum of the Bulk Shielding Reactor, Part |, ORNL CF
53-5-105 (to be published). '

5R. G. Cochran et al., ANP Quar. Prog. Rep. Sept. 10,
1953, ORNL-1609, Fig. 11,1, p. 127.

TR- G, Cochran and K. M. Henry, A Proton Recoil Type
FastNeutron Spectrameter, ORNL.-1472 {Apr. 2, 1953).

8p. 6. Cochran and K. M. Henry, Fuast Neutron Spec-
trum of the Bulk Shielding Reactor, Part 1I, ORNL CF
53-11-45 (to be published),

117
ANP QUARTERLY PROGRESS REPORT

In addition to the proton-
data obtained with the
have been used to extend
the spectrometer data from 1.3 Mev down to 0.5

Mev.

is shown in Fig. 11.3.
recoil measurements,

nuclear plate camera’®

Meoswements were also mode out to 50 em from
the reactor. The spectrum measured at 50 cm is
shown in Fig. 11.4; the statistics on these data
are quite poor, since the neutron intensity was
approximately at the limit of sensitivity of the

speciromater.

An attempt was also made to measure the spec-
trum of neutrons emerging from the side of an air-
craft divided-shield mockup.'® The mockup con-
sisted of a slab of lead 4% in. thick placed 16 in.
from the reactor. The shield therefore extended
out through the water to a position 113 cm from the
reactor. Unfortunately, the neutron spectrometer
was not sufficiently sensitive to detect neutrons
at 113 em; however, a spectral measurement was
made at 70 cm, but the statistics (Fig. 11.5) were
rather poor.

A measurement of the angular distribution of
the fast neutrons in water was made during the

 

M. P. Huydon, E. B. Johnson, and J. L. Meem,
Measurement of the Fast Neutron Spectrum of the Bulk
Shielding Reactor Using Nuclear Plates, ORNL CF
53-8-146 {10 be published).

Repori‘ of the ANP Shielding Boord for the Aircraft
Nuclear Propulsion Program, ANP.53 (Oct. 16, 1950).

L
DWG 2184C

S
O RECOIL SPECTROMETER DfxTA END OF *

 

 
 
 
  

COLLIMATOR AGAINST REACTOR
O NUCLEAR PLATE CAMERA DaTA

|

]

PLED (nsuhons/cm?/sec/woH/Mev 103

 

& 7 9 10
ENERGY {(Mev)

Fig. 11.3. Meutron Spectrum ot the Face of the
B3R,

118

quarter. H

The experimental setup for this meas-
urement is shown in Fig. 11.6, and the spectra vs.
angle data are plotted in Fig. 11.7. At neutron
energies above 8 Mev, the intensities af the various
angles converge. This effect is probably due to
neutrons which bypass paort or all of the collimator.
A distance greater than 10 cm from the reactor
would be more desirable for this measurement, but
then it would be difficult to obtain sufficient in-
tensity. Figure 11.8 shows the variation of flux
with direction for each of several neutron energies.

 

]R. G. Cochran and K. M. Henry, Angulor Distriby-
tion of Fast Neutrons in Warer, ORNL CF 53-11-46 (ic
be published).

DWG. 24921
e

 
 

. . .o - e
50 cm Hy(? BETWEEN I

IREACTOR AND COLLIMATOR, |

4-ft COLLIMATOR |

&(£) ineutrons/cm2/sec/ watt/Mev X 1073)

 

ENERGY [Mev)

 

    
 

 

Fig. 11.4. Neutron Spectrum 50 cm from the BSR,

1 ey e _,,,,DWG.,EE?Oi
; | 70 em W0 BETWELN ! E XPERIMENT i

" |1 | REACTOR AnD COLUMATOR. . en @ LEAD

=7 2-ft COLLIMATOR T e o

o | : ' : 3SR 4 1,0 r———ij

2 | ‘ b ST .\\

S ; NEUTRON /|

g 8 SPECTROMETER” |

=~ [

g ‘

< T

E <

=

g

5

g,

é 4

=

E '

w2 !

 

ENERGY (Mev)

Fig. 11.5. Nevtron Spectrum 70 cm from BSR.
FUEL ELEMENT -~ _

10 ¢m

per—-

 

 

 

 

 

 

BULK SHIELDING
REACTOR

 

 

 

 

 

PERIOD ENDING DECEMBER 186, 1953

 

 

 

 

 

 

OWG, 22199
4 ft
emaasa -- e e - - e
f/" 1-in. C_OLLIMATOR NEUTRON
T e SPECTROMETER
N HOUSING
\\ . e e
\\\ ; Yol T - ’Q deg
e T T e
-~ - \\\\
\\ - S - 2
T L0
. £
~. =5 o
A
.k ",
\EQN
™~
DIRECTION ©OF
ROTATION
TOP VIEW

Fig. 11.6. Experimental Arrangement for Me asurement of Angular Distribution of Neutrons from BSR in Water.

$EY { PEUITORS / GMT/gec/ watt/ Mey i

END OF COLLIMATOR 10 cm
FROM :REACTCR

 

NEUTRON ENERGY {Mev)

 

A
DWG. 21843

Fig. 11.7. Angulor Distribution of Neutrons from
BSR in Water,

  

W
DWG. 21842

b (£) (neutrons /cm/sec/ watt/ Mev)

 

>/ / "; L /z, i
EAP - \\ /\ // \/ /// // - li i’:"t‘i»-/_‘ ’f /:’/ 207
[REACTOR N\ T S o / 7\./ /7T
o0° 50° 30° 25e

Fig. 11.8. Energy Angular Distribution of Neutrons
from BSR in Water,

119
ANP QUARTERLY PROGRESS REPORT

12. LID TANK FACILITY

C. L. Storrs -
GE-ANP

G. T. Chapman
J. M. Miller

D. K. Trubey
F. N. Watson

Physics Division

Tentative results of a preliminary experiment
performed at the Lid Tank indicate that most of the
air-scattered neutrons that penetrate an aircraft
crew shield follow the slant path. In another
experiment, the study of neutron streaming through
air ducts was resumed; data from the initial experi-
ments have presented on important confirmation of
the Simon-Clifford theory. A re-evaluation of all
measurements for the determination of removal cross
Refer~
ence to Lid Tank neutron and gamma data is facili-
tated by a recent survey of the first 50 experiments.

Approval has been received for the use of enriched
vranium for a new Lid Tank source plate, and a
design has beendeveloped which will make possible

sections has improved some of the values,

a far more accurate power calibration. The source
plate, as well as its appurtenant power-measuring
equipment, is at present being constructed in the
ORNL Research Shops. It is planned to install the
source plate in January.

SLANT PENETRATION OF NEUTRONS
THROUGH WATER

A preliminary experiment has been performed to
determine the slant penetration of fission neutrons
through water. An air-filled cluminum duct, 40 in.
long and 2 in. in diameter, was used as a collimated
“point’’ source of fast neutrons, and the fast-
neutron dose was determined in the water around
the end. The duct was tilted from the normal to the
source plate to reduce background of fast neutrons
from the ORNL Graphite Reactor. Since the fast
neutrons are collimated in the outward direction, if
not reduced they would overshadow the fission-plate
neutrons which are, of course, isotropic. The data,
plotted as isodoses oround the duct, are shown in
Fig. 12.1.

By integrating the dose along planes through the
radiation
thicknesses of water at various angles was de-
termined (Fig. 12.2). This information must be
regarded as tentative until the dose determinations

field, the dose penetrating different

120

have been repeated with more care, but the results
obtained so far indicate that most of the dose comes
from neutrons that penetrate the shield along rthe
slant path. This indication is similar to that
obtained for gamma rays for comparable attenu-
ations.!

The observed relaxation lengths range from 4.9
to 6.3 cm. As would be expected, these values are
smaller than those obtained with thick shields,
since a fission spectrum was used as a source and
the shields were quite thin.

 

IF. 5. Kirn, R. ). Kennedy, and H. 0. Wyckoff, Oblique
Attenuation of Gamma-Rays from Cobalt-60 and Cesium-
137 in Pclyethylene, Concrete and Lead, MNBS-2125
(Pec. 23, 1952).

RN
DWG. 2iB74

 

 

No |- e

z, DISTANCE FROM SOURCE (cm)

 

 

T
9o
|
oot o
40 30 20 o
Fige 12,1. Fast-Neutron Dose Distribution Off

the End of a Duct 101.6 cm Long by 5,08 e¢m in

Diameter.

7
s—
DWE. 21876

 

 

 

 

 

 

 

RELATIVE NEUTRON DOSE (orbitrary units)

 

 

 

 

 

 

 

 

 

7, SLANT THICKNESS (em)

Fig. Slant Penetration of Neutrons in

Water.

12.2.

To determine the total air-scattered dose in a
divided shield, it is necessary to integrate over
all angles. Such an integration would presumably
give a straight-through relaxation length comparable
to that observed in the BSR air-scattering experi-
ment.? 1t would be desirable to extend the measure-
ments to greater thicknesses of water and to harden
the spectrum by putting water between the source
and the duct entrance. Intensity problems limit the
extent to which this can be accomplished, however,

 AIR DUCT TESTS
The use of the Lid Tank Facility for the study of

neutron streaming through air ducts has been
resumed with emphasis on fast-neutron measure-
ments. A flow-type dosimeter with somewhat
greater sensitivity than that available heretofore
was used to determine the dose around the ends of
several ducts. One of these ducts was taken from

 

2H. E. Hungerford, The Skyshine Experiments at the
Bulk Shielding Facility, ORNL-1611 (to be published).

PERIOD ENDING DECEMBER 10, 1953

the 19-duct array studied a year ago.? |t had three
sections, 22 in. long and 3%
joined at angles of 45 degrees. A second duct
consisted of two such sections. The slant pene-
tration experiment described in the preceding sub-
section provided a third measurement.

Putting the data from these three measurements
into the Simon-Clifford formula? and solving for the
reflection factor, a, gave values of 1.0 and 0.9,
respectively, for the first and second bends. These
values are in agreement with the velue of about 1
obtained from the previous thermal measurements®
made in water at a distance of 30 c¢m from the ends
of the ducts (thus the emergent flux was filtered and
was representative of the hard neutrons that de-
termine the dose), and they constitute an important
confirmation of those measurements. The value of
1 has been assumed in recent direct-cycle design
studies. It is now felt that engineering designs
based on the Simon-Clifford theory are on a sound
basis when applied to a single duct, especially
when the dimensions are similar to those studied.

Since it was observed that the peak values of
dose measured at the end of the 19-duct array were
a factor of approximately 10 higher than those ob-
served with the single duct, a study of interference
between adjacent ducts was indicated. This study
has been started, with a single section and o
three-section duct being used. In the first con-
figuration, one of the possible streaming paths was
reproduced in the array by placing the short duct
parallel to the first section of the long duct and
12!{2 in. from it. In this position, it lined up,
approximately, with the last section. The result
was an increase by a factor of 4,5 in the dose at
the end of the long duct. 4

The short duct was then moved to a position that
was again parallel to the first section but con-
siderably closer (§ in.) and above it in a different
plane so that there was no direct streaming path.
The data (which must be considered tfentative)
showed no increase in the peak dose beyond the
long duct but did show a broadening of the emergent
nevtron beam. The results obtained from these
experiments are shown in Fig. 12.3.

in. in diameter,

 

3,. . Flynan and C. L. Storrs, Radiation Measurements
Aroundan Array of Cylindrical Ducts, ORNL CF-52-12-187
(Dec. 15, 1952). 7

4A, Simon and C. E. Clifford, The Attenuation of
Neutrons by Air Ducts in Shields, ORNL-1217 (Nov. 11,
1952).

SC. E. Clifford, private communication.

121
ANP QUARTERLY PROGRESS REPORT

DWG. 21918

current C F]

measurements, gives a removal cross

section of 1,31 barns/atom for fluorine as compared
with the value of 1,36 barns/atom reported in the

previous report.

A re-examination of all the Lid Tank measurements
for the determination of remeval cross sections has
improved some of the values. A summary of the
presently accepted values obtained from direct
measureaments is given in Table 12.1. Additional
values based on the direct measurements are given

in Table 12.2.

TABLE 12.1. EFFECTIVE FAST-NEUTRON REMOVAL
CROSS SECTIONM VALUES OBTAINED FROM LD
TANK DIRECT MEASUREMENTS

 

EOSL oot T
0.
E ,,,,,,
E
L 2 e
w
Q
O

.».1 ~
z 10 bt
o
=
-
w5 TWO-BEND DUCT A+ - —— 1+ ..
= WITH SINGLE SECTION
g DUCT IN LINE /T ———T7

2 b

TWGO-BEND DUCT
1072 WITH SINGLE SEC--
. TION DUCT NOT
IN LINE
TWO-BEND DUGT
ol L b
1072

-20 -18 -2 -6 o 6 12 18 20
». DISTANCE FROM CENTER OF END OF DUCT (cm)

Fig. 12.3. Fast-Neutron Dose Beyond Yarious
Configurations of Ducts with 45-deg Bends.

Further tests are planned to give more quantitative
information on the inferaction between neighboring
ducts and possibly to improve the formula for
streaming through ducts.

REMOVAL CROSS SECTIONS

The measurement of the effective fasteneutron
removal cross section of C_F . has been repeated
with the use of 15 in. of the fiquicl instead of 28
in. as in the previous measurement.® Despite this
large difference in thickness, the measured values
of the molecular removal cross section agreed to
within 5%, which is within the experimental error,
and thus they provide an interesting check on the
validity of the geometrical corrections that were
made. The revised value for carbon, 0.80 barn/atom
instead of 0.84 barn/atom, when used with the

 

6C. L. Storrs et al., ANP Quar. Prog. Rep. Sept. 10,
1953, ORNL-1609, p. 145-148.

122

 

 

 

MATERIAL gn
Al 1.31 barns/atom
Be 1.05
Bi 3.43
C 0.80
Cu 2,08
Fe 1.95
Ni 1.85
W 2.6 + 0.2
C7F16 26.6 barns/molecule
C,FyCi 6.50
34(: 4,26
Hzo* (2.90)
520* (2.78)

 

*

Neither hydrogen nor deuterium haos, properly, an
effective remova! cross section, since the total cross
section is not nearly constant in the high-energy region.
Rough values have been obtained for H20 and 020

bosed on 140~cm shields.

rentheses,

These are reported In pa-

TABLE 12,2. EFFECTIVE FAST-NEUTRON REMOVAL
CROSS SECTIONS BASED ON LID TANK DIRECT

 

 

MEASUREMENTS
MATERIAL URI {(barns/atom)
B 0.87
D* (0.92)
F .31
H* (0.98)
0 0.94

 

 

1'rCf., footnote to Table 12.1.
w

SURVYEY OF LID TANK EXPERIMENTS

Shielding experimentation at the ORNL Lid Tank
Facility began in June 1949, and since that time
approximately 60 major experiments have been
performed. In view of the large role this facility has
had in the development of various reactor shield
designs and because of the constant need for
reference to the data, it was felt that the experi-
mental resuvlts should be compiled in one publication.
Accordingly, o summary of the first 50 Lid Tank
experiments will socon be presented by E. L.
Czapek.”:8 ‘

As described in several previous reports, the Lid
Tank Facility consists of a large water tank
adjacent to an opening in the concrefe shield of the

 

TORSORT student from Electric Boat Division of
General Dynamics Corporation.

8. L. Czapek, General Survey of Lid Tank Experi-
ments 1 Through 50, ORNL-1636 {to be published). '

PERIOD ENDING DECEMBER 10, 1953

ORNL. Graphite Reactor (Fig. 12.4). A converter
plate of utanium covers the opening between the
water tank and the reactor shield. This plate
absorbs most of the incident thermal neutrons
emerging from thereactorand causes fission, which,
in turn, produces the known uranium spectrum of
neutrons, as well as gamma rays of various energies.
Shielding samples are inserted in the tank, and the
transmitted radiation is measured behind the shield.
The geometry of the Lid Tank permits the use of
very large samples, which, in turn, help to reduce
uncertainties due to boundary conditions. The water
in the tank protects personnel from radiation and
allows easy positioning of test samples.

The Lid Tank Facility was built to test both
compositioris and geometries of shields. With the
resuiting experimental data, difficult attenuation
problems could be quickly solved and shield thick~
nesses couid be predicted {to within an accuracy of

 

Fig. 12.4. View of Lid Tank Focility from Above.

123
ANP QUARTERLY PROGRESS REPORT

+5%) without having to rely too much on a knowledge
of the attenuation processes.

Experimentation at this focility has primarily been
limited to investigations of the ‘‘unit shield’’ in
which all material that functions primarily as
shielding is located around the reactor. All the
experiments covered in the survey can be grouped
as follows:

1. Water Shield (experiments 1, 4, 10, 13)
2. Preliminary |nvestigations” (experiments 2, 3,
5,6, 8,10)
a. Lead-water shield
b. lron-water shield
3. Boron Carbide Shields (experiments 11, 12)
4. Thermal Shields? (experiments 10, 20, 36)
a. lLead-water shield
b. lron-water shield
5. Shield Mockups (experiments 7, 14, 15, 16, 17,
18, 29, 45)
d. Submorine Thermal Reactor
b. Submarine |ntermediate Reactor
c. Reflector-Moderated Reactor (ANP)
6. Shield OQOptimizations (experiments 9, 13, 24)
a. Two-component shield of lead and borated
water
b. Suppression of capture gammas in lead-water
shield
c. Optimum position of a lead-iron gomma shield
in relation to pressure shell

 

YA small amount of boron was present in a majority of
these shield mockups.

124

7. Shield Penetrations (experiments 19, 21, 22, 25,

26, 37, and 38 and special experiments)

a. Cylindrical air ducts with or without bends

b. Array of cylindrical air ducts with or without
bends

¢. Annularduct (specialized mockups simulating
GE-R1 air outlet and inlet sections)

d. Structural members in shields (streaming
effects)

e. Instrument plug (modification in shield to
improve control instrumentation)

8. Removal Cross Sections (experiments 23, 30, 31,

32,33, 34, 35, 39, 40, 41, 42, 43, 46, 47, 49, 50)

9. Special Investigations .

a. Effects of boration of water (experiment 10)

b, Effects of water and air gaps in shields
(experiment 48)

c. Induced radiocactivity outside an air duct
(experiment 27)

d. Activation expected in a coolant (sodium)
at various locations in a boren carbide shield
(experiment 10)

e. Gamma scattering around submarine bulkhead
(sodium source) (special experiment)

The situation at present in regard to unit shields
or reactor parts of divided shields appears to be
that the basic mechanisms are understood and that,
except for the very simplest problems, shield
mockups are essential, although the effects of
small changes in design can be quite well predicted.
The Lid Tank method of mockup testing is highly
developed, and large amounts of data can be col-
lected in relatively short times.
PERIOD ENDING DECEMBER 10, 1953

13. TOWER SHIELDING FACILITY

C. E. Clifford

T. V. Blosser

L. B. Holland

Physics Division

The steel structure of the Tower Shielding Facility
has been erected (Fig. 13.1), and the ground
structures are approximately 75% complete. The
underground building for the reactor controls and
for personnel will be available for occupancy by
December 10; access to the control area of the
building was authorized in November.

The mechanical components of the reactor have
been constructed and are now being assembled, and
the reactor tank has been received. Construction
of the crew-compartment tank is nearing completion.
The last hoist is scheduled for delivery by January
1, 1954. :

 

The optimistic date for completion of the facility
is January 15, after which there will be a one-month
period for critical experiments and shakedown
operation. This will be followed by differential
shielding measurements in the reactor and the
crew-compartment water tanks.

A mockup of a GE airecooled reactor shield
design is scheduled for delivery by March 15 by the
General Electric Company, and for testing in
conjunction with this, a crew-shield mockup will

be supplied by ORNL.

PHOTO 11959

Fig. 13,1. Tower Shielding Facility, November 13, 1953.

125
14. SHIELDING ANALYSIS

E. P. Blizard
J. E. Faulkner

M. K. Hullings
F. H. Murray

Physics Division

H. E. Stern
Consolidated Vultee Aircraft Corporation

Much of the shielding analysis work for this
quarter was reported in the Physics Division Semi-
annual Report! and is not repeated here. The work
reported in the Division report included two treat-
ments of geometrical transformations with methods
of estimating reactor leakage. |n addition, some
calculational methods for obtaining neutron at-
tenuation were presented. On comparison of the
calculated attenuations with measured attenuations,
it was possible to derive ‘‘effective removal cross
sections’’ that when compared with other measure-
ments showed reasonable agreement. In addition,
gamma-ray absorption coefficients which have been
mecsured for some elements have been used to
interpolate the coefficients for other elements.

The visible light from an aircraft nucleor power
plant has been estimated, and the calculations will
be compared with measurements made on the BSR.
The measurements of visible light are made in air
and are not to be confused with the Cerenkov
radiation (blue glow) which is so obvious in water.
An estimate of neutron albedo was obtained to
supply a simple model for ground and structure
scattering problems.

VISIBLE LIGHT FROM A NUCLEAR
POWER PLANT

The results of some experiments carried out at
l.os Alames on the visible light from a polonium
source? have been used in obtaining an estimate of
the ionizing radiotion that would have to be present

T, P. Blizard et al., Phys. Div. Semiann., Prog. Rep.
Sept. 10, 1953, ORNL-1630 (in press).

126

around the aircraft reactor shield to just make it
visible. If the airplane were approximately 200 ft
from the observer, the ionizing radiation near the
shield would have to be about 5 x 10% r/hr. If the
airplane were very far from the observer, say
50,000 ft, the ionizing radiation would have to be
about 10% r/hr. These estimates will be compared
with measurements made on the BSR.

NEUTRON REFLECTION COEFFICIENT
FOR WATER

An estimate of the neutron reflection coefficient
for water has been obtained by using a simple
model with a coincident source and a nondirectional
receiver located above the water surface, If iso-
tropic oxygen scatter and attenuation governed by
the removal cross sections are assumed, the re-
flected flux F at height 5 for source strength N is

 

NO scotter
F = em———— ||
48”}.72 removal
The albedo appreach gives
No a
F = —
10752

for a re-emission distribution between isotropic and
cosine, where a is the reflection coefficient. A
value of a = 0.085 is thus obtained which is in
fair agreement with the experimental value of 0,08.

2, G. HoHman, Ruadiation Doses in the Pajarito
Accident of May 21, 1946, L A-687 (May 26, 1948).

 
 
 

4%
REPORT NO,

CF 53-9-15
CF 53-9-53

CF 53-10-62
CF 53-11-95

CF 53-12.9
ORNL-1650

CF 53-9-32
CF 53-9-102

CF 53-9-180
CF 53-10-208

CF 53-9-19
ORNL-1615

ORNL-1634

CF 53-5-239

CF 53.9.4

CF 53-9-16

CF 53-8-146

CF 53-9-161

CF 53-10-1

CF 53-10-16

CF 53-11-2
CF 53-11-45

15. LIST OF REPORTS ISSUED DURING THE QUARTER

TITLE OF REPORT
l, Aircroft Reactor Experiment

Experimental Procedure on the ARE (Preliminary)

Supplement to ARE Hazards Summary Report (ORNL-1407)

ARE Fuel Recovery

Report and Recommendations of Second ARE Cleaning
Committee

ARE Design Data
Stress Anclysis of the ARE

il Experimental Engineering

ANP Experimental Engineering Quarterly Status Report

Design Data and Proposed Test Schedule for Sodium-to-
Air Radiators

LITR Fluoride Fuel Test Loop

A Flat Plate Heat Exchanger for Reactor System Uses

1. Critical Experiments
Static Analysis of the ARE Criticciit):r Experiment
Critical Experiments on Reactor

Preliminary Critical Assembly for the Aircraft Reactor
Experiment

IV. Shielding

Nuclear Dose Measurements on the Divided Shield Mockup
at the Bulk Shielding Facility

Lid Tank Shielding Test of the Reflector-Moderated Reactor

Minimum Shield W_eighf Penalty for Air Ducts

Measurement of the Fast Neutron Spectrum of the Bulk
Shielding Reactor Using Nuclear Plates

An Estimate of the Neutron Reflection Coefficient, Using
the Concept of Removal Cross-Section

Computation of Effective Removal Cross-Section Meaosured

at ORNL Lid Tank :

Spectrum of Gamma Rays Emitted by the Bulk Shielding
Reactor

The Shielding of Nuclear Radiations, Lecture |

Fast Neutron Spectrum of the Bulk Sh:iellding Reactor, Part 11

AUTHOR (s)

jc Ln Meem

£. S5, Bettiz
W. B, Cottrell

F. N. Brewder

Gs A- Cristy

L. A. Mann

F. F. Blankenship
G, M. Adamson

W. B. Cottrell

R. L. Maxwell
Jl w. Wfliker

H, W. Savage
H.s Jo Stumpf

O, Sisman

R. W, Bussard .

Cq Ba Mi“S

D. Callihan:
Rl Cc Keen

. Callithan
D. Scott

Hs E. Hungerford

Fc No WQ‘Ison
E, P, Blizard
M. P. Haydon
E. B. Johnson

J. L. Meem

Ho El Sfefn.

L. S. Abbott

F. C, Maienschein
Te Al Love

E. P. Blizard

R, G, Cochron
K. M. Henry

DATE ISSUED

9-3-53
9-25-53

10-5-53
11-10-53

to be issued

to be issued

9-2-53
9-8-53

9-25-53
10-26-53

9-1-53
10-22-53

10-28-53

5-20-53

9-30-53
9-2-53

to be issyed

9.29-53

no date

fe be issued

11-2-53.

te be issued

129
ANP QUARTERLY PROGRESS REPORT

REPORT NO,

CF 53-11-46

CF 53-11-53
CF 53-11-54
ORNL-1636

CF 53-10-117

CF 53-10-228

ORNL-1565
ORNL-1633
ORNL-1647
MM-147

MM-151

MM-157

CF 53-10-78
ORNL-1626
UA-PR-13

BMI-852

BMI-864

CF 53-8-106
CF 53-8-217

CF 53-9-98

CF 53-10-86

CF 53-11-128

ORNL-1624

130

TITLE OF REPORTY
IV, Shielding (continued)

Angular Distribution of Fast Neutrons in Water

The Effect of Some Liquid Metal Ducts on Reactor Shislds
Gamma-Ray Spectrum of the Bulk Shielding Raactor

General Survey of Lid Tank Experiments 1 Through 50
V. Metallurgy

Excomination of LF Pump Loep

Metallographic Examination of Second Heat Exchonger from
Bi-Fluid Pump Loop

Scaling of Columbium in Air

Fabrication of Spherical Particles

Interim Report on Static Liquid Metal Corrosion

Third Progress Report on the Flash Welding of Malybhdenum.
Part | — Temperature Distribution During the Flashing
Cycle

Redrawing of Spacial High Purity Incone!

Progress Repori and Final Report to Carbide and Carbon
Chemicals Company

Yl. Chamistry
Fused Salt Compositions
Determination of Zirconium by the Chloranilic Acid Method

Progress Report for the Period April 1, 1953 Through
June 30, 1953

Vapor Pressures of Beryllivm Fluoride end Zirconium
Fluoride

Potential Liquid Fuels for Nuclear Power Reactors

VIl. Heat Transfer and Physical Properties

Preliminary Results on Flinak Heat Transfer

Preliminary Measuremants of the Density ond Viscosity of
Fluoride Mixture No, 44

Heat Copacity of Two Samples of 310 Stainless Steel and
of a Brazing Compound

Preliminar y Measurement of the Density and Viscosity of
Composition 43 (NazUFé)

Heat Capocity of Fuel Composition No. 23

The Nature of the Flow ot Ordinary Fluids in @ Thermal
Convection Harp

AUTHOR (s)

R. G. Cochran
K¢ M: Henry

Mo K. I'iU“ingS
F. Ci Maienschein

El Ln Czapek

G. M. Adamson
R. S Crouse

Gc Mn Adqmson
R. 5. Ciouse

H. lnouye

A, Levy

W. D. Manly

Rensselaer
Polytechnic

Institute

Superior Tube
Company

Commonwealth

Engineering
Co,. of Ohie

C, J. Barton
0. Menis

University of
Arkansas

Battelle Memorial
Institute

Baitelle Memorial
Institute

H. W. Hoffinan

S. Is Cohen
T+ N. Jones

W. D, Powers
Go Ca B]OIOCk

S l. Cohen
Te N. Jones

W. D, Powsers
G. C. Blalock

D. C. Hamilton
F. E, Lynch
L. D. Palmer

DATE ISSUED

to be issued

to be issued
11-8-53

to be issued

10-6-53

10-27-53

?-1-53

11-12-53

to be issued

9-30-53

10-7-53

11-9-53

10-9-53
10-28-53
7-30-53

7-13-53

9-8-53

8-18-53
8-31-53

92-18-53

10-14-53

11-23-53

to be issued
REPORT NO.

CF 53-9-8

CF 53-11-147
ORNL-1609

ORNL-1632

PERIOD ENDING DECEMBER 10, 1953

TITLE OF REPORT

Yitl. Miscellaneocus

The Small Hydrogen Moderated Reactor

Directory of Active ANP Reseorch Projects at ORNL

Aircraft Nuclear Propulsion Project Quarterly Progress
Report for Period Ending September 10, 1953

ORNL Contrel Computer

 

AUTHOR (s) DATE (SSUED
C. B, Mitis G-2-53
W. B. Cottrell 11-27-53
W. Bo Come“ 10‘15"53
Je J. Stone to be issved

16. DIRECTORY OF ACTIVE ANP RESEARCH PROJECTS AT ORNL

NOVEMBER 27, 1953

l. REACTOR THEORY AND COMPONENT DESIGN

A,

B.

E.

Aircraft Reactor Design

L
2,
3.
4,

Reflector-Moderated Reactor Studies

Fluid-Fuel Flow Studies
Gamma Heating of the Moderator

Design Consultants

ARE Design

1.
2,
3.
4,

Building

Electrical Power Circuits
Fluid Circuit Design
Fuel Recovery Equipment

ARE Installation

1.
2
3.
4,
5.
6.

Plumbing
Electrical System
Controls
instrumentation
Expeditors

Coordinators

ARE Operations

1.
2.

3.

Staff
Day Shift

MNight Shif?

ARE Control Studies

.‘.

High-Temperature Fission Chamber

9704-1
9704-1
9704-1
9704-1

1000
1000
7503
9201-3

7503
7503
7503
7503
7503
7503

7503
7503

7503

4500

Fraas, L.aVerne, Bussard

Stumpt

Bussard, Wilner

Wislicenus, JHU; Maines;
Chambers, U.T.

Browning
Walker
Cristy
Estobrook

Webster et ol.
Packard et =i,
Epler, Mann

Affel

Perrin, West
Wischhusen, Watts

Bettis, Meem, Cristy, Mann

Whitman, Cottrell, Grindell,
Croley, Schafer, Williams,
Southern

Affel, Perrin, Cobb, Leslie, Allen,
Reid, Morrison, Huntley,
Gregory, Addisen

Hanauer

131
ANP QUARTERLY PROGRESS REPORT

2, Control System Design

3. Control Studies on Simulator

F. Reactor Physics
1. Anclysis of Critical Experiments
2. Kinetics of Circulating=-Fuel Reactors
3. Computation Techniques for ARE-Type Reactors
4. Statics of Reflector-Moderated Reoctors

5. Machine Computations

6. Interpretation of Critical Experiments

G, Critical Experiments

1. Reflector-Modzrated Reactor Critical Assembly

2. Preliminary Assembly of Supercritical-Water
Reactor (Prott & Whitney)

3., Preliminary Assembly of AC-1 (General Electric
ANPP)

H. Besaring, Secl, and Pump Development
1. Mechanical Pumps for High-Temperature Fluids
2, High-Temperature Seals for Rotary Shafts
3. Pump for Hydroxide Systewms
4, Seals for Hydroxide Systems
5. Bearings for High-Temperature Application
6. Canned-Rotor Pump Development

7. Materials Compatibility in Molten Salts at
High Temperatures '

. Valve Development
1. Valves for High-Temperature Fluid Systems
2. VYalves for FHydroxide Systems

J. Heat Exchanger and Radiator Development

1. High-Temperature Fluid-to-Air Heat Exchange Test
2. Fluoride-torLiquid Metal Heat Exchange Test

3, Heat Exchanger Fabrication
4, Boeing Turbojet with Sodium Radiator
5. Radiator and Heat Exchanger Design Studies

6. FulleScale Test Facilities

K. Fluid Dynamics of Liiquid Fuel Reactors
1. Pump-Core Shell Component Test

.. Instrumentafion
1. Pressure-Measuring Devices for High-Temperature

Fluid Systems

 

]Buflelle Mewmorial Institute,

132

4500

4500

9704-1
9704-1
9704-1
9704-1
4500

4500

9213

9213

2213

9201-3
9201-3
BMI!
BMI
9201-3
9201-3
9201-3

9201-3
BMI

9201-3
9201-3
9201-3
9201-3

9201-3

9201-3

9201-3

9201-3

Epler, Ruble, Mann, Green,
Honauer

QOakes, Mann, Green

Mills

Ergen

Bengstan, Ergen

LaVerne, Burtnette, Spencer

Charpie, Coveyou, Sangren,
Gerberich, Given, Atta

Prohammer

Callihan, Williams, Scott,
Zimmerman, Lynn, Noakes (P &W)

Callihan et al,

Callihan et al,

McDonald, Cobb, Huntley,Grindell

McDonald, Tunnell, Smith, Mason

Dayton

Simons, Allen

Stair, Tunnell, Smith, Mason

McDonald, Stair, MacPherson,
Grindeil

Stair, Tunnell, Smith, Masen

McDonald

Simons, Allen et al.

MacPherson, Salmon, Frasos,
Stumpf

Salmon, Fraas, MacPherson,
Bussard

Petersen, Fiaas

Fraas, Coughten

Fraas, Stumpf,

Bailey, U.T.

Coughlen

Bussard, Gallagher

MecDonald, Smith
2. Flow-Measuring Devices for High-Temperature

3.
4.

Fluid Systems
High-Temperature Strain Gage
Highe Temperature Calibration Gage

Il. MATERIALS RESEARCH

A. Phase Equilibrium Research

1.

2,
3‘

4,
3.
6.
7

8.

9.
10.

Thermal Analysis of Fluoride Systems

Therma!l Analysis of Chloride Systems
Phase Equilibria Among Fluorides by Quenching

Equilibria in BeF -Bearing Systems by Quenching

Differential Thermal Analysis of Fluorides

Differential Thermal Analysis of BeF -Bearing
Systems

Equilibria Among Fluorides and Chiorides by
Filtration

X-Roy Diffraction Studies of Complex Fluorides

Phase Equilibria in Hydride-Hydroxide Systems
Phase Equilibria Consultant

B. Chemistry of High-Temperoi:fure Liquids

1 -
2,
3.

10.
1L

12.
13.

15.

Equilibrium Pressures iin Hydroxide-Metal Systems
Equilibrium Pressures in Hydroxide-Metal Systems
Equilibrium Pressure in Hydride-Halide Systems
Chemical Reactions in Fused Salts

Solutions of Metals in Fused Salts
Thermagalvanic Potentials in Fused Systems
Decomposition Potentials in Fused Systems

Spectrophotometric Investigations of Fused Salts

Preparation, Phase Behavior, and Reactions of UF,

Rates of Reduction in Fused Salts

Solubility of Uronium Compounds in Hydride-
Mydroxide Systems

Frae Energy of Fluorides

Preporation of Pure Hydroxides

Preparation of Complex Fluorides

Chemistry Consultants

C. Fluoride Production and Handling

1.

2,

PiloteScale Production of Fluorides

| aboratory-Scole Fluoride Production

 

2Ba|dwin-l..ima Hamilton Corp.
IMetal Hydrides, Inc.

PERIOD ENDING DECEMBER 10, 1953

9201-3
BLH?

Cornell

9733.2

9733.2
9733-2

Mound
9733-2

Mound

9733-2
4500
9766
MHI®

U. Ark.
9766
Mound
9733-3
4500
9733-3
97333
9732-3
9733-2

97332
aml!
9201-3

9733.3
9733-3

9928

97332

McDonald, Smith

Insley, Barton, Bratcher,
Truitt

Borton, Boyer

insley, Moore, Metcalf, Truitt,
Thoma, White

Heik, Orban, Joy et al.

Bolomey

Orban et of.

Barton, Sheil
Bredig
Agron, Thoma
Banus et al.

McVYay, U. Ala,

Smothers et al.

Knox, Kertesz

Otto, Orban

Overholser, Redman, Weaver

Bredig, Johnson, Bronstein

Cuneo

Topol, Overholser

Friedmon, Blankenship

Barton, Truitt, Bloed, Watson,
Blankenship

Blood, Watson, Blankenship

Simons et al.

Maan, Petersen, Cisar
Ketchen, Overholser
Sturm, Overholser
Gibb, Tufts U.

Hill, Duke U,

Carter, Carter L ab.

Nessle, Blakely, Eorgan, Croft,
Didloke

Blood, Watson, Blankenship,
Boody

133
ANP QUARTERLY PROGRESS REPORT

G,

H.

134

3.
4,
5.
5.
7.
8.

Preparation of Materials for Radiation Damaoge
ARE Fuel Carrier Production

ARE Fuel Concentrate Production

ARE Filling Operation

Sampling Techniques

Sampling Techniques

ARE Fuel Reprocessing

1.
2,
3.
4,

5.

Fuel Recovery Process Development
Fuel Recovery Design Studies

Unit Operotions Studiss

Pilot Plant Qperations

Fuel Recovery Process Studies

Corrosion by Liquid Metals

1.
2
3-

=

4
Sl

6.
7.
8.
9.
10.
11,
12.

Static Corrosion Tests

Dynamic Corrosion Research in Convection Loops
Effect of Crystal Orientation on Corrosion

Effect of Carbides

Dynamic Corrosion by Liquid Lead

Diffusion of Molten Media into Solid Metals
Structures of Liquids by Diffraction Techniques
Protective Coatings for Corrosion Resistance
Corrosion Inhibitor Studies

Handling of Liquid Metal Samples
Compatibility of Cermets in Liguid Metals

Mechanism of Isotherma! Mass Transfer

Corrosion by Fluerides

—

2,
3.

4,

S
6.

Static Corrosion of Metals
Static Corrosion of Metals

Corrosion in Small-Scale Dynamic Systems

Dynamic Corrosion Tests in Thermal-Convection
Loops
Corrosion of Nonmetallic Materials

Forced-Convection Corresion Loop

Corrosion by Hydroxides

1.
2,
3.
4,

Static Corrasion of Metals and Alloys
Physical Chemisiry of Hydroxide Corrosion
Static and Dynamic Corrosion

Static and Dynamic Corrosion by Hydroxides

Physical Properties of Materials

1.
2.
3.
4.
5.
6.
7.
8.

Density of Liquids

Yiscosity of Liquids

Therma! Conductivity of Selids
Thermal Conductivity of Liquids
Specific Heat of Solids and Liguids
Electrical Conductivity of Fluorides
Surface Tension of Fluorides

Viscosity of Fused 5alt Systems

97332
9201-3
9212
7503
9201.3
9766

4500
4500
4505
3019
3505
7503

2000
9201-3
2000
2000
2000
9201-3
2000
2000
2000
9201.3
9201-3
2000
2000

2000
9766
2000

9201.3
9766
9201.3

2000
2000
9766
BMI!

9204-1
9204-1
9204-1
9204-1
9204-1
9204-1
9204-1
9766

Blankenship, Boody

Messle, Eorgan, Blond, Thoma
Nessle, Forgan

MNessle, Forgan

Mann, Blakely

Kertesz, Meadows, Didlcke

Ferguson, Cathers

Goeller, Browder, Ruch
Eister, Nurmi, Kackenmaster
Jackson

Lewis, Matherne

Messlz, Blaksly, Croft, Eorgen

Yreeland, Hoffman
Adomson

Smith, Cathcart, Bridges
Vreslond, Hoffman
Cathcort

Adamson

Smith, Cathcart
Smith

VYreelond, Hoffman
Adamson

Ketchen

Yreeland, Hoffman
Cathcart

Yreeland, Hoffman
Kertesz, Buttram, Smith, Meadows
¥reeland, Trotter, Nicholson,

Hoffman

Adamson
Kertesz, Buttram, Meadows, Smith

Sc‘.‘ll maon

Vreeland, Smith, Hoffman, Cathcart
Smith, Steidlitz, Boston
Kertesz, Buttram, Smith

Simons et al.

Joines, Cohen

Jones, Cohen

Powers, Burnett
Cloiborne, Fowers, Burnett
Powers, Blalock

Greene

Cohen, Jones

Kertesz, Knox
9. Vapor Pressure of Fluoride Systems
10, Vapor Pressure of BeF ,-Bearing Systems
11. Vapor Pressure of Fluid Systems

i» Heat Transfer
1. Heat Transfer Coefficients of Flucride and
Hydroxide Systems
2. Heat and Momentum Trapsfer in Convection Loops
3. Heat Transfer in Circulating-Fuel Systems
4. Forced Convection in Annuli
5. Transient Boiling Research
6. Fluid-Flow Phenomena

J. Metals Fabrication Methods
1. Welding and Brazing Techniques for ARE

2. Molybdenum Welding Research

3. Molybdenum Welding Research

4, Resistance Welding for Molybdenum and Clad
Metals |

5. Welds in the Presence of Various Corrosion Media

6. Nondestructive Testing of Tube-to-Header Welds

7. Basic Evaluation of Welds Metal Deposits in
Thick Plates

8., Development of High-Temperature Brazing Alloys

9. Evaluation of the High-Temperature Brazing Alloys

10. Application of Resistance-Welding Methods to
ANP Materials

K. New Metals Development
1. Molybdenum and Columbium Alley Studies
2. Heat Treatment of Alloys for Special Projects

3. Alloy Development of New Container Materials

L. Solid Fuel Element Fabrication
1. Solid Fuel Element Fabrication
2. Diffusion-Corrosion in Sclid Fuel Elements
3. Determination of the Engineering Properties of
Solid Fuel Elements
4. Inspection Methods for Fuel Elements

M. Ceramics and Metals Ceramics

1. Metal Cladding for Beryllivm Oxide

2. Hot Pressing of Bearing Materials

3. High-Temperature Firing of Uranium Oxide to
Produce Selective Powder Sizes

4, Development of Cermets for Reactor Components

3. Ceramic Cooatings for ANP Radiators

 

4Massachuseh‘s Institute of Technology.

SRensselaer Pol ytechnic Institute.

PERIOD ENDING DECEMBER 10, 1953

9733-2
Mound
BMI]

9204-1
9204-1
9204-1
9204-1
9204-1
9204-1

2000

BMmi!
miT4

RPI®
2000

2000

2000
2000
2000

2000

2000
2000
2000

2000
2000
2000

2000

Gerity
MiChn
2000

2000
9766
9766

Moore
Orban et of.

Simons et ol.

Hoffman, Lones
Hamilten, Palmer, Lynch
Poppendiek, Winn, Palmer
Bradfute

Rosenthal

Palmer, Winn, Poppendiek

Manly, Housley, Slaughter,
Patrigrca

Russell

Wulff

Nippes, Savage
Vreeland, Slaughter Hoffman,
Patriarca

Slaughter, Patriarca

Slaughter, Gray, Patriarca
Patriarca, Slaughter

Slaughter, Potriarca

Slaughter, Patriarca

Incuye
Bomar, Coobs

inouye

Bomar, Coobs

Bomar, Coobs

Bomar, Coobs, Woods, Oliver,
Douglas

Bomar

Graaf
Bomar, Coobs
Bomar, Coobs

Johnson, Shevlin, Taylor, Hamner

White, Griffin

135
ANP QUARTERLY PROGRESS REPORT

6., Ceromic Valve Parts for Liquid Metals and
Fluorides
7. Ceramic Reflactor

8. Ceramic Coatings for Shielding

9. Fabrication of Be( Shapes

10, Cosrosion of Ceramics

11. High-Density Graphite

N. Strength of Materials

1. Creep Tests in Fluoride Fuels

2, Creep and Siress Rupture Tests of Metals in
Vacvum Gaseous Environments and in Fluid Media

3. High-Temperature Cyclic Tensile Tests

4, Tube-Burst Tests

5. Relaxation Tests of Reactor Materials

6. Evaluation Testing for Other Development Groups

7. Fatigue Testing of Reactor Materials

0. Radiation Damage

1. Fused Salt Fuels lrradiation in LITR

2. Fused Salt Fuels lrradiotion in MTR
3. LITR Dynamic Corrosion L.oop
4. Fluoride Fuel Loops for LITR and MTR

5. Engineering Properties of lrrodiated Metals

6. Meutron Spectrum of LITR
7. Creep of Metals and Stress Corrosion in ORNL
Graphite Reactor and LITR
8. Creep of Meials in MTR
9. Thermal Conductivity of Metals in ORNL Graphite
Reactor and LITR
10. Remote Metallography for GE-ANP, Pratt & Whitney,
and In-Pile Corrosion Tests

11. Radiation Damoge Consultants

P. Materials Analysis ond Inspection Methods

136

1. Determination of the Oxidotion States of Metallic
Corrosion Products

2. Determination of Trivalent Uranium in Fuels

3. Anclysis of Reactor Fuels and Coolants

4, Determination of Reducing Power of Fuels and
Coolants

5. Determination of Oxides in Reactor Fuels

6. High-Tempercture Mass Spectrometry

7. X-Ray Study of Complex Fluorides

9766
9766
9766
U. Ala,
9766

9766
2000
9766

9201-3

2000
2000
2000
2000
2000
2000

3550
3005

3550
3550
3550
3005

3550

3005
3025
3001
3025
3001
3005
3025
3001

9733-4
9733-4
9733-4

9733-4
9733-4
9735
3500
9766

Shevlin, Johnson, Taylor
Johnsen, White

White, Shevlin

Handwerk

Doney

Curtis, Kertesz, Buttram
Vreeland, Hoffman

Doney, Corey

Adomson

Oliver, Douglas
Otliver, Woods
Olivar, Dougles
Oliver et al.

Oliver, Woods, Reber
Qliver, Reber

Keilholtz, Morgan

Robertson, Webster, Robinson,
Willis, Browning

Keilholtz et al.

Keilholtz et ol.

Sisman, Baumann,

Carroll, Brundage, Blacksher,
Parkinson, Ellis, Olson, Morgan

Sisman, Parkinson, Baumann
Carroll

Trice, Sismon

Wilson, Zukas

Davis

Wilson, Hinklz

Cohen, Templeton

Weeks

Feldman, Richt

Acton, Schwartz

Ruoark, U. Ala,

Smith, Cornell U.

White, Manning
White, Ross

White, Ross, Roemer, Manning

White, Manning

White

Baldock, Sites

Agron, Johnson, Bredig
Thoma

'
8.
9.
10.
11.

12,

13.

14.
15.

Petrographic Examination of Fuels

Chemical Methods of Fluid Handling

Metallographic Examinotion

ldentification of Corrosion Products from
‘Dynamic Loops

Assembly and Interpretation of Corresion Data
from Dynaomic Loop Tests

Metallurgical Examination of Engineering Ports

identification of Corrosion Products

Experimental Material Distribution

lil. SHIELDING RESEARCH

A, Cross-Section Measurements

Te

Effective Removal Cross Sections

B, Shielding Measurements

1.
2,
3.
4.
5.
6.
7.
8,
?.

Slont Penetration of Neutrons

Air Duct Tests — Fast Neutrons

Unit Shield Optimization

Mockup Tests for Shield Design
Divided-Shield Mockup Tests (GE-ANPP)
Duct Tests (GE-ANPP)

Gomma Rays from Fission

FulleSize Mockup Tests of Aircraft Shields
Design Parameter Studies {Divided Shield}

C, Shielding Instruments

1.
2.

3.
4.
5.
6.

Gammoa-Ray Spectroscopy and Calculations

Short-Period Fission-Product Gamma-Ray Spectra

Fast-pleutron Dosimeter
He® Counter for Neutron Spectrometry
Neutron Spectroscopy ~ Proportional Counter

Neutron Spectroscopy — Photographic Plotes

D. Shielding Analysis

1.
2,
3.
4.

Development of New Calculational Methods
Application of Stochastic Methods
Calculation of Neutron Attenuations

Calculations of Gammua-Ray Attenuation

E. Experiments on Shielding Facility Reactors

1.

2,

Miscellaneous Reactor Experiments { Temperature
Coelficient, Xenon Poisoning, etc.)
Bulk Shielding Reactor Power Colibration

 

St ower Shielding Facility,

yNuclaur Development Associates, Ine.

PERIOD ENDING DECEMBER 10, 1953

9766
9201-3
2000
9733-3
2000
9201-3
2000
9201.3
2000
9766

3001

3001
3001
3001
3001
3010
3010
3010
TSFS
TSF®

3010
3010
3001
4500
3010
3010
3006

4500
4500
NDA7
NDA’

3010

3010

McVYay, White

Mann, Blakely

Gray, Krouse, Roeche
Blankenship

Smith, Borie, Dyer
Adamson, Blankenship
Smith, Vreeland

Adamson, Gray, Vreeland
Y akel

Kertesz, Meadows, Didlake

Storrs, Trubey, Watson

Storrs, Chapman

Storrs, Miller

Storrs and crew

Storrs and crew

Cochran and crew

Cochran, Fiynn, Hull, Hungerford
Maienschein, LLove, Cochran, Henry
Clifford, Blosser, Hollond, Watson
Clifford, Blosser, Holland, Watson

Maienschein, Love
Maienschein

Campbell, Love

Blosser, Hurst

Cochran, Henry

Cochran, Henry

MHaydon, Johnson, Cochran

Murray

Favlkner, Zerby

Goldstein, Aronson, Certaine
Goldstein, Wilkins

Cochran, Meem, Cole, Johnson,
Flynn

Johnson, Cochran

137
 

CHART OF THE TECHNI{CAL ORGANIZATION OF THE AIRCRAFT NUCLEAR PROPULSION PROJECT
THE OAK RIDGE NATIONAL LABORATORY

AT

DECEMBER 1, 1953

 

 

ANP PROJECT DIRECTOR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SOVIIND
SRR —

“*Thin material comtaing informotion gHwcting the
of tha Unitad Statan within the
mucning of the aupi rla 18, U.S.C.,
Seca, 793 ond 794, the mansmission or revelotion

natianel dufm

    

o

ianage laws.

of which in any maner o on unauthorized person

is prohibited by law.**

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R. C. BRIANT RD
D, HILYER,* SEC. ANP
ASSISTANT DIRECTOR
A4 MILLER ANP
TECHNICAL ASSISTANT
ASSISTANT TO DIRECTOR
W, B, COTTRELL ANP
L. APA. fiAogKmu e m: A. W. SAVOLAINEN ANP
: i 3EC J. M. CISAR®
ANP LIBRARY P. HARMAN, SEC. AN:‘.
BYILDING §704-1 '
M. CARDWELL RD
E. WEBSTER RD
STAPR ASSISTART FOR SHEELDING STAFF ASSISTANT FOR PRYSICS A ) MiLLER ANP STAFF ASSISTANT FOR DESIGN ARE PROJECT ENGINEER STAFF ASSISTANT FOR METALLURGY STAFF ASSISTANY FOR CHENISTRY
D. HILYER,* SEC, ANP
E. P. BLIZARD P W, K. ERGEN ANP A. P. FRAAS ANP E. 5. BETTIS* ANP R. . BRIANT RO W. D. MANLY M W. R. GRIMES MC
SHIELDING RESEARCH REACTOR PHYSICS RADIATION DAMAGE GENERAL DESIGN ARE CPERATIONS EXPERIMENTAL ENGINEERING HEAT TRANSFER AND PHYSICAL METALLURGY CHEMISTRY
BUILDING 4500 BUILDING 9704-1 BUILDING 3025 BUILDING 9704-1 BUILDING 7503 BUILDING $201-3 PROPERTIES RESEARCH BUILDING 2000 BUILDING 97333
£. P. BLIZARD P ¥, K, ERGEN ANP D. 5. BILLINGTON* 55 A, P. FRAAS ANP E. 5. BETTIS ANP H, W, SAVAGE ANP BUILDING 5204-1 W. D. MANLY M W. R, GRIMES MC
4. B. TRICE 55 J. L. MEEM ANP H. F. FOPPENDIEK REE
L. 5. ABBOT P 1, BENGSTON ANP C. 5. BURTNETTE USAF G. M. ADAMSON M E. S. BOMAR M P. A. AGRON c
1. E. FAULKNER P €. 8. MLLs ANP B O N = R, . BUSSARD ANP R. G. AFFEL ANP J. P. BLAKELY He 1. 0. BRADFUTE REE C. R, BOSTON " C. 1 BARTON e
. X. BULLINGS P D. HILYER,* SEC. ANP " 1, 6. GALLAGHZR ALT J. W, ALLEN ANP C. P. COUGHLEN ANP 5. 1. COHEN REE W, H, BRIDGES M F. F. BLANKENSHIP MC
, T, BERESOVSKI PW G,
F. K. MURRAY p ¥, E. BROWNING by 8. L. GREENSTREET ANP . G. COBB ANP J, E. EORGAN MO N. D. GREENE REE 4. ¥. CATHCART M C. M. BLOOD MC
H, E. STERN cy COMPUTERS W E. BRUNDAGE s M. E. LAYERNE ANP W, B, COTTRELL" ANP W. C. GEORGE Y-12 0. €. HAMILTON REE J. M, CISAR* ANP R. A, BOLOMEY MC
R, RICKMAN, SEC, P A. FORBES ANP R, M. CARROLL bl W. L. SCOTT ANP G. A. CRISTY EM R.HELTON ANP H. W, HOF FMAN REE J. H. COOBS M F.P. BOODY MC
M. TSAGARIS ANP W W, DAYIS 38 R, L. SPENCER USAF J. Y. ESTABROOK ANP E. M. LEES Anp F. £, LYNCH REE 0, A, DOUGLAS M 5. A, BOYER MC
CONSULTANT M ). FELDMAN o . J. STUMPF ANP A. G, GRINDELL ANP R, E. MACPHERSON ANP L. D. PALMER REE E. E. HOFFMAN M L. M. BRATCHER MC
H. A. BETHE, CORNELL UNIVERSITY R W. HALL NACA B. M. WILNER AGC W, R, HUNTLEY ARF L. A MANK ANP . D. POWERS REE H. INQUYE o D. R, CUNED MC
- A . N E. HINKLE se D. HILYER,* $£C. ANP J. K. LESLIE ANP R. N. MASCN EM M., ROSENTHAL REE R.B. OLIVER M H. A. FRIEDMAN MC
. W, KEH HOLTZ s E. R. MANN® i #. B. MCDONALD ANP T. SUTTON, SEC. REE P, PATRIARCA M E, E. KETCHEN W
CONTRACTOR CRITICAL EXPERIMENTS 1 6. MORGAN s CONSULTANTS E. L. MORRISON USAF G. J. NESSL MC K. W. REBER M R.P. METCALF MC
NUCLEAR DEVELOPMENT ASSOCIATES, INC. BUILDING 9212 Mo T, MORCAN s A. H. FOX, UNION COLLEGE E. B. PERRIN AP o 'F‘- N ::; TECHNICIANS G. M. SLAUGHTER M R. E. E MC
H. GOLDSTEIN A D. CALLIMAN* p A, S. OLSON 5 R. L., MAXWELL, UNIVERSITY OF TENNESSEE A. L. SOUTHERN ANP g €. 6. BLALOCK REE G. P. SMITH M L. G. OVERHOLSER MC
H. 6o . D, L 5 G B WISLICENOS, JORNS HORKINS UNIVERS:TY H. L. WATTS REE D.F. ANP R. M. BURNETT REE . €., STEIDLITZ M J. D. REDMAN MC
- ARONSON V. 6. HARNESS* B W, PARKINSON 5 C. F. WEST ANP P. G, SEXTON Y-12 5. J. CLAIBORNE REE 0, C, YREELAND M R. J. SHEIL MC
J. J LYNN P A E. RICHT 55 G. D. WHITMAN ANP P\ G, SMITH ANP T. N. JONES REE J. W. WOO0DS M 8. J. STURM MC
J. H. MARABLE" P | H. E. ROBERTSON 55 E. WISCHHUSEN EM W, C. TUNNELL ANP J. LONES REE 1. THOMAS, SEC. M R, E, THOMA, JR, MC
1. W. NOAKS PW " T' ROBINSON s P. HARMAN," SEC. ANP D. R. WARD ANP G. M. WINN REE L, E. TOPOL MC
E. R. ROHRER* p O SISMAN s D. ALEXANDER, SEC. ANP TECHNICIANS JACK TRUITY NC
BULK SHIELDING FACILITY . TECHNICIANS D. STOREY, REC. CL. ANP . M. WATSON e
D, SCOTT, JR, ANP W, J. STURM 55 b PR 'ST N 6. D, BRADY M
BUILDING 3010 D. V. P. WILLIAMS P L. C TEMPLETON = 1. 5. ADDISON ANP . . STEND, 1 T. EAST M C. F. WEAVER Me
R. G. COCHRAN P £. L. ZIMMERMAN P P. M. UTHE USAF T. E. CRABTREE ANP TECHNICIANS G. M. GONZALEZ M 0. E, CALDWELL, SEC.  MC
F. C, MAIENSCHEIN® P M. L. RUEFF,* SEC. P C. C. WEBSTER 55 J.R. CROLEY ANP CERAMIC RESEARCH L. L. HALL M
4 R, A WEEKS % T. L. GREGORY ANP G. b. BRADY M BUILDING 5764 1. . HUDSON M CORROSION LABORATORY
1D, FLYNN P W. R. WILLIS 5§ R. REID ANP G. 5. CHILTON ANP 1. M. WARDE" M R, W, JOHNSON " BUILDING 9766
B HAYDON* I e WILSON b F. ). SCHAFER ANP J. M. COBURN ANP V. G LANE "
M P. HAYDON P . R CROFT e F. KERTESZ MG
K. . HENRY e 1 C. ZUKAS 55 B. C. witLL1AMS ANP . R, C. E. CURTIS* M B. MCNABE M H. J. BUTTRAM MC
H, E. HUNGERFORD p REACTOR CALCULATIONS M, STEWART,* SEC, 55 4. M. CUNNINGHAM ANP L. M. DONEY* M J. €. POPE M F. A KNOX MC
E. B. JOHNSON P BUILDING 4500 CONSULTANT é h: %lgsl-sAKE fip 5, D, FULKERSON® M C. E. SHUBERT M R E. MEADGWS uC
G. M. MCCAMMON P TECHNICIANS J. H. BUCK, WELL SURVEYS. INC. R. k. HAMNER® M L. R TROTTER M
< %unos, SEC. P R4 CHARRE o C.ELLIS 5§ BT L Anp M. P, HAYDON® P €. K. THOMAS M No Vo SMITH Me
5. B ATTA" " F. M. BLACKSHER 5 R A m;ch ANP L J. R, JOHNSON M CONSULTANTS
TECHNICIANS R. R. COVEYOU* ANP B L. JOHNSON ANP e ; L‘:'#g“ " METALLOGRARHY J. M. CARTER
H. JARVIS Ic C. L. GERBERICH* WP CONSULTANTS 5. W. KINGSLEY ANP " A 0B85 * SEC " R, J. GRAY" u T. R. P. GIBE, WILLIAMS COLLEGE
D. J. KIRBY P N, D, GIVEN* MP J. E. GOLDMAN, REACTOR CU“STRDI- D. £, MCCARTY ANP . ‘ . "R, 5. CROUSE* M D. G. HILL, DUKE UNIYERSITY
R. M. SIMMONS P F. G. PROHAWMER* ANP e BUHLDING 4500 G. E. MILLS ANF TECHNICIAN : , H. INSLEY, BUREAU OF STANDARDS
G. STOUT K W. C. SANGRENY MP CARNEGIE INSTITUTE OF TECHNOLOGY E. P, EPLER* ic . A, REDDEN ANP LA GRIFF:«IC' < ¥ g iggfgv, . T. N MCVAY, UNIVERSITY OF ALABAMA
H. WEAVER P D. F. WEEKES, TEXAS A & M COLLEGE D. £, TIDWELL Abie - A M + M
; E SS;EN :g R.G MLEY ANP CONSULTANTS TECHNICIANS CONTRACTORS
1. C. GUNDLACH® ic CONSULTANTS T, N, MCYAY, UNIVERSITY OF ALABAMA M. D. ALLEN® M BATTELLE MEMORIAL INSTITUTE
_ S H. HANAUER® i T. 5. SHEVLIN, OHIO STATE UNIVERSITY N. M. ATCHLEY! M 1. W. CLEGG
YONER SHIELDING FACILITY E. R MANNE i© J, B, BAILEY, UNIVERSITY OF TENNESSEE E. R. BOYD M R, ¥, DAYTON
. W. R. CHAMBERS, UNIVERSITY OF TENNESSEE W, H, FARMER* M
C. E. CLIFFORD P . 1.8, RUBLE ic . R, I. JAFFEE
sdbd L. C. OAKES* ic J. F. HAINE ®, L. FITZGERALD M H. A. PRAY
Y. v, BLOSSER P C. 5. WALKER* i< W. K. STAIR, UNIVERSITY OF TENNESSEE J. €. GOWER® M " AND OTHERS
o P. GROOVER, SEC, ic E. P. GRIGGS" H
'j' E' :Sttmb : OOVER, R. M. WALLACE* M CARTER LABORATORIES
F. N. WATSON P C. E. ZACHARY* M 2 #. CARTER
TECHNICIANS MOTE: THIS CHART SHOWS ONLY THE LINES OF TECHNICAL COORDINATION OF THE ANP PROJECT. THE COMSULTANTS MEl;AS rgrf&snss, INE,
L. N, MONEY P VARIGUS INDIVIDUALS AND GROUPS GF PEOPLE LISTED ARE ENGAGED EITHER WHOLLY OR PART TIME ON FUEL REPROCESSING N. CABRERA, UNIYERSITY OF VIRGINIA e s
D, SMIDDIE ic RESEARCH AND DESIGN WHICH IS COORDINATED FOR THE BEMEFIT OF THE AMP PROJECT N THE MANNER BUILDING 4550 N. J. GRANT,
INDICATED ON THE CHART. EACH GROUP, HOWEVER, 15 ALSQ RESPONSIBLE TO IT$ DIVISION DIRECTOR F.L. CULLER® et MASSACHUSETTS INSTITUTE OF TECHNOLOGY MOIJJN-:) ;S:gié‘E’ORY
FOR THE DETAILED PROGRESS OF ITS RESEARCH AND FOR ADMINISTRATIVE MATTERS, F. R. BRUCE cr J. L. GREGG, CORNELL UNIVERSITY "AND OTHERS
E.F. NIPPES,
F. N. BROWDER cT
¥ AS
LD TANK THE KEY TO THE ABBREVIATIONS USED IS GIVEN BELOW. IT SHOULD BE NOTED THAT SEVERAL OUTSIDE G. I CATHERS cr RENSSELAER POLYTECHNIC INSTITUTE R T KANS
ILDING W. F. SAVAGE e
RGANIZAT| VE PERSONNEL PARTICIPATING IN THE ORNL ~ANP PROGRAM . AND OTHERS
C. L. STORRS GE ORGANIZATIONS HAVE FERSONKEL PARTIC ORNL RENSSELAER POLYTECHNIC INSTITUTE
E. C. WRIGHT, UNIVERSITY OF ALABAMA
G. T. CHAPMAN P AC ANALYTICAL CHEMISTRY DIVISION — ORNL
R £ AGC  AEROJET-GENERAL CORFORATIGN CONTRACTORS
- ALC  AMERICAN LOCOMOTIVE COMPANY BALDWIN-LIMA-HAMILTON CORFORATION ANALYTICAL CHEMISTRY
TECHNICIANS F. G, TATNALL BUILDING 57334
ANP  AIRCRAFT NUCLEAR PROPULSION DIVISION ~ ORNL AND OTHERS C. D, SUSANG ac
W, 4, MAYNARD ; c CHEMISTRY DIVISION ~ ORNL B, E. YOUNG,” SEC, AC
i. R, TAYLOR BATTELLE MEMORIAL INSTITUTE
. W, WAMPLER P cr CHEMICAL TECHNOLOGY DIVISION — ORML (AEC CONTRACTOR}
J. SELLERS ic cv CONSOLIDATED VULTEE AIRCRAFT CORPCRATION T. M. PARKE ch :;ICT%TCHE"' :g
EM ENGINEERING AND MAINTENANCE DIYISION ~ ORNL - H. SALLEY E .
7 SCHOFIELD E. M. ZARZECKI,* SEC. AC
GE GENERAL ELECTRIC COMPANY :
AND OTHERS BUILDING 9201-2
ic INSTRUMENTATION AND CONTROLS DIVISION — CRNL ERITY MICHIGAN COMPANY
_ A. F. ROEMER, JR. AC
M METALLURGY DIVISION - ORNL GRAFF R, as
MC MATERIALS CHEMISTRY DIVISION ~ ORNL AND OTHERS AND OTHERS
WP MATHEMATICS PANEL - ORNL MASSACHUSETTS INSTITUTE GF TECHNOLOGY CONSULTANT
NACA  NATIOMAL ADVISORY COMMITTEE FOR AERONAUTICS T WULFF M L WILLARD. UNIVERSITY OF MICHIGAN
B FHYSICS DIVISION — ORNL AND OTHERS ’
PW PRATT AND WHITNEY AIRCRAFT DIVISION — UAC SPECTROGRAPHIC ANALYSIS
R
KD RESEARCH DIRECTOR'S DEPARTMENT — ORNL B pe Y TECHNIC INSTITUTE BUILDING 9734
REE  REACTOR EXPERIMENTAL ENGINEERING DIVISION - ORML AND OTHERS J.R. MCNALLY sl
] STABLE ISOTOPES DIISION — DRNL SUPERIOR TUBE COMPANY J. A NORRIS* H]
. . AND OTHERS
55 SOLiD STATE DIVISION - CRNL HUGH COOFER
USAF  UNITED STATES AIR FORCE MASS SPECTROME TRY
¥-i2  CARBILE AND CARBOK CHEMICALS COMPANY {Y-12 SITE) BUILDING 9735
PART THHE C. R. BALDOCK* ]
1. R. SITES H]

 

 

 

 

 

 

 

 

 

138