ORNL-0858.txt

 

-

r

—
Iq.:} s

 

    
   
   
   
  

 

/”?y;;7 4%Z4;;¢aa§?—f~

Progress Report:

N PY 127

 

I | (9,.;

nw.) Z° g@'

&y’ =

 

 

 

3 Y456 D56L2A8 7

      

 

- i |\
. : ' /) o)
. {¥ ) ) " "1;' 4 ' o
) 3 e A dffidlfifl Iluflff(_- j:?;tftur fthz &g
at L{T'L-Qé}f{}-fig_ e c /?[(f "f{ A& L-f.-t ,E. &
%.Ei;tfifi.héfina—-m——m'” //f/”?(/
RABORAY _
i h_f??fif21f ffiffi; : -

THE AIRCRAFT NUCLEAR PROPULSION
PROJECT QUARTERLY PROGRESS REPORT |
FOR PERIOD ENDING AUGUST 31, 1950 ‘g

{

@

Py - 5

45 4. 2
i e
tu&.ihflfw TR ok L e 'I-fl-,-."u fliM\w

a9y ad
L O«

f?>é?”?f

OAK RIDGE Ni‘l’ éNAL LAEORATORY

Wi il

CATION Eflflqpfilil,

H‘T"’;(

b
JUL

     

J__F\
#
o ‘.r_
/. e '.-"q " ,/
N TN ‘j'zi---"
BT B iy "
LR i PER
X LOE g
L' M

vl 2T
& "y E-r'-|‘ E'I:j-??

T

ff:i' :{J]d"{,t 3 “d
l‘iL & U’F-fl"‘- f

S g,

OPERATED BY

. IDE_AND CARBON CHEMICALS DIVISION
SLAS3EIp UNIG 'CARBIDE AND CARBON CORPORATION

POBT OFFICE BOX P
OAK RIDGE, TENNESSEE

a
~SECRET

 

 
 

ORNL 858

This document consists of 141 pages.
CopyazLfl7 of 182. Series A

Contract No. W-7405, eng 26

THE AIRCRAFT NUCLEAR PROPULSION PROJECT

QUARTERLY PROGRESS REPORT

for Period Ending August 31. 1950

Edited by
C. B. Ellis and W, E. Thompson
 

 

 

ORNL 858

This document consists of 141 pages.

Copy 2.7/ of 182

Contract No. W-7405, eng 26

THE AIRCRAFT NUCLEAR PROPULSION PROJECT

QUARTERLY PROGRESS REPORT

for Period Ending August 31. 1950

Edited by
C. B. Ellis and W, E. Thompson

DATE 1SSUED: €6 4 1950

OAK RIDGE NATIONAL LABORATORY
operated by
CARBIDE AND CARBON CHEMICALS DIVISION
Union Carbide and Carbon Cerporation
Post O0ffice Box P
O0ak Ridge, Tennessee

Series A

 
  

—

INTERNAL DISTRIBUTION

ORNL 858

Reactors

Progress Report

Jones
Miller
Billington
Blizard
Breazeale
Ergen
Ciifford
Nelson
Shipley
Clewett
Keim

@

>
2

isman
Callihan
Livingston
Susano

Taylor

Manly

» Meem

ANP Library

Central Files (0.P.)

>

“EMODFOOOMEIONET=EOEW

o

1. G. T. Felbeck (C&CCD) 25. A. Hollaender 44.
2-3. 706-A Library 26. F. L. Steahly 45.
4, 706-B Library 27. K. Z. Morgan 46.
5. Biology Library 28. D. W. Cardwell 47.
6. Health Physics Library 29. M. T. Kelley 48.
7. Metallurgy Library 30. W. H. Pennington 49.
8-9. Training School Library 31. C. E. Winters 50.
10-13. Central Files 32. J. A. Lane 51.
14, C. E. Center 33. M. M. Mann 52.
15. C. E. Larson 34. G. E. Boyd 53.
16. W. B. Humes (K-25) 35. R. W. Stoughton 54.
17. W. D. Lavers (Y-12) 36. F. R. Bruce 55.
18. A. M. Weinberg 37. H. W. Savage 56.
19. J. A. Swartout 38. W. K. Eister 57.
20. E. J. Murphy 39. A. S. Householder 58.
21. F. C. VonderLage 40. C. B. Graham 59.
22. R. C. Briant 41. R. E. Engberg 60.
23. C. B. Ellis 42. R. N. Lyon 61.
24. A. H. Snell 43. W. R. Gall 62-71.
72-77.
EXTERNAL DISTRIBUTION
78-82. Air Force Engineering Office, QOak Ridge
-83-94. Argonne National Laboratory
95-102. Atomic Energy Commission, Washington
103. Battelle Memorial Institute
104-107., Brookhaven National Laboratory
108. Bureau of Aeronautics
109. Bureau of Ships
110-111. Chief of Naval Research
112-115. General Electric Company, Richland
116. Hanford Operations Office
117-120. Tdaho Operations Office
121. Towa State College
122-125. Knolls Atomic Power Laboratory
126-128. Los Alamos
129. Massachusetts Institute of Technology (Kaufmann)
130-131. National Advisory Committee for Aeronautics, Cleveland
132. National Advisory Committee for Aeronautics, Washington
133-151. NEPA Project
152-153. New York Operations QOffice
154. North American Aviation, Inc.
155-160. Oak Ridge National Laboratory, Y-12 Site
161. Patent Branch, Washington
162-176. Technical Information Service, Oak Ridge
177-178. University of California Radiation Laboratory
179-182. Westinghouse Electric Corporation

  
 

TABLE OF CONTENTS

SUMMARY
INTRODUCTION

PART 1. RESEARCH CONTRIBUTING TO THE ARE

SHIELDING' RESEARCH
‘Bulk Shielding Measurements--Lid Tank

A method for experimental shield optimization

33% Pb--67% H,0 neutron data

Efféct of addition of boron to water

Gamma attenuation as a function of lead disposition

Fast neutron dosimeter

Shield for reactor with Fe reflector

Interpretation of dosimeter data

Neutron energy spectrometer

Measurements of fast neutrons in the lid tank us1ng sulphur
as a detector

‘Liquid Metal Duct Tests in the Thermal Column

Shield Calculations

Calculations of neutron attenuation
Analysis of lid tank data

Heat generation in shields

Shield calculations

Large air ducts in shields

New Shield Test Facility
CRITICAL EXPERIMENTS
EXPERIMENTAL ENGINEERING

HEAT TRANSFER

Heat Transfer Theory

Forced convection entrance solutions
Heat and momentum transfer in ducts whose cross sectlons are
irregular or annular

Experimental Heat Transfer Equipment

3

/

11

16

17

17

17
18
18
25
28
28
28
34

35
38

38

38
4]
41
41
41

42

 
 

J
Boiling Liquid Metals
High Heat Transfer Coefficients
Pump Development for Experimental Systems
Physical Property Determinations

Liquid Metals Handbook
Liquids Metals In-pile Experiment

METALLURGY AND MATERIALS

Static Corrosion Testing
Stainless steels in lithium
Special tests
Pure metals in lithium
Refractories in lithium
Pure metals in lead
Materials in sodium hydroxide
Materials in sodium

Dynamic Corrosion Testing

Thermal convection loops
Creep-Rupture and Stress-Corrosion Tests'

~ Assembly of Fuel Elements

RADIATION DAMAGE

'Y-12 Cyclotron Experiments

‘North American Aviation, Inc.: Lithium-Iron Accelerator Corrosion
Experiment

Purdue University

In-pile Creep

NUCLEAR MEASUREMENTS

Mechanical Velocity Selector
Xenon Cross-section Measurements

Molybdenum Cross-section Measurements

REACTOR PHYSICS
ARE Calculations

Multi-group calculations
IBM calculation

55
56
56
58
60
60

61

61

61
71
71
78
78
78
81
82

82
83
83

85
85

86 .
86
87

88

88
88
88

90
90

90
91

 
 

s

Reactor constants , 91
Adjoint calculations | 92
Bare reactor calculations | 92
Multi-group calculations in cyclindrical geometry 92
Energy Dependent Pile Equations | 93 
DESIGN,OF THE AIRCRAFT REACTOR EXPERIMENT 95
~ Fundamental Design Principles 95
'Dqtailed Design and Construction 96
PART II. LONGER RANGE ACTIVITIES 99
- INTRODUCTION 100
COMPRESSOR -JET CYCLES ' 101
. CIRCULATING FUEL REACTORS | - 102
Circulating Fuel Reactor Analysis 102
Circulating Fuel Reactor for Supersonic Plane 103
Cheniistry of Liquid Fuel Systems 104
General characteristics of fuel system 105
Choice of uranium compound 105
Choice of vehicle for self-moderating systems 105
Choice of vehicle for unmoderated systems 107
Systems containing sodium hydroxide - 107
Low melting fluoride systems-—thermal analysis 110
FIXED FUEL CIRCULATING MODERATOR REACTORS 117
'LITHIUM ISOTOPE SEPARATION 121
Ion-Exchange Separation of Lithium Isotopes 121
Large-Scale Lithium Isotope Separation Methods 121
Lithium amalgam-aqueous hydroxide system 122
Organic-organic systems 127
Molecular distillation 130

 
——

-

la
'1b

T e W N

10a
10b
11
12
13
14
15
16
17

LIST OF ‘TABLES

Neutron Centerline Measureménts

‘Thermal Neutron Centerline Measurements

Neutron Centerline Measurements
Effect of Water Boration on Gamma Centerline Measurements

Gamma Measurements

" R?1 for Non-uniform Pb-H,0

Effect of Moving Single Pb Slab on R?l - All.Qther Slabs
Remaining Constant

Effect of Moving Pb Slab on B2l - All Other Slabs Remain-

ing Constant

Effect of Moving Pb Slab on R21 - All Other Slabs Remain-
ing Constant

Gamma Centerline Measurements

Fast Flux Measurements in the Lid Tank

Corrosion Results of Stainless Steels in 1000°C Lithium
Corrosion Data on the Pure Metals in 1000°C Lithium
Aircraft Reactor Specifications

Uranium Compounds

Chief Characteristics of‘NaOfi-Cooled ARE Reactor
Organic Solvents for Lithium Exchange

Non-oxygen Coordinator Solvents

Standard List of Organic Solvents

Data for Lithium Runs in the Molecular Still

20
23
24
26
27
29

30

31

32
33
36
63
79
97
106
118
131
131
132
133

 
U

O O NN N A W

10-16
17-20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

LIST OF FIGURES

Effect of Pb Addition on Thermal Neutron Distribution

Attenuation of Fast Neutrons in Lid Tank Using Sulfur Threshold
Detector o

Liquid Metal Duct Mock-up~—First Leg

Liquid Metal Duct Mock-up—-Second Leg and Elbow
Convection Loop

Figure "8" Loop

SS Type 304 Harp in Qperation

Cross Section of Hydraulic Bearing .

Completely Sealed Rotary Pump

Stainless Steels Exposed to Lithifim at 1000°C for 40 hours
Effect of Time of Exposure to Lithium at 1000°C

Effect of Time on Lithium Corrosion of Stainless Steeis
Effect of Graphite Addition on Lithium Corrosion of Stainless Steel
Tentative General Core Arrangement for ARE

Filtration Apparatus

Crucible Assembly

Thermal Analysis Assembly

NaF-UF, Phase Diagram

KF-UF, Phase Diagram

Exchange Cells

Counter Current Lithium Exchange

Exchange Cells in QOperation

Madorsky Type Still |

Condenser Plates for Madorsky Still

Evaporator for Madorsky Still

Vertical Still

19

37

‘39

40
A7
48
50

37

59
64-70
72-15
76

77

98
109
111
113
114
115
124
125
126
136
137
138
141
Jr
ol

. SUMMARY

Shielding. Measurements in the lid tank on lead and water combinations
are continuing, The optimum distribution of lead has cértainly‘not yet been
achieved. Neutron measurements show that lead is slightly better than water
for a neutron shield. on a volume basis. Boron has now been added to vhe water
with a consequent reduction in calculated shield weights. For one‘of‘the
measured lead-water shields, the boron experiments show that the number of
capture gammas had been six times ‘as great as the number of gammas from all
other sources. One arrangement of lead and boratéd watqr appeared at the end
of the gquarter to yield a shield to surround a four-foot spherical reactor
which weighs only 54 tons, if the reactor-crew separation is taken as 100 feet.

Further measurements to check this calculated weight are still in progress.,

A fast neutron dosimeter has been installed ip the lid tank, Its measure-
ments represent an improvement ever the previous system of calculating fast

flux from the measured thermal flux.

In one Pb-H,0-B arrangement, an iron slab was placed on the inside of the
shield. The iron proved to be as efficient for attenuating neutrons and

gammas in this position as the lead-water it replaced.

A new method is being developed for measuring fast neutrons in the lid

tank which used sulphur as a detector.

A liquid metal duct test is being set up in the thermal column 1in the

ORNL reactor.

The new Bulk Shield Test Facility is expected to be completed by the end
of October. |

critical Experiments. The Oak Ridge Area Critical Mass Laboratory was
turned over to the operating staff on August 21, 1950, Experimental equipment

is being installed,.

Experimental Engineering. Stainless steel harps have now been operated
both under thermal convection and with an electromagnetic pump using Na and

NaK. The tests have reached 1700°F with Nak.
‘Heat Transfer. Experimental equipment for measuring molten metal heat
transfer coefficients was essentially completed during this period. The
experimental program will be started in the near future. Equipment for
measuring the heat transfer to boiling liquids is being designed. Preliminary
design of a system for obtaining high coefficients of heat transfer to molten
sodium has been started. Equipment for the determination of the thermal

conductivity of metals at temperatures up to 1800°F is being construcced.

‘Metallurgy and Materials. Experimental evidence to date indicates that
the ferritic stainless steels are more resistant to lithium at 1000°C than are
the austenitic alloys for short exposures. Beneficial results of adding
graphite to the molten metal were noted for the 310 and 316 stainless steels.
Longer tests support earlier evidence that iron, zirconium and columbium show
good resistance to lithium at 1000°C. Only moderately good resistance to
lead at 1000°C was noted in tungsten, zirconium, iron and tantalum; molybdenum,
columbium and beryllium were fair; titanium and nickel poor. Visual inspec-
tion of samples indicates that most materials, especially the austenitic
stainless steels, are more resistant to sodium at 1000°C than to lithium,
Nickel is found to be quite resistant to sodium at 1800°F. Various metals and
refractories have been tested at 1800°F in NaOH. Nothing except nickel has

yét been found which has not been drastically attacked.

_ Radiation Damage. Plans are being made to irradiate copper and 316
stainless steel in the 20 Mev proton beam of the Y-12 cyclotron. The in-pile

creep apparatus has been tested and placed in operation in the ORNL reactor.

Nuclear Measurements. The possibility of extending earlier thermal
xenon cross-section vs. energy measurements to higher neutron energies is being
considered. Two possible methods are being studied, the epi-cadmium deactiva-
‘tion at high flux and danger coefficients in intermediate critical assemblies.
‘Intermediate cross-section measurements on molybdenum are underway at Columbia
‘University.

Reactor Physics. Thirteen-group calculations, patterned on the GE

routine, are in progress for the proposed ARE reactor. The calculation system

is now being adapted to the IBM computing machines.
ARE Design. A preliminary report, Y-F5-15,with Supplements I to VIII,
has been completed describing the general features of the proposed ARE reactor.

%,
9
Joint ORNL-NEPA groups have been established to develop the details of the
shielding and control system designs. A preliminary design of the building to

house the reactor has been drawn up and a site selected.

Circulating Fuel Reactors. The H. K. Ferguson Company has prepared a
preliminary report on an integrated design for a B-52 aircraft powered by a
homogeneous reactor which uses a uranium suspension in NaQOH. Researci at Y-12

is in progress on possible solutions and suspensions of uranium compounds in
NaOH.

Circulating Moderator Reactors. The H. K. Ferguson Company has prepared-

a preliminary studyof a fixed-fuel circulating-moderator design as an alternate
ARE.

Li’ Separation. Research continues on the ion-exchange, the molecular
distillation and the liquid-exchange column methods of lithium isotope sepa-
ration. Enrichment up to 94% Li’ has been achieved in a system using lithium

amalgam and aqueous lithium hydroxide flowing countercurrently.

10
INTRODUCTION

During the past quarter the Aircraft Nuclear Propulsion (ANP) projecthfit
the Oak Ridge National Laboratory has continued to expand. The work has fallen
into three main categories: (a) sponsorship of the ANP Technical Advisory
Board, and cooperation in its studies to delineate the most feasible typeslof
supersonic aircraft nuclear power plants, (b) research aimed at solving the
problems involved in the liquid-metal-cooled type of reactor with the aim of
constructing a 1000 kw prototype in Oak Ridge (the ARE), and (c) exploratory
research on problems of other reactor cycles which might be employed in future

mode ls.

The Technical Advisory Board. During June and July, 1950, the Oak Ridge
National Laboratory, cooperating withNEPA, brought to Oak Ridge a distinguished
group of consultants to form the Technical Advisory Board (TAB) to the nuclear
aircraft program. The TAB explored analytically a great many possible reactor
and power plant arrangements for supersonic nuclear aircraft. Its members
included specialists in metallurgy, reactor theory, shielding theory, reactor
engineering, heat transfer, radiation damage, turbojet design, and aircraft
design. All of these fields of effort were involved jointly in the considera-
tions of the TAB and in its evaluation of the relative merits of the various
possibilities for a supersonic craft. A considerable fraction of the effort
of the Laboratory during the past quarter was devoted to consulting with and
assisting the members of the TAB in their work. At the close of its session,
the TAB prepared a final summary report which is now in press and will shortly
receive a wide distribution. The following section is a reprint of parts of
the chapter entitled "Conclusions and Recommendations™ from the final report

of the TAB:

"The TAB, in course of its session, has worked up steadily to the con-
clusion that there is a good chance for realization of a subsonic nuclear
plane in the near future. Such a plane is quite likely to attain a speed and

altitude comparable to those of the most advanced designs for large airplanes.

"Obtaining supersonic performance later may be expected as the result

of further development. It depends on the solution of crucial problems with

11
respect to materials, fabrication, and airplane and engine developments. The

time scale for solution of these problems cannot now be predicted.

"The TAB is of the opinion that a realistic program directed immediately
towards a tangible goal will greatly enhance the chance of obtaining an ultimate
high performance plane. An attempt to develop the ultimate plane without
intermediate stages 1s a proposition that would seem likely to lead only to

confusion and delays.

"These conclusions are based on the opinion that a sodium-cooled reactor
with maximum wall temperatures in the neighborhood of 1500°F and linear
dimensions of the order of 3.5 feet can probably be built to give a po&er out -
put of several hundred megawatts. With such a reactor anda fairly conventional
propulsion system the attaining of Mach 0.8 at 35,000 feet with a B-52 may be

set as- an early objective.

"The TAB is furthermore of the opinion that a homogeneous reactor would
have outstanding advantages. A major effort to determine its feasibility

should be made.. ... vovvan..

"As a means of concentrating effort on the more promising alternatives,

a program based on the following recommendations should be adopted.

A. Continue to give top priority to the liquid metal reactor with
sodium as a coolant.

(a) For the present, base the design studies on a con-
servative extrapolation of present-day knowledge of
materials and parts.

(b) Make integrated design studies for airplanes and
engines in conjunction with this reactor, both for
Phase I and, to the extent indicated in Recommendation

2, for Phase II.*

(¢c) Continue vigorously such research on materials for
the liquid metal reactor as will lead to eventual
operation at higher temperatures, possibly as high as
1800°F, but do not, for the present, make either the
reactor or the integrated design contingent upon the
success of this research.

Metallurgical and materials problems connected with
fuel elements for high heat flux should be vigorously
attacked until a practical solution is attained.

¢ pPhase I and Phase II refer to the subsonic snd supersonic nuclear aircraft,
respectively, :

12
B. Make a strong and sustained effort tosolve the specific problems
of the use of NaOH; that is, its corrosive properties and its
ability to carry fuel either in solution or as a dispersion;
also, to the extent possible, its stability against radiation
damage. The present design studies on the circulating moderator
reactor should be continued and, if the results of materials
research are encouraging, should be accelerated. If fuel carry-
ing also seems promising, major effort should be directed to the
homogeneous reactor.

C. Retain a small group to study other reactor types, not spec. fi-
cally referred to in A or B, but put no major effort on these
unless facts are discovered which alter the situation materially.
Continue some research on the crucial materials problems for the
air cycle in the hope of approaching a conclusion.

D. Avoid any expensive commitments for the separation of lithium-T7.
Continue developmental work on the molecular distillation and
chemical exchange methods at the present modest level, at least
until the high-temperature corrosion situation is clarified.
Drop research on the relatively unpromising electromagnetic
method.™

Research for the Aircraft Reactor Experiment. As' has been diSplayed in
previous quarterly reports, the Oak Ridge National Laboratory has research
underway in the following fields which touch upon problems of the liquid-metal-

cooled aircraft reactor:

Shielding,

Radiation Damage,

Heat Transfer,

Critical Experimentation,
Experimental Engineering,
Metallurgy, and

General Reactor Design.

During the past quarter the work on all of these groups has been more closely
pointed toward solution of problems for the 1000 kw liquid-metal-cooled
prototype reactor (ARE) proposed for Oak Ridge. The shielding group has
continued its tests in the lid tank and on combinations of iron, lead, and
water, and has extended them to include borated water. The radiation damage
group has continued work towards placing a creep test in the ORNL reactor, as
well as work on the corrosion of liquid metal containers under radiation. The

critical experiment group is preparing to test configurations appropriate to

13
the ARE. The experimental engineering group is beginning to circulate liquid
metals at high temperature in various laboratory-scale loops. The heat
transfer group is developing the theory of transfer to liquid metals in pas-
sages whose shapes are appropriate to ARE design. The metallurgy group is
continuing corrosion tests of probable container materials in liquid metals
of interest for ARE. Meanwhile, the ANP general design group has developed a
preliminary design covering the general characteristics of the ARE reactor.
Detailed design of the building and of the reactor proper are now underway.
The completion of the design will proceed as rapidly as answers to the many
problems involved come in from the research groups. If permission to build
this reactor is granted by the AEC, it is hoped that it may be put into
operation by January 1, 1952.

Previous to this quarter, most of the design thinking had been con-
centrated upon a reactor to power a highly supersonic aircraft. However, a
gradual lowering of the predicted shield weights resulting from the latest lid
tank results, and the change in the current international situation, have led
the Laboratory to place closer stress upon the possibility of attaining at
least subsonic nuclear flight at a much earlier date than previously visualized.
Therefore, some of the design specifications for the ARE are now being altered
so that it may serve as a prototype for the so-called "War Model" reactor for
a subsonic plane. The most notable change in this respect has been the de-
crease in the proposed fuel wall temperature from 1800°F to 1500°F, as rec-

ommended by the TAB.

Alternate Approaches for an Aircraft Reactor. Although, as just de-
seribed, the major part of the ANP effort of the Laboratory is now pointed
towards development of the fixed-fuel liquid-metal-cooled"type of reactor, a
considerable amount of research is also going into other types which might
offer promise for future models. Under contract with the Laboratory, the
H. K. Ferguson Company is continuing its study of the circulating-fuel reactor
possibilities, principally involving uranium suspensions and sodium hydroxide,
with the aim of presenting a complete design for operating a B-52 airplane
with such a reactor. Meanwhile at ORNL, the ANP chemistry group is attempting
to form a stable suspension of some uranium compound in molten sodium hydrox-
ide. This group is also searching for liquids, such as perhaps sodium fluo-

ride, in which uranium might be dissolved in sufficient quantities to provide

14
a liquid reactor fuel. As another line of attack for the future, the isotope
separation group is continuing the exploration of methods of separating
lithium-7 economically on a large scale for use as a future high-performance
coolant. Under direction of the Laboratory, North American Aviation, Inc., is
analyzing the possibilities of various compressor-jet cycles which might in-

volve such substances as sodium and mercury vapors as working fluids.

Conclusion. In general, the Oak Ridge National Laboratdry roncurs with
the findings of the Technical Advisory Board: (a) a reasonable amowu.t of re-
search should now solve the problems necessary for attaining subsonic nuclear
flight in any one of several ways, the most promising of these at present
involving a sodium-cooled reactor whose maximum temperature is 1500°F; and (b)
by extension of lines of research now underway there is promise of eventual

achievement of a supersonic nuclear aircraft reactor.

15
PART I. RESEARCH CONTRIBUTING TO THE ARE

16
SHIELDING RESEARCH

E. P..Blizard, Physics Division

Shielding Research at the Oak Ridge National Laboratory has made consider-
able progress during the past gquarter. Further results were accumulated
on combinations of lead, iron, boron and water in the lid tank, and the
instrumentation of that facility was improved. A new experiment on the effect
of liquid metal ducts is planned, which will go into a tank above the thermal
column on top of the reactor. Various calculations of neutron and gamma
attenuations have also been completed and the construction of the new shield
test building and pool is well underway. It is believed that almost enough
data arenow on hand to permit intelligent design of an engineered unit shield
for an aircraft reactor with the plumbing embedded within the shield in the
proper places. Plans are underway for construction of a moek-up of the first
engineered aircraft shield which should be installed in the new test facility
toward the end of the next quarter, The following sections describe the

results in these various shielding activities during the past three months.

BULK SHIELDING MEASUREMENTS--LID TANK

‘A Method  for Experimental Shield Optimization (E. P. Blizard). After
several uniform lead-borated water shields had been tested it became clear
that the optimum distribution of lead had not been achieved. A criterion was

set up for determining the optimum arrangement as follows:

(1) Tt is assumed that lead and water are equivalent on a volume
basis for neutron attenuation;

(2) A gamma detector is set up at the outside of the shield mock-up
and the effect of inserting a slab of lead is observed on this
instrument., A quantity [ is then defined as follows:

 

 

17
where

t = thickness of slah inserted,
[' = gamma intensity, slab in,
'y = gamma intensity, slab out.

The weight of the slab is proportional to R?t, where R is the radius
(on a spherical reactor) at which the slab is inserted. A given attenuation
can obviously be achieved most economically (weight basis) where R*l is a
constant. Since the | observed is a function of the whole shield, the optimum

must be approached by successive approximations.

The method has also been worked out for the more general case when the
two materials are not equivalent for neutrons. In addition, the optimum choice
of gamma and neutron dosage distribution has been shown to be, for the lead-

water shield, ¥ neutrons and % gammas.

For the case in which the reactor is not a sphere but a square cylinder,
the R used above loses meaning. R? must be replaced by the actual surface
of each layer. However, it has been shown that the square cylinder surfaces
for a 3-ft square cylinder coincide very closely with those for a 3.8-ft
sphere. Thus, the optimization already carried out in the lid tank can be
expected . to apply for both cases.

All the above discussion applies primarily to unit shields,

33% Pb - 7% Neutron Data (C. E. Clifford, E. P. Blizard, J. D. Flynn,
T. V. Blosser, L. H. Ballweg,* R, H, Lewis,** M, K. Hullings, K. Martin).
‘Neutron measurements were made for a 1/3 Pb, 2/3 H,0 shield mock-up. Data are
recorded in Table 1 showing the thermal neutron flux in.water behind success-
ively increasing numbersof slabs., The slab next to the source was inserted
first and the shield was built up regularly of l-in. Pb slabs every 3 in.,
Since the usual plot of flux vs. distance shows many lines superimposed, for
different numbers of lead slabs, Fig. 1 is used to show the effect of lead
insertion. It is seen that in .general the lead is slightly better than water
on a volume basis. The upturned ends for large numbers of lead slabs demonstrate

the relative thermal neutron transparency of lead.

Effect of Addition of Borom to Water. For the above configuration, with
nine Pb slabs inserted, B203 was added to.the water in several steps. The

* usar,
s+ NEPA,

18
6l

THERMAL NEUTRON FLUX RELATIVE TO PURE WATER DATA

DW!.9672

 

 

 

 

 

 

 

 

'fi | | I | | | | I | l l
“ FIG. |
EXP "8 -33% Pb-67% H,O NEUTRON ¢
1.3 EFFECT OF Pb ADDITION ON THERMAL NEUTRON
DISTRIBUTION
.2}
1.1
1.0¢ ®
o9}l - - Z:100
2= 110 J
0.8
e Z2=90cm
o Z=100c¢cm
o7 A Z=1i0 cm
a Z=120cm
8 Z=130cm
o Z=(40cm
0.6}- x Z=150¢m
0.5
I I | I I | | | | | I I
0 2 3 4 5 6 T 8

NO. OF 1"LEAD SLABS

 
0¢

TABLE 1a

-Experiuent‘s;- 33% Pb - 67%5520 Neutron Centerline. Measurements

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

25 in. BF,
" NUMBER OF SLABS T . . '
"IN PLACE Z = 80 Z = 90 ‘Z = 100 Z = 110 Z = 120 ‘Z = 130 Z = 140 Z = 150
0 396,580 121, 406 36,920 11,037 ‘3,665 1 336 '398 145
0 392,008 118,766 35,177 11,128 ; | |
0 397,422 119,003 35,594 10,994 3,620 1,141 391
0 418,488 118,284 34,810 11,108 3,546 1,151 ‘386
Average ¢/m 401,124 119,365 35,625 11,067 3,610 1,143 392 145
NV 82.792 24.69 7.67 |- 2.50 .792 .272 -100
1 400,934 121,079 36,085 10,879 3,441 1,103 353 130.7
1 406,588 121,002 36,346 11,134 3,545 1,148 372
Average 403,761 121,040 36,215 | 11,007 3,493 1,125 362 130.7
NV 83.88 25.1 7.62 2.42 779 - .251 .091
2 409,170 119,675 36,005 10,712 3,499 1,095 362
2 406,993 120,823 35,533 10,791 3,398 1,135 362
Average 408,081 120,024 35.769 10,751 3,448 1,115 362
NV 83.2 24.8 7.45 2.39 773 .251
3 117,862 33,517 10,159 3,299 1,045 344
3 114,192 33.283 10,277 ‘3,210 1,047 347
Average 116,027 33,400 . | 10,218 3,220 1,046 346
NV 77.3 23.1 7.08 2.23 . 125 : 240
4 113,816 32,704 10,102 3,095 963 330
4 115,658 33,924 9,922 3,158 982 337
Average 114,740 33,314 10,012 .. 3,127 972 334
NV 79.5 23.1 6.94 2.17 673 .231
(117,856 |
5 (115,362 33,016 9.955 3,156 973.9 321.3
5 115,076 32,858 9,865 3,043 948.1 335.7
Average 116,098 32,937 9,910 3,099 943.0 329 0
NV 80 4 22.8 6.86 2.15 .654 .228

 
1¢

‘TABLE 1a (Cont’d)

 

NUMBER OF SLABS

z =

Z =

 

 

 

 

 

 

 

 

 

 

 

.IN PLACE 80 90 Z = 100 110 Z = 120 Z = 130 -Z = 140 Z = 150
6 117, 642 32,760 9,525 2,955 940.0 304
6 (113,196 32,659 9,520 2,991 938 311
(114,185
Average 115,008 32,710 | 9,523 2,973 939 308
NV 79.7 22.7 6.60 2.06 .651 .21
7 113,134 31,707 9,195 2,875 874 286
7 (115,045 131,593 9,298 2,836 921 304.2
(114,160 32,429 9,351 2,897 905
‘Average 114,113 31,910 9,281 2,869 903 295
NV 79.1 22.1 6.43 1.98 .625 .204
8 118,988 32,291 9,146 2,844 886 283 92
8 118,914 31,676 9,076 2,829 869 286 --
Average 118,951 31,984 9,111 2,837 878 285 92
NV 82.4 22.2 6.31 1.466 .608 198 .064
9 (124,662
(122,868 31,271 ‘8,915 2,752 824 264. 1 94
9 125,738 31,787 8,880 2,636 830 266.8
Average 124,423 31,529 ‘8,898 2,694 827 '265.4 . 94
NV 86.2 21.8 6.17 . 1.87 .573 ©.184 .065
10 (141,826 32,822 9,207 2,688 820 2582
- (143,221 | |
10 141,446 33,422 8,927 2,684 841 263.0
Average 142,164 33,122 = | 9,067 2,686 830.5 260.6
NV ' 98.5 22.95]  6.28 1.86 .575 .181
11 177,562 35,907 9,027 2,662 787.4 251 79
11 169,122 35,004 8,943 2,589 777

 

 

 

 

 

 

 

267

 

 
A4

TABLE la (Cont'd)

 

.NUMBER OF .SLABS

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IN. PLACE Z = 80 Z = 90 zZ = 100 110
Average 173,300 35,500 8,980 2,625 782 259 79
NV .120.0 24.6 6.22 1.02 .542 .179
12 (43,545 9,446 2,617 776.4 238.5
(44,982 o
12 43,544 9,905 2,638 792.6 244 .4
Average 44,023 9,676 2,628 784.5 241.5
NV 30.5 6.71 1.82 .544 . 167
13 11,204 2,707 773.8 229.1 17
13 11,014 2,701 773 237
Average 11,109 2,704 773 233 77
NV 7.70 1.87 .536 .161 .053
14 14,282 3,084 769 --
14 14,804 3,102 785 231 75
Average 14,543 3,093 177 231 75
NV 10.1 2.14 .538 .160 .052
Note: Thermal flux (NV) = 0.693 x 10°3 X counts per minute.
TABLE 1b

Experilent_bu- 33% Pb - 67% H,0 Thermal Neutron
Centerline Measurements

Indium Foils.- Between Pb Slabs

 

 

éfi:lzgo '?éfis) NV
1 5 5.355 x 107
2 13 1.790 x 107
3 20 7.20 x 10°
4 28 1.512 x 106
5 35.5 4.678 x 10°
6 43 1.387 x 10°
7 51 3.343 x 10*
8 58.5 1.248 x 104
9 66 3.439 x 103
10 73.5 1.052 x 103
11 81.5 3.96 x 102
12 89 1.205 x 102
13 96.5 5.26 x 10!

 

 

 

23
¥e

TABLE 1c

25 in, BF,

Normalized to Pure Water

;Experiment 8 - 33% Pb: - 67$ H,0 Neutron Centerline Measurements

 

 

mmfl?}fikiums Z = 80 Z = 90 Z = 100 Z = 110 Z = 120 Z = 130 Z = 140 Z = 159—
0 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
1 1.0066 1.0140 1.0166 0.9946 0.9676 0.9842 0.9234 0.9013
2 1.0173 1.0055 1.0040 .9714 ~.9551 .9755 .9234
3 0.9720 0.9375 .9233 .8920 0.9151 0.8827
4 961 .935 905 8662 .850 .852 .823
5 973 .924 .895 .858 .825 .839
6 .963 .918 860 824 822 . 786
7 1956 .896 .839 795 .790 753
8 .9965 .8978 0.8233 .7859 .7682 .7270 .6345
9 1.042 .8850 8040 7463 .7235 677 .648

10 1.191 .930 .819 744 :.7265 . 665
11 - 1.4518 .9965 0.8114 0.7271 0.6842 0.6607 .545
12 1.236 0.8743 .7297 0.6864 0.6161
13 1.0038 0.7490 0.6763 0.5944 0.5310
14 1.3141 0.8568 0.6798 0.5893 0.5172

 

 

 

 

 

 

 

 

 
gamma flux at several thickness of the shield mock-up was recorded as a function
of boron added (see Table 2). A plot of gamma intensity vs. fraction of
captures in hydrogen and lead givés the ratio of gammas due to capture to
those due to all other sources. For the Pb-H,0 shield in question, this ratio
was about six. After sufficient boron was added to reduce captures to one-
twelfth, it was considered unnecessary to add more, and this concentration has
been maintained in the lid tank since that date. It is possible that further
addition might be advantageous for the recent measurements involving iron. The

water is, however, essentially saturated with the present concentration.

Gamma Attenuation as a Function of Lead Disposition. Two shield mock-ups
were measured for which the lead-water spacings were uniform. For the 33% Pb
(volume) case two concentrations of boron were used successively, showing only
a slight improvement for an increase from 0.2% B to 0.4% B. For the 17% Pb
case the measurement at lower boron concentration was omitted. The "effective
attenuation lengths, 1™ as described above were measured for these configu-
rations, and since R?l (R is the radius for a shield around a 4-ft spherical
reactor) is not uniform, it is evident that improvement is possible. Accord-
ingly, the lead slabs were moved successively to make R?] as nearly constant
as possible. For all these measurements the object was to specify experimentally
‘the amount of the lead required to effect gamma tolerance at the neutron-
determined shield outside. It is assumed that lead and water are equivalent

for neutrons.

Table 3 shows the values of | and R%2l;for these experiments. Shield
weights reported are in short tons and for highly idealized shields, being
valuable for comparison only. The method of computation is outlined in the

last ANP Quarterly Report (ORNL 768) and is based on the following assumptions:

Reactor power, 400 megawatts

Reactor diameter, 4 feet

Reactor leakage, 20%

‘Tolerance, l’B/hr, 3/4 gémmas, 1/4 neutrons
Reactor-crew separation, 100 feet

No ducts, heat exchangers

No shadow shielding

25
- TABLE 2

Effect of Water Boration on Ganma Centerline Measurements

 

'~ 33% Pb, 10 1+IN. SLABS, MODIFIED BY REMOVAL

OF . 2ND: SLAB. FROM' SOURCE

33% Pb; 11 SLABS, MODIFIED BY

|JREMOVAL OF 8TH' AND 10TH SLABS

 

100 1b B,0,

 

200 15 B,0, | 400 Ib B0 400 1b B0, 600 1b B,0,
4 0.1% B 0».,2% B 0.4% B 0-4% B 0.6% B
80 4952 3710 2915
90 2739 2063 1664
100 1523 1133 906 861
110 879 631 511 463
120 525 384 307 342 301
130 313 236 183

 

 

 

 

 

 

26
K

Gamma Measurements

TABLE ‘3

R2%l and 1 for Pb, H,0 and B Composite Shields of 1l-in. Pb Slabs

 

v

N

 

 

| uu?rgng;;:xngzag 67% H,0 UNIFORM 175 Pb - 835 H,0 UNIFORM 33% Pb- - 675 H,0 NOK- UNIFORM
‘RADIUS| I+ Rl x 1074 RADIUS 1 R%1 x 107% RADIUS 1 R%1 x 10™* | RapiUs 1 R2} x 1074
¢in.) | (em) (cnd) (in.) (cm) (cnd) (in.) (cm) (end) (in.) (em) (end)
25 neg. 25 3.61 1.46 25 40.35 16.27 '25.31 13.8 5.73
28 30.88 2.47 1.52 28 18.39 9.30 26.6**
‘31 36.76 '2.66 2.32 31 17.04 10.57 27.9 10.10 5.05
34 6.09 4.54 42 .64 - 34 12.62 9.41 29 3*+*
37 48.52 2.59 3.94 37 30.8 8.77 5.23
40 6.37 6.58 54.40 2049 4.75 40 11.0 11.34 32.3%*
43 60.78 2.33 5.46 43 33.9 6.32 4.56
46 |5.18 7.07 66.16 2.35 6.63 46 6.71 9.15 |35.3%*
49 49 37.4 4.72 4.15
52 3.66 6.38 52 4.42 7.7 41.8 3.70 4.08
47.1 3.56 4.97
54.0 2.73 5.15

 

 

 

 

 

 

 

 

 

Weight Shield 59.5 Ton%

Weight Shield 66 Tons

Weight Shield 59.1 Tons

‘Weight Shield 54 Tons

 

Note:

*

Shield weights were calculated for a 118-cm

 

Second slab removed when these l’'s were measured.

** 3/16~in. Pb slabs.

 

shield, 24-in.«radius spheri-

 

i reactor, 400 mw power.
Table 4 shows in more detail the method of computation of R?], using as

an example a non-uniform shield with 0.4% B in the water.

Table 5 shows the effect of moving a single slab with all others in
nehrloptimum positions. It has been found that it is most expeditious to
adjust the outermost slab first, since the R?l for this slab is less sensitive
to the position of the other slabs. This slab is also the most important from
the point of view of both weight and attenuation. Other slabs are then adjusted
to give the same value. Table 6 shows essentially the same data as= Table 5
for a different distribution of intermediate slabs, demonstrating the relative

‘insensitivity of the optimum position for the outermost slab.
The minimum in R%21 for an intermediate slab is shown in Table 7.

Further experiments and calculations to check the shield weights quoted

here are still underway..

Fast Neutron Dosimeter. This instrument, described in previous reports,
was developed by the OBNL Health Physics and Instrument Divisions. It has
been made water-proof and used in the lid tank for a few representative read-
-ings. Comparison with fast neutron estimates from thermal flux measurements
shows the latter to have been somewhat low, so that the shield thickness
estimates have been increased. The reasons for the discrepancy are discussed

below separately.

Shield for Reactor with Fe Reflector. Two and five-eights inches of Fe
were inserted at the inside of a Pb-H,0-B mock-up in order more closel; to
approximate an engineered installation. The attenuation was not changed
appreciably, indicating that Fe is approximately as efficient for neutrons and
gammas as the Pb-H,0 it replaced. The data, including an ideal shield weight

calculation for 4-ft diameter outside the iron, are shown in Table 8.

Interpretation of Dosimeter Data (E. P. Blizard). Discussions with
H. A. Bethe, F. Friedman, and M. Deutsch served to clarify an apparent dis-
crepancy between the dosimeter determination of fast current and that inferred
from thermal neutron flux measurements (see ORNL 768). The dosimeter indicated
fast flux to be about three times higher than is indicated from a simple

interpretation of thermal flux data. The difference is attributed primarily

28
Y4

TABLE 4

Rzl for Non-Uniform. -

0.4% Boron - 9 in. Pb

Pb-H,0

 

 

 

Semwovep | Gy R’ RR? | aizoems | B/T | WnTIT | ciom @y | rrmg | FEASY
none 390
1 25 625 1.00 A71.9 1.21 .1908 2.75 | 14.41 14.41 5.81
2 27.94 780.6 1.25 458.4 1.1753 | .1615 2.73 | 16.9 21.10 8.51
3 30.88 953.6. 1.53 541.9 1.389 .3282 2.57 7.8305 | 11.98 4.83
4 33.82 1143.8 1.83 590.3 1.513 .4121 2.73 6.625 12.12 4.89
5 36.76 1351.2 2.16 593 1.520 .4125 2.53 6.133 | 13.25 5.34
6 39.70 1576.1 2.52 647 .4 1.659 .506 2.72 5.375 | 13.545 5.46
7 42.64 1818.1 2.91 675.1 1.731 .5481 2.51 4.579 | 13.32 5.37
8 48.52 2354.2 3.77 785.8 2.015 .700 2.53 3.614 | 13.63 5.49
9 54.40 12959.4 '4.73 875. 2.243 .806 2.54 3.151 | 14.90 6.01
SR2%13,562

 

SR =338

 

 

 

 

 

 

 

 

 

 

Shield Weight Calculation:

Weight Pb = (1

A reena e DU,
B T

3,562-338) x 2.346 x 10°3 = 31.02 tons of Pb
25.45 tons of~H§O

56}47 tons. - Total
- TABLE 5

Effect of Moving Single Pb Slab on R3] - All other Slabs Remaining Constant

Total 9 in. Pb - 0.6% Boron

 

 

 

. ACTUAL DATA PATA 33:%2%:;%? o po'srfmu
R ' l B2l x 107 1 R21 x 1074 | OF SLABS
(in.) (cm) (en®) (cm) (en®)
50.03 |  3.49 5.63 3.61. 5.82 25
51.40 3.36 5.73 3.30 5.63 27.9
52.40 3.07 5.44 3.12 5.52 30.9
53.40 2,97 5.46 2.94 5.41 33.8
54.40 2.95 5.63 2.79 5.32 36.8
55.40 2.77 5.48 2.68 5.31 39,7
56.40 2.59 5.32 2.60 5.34 42.6
57.40 2.46 5.23 2.51 5.34 48.5
58.40 2.49 5.48 2.45 5.40 54.4%
59.40 2.44 5.56 2.42 5.51
60.40 2.37 '5.58 2.38 5.61
64.40 2.08 5.59

 

 

 

 

 

 

'* Slab moved in this experiment,

30
Effect of Moving Pb slabudn R’l - All Other Slabs Remaining Constant

‘TABLE 6

Total 9 in. Pb - 0.6% Boron

 

" DATA FROM: SWOOTH CURVE: OF

POSITION OF SLABS

 

 

 

 

 

 

 

 

. ACTUAL.  DATA " COUNTS VS«RADIUS
(13.‘) - (éi) Rzl&.%?-‘ | (i-) R?:c:‘x)b" ; éifi ) ktsrll‘-gi
46.15 5.63 7.74 5.93 8.16 25.00 1
48.15 4.41 6.60 4.11 6.15 27.94 2
50.15 3.38 5.48 3.36 5.44 30.88 3
52.15 3.20 5.61 2.95 5.17 33.82 4
54.15 2.64 4.99 2.73 5.16 36.76 5
56.15 2.58 5.24 2.59 5.27 39.70 6
58.15 2.54 5.54 2.54 5.54 43.75 7
45.25 8
54.40° 9

 

¢ Slab moveds

31
Effect of Moving Pb Slab on R?[ - All Other Slabs Remaining Constant

TABLE 7

Total 8 inn pb - 016% Boron

 

DATA FROM SMOOTH CURVE OF

POSITION OF pb SLABS

 

 

. . ACTUAL. DATA COUNTS VS, RADIUS
R 1 Ml x 107 1 R31 x 10 | gaprvs SLAB
(in.) (em) (end) (em) (end) (in.) NUMBER
34.87 4,93 3.88 5.01 3.94 25.00 1
36.00 4.82 4.02 4.85 4.05 27.94 2
37.25 4.66 4.17 4.46 4.00 30.88 3
38.25 3,98 3.75 4.06 3.83 33.82 4
39.25 3.88 3.86 3,87 3.86 36.76% 5
40,25 3.85 4.02 3.85 4.02 43.75 6
41.25 3.99 4,37 3.94 4,31 45.25 1
42,00 4,03 4.59 4.05 4.61 54.40 8

 

 

 

 

 

 

 

® :Slab moved.,

32
TABLE 8

Gamnma Centerline Measurements

9-5/8 in. Fe Thermal Shield and 7-3/16 in. Pb.- Total
Thickness Neutron Shield 142 cm

Total Weight Shield Including Fe Reflector =61.0 Tons*

 

 

 

 

 

 

 

 

 

z MATERIAL ‘R (in.) GAMMA CENTERLINE MEASUREMENTS: - GM TUBE
(In. from Source to (48 in. Sphere Re-) .
to'Back of 8lab) flector Inside) (cms) (¢/m) R/hr x 10
1 7/8 in. Fe 22 135 '502.1 '5.52
2 7/8 in. Fe 23 120 971 10.68
3 7/8 in. Fe 24 110 1551 17.06
5.06 1 in. Pb 26.06
7.00 1 in. Pb 28.00
9.31 1 in. Pb 30.31
11.75 1 in. Pb 32.75
15.13 1 in. Pb 36.13
14.13 1 in. Pb 40.13
21.25 3/16 in. Pb 42.25
24.4 1 in. Pb 45.4 .
(1) We. H0 142 _ | .
inches H,0 = 351 = 55,9 in. + 24 = 80 1in.

(2) We. 1 in. Pb-wt. H,0 displaced = 8142 x 2.346 x 1073

wt. H20 =

3

47 (803 - 243)
2000

(5.12 x 105 - .1382 x 10%) 7.56 x 10°°

(3) Wt. 3/16 .in. Pb=(41.24)2 x 3/16 x 2.346 x 1072

(4) Wt. Fe

1701 x 3/16 x 2.346 x 1073

7 4.188 (243 - 213) 7.86 x 62.4

 

— X
8

2000

.520 x 10-% (4,563)

33

1728

.503 x 10-3 (13,824 - 9,261)

*Total

-

it

,0361 = (80% - 243) 7.559x10°% =

38.6 Tonsifizo

-19.22 Tons Pb

.75 Tons Pb

2.37 Tons Fe

——————————————————

61.0 short tons
to two factors:

(1) The dosimeter is non-directional, hence weights all fast neutrons
equally regardless. of direction, whereas the thermal flux indi-
.cates fast current, weighting individual neutrons with the
.cosine of their angle to the outward mormal. This discrepancy
is hard to define quantitatively, but Wick’s method will be
applied to obtain an order of magnitude correction.

(2) Although there is a dominant high energy which determines the
local relaxation length, there is considerable extra lower
energy neutron component for which the age to thermal is con-
siderably smaller. This indicates that the "displacement”
distance between the fast collision and thermal absorption will
be about one-half of that predicted by the simpler treatment.

That this is so is indicated by an exact calculation of the density of
second collisions in hydrogen which has recently been carried out by Nelson,
Albert, and Mulliken of the OBRNL Mathematics Panel. (1) Their results are as
follows:

¢

(Ratio of second to first collision density in~hydrogeny

 

 

t - 1 v51.222 | v o=.1.5
10 2.97 '9.83 2.67
15 3.35 3.17 2.98
20 3.62 3.41 3.20
25 3.83 3.61 3.37

 

 

 

 

 

 

~where t is the distance from source in mean free paths and v describes the

variation of mean free path with energy of scattered neutrons as follows:

Neutron Energy Spectrometer (F. J. Muckenthaler,* K. Henry*). During
the past quarter some changes were made .on the proton recoil fast neutron

spectrometer. The glass sleeve on the Kovar seal was shortened, and the

(1) Memo, Nelson to Blizasd,. August 31, 1950, Second Collision Density in Hydrogen.
® NEPA.

34
sensitive region of the tungsten anode was then restricted by enlarging the
diameter of the anode wire up to this region. This reduced the effect of static
charges forming on the glass. Various gas mixtures have been tried but the

data remain to be analyzed.

Emphasis has been changed from the single wire counter (ORNL 711) to a
two anode coincident arrangement. This has worked very well in preliminary
tests. A Po-Be spectrum indicates a better resolution than that obtained with
the single anode. Several points of the U235 spectrum were obtained using the
thermal column water tank, and these points give a spectrum similar to that
found by B. E. Watt (LA-718).

‘Measurements of Fast Neutrons in the Lid Tank Using Sulphur as a Detector
(H. E. Hungerford). Measurements of fast neutrons have been made using sulphur
as a detector in conjunction with the 33% Pb-H,0 experiment in the lid tank.
The reaction used is a S32(n,p)P32 reaction, which has a threshold of about

1 Mev. The half-life of the beta-active product is 14.3 days.

Sulphur in the form of ammonium sulphate, (NH4)2SO4, was packed into
a thin layer in appropriate containers and exposed between the lead slabs in
the l1id tank. After exposure the active phosphorus was chemically separated
from the inert material using molybdic acid as the reagent. The yeilow-green
precipitate, ammonium phospho-molybdate, (NH4)3PO4°12M003, was washed, dried,
and mounted: on suitable cards and counted. The counters were calibrated
against the gtandard pile by using both indium foils and phosphorus: in the

form of monobasic ammonium phosphate, NH4H2PO4,1suitably mounted.

The results of the measurements appear in Table 9 and Fig. 2. Results
are reported on measurements taken out to 42 cm from the source plate. This
distance was found to be the practical limit for reliable measurements. In
‘these calculations the cross section for the S(n,p)P reaction used was 285
‘millibarns (AEC Report NP-1254) and thermal activation cross sections for P
of .23 barns and for indium of 138 barns as reported in Phys. Rev. 72, 888
(1947).

The measurements are fairly reproducible, but lack of identical conditions
infphe lid tank, uncertainty of the standardization of the counters, self-
absorption in the samples, and losses during chemical separation have all
contributed to errors in the reported results. Work on the standardizing of

the counters is still in progress. Because the indium betas are of lower energy

35
TABLE 9

Fast Flux«leasurenénts in the Lid Tank

 

"DISTANCE OF ' SAMPLE

" 'FAST FLUX

FAST FLIX

 

;gg{;ggn FROI:S?S:?E-PLATE (Indf::SStnnd) ‘Piffii?d’
Open 4.1 2.75 x 10° 1.08 x 108
Open 4.7 3.31 x 10° 1.30 x 10°
Open 4.7 2.55 x 10° 1.00 x 10°
Open 4.7 2.05 x 10° 8.07 x 10°
Open 12.0 9.48 x 10° 3.73 x 10°
Open 18.8 1.66 x 103 6.53 x 10*
Open 19.6 3.78 x 105 1.4b x 10°
Open 19.6 3.09 x 10° 1.22 x 108
Open 30.3 7.00 x 104 2.76 % 10*
Open 41.5 5.64 x 103 2.22 x 103
Open 42.0 3.78 x 10* 1.49 x 10*
Open 42.0 2.23 x 104 8.78 x 10°
Closed 4.1 2.18 x 10° 8.58 x 104
Closed 18.8 1.65 x 104 6.50 x 103
Closed 34.2 1.93 x 103 7.60 x 102

 

 

 

 

.36
FAST FLUX, nv

 

WG. 9673

 

 

  
 

 

   
 

 

  

 

 

 

 

7
oF—T1— 71 1 T 1 T T T T T 1T 3
~ FIG. 2 _
- ATTENUATION OF FAST NEUTRONS :
"\ IN LID TANK USING SULPHUR
-\ THRESHOLD DETEGCTOR —
\\
- \ —
N\
6 R
10 S 7
- N\ -
_ \ SHUTTER OPEN 7]
_ “«> INDIUM STANDARDIZATION _
— \\ -
5
10 |- < -
[ \ —
\ —_
_ \ _
_ SHUTTER OPEN, _
_ PHOSPHOROUS _
i . STANDARDIZATION ]
4
10| ]
- . 3
-
_ . _
- n
SOURCE CURVE
S.0.-S.C.
3
10 | -
- SHUTTER GCLOSED,  —
- PHOSPHOROUS ]
_ STANDARDIZATION |
I | I I | I | | I I I
0 5 10 15 20 25 30 35 40 45 50 55 60

DISTANCE FROM SOURCE PLATE, cm.
37
than those of phosphorous, the results reported using the indium standardi-
zation are probably too high. At present a known quantity of P32 obtained
from the Isotopes Division is being processed for counting on the same apparatus

to obtain an independent calibration.

LIQUID METAL DUCT TESTS IN THE THERMAL COLUMN

R. H. Lewis* C. E. Clifford
H. K. Marks M. K. Hullings

Transmission of radiation through a liquid metal filled duct will be
measured in a water tank above the thermal column on top of the X-10 pile.
- This experiment is designed primarily to determine the activation of sodium
in a secondary coolant cycle for the case in which the liquid metal duct
constitutes the primary neutron path from reactor to heat exchanger. Since
this is of immediate interest to the KAPL submarine reactor project, that
organization will supply the duct mock-up. Measurements may be made with the
liguid metal duct empty as well as full, to check simple duct theory. In
addition, the effect of varying the wall material from boron carbide to water

will be investigated. Figures 3 and 4 show the duct and foil holders.

The source box and source box locator have been received from KAPL and
are being installed in the tank. There are 24 X-10 slugs in the box. This
secondary source is estimated to give approximately 107 fast flux. Background
measurements of the thermal and epi-cadmium flux are being taken at various
positions in the tank. A pile monitor for the thermal flux entering the source
box has been installed and is being calibrated. The tube in the monitor is a
boron lined ion chamber. Plans for a platform to be built around the water
tank have been completed and construction will begin September 18. The

first leg of the duct is expected to arrive from KAPL by September 30.
SHIELD CALCULATIONS

Calculations of Neutron Attenuation (F. H. Murray). Calculations for
neutron attenuation in a shield composed of lead and water were made by the
method outlined in ORNL 748. This method is based entirely on cross-section

information, without the use of lid tank results. The calculated curves for
* NIPLL

38
 

 

 

 

 
 

 

 

 
17% lead in water werefound to agree quite well with the Hurwitz-corrected 1lid
tank data.

The calculations also showed the curves for 17%, 20%, 30%, and 40% lead

in water to be nearly coincident.

Analysis of Lid Tank Data (S. Podgor*). As a first step in the analysis
of existing data on lead-water and iron-water systems, effective absorption
cross sections for lead and iron are being deduced from the experimental
attenuation curves. A preliminary value of 3.8 barns has been obtained for
lead, but further geometric corrections will probably reduce this value some-

what.

Heat Generation in Shields (T. Welton). Some simple calculations have
been made of the heat sources in the reflector and shield of a typical reactor.
The general conclusion reached is that a steel reflector, of three inches
thickness and cooled by reactor coolant, can be followed almost directly by
‘thermal insulation and a hydrogeneous layer without dissipating more than one
percent of the reactor power in the hydrogeneous layer. Detailed considerations
are presented in a memorandum, ANP-5015, entitled Distribution of Heat Sources
in Reflector and Shield.

Shield calculations (K. Keyes). The methods of shield computation devised
during the summer by members of the TAB are being checked and extended with
the aim of placing at the disposal of design people a flexible and reliable
recipe for calculating shield weights. It is hoped that such computations can

also give effective guidance in some phases of the lid tank program.

Large Air Ducts in Shields (W. K. Ergen*). A memorandum (CF 50-8-1) was
written while the TAB was in session. It was intended to summarize the
presently available information on the shielding for the air cycle. The main
difficulty is the decrease in shield efficiency due to the enormous air ducts.
Shielding considerations cannot eliminate the air cycle because there are
numerous ways (some of which were listed) to circumvent the shielding diffi-
culties, but each of these ways would introduce undesirable features into the

air cycle and make it less competitive. -

Three theoretical approaches to the duct problem are available: the
"distributed density approach," the "line of sight approach," and the following
of neutron histories including those which involve scattering. All these

approaches yield essentially lower limits for the shield fieights, but the¢se

SNEPA.

41
limits are different for these approaches and increase in the order given
above, The first two approaches have been used-at NEPA and gave widely different
results, mainly because of different engineering assumptions, and only second-

arily because of the difference in estimation of the influence of the ducts.,

No practical way of gamma shielding the air cycle reactor has been pro-
posed and it appears that the shield weights would be prohibitive, even if
only partial gemma shielding at the air cyclé reactor. were attempted.

The necessary large air planes of the air cycle allow the use of r2
attenuation, but this aggravates the problem of neutron-induced activity in

the material surrounding the reactor.

NEW SHIELD TEST FACILITY

‘W. M, Breazeale and J, L. Meem

The work on the building and pool is well underway. The contractor now
expects to be out of the building by Cctober 20. laboratory employees will
then move in and install the reactor, measuring equipment, etc. Instrument

development for this facility is a joint effort between NEPA and ORNL.

A schedule (CF 50-9-16) has been drawn up covering the installation and
initial "shake down" period totaling three and a half months. After this a
mock-up of the first shield will be tested. Design and construction of this
mock-up will be started when the ANP Shielding Board finishes the preliminary
design of an aircraft shield (October 16).

‘The neutron camera discussed in the last quarterly report has been built
and tested with U?3% fission spectrum. The results are in accord with the
Los Alamos experience. A coincidence circuit has bedn added to the proton
recoil counter(z) which substantially reduced the background. The flux re-

quired for satisfactory operation is still large, about 10* nv.

A large NalI-TlI crystal has been received by P. R. Bell and is being
incorporated in a scintillation spectrometer which he will test. It is hoped

that the resolution and sensitivity will be sufficient so that similar equipment

42
can be used in shielding work. A Compton recoil coincidence spectrometer has
been built and bench tested. Inherently its resolution is not as good as the

crystal spectrometer,

43
CRITICAL EXPERIMENTS

A. D. Callihan, Physics Division
J. F. Coneybear, NEPA

Construction of the Oak Ridge Area Critical Mass Laboratory, Building
9213, isnow essentially complete. The building was turned over to the operat-
ing staff on August 21, 1950, and experimental equipment is now being 1in-
stalled. Operations will commence sometime in October unless the procurement
of essential experimental materials, such as graphite, beryllium, and uranium,

causes a delay.

Each assembly of fissionable and moderating material will be made in a
supporting grid of square aluminum tubing. A sample section of sixteen tubes
welded together has been received but preliminary tests showed that excessive
warping was introduced in the welding. It appears necessary that the matrix
be built with individual tubes held in place by an external frame. The re-
quired aluminum tubing has been received. Control and safety rods have been

ordered and are scheduled for delivery early in October.

It has been agreed that among the early experiments will be an inves-
tigation of heavy metals as reactor reflectors. Specifications have been pre-
pared of Type 310 stainless steel requirements in order to completely enclose

the core by approximately eight inches of reflector.

A report, Initial Program of ANP Critical Experiments (NEPA-1522), was
issued stating details of the first experiments and some of the techniques to

be used.

44
EXPERIMENTAL ENGINEERING

H. W. Savage, ANP Division

As mentioned in the last ANP Quarterly Beport, a new group of the Labo-
ratory has been organized to do experimental engineering work for the ANP pro-
gram. It is the function of the Experimental Engineering Group to develop the
technology of large-scale handling of the coolants of the aircraft reactor,
and to make mock-ups and engineering studies of the proposed components. It
is intended that the fundamental research and initial development on all such
matters as fuel elements, materials for containing corrosive liquids, and heat
transfer properties of materials will be done by the other groups of the pro-
ject whose work is described elsewhere in this report. The Experimental
Engineering Group will take the basic information and design suggestions and

will reduce them to engineering practice for the ARE.

Present personnel with the Experimental Engineering Group number 19, of
whom four are technicians and the balance professional. Several additional
technicians are to be added in order to permit 24-hour operations when neces-
sary. During the past three months, attention has been directed principally
toward program definition, laboratory installation and personnel acquisition.

However, a few liquid metal loops have already been operated.

Because a primary requirement of the ARE is to establish the compatibility
of materials, the experimental program is being carefully coordinated with the
research programs of the Metallurgy Group, the Heat Transfer Group, the
activities of the ANP General Design Group, and the corresponding activities
of NEPA. All of the present work is deveoted to establishing the compatibility
of various metals with sodium and lithium over a range of temperatures reach-

ing somewhat above 1500°F.

‘The subjects which will be considered in the course of the development of
the project are listed below. Many of these tests will be of actual components

of the ARE.

1. Heat transfer—-specific configurations (not fundamental).

2. Hydrodynamics-—free convection, pressure and velocity distribution,
etc., header geometry, dynamic loads, etc.

3. Chemistry--corrosion, erosion, solubility.

45
4 Mechanical elements —pumps, valves, seals, joints, etc.
5 Purification-—working fluids, system cleaning.

6. General handling-—filling, draining, warming up.
7

. Structures —fuel elements, moderator, containers, shield, plumb-
ing.

8. Controls-—reactor, system controls, etc.
9. Cycle tests—small scale experiments on various cycles.

10. Fabrication techniques-—joining, assembling, maintenance and re-
pair.

11. Experimental techniques--sampling, test rig designs, etc.
12. Instrumentation-—temperature, pressure, flow, etc.
13. Effects of component failures-—leaks, fire hazard, sudden rupture.

14. HRemote handling—-opening and sealing system, replacing components,
handling fuel elements, etc.

The laboratory installation in Building 9201-3 is progressing. In ad-
dition to the necessary office space there is included an area in which a
number of small experiments can be conducted, another area in which experi-
ments involving extensive equipment can be conducted, an area for the disposal
of used materials, particularly sodium and lithium, and an area in which the
equipment can be serviced and small parts built. Tt is also expected that in
an adjacent area will be placed the lithium isotope separation equipment being

developed by the Isotope Separation Group.

In the experimental work a convection loop, Fig. 5, has been placed in the
shops for fabrication from stainless steel Type 321, and a "Figure-Eight Loop,"
shown in Fig. 6, which will include a pump and heat exchanger, is being de-
signed. Six convection harps and sodium loading and filtering equipment have
been received from the Metallurgy Group and these, together with a number of
others to be obtained shortly, will be subjected to 1000-hour continuous tests
under specifications determined by the Metallurgy Group. The materials for
the first six harps are (1) 347 stainless steel, (2) 304 stainless steel, (1)
low carbon Fe, (1) nickel--grade A, and (1) L605 alloy, and the liquid metal

involved will be sodium under the following conditions:

46
DWG. 99i2
VENT NOT CLASSIFIED

GAS *

VALVES $

SYSTEM SPARK PLUG
— —DG—H / LIQUID LEVEL

 

 

 

 

 

HEATER
_\

—— -
:
—_—— e ——)

 

 

SPARK PLUG
LIQUID LEVEL
GAS PROBE

—

PRESSURE TO

HOLD LIQUID |
l «— FLANGE TO

EM
IN SYST '] FIT TRANSFER
I TANK FOR

Iy FILLING
L

 

 

 

 

 

 

FIGURE 5 GCONVEGTION LOOP

a7
DWG. 9913
VENT NOT CLASSIFIED

GAS VALVES 2 SPARK PLUG LIQUID
LEVEL PROBES, T C.
WELL, SAMPLING PORT

 

 

 

 

 

 

SYSTEM
FLANGE TO FIT
PRESSURE /TRANSFER TANK
FOR FILLING
™
XCOLD SECTION
VALVE FOR VENTING Al PUMP SECTION

 
 
  

SYSTEM TO DROP ALL 1__[ H
LIQUID TO SUMP TANK

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

wl
>
-t
e §
z !
3 —-Dq—"l'j-_ o
e o Z
m — —
I GAS 5 3 METALLURGICAL INSERTS
c . a
% PRESSURE| S = T
o TO own || HEATER SECTION
- HOLD QO :
o LlQuiD )
IN o
£ SYSTEM || <3z
S 2> .«
J aw©o
g O J- Qo
n ( '
»
\ |
1
? ‘X u
@ DRAIN ZSUMP
o FI1G.6

FIGURE "8" LOOP

48
MATERIAL TEMPERATURE

(°F)
1 - 347 1350
1 - 304 1350
1 - L605 1600
Low carbon Fe 1200
Nickel 1500

In order to ascertain the power requirements for this type of equipment
it was decided to run one harp (Fig. 7) using sodium-potassium eutectic alloy,
first uninsulated and then insulated. The results of these tests indicate
that a temperature of approximately 900°F was readily obtained without in-
sulation and approximately 1700°F with two types of standard rock wool in-
sulation. A temperature difference of 180°F was observed in both types of

systems.

An electromagnetic pump is being fabricated for use with these loops and
the first of these is expected to be tested during the week of September 11.
A mechanical pump development program is also underway. In this connection a
centrifugal pump having vanes on the back of impeller has been designed and is

being fabricated for tests.

Consideration is also being given to the following mock-ups of proposed
AKE components, and fabrication of certain of these will begin in the next

quarter. These are:

. Sample fuel element bundle.
. Control mechanism.

. Hydraulic mock-up.

1
2
3
4. Shield mock-up.
5. Bemote handling equipment.
6

5. Fictorial mock-up.

49
0¢

 

 

(V-2 PHOTO 10190

 

 

i SS TYPE 304 HARP IN OPERATION

 
HEAT TRANSFER

R. N. Lyon, Reactor Technology Division

HEAT TRANSFER THEORY
H. F. Poppendiek

The Theoretical Heat Transfer Group of the liquid metals section has con-
centrated its efforts on the inves?igation of (1) forced convection ueat trans-
fer in the entrance region of flow systems,* and (2) forced convection heat
and momentum transfer in ducts whose cross sections are irregular or annular
in a region beyond the influence of the entrance secEion. It is believed that
if the theoretical solutions developed during the course of these investigations
are complemented by fundamental experimental information, it will be possible,
in a larger number of cases than previously, to predict forced convection heat
trans fer in nuclear reactors and their auxiliary heat exchangers. Special
emphasis is being given to the specific liquid metal flow systems that are at

present being considered for the ANP and ARE reactors.

Forced Convection Entrance Solutions. One phase of the forced convection
entrance studies consists of a compilation of all pertinent laminar and turbu-
lent flow heat transfer analyses that have appeared in the literature. These
analyses pertain to flow between flat plates and in pipes. Although a series
of laminar flow solutions for specific boundary conditions are available, only

a few turbulent flow solutions were found in the literature.

Another phase of the forced convection entrance studies consists of the
derivation of heat transfer solutions for systems that have not yet been in-
vestigated theoretically. A laminar flow solution has been derived for a
system in which (1) a fluid is flowing uniformly and unidirectionally (slug
flow) between two infinite parallel plates, (2) the initial fluid temperature
is uniform, and (3) the fluid suddenly flows over surfaces which are emitting
a uniform heat flux. A turbulent flow solution has been developed for a system
in which (1) a fluid is flowing uniformly and unidirectionally between two in-
finite parallel plates, (2) the initial fluid temperature is uniform, (3) the

fluid suddenly flows over surfaces which are at some new uniform temperature,

 

#The region in which thermal and hydrodynamic boundary layers are being initiated.

‘51
and (4) the eddy diffusivity is postulated to be a linear function of the dis-
tance from the duct wall (the eddy diffusivity becoming the molecular diffusi-
vity at the duct wall). Another turbulent flow solution that has been derived
pertains to a system in which (1) avelocity profile characterized by a laminar
sublayer and a turbulent core flow over an infinite flat plate, (2) the laminar
sublayer is postulated to have a uniform velocity, (3) the turbulent core 1is
postulated to have a uniform velocity (magnitude different than the laminar
sublayer velocity), (4) the eddy diffusivity in the turbulent core is postu-
lated to be uniform, (5) the initial fluid temperature is uniform, and (6) the
fluid suddenly flows over a surface which is emitting a uniform heat flux.
Several additional turbulent flow solutions are in the process of being com-
pleted; these entrance solutions pertain to systems with somewhat more complex

boundary conditions than those described above.

Another phase of the forced convection entrance studies consists of the
Jetermination of numerical solutions of specific flow systems. The system
that is currently being investigated is one in which (1) a fluid flows from a
large header into a pipe, thus initiating a hydrodynamic boundary layer, (2)
the fluid is initially at some uniform temperature, (3) the pipe wall is
emitting a uniform heat flux, and (4) the thermal and hydrodynamic boundary
layers are characterized by turbulent flow. These numerical solutions can be
used to study the limitations which are associated with some of the idealized
theoretical solutions as well as to analyze complex heat transfer-fluid flow

systems.

Heat and Momentum Transfer in Ducts Whose Cross Sections Are Irregular or
Annular. O(One phase of the studies of heat and momentum transfer in ducts whose
cross sections are irregular or annular consists of the compilation of all
pertinent laminar flow solutions that appear in the literature. Some of the
irregular cross sections that are being tabulated are the (1) square, (2)
rectangle, (3) equilateral triangle, (4) right isosceles triangle, (5) ellipse,

and (6) the rector of a circle.

Another phase of these studies consists of the determination of laminar
and turbulent flow solutions which are now not available. A turbulent flow
solution for established flow in a rectangular duct has been obtained with the
cooperation of the Mathematics Panel. This solution was based on the postulate

that the eddy diffusivity is proportional to the fluid velocity itself. A

52
study of the laminar velocity distribution has been initiated for the case
where a fluid flows between an infinite number of solid cylinders of uniform
diameter and uniform spacing; the fluid flow i1s parallel to the axes of the

cylinders.

A third phase of these studies pertains to the investigation of the few
sets of basic experimental velocity data for turbulent flow in irregular cross
sections that are available 1n the literature for the purpose of attempting to
generalize them much in the same manner as has been done for the case of flow

in a pipe.

EXPERIMENTAL HEAT TRANSFER EQUIPMENT

C. P. Coughlen

During the past quarter the major components of the equipment for measur-
ing molten metal heat transfer coefficients were assembled, equipped with
Calrods and the system pickled with a strong hot mixture of nitric, hydrochloric
and hydrofluoric acids. The equipment is designed to permit experimental de-
termination of heat transfer coefficients under a wide variety of conditions
including various shapes of conduit and various lengths of tube with both

sodium and lithium.
Specifically, the following was accomplished:

1. A new, smaller catch tank was built and calibrated so that at
500°F volume will be known within * 2, 5%, and probably within
t 1%.

9. The lithium melting tank was fabricated, equipped with heaters
and a small filter, lagged, and pickled. It is now ready for

charging.
3. Large filters for purifying the circulating charge have been
fabricated. It was found necessary to tack weld the micro-

metallic material to a supporting ring and then to seal around
the filter by metal spraying.

4. An automatic control to maintain a positive pressure of inert
gas in the system was constructed and tested. It requires no
electrical power for operation and appears to be extremely re-

liable.

.53
5. The inert gas cleaning system was constructed. This includes a
NaK scrubber for removing oxygen and water vapor, and a molten
lithium scrubber for removing nitrogen.

6. The flow measurement sections with orifices and gas pressure
transmitters were constructed, calibrated with water and 1in-
stalled.

7. Panel board construction in the shop was about 90% completed.

8. Lagging of the system with thermal insulation was about 70%
completed.

9. The cooling water system was constructed and gave satisfactory
results in tests.

10. - The liquid level indicators were successfully tested with Nak.
These indicators are designed to start and stop an electric
timer as the catch tank starts and completes filling. The
liquid metal provides a low resistance shorted secondary for a
trans former core welded into the piping. The primary coil 1is
external to the piping. The increase in primary current at a
constant impressed voltage is several times greater than that
required to operate switching devices.

11. Water circulation tests were made. Flow measurement sections
operated successfully. It was during these tests that a smaller
catch tank was found to be desirable and also that it was found
necessary to relocate the pump so the full capacity of the sump
could be used.

12. Pure magnesium oxide was found to be effective in smothering
lithium fires. Pyrene Gl also was found to be fairly effective.
A serious difficulty with the magnesium oxide 1is that,because 1t
is an excellent thermal insulator and has a low specific heat,
it prevents rapid cooling of the melten lithium. For this
reason, the smothered fire may reignite a considerable time
after the first fire was extinguished, if the loose smothering
coat opens at any poilnt..

13. A pure magnesium oxide plaster with water was found to harden
into an adherent coating for pipe insulation and other more
reactive construction materials. It will be tested with burn-
ing lithium to determine its protective ability.

Work required before the equipment can be filled and data obtained 1is as

follows:

1. Top for the sump must be fabricated.

9. Instrument lines must be 1nstalled.

54
Electrical lines for the heaters must be installed.
Fabrication of the first test heat exchanger must be completed.
The instrument panel must be completed and installed.

The remaining thermal insulation must be applied.

-~ N U B W

Protection for the concrete floor from the hot lithium must be
arranged. This protection will probably take the form of a
thick layer of sand or some other insulating powder, which the
lithium will not penetrate deeply.

BOILING LIQUID METALS

W. S. Farmer

Equipment  for the experimental determination of heat transfer to boiling
liquid metals is now being designed. Because of the high temperature and
moderate pressures at which heat can be conveyed in gaseous metals, and because
of the large heat of vaporization of many metals, it is believed that this
heat transmission method may be useful in solving some of the serious heat
trans fer problems in the design and development of nuclear reactors, particu-
larly where useful power is to be extracted. It is not expected that boiling
liquid metals will be used in the initial aircraft reactor, but they are being

investigated as a possible long-range alternative.

The current design embodies a small tank, partly filled with the boiling
liquid. Heat will be supplied to the liquid py a submerged one-inch tube
through which hot molten sodium is pumped, or which is heated by an enclosed
high frequency coil. Condensation will take place in the upper portion of the
tank on a tube which will be cooled by liquid sodium or some other appropriate
fluid. A baffle between the boiling and condensing sections of the tank will

reduce the transport of entrained liquid into the condensing section.

A heat balance between the heating end cooling fluids will enable the
heat fluxes tobe calculated. The rate of condensate flow may also be measured.
An attempt may be made to install thermocouples in the heat transfer tubes,
butif this fails, the thermal resistance between the boiling liquid and the
heating tube surface and the resistance between the condensing liquid and
the cooling tube surface can be determined by subtracfiing the resistances of

the tube and the sodium from the overall thermal resistance.

55
Tests will be performed over a range of pressures and temperature drops.
The effect of varied submergence of the heating tube will also be examined.
In initial tests, potassium will be used as the boiling metal, but in later

tests mercury may be used as well.

HIGH HEAT TRANSFER COEFFICIENTS

W. B. Harrison

Preliminary design has been made of a system for obtaining high co-
efficients of heat transfer to molten sodium. Characteristics of the design
which are expected to influence favorably the magnitude of the coefficients

are.

1. Small diameter flow channel in the test section.
2. High velocity stream.

3. Short heat length.

The main element of the test section is a circular copper disc, three
inches in diameter and 1/8 in. thick. Around the periphery of the disc is
soldered a cooling tube which will be maintained at almost constant temperature
by a stream of water. The sodium stream flows at about 80 fps through a 1/32-
in. diameter hole in the center of the disc. Thermocouples will be located at
different radial positions on the disc and the sides will be insulated so as
to insure radial heat flow. After steady state is attained, the measured
radial temperature gradient in the disc will be used to compute the heat flux
and temperature at the heat transfer surface. Sodium temperature will be

measured above and below the test section.

Barring unforeseen difficulties, it 1s expected that coefficients of the

order of 200,000 Btu/hr ft? °F (or better) will be attained.

PUMP DEVELOPMENT FOR EXPERIMENTAL SYSTEMS
A. R. Frithsen®*

Pump development for experimental heat transfer systems has been concen-

trated on a completely sealed rotary type. It is believed that this type shows

SUSAF
56
 

promise of considerably greater flexibility for experimental work than do other
types. Hydraulic bearings have been tested withasealed rotor motor in water,
and it is now planned to build a sealed 10-gpm, 200-ft head pump with a turbine
type impeller and a 3-hp motor.

Figure 8 is a photograph of a hydraulic bearing which has been cut along
its longitudinal axis. This bearing is a modified type‘of journal bearing
which derives its load carrying capacity from high pressure fluid introduced
in small orifices into longitudinal slots facing the shaft. The shaft is held

away from the surface of the bearing by the escaping fluid.

 

e T Py

ad -F e =1

FIG. 8

Cross Section of Hydraulic Bearing

57

 
Figure 9 is a photograph of an assembled unmounted pump with the sealed
rotor motor and with bearings at either end. This pump was tested in water
without its impeller, and the shaft was found to rotate for several minutes
after current to the mator was stopped, indicating exceptionally low friction
in the bearings despite the use of a stainless steel shaft in stainless steel
bearings. As was expected, the liquid around the motor armature caused rather
large windage losses. Tests on the pumping characteristics of this test model
with water gave results which were predicted by design calculation:. Since
the pressure produced was rather low, no attempt was made to operate the bear-
ings from the pump outlet. Additional studies of the hydraulic bearings will
be made to determine the best compromise between simple construction and
efficient operation. :

A hydraulic bearing pump for liquid metals at temperatures above 1000°F
with a capacity of 10 gpm at 200 ft has been designed, and construction will
start shortly. Since this high a head cannot be obtained efficiently at such
a low flow rate using a centrifugal pump, a turbine impeller pump will be
used. A 3-hp motor is already being fitted with a sealed rotor for use with

this pump.

PHYSICAL PROPERTY DETERMINATIONS
A. R. Frithsen®*

During the past quarter equipment was desiéned, and construction started
on equipment for the determination of the thermal conductivity of metals at
temperatures up to 1800°F and for the determination of the specific heats of
both liquids and solids in the same temperature range; the design of equipment
for measuring the thermal diffusivity of high temperature liquids was begun;
and preliminary arrangements were made with the Office of Air Research for
assistance in setting up a program outside the Laboratory for measuring the
physical properties of some of the materials of interest to the ANP. Thermal
property information is urgently needed on a wide variety of materials for the
design of the ANP reactor. The current dearth of information is due in part
to the high temperatures involved, and in part to the unusual materials being

considered.

*USAF
‘58
 

 

Y12 PHOTO 6 1445 |
NOT GLASSIFIED

 

 

R INLET
5 FROM
EARING

FOR
PUMP HOUSING———,  "ThRUST B

  
   
   
 
 
 

 

HYDRAULIG BEARING

, I i - 3
© (L P

- i ‘
= S ] —
B | -
S | I "
™ . -. | | l :
- | I
SRS e | . l
= - R 5 =
Ty - e 9 r
2k Yo e THe . e . ..
=t s Yy =E ’ | - :
; - S s
; 3
-
- o
i =
1
= 4
2

e Py o et L gta s
5P ot T';';'jfif't‘.{':‘f 0 i

| HYDRAUL\& BE
S—— T

P

CROSS OVER QUTLET

M
FLUID FRO
EI’?&UST BE ARING

  

ARING

    
    

—

[

sl

 

 

 

FIG. 9. COMPLETELY SEALED ROTARY PUMP

 

 

 
LIQUID METALS HANDBOOK
R. N. Lyon

A Liquid Metals Handbook was published by the Government Printing Office
and is now on sale by the Superintendent of Documents. The major editorial
responsibility for the handbook has rested with ORNL. Because of the coopera-
tion of other AEC and Navy installations this manual is the first complete

compilation of liquid metal date and experience.

A proposal for effecting a continuously up to date, loose leaf liquid
metals handbook was drawn up and preliminary negotiations for establishing such
a handbook were made with the Technical Information Division, NEPA, and the

Committee on the Basic Properties of Liquid Metals.

LIQUID METALS IN-PILE EXPERIMENT

0. Sisman

The mechanical components of the test loop have been built and partially
tested. Heating tests with the loop contained in stacks of graphite and con-
crete blocks indicate that heaters to reach the required 1800°F present no
particular problem, but that a cooling jacket may be required to prevent over-

heating of the graphite and the concrete shield in the X-10 reactor.

Tests on the G. E. electromagnetic pump with lithium have begun, and
equipment has been set up to calibrate the electromagnetic flowmeter with both

lithium and sodium.

The first experiment planned for the in-pile equipment will be a study of
Brehmsstrahlung from the B rays produced by lithium. This experiment is ex-

pected to require only a short time, once the equipment is in place.

When the lithium studies have been completed, it 1is planned to study

radiation damage and corrosion with the system refilled with sodium.

.60
METALLURGY AND MATERIALS

E. C. Miller, Metallurgy Division

The principal efforts of the Metallurgy Division in the ANP Program are
being directed toward a solution of the more immediate problems of the ARE.
The corrosion testing work 1s now taking the form of an investigati .n of the
resistance of materials to sodium at temperatures ranging from about 1200°F to
1800°F. However, since this redirection of effort has only occurred recently,
relatively little information oncorrosion in sodium is available in reportable
form, and the work reported herein deals primarily with the earlier work with

lithium and, to a lesser extent, sodium hydroxide and lead.

There is, fortunately, some evidence available from earlier work at other
sites that the austenitic stainless steels, 304, 316, 310, and 347, possess
good resistance to sodium and NaK up to the boiling points of the liquids.
Because of their good mechanical properties at elevated temperatures, a special
effort is being made to investigate types 310 and 316 stainless steels to
determine their suitability for reactor construction from other standpoints in
addition to the one of corrosion resistance. These include creep-rupture (both
in and out of liquid metals environments), welding and fabrication into plumbing

systems, and feasibility of incorporation into suitable fuel elements.

STATIC CORROSION TESTING

This phase of the liquid metal corrosion testing program is designed to
find suitable high temperature materials of construction for use in contact
with various coolants. Sorting tests on lithium, bismuth, and lead have
continued into this quarter, and these are being extended tocover sodium, Nak,
and sodium hydroxide in the 600° to 1000°C temperature range. A more extensive
study of the corrosion behavior of the stainless steels in molten lithium is

in progress.

Stainless Steels in Lithium. Stainless steel alloys were corrosion tested
in 1000°C (1832°F) lithium for exposures of 4, 40, and 400 hours. The alloys
tested included the ferritic chromium steels, 405, 430, and 446, and the
austenitic grades 304, 310, 316, and 347.

61 -
Tests were conducted in the usual manner: sealing the test specimen
with a controlled amount of lithium in an evacuated Armco iron capsule, heating
to the desired temperature, and holding at temperature for the specified time."
The ratio of test specimen surface to lithium volume varied between 0.5 and

0.7. All tests were run in triplicate.

The 4, 40, and 400 hour results are tabulated inTable 10a. Weight change
(mg/cm?) data are reported as averages of three specimens tested. Pen. cration,
as measured during metallographic examination, is reported on the photomicro-

graphs. Analysis of the bath metal after test was made spectrographically.

To aid in understanding the nature and depth of attack and to determine
the degree of selective leaching of constituents in the bath metal, an exten-

sive metallographic examination of the 40-hour test specimens is in progress.

All specimens were cut diagonally and mounted with protection bars at the
exposed surface to reduce edge rounding during polishing. Thickness measure-
ments were taken at the center area to compare with similar measurements made
on the specimens prior to test. Maximum corrosion depths reported on the
photomicrographs refer to penetration measured beyond the exposed surface from
one side of the specimen.v Six fields were measured to obtain a representative

value.
The following photomicrographs were prepared for study:

1. Etched transverse section of the specimen, showing surface con-
dition and structure prior to testing.

2. Unetched transverse section of corroded specimen showing con-
dition of metal at the exposed surface. Examination of the
specimen prior to etching was considered necessary because
etching might destroy films or other corrosion effects at the
interface.

3. Transverse section of corroded specimen etched by alkaline
potassium ferricyanide to bring out the carbide in the structure.

4. Transverse section of corroded specimen etched with oxalic acid
to resolve the grain structure.’

The photomicrographs for each of the seven stainless alloys tested are shown

in Figs. 10 through 16.
62
TABLE 10a

corrosion Results of Stainless Steels in 1000°C Lithium

Specimen Size 1 in. X ¥ in. X. ) in.

 

 

 

 

 

 

 

 

NUMBER TEST AVERAGE WEIGHT BATH ANALYSIS
STAINLESS SAMPLES DURAT ION CHANGE N AFTER TEST REMARKS
STEELS TESTED (Hours) mg/cm (ppm)
304 3 4 -1.39
3 40 -5.62 Fe 20-90 Moderate resistance; 40 hour micro shows inter-
Ni 90-250 crystalline attack to a depth of two mils and
Cr N.D. some evidence of a phase transformation near
3 400 -16.06 the attacked grain boundary.
310 3 4 -4.08 Fe 20-70
Ni 140-250
Cr N.D. Fair resistance; 40 hour micro shows heavy inter-
3 40 -30-75 Fe 25-54 crystalline attack to a depth of one mil.
Ni 310-450
Cr N.D.
3 400 -64.76
316 3 4 -1.56 -
3 40 -6.70 Fe 23-71 Moderate resistance; 40 hour .micro shows inter-
Ni 91-270 crystalline type attack to a depth of four mils
Cr N.D.: and some evidence of a possible phase transforma-
3 400 ~-25.27 tion near attacked grain boundary."
347 3 4 -1.97 - -
3 40 -10.49 Fe 25-60 Moderate resistance; micro shows intercrystalline
Ni 170-200 attack to a depth of three mils.
Cr and Cb N.D.
3 400 -27.24
405 3 4 -1.09
3 40 -3.44 Fe 73-150 Good resistance; micro showed practically no
Cr N.D. intercrystalline atta x«.
3 400 -8.36
430 3 4 -1.26 Good corrosion resistance; micro showed practical-
3 40 -4.13 ly no evidence of intercrystalline attack.
3 400 -9.97
446 3 4 -3.24 ~
3 40 -9.79 Fe 50-150 Moderate resistance; micro showed decarburization
Cr N.D.. and grain growth at exposed surface.
3 400 -44.51 '

 

 

 

 

 
 

 

   
  

 

 

 

" AFTER EXPOSURE 250X UNETCHED

~

250X ETCH OXALIC ACID

 

SPECIMEN THICKNESS BEFORE TEST---0.572 cm.
SPECIMEN THICKNESS AFTER TEST---0.569 cm.

 

 

 

 

 

 

 

 

0.020

|

o
4 CENTER OF SPECIMEN MAX. CORRQSION DEPTH=-=========-=== (0 O|Bcm.
PER CENT COMPOSITION
E' msPECTMHOGRnF;it u_h Aj G_ -H:nC‘I.HJHQi FUSION—
3 Mn 0 -
54.4 z%a 9| 0I5 - - - Ol4 005 002 <000
SINGLE SAMPLE
0.020cm |25
LOCATION
OF
EXAMINATION
254¢
/7‘
1 200cm
o
063cm

TYPICAL TEST SPECIMEN

TYPE. 20

NOT CLASSIFIED
PHOTO NO. Y-2609

SURFACE

 

 

 

  

- -

AFTER EXPOSURE

250X ETCH: ALKALINE
POTASSIUM
FERRICYANIDE

SURFACE

 

 

 

STAINLESS STEEL EXPOSED TO MOLTEN LITHIUM

1000 °C FOR FORTY HOURS

Fig. 10

R.J. Gray

 
 

SURFACE SURFACE nNoT CLASSIFIED
PHOTO NO- Y-2537

  
   

   

—

BEFORE EXPOSURE 250X ETCH GLYCERIA REGIA AFTER EXPOSURE 250X UNETCHED AFTER EXPOSURE 250X ETCH GLYCERIA REGIA

 

 

 

 

 

 

 

 

CENTER OF SPECIMEN

 

 

a
- "—‘ n o B | SPECIMEN THICKNESS BEFORE TEST---0.584 cm.
_. 3 \ \ | SPECIMEN THICKNESS AFTER TEST---0.583 cm.
1 = o - | MAX. CORROSION DEPTH====--- ~0.0I6 ¢cm.
x % - ) = 53" 0.005¢cm
- . - ' PER CENT COMPOSITION
| - e e, 7 o e -SPECTROGRAPHIC ——— ~VACUUM  FUSION-—
0.020em| . N "ol ")~ | F Cr N M Mo Nb A C N2 Q2 H2
- b e 690 68 ILT 02 2.32 = - 0046 00IB 00I5 <000l
b S N ™ | sinGLE sawepe
l - ! o - “"‘"fi

|

L : v i LOCATION
{. . e - | OF
e—— —  — e = ATl
AFTER EXPOSURE 250x ETCH GLYCERIA REGIA EXAMINATION

--or"""

 

 

| ot el B\
~1 < ,fl,n»**'?)J

F i YCERIA REGIA
TYPICAL TEST SPECIMEN AFTER EXPOSURE 1000 X ETCH GLYCE

TYPE 3l

STAINLESS STEEL EXPOSED TO MOLTEN LITHIUM
I000°C FOR FORTY HOURS R 6oy

FIG. 1l

 

 
SURFACE

 

S SR e DT

\y =N NS LG o SRR
e = AN Y [N e .‘?’*-:-a&:»..[

 

 

BEFORE EXPOSURE 250X ETCH OXALIC ACID

 

 

 

 

0.020c¢m. |

__. .q_ 3 | l

 

 

 

 

wn
=

RFACE NOT CLASSIFIED
PHOTO NO. Y-2610

   

 

AFTER EXPOSURE 250X UNETCHED

SPECIMEN THICKNESS BEFORE TEST---0.585cm,

 

 

SURFACE

 

 

SPECIMEN THICKNESS AFTER TEST---0.584cm.
> CENTER OF SPECIMEN MAX. CORROSION DEPTH-==-====-=-=== 0,013 ¢m.
g T LG | PER CENT COMPOSITION
AN \ J | m——————SPECTROGRAPHIC———— ~VACUUM FUSION—
{ J'=-._\\ ~“ 'Fe Cr Ni M™Mn Mo Nb Al C N2 Ha
SN \ 722 I77 976 035 - - - 0078 QOI9 004 {000l b
X uI,’ \ SINGLE SAMPLE 0.005¢cm.
| [ - . - .
0.020¢m L /‘;{: ' (_/-“' e LOCATION
' [ TRy s / ' o
| 'L,. - | J ":\J# o :‘l"m . EthlNATlDN
Y | L j ;
-1“__: I {T -
SRR : ol |
3 AN 2O ) 7| 200cm "
| AFTER EXPOSURE 250 ETCH OXALIC ACID / AFTER EXPOSURE
| -
. 063cm

STAINLESS

TYPICAL TEST SPECIMEN

TYPE 304

STEEL EXPOSED TO MOLTEN LITHIUM
000 °C FOR FORTY HOURS

Fig. 12

1000 X ETCH OXALIC ACID

R.J. Gray
 

 

 

 

 

 

 

SURFACE SURFACE
. = -
. -
"““ :-; '-- : "'-. - ’
vy ' e S
N L)
- = ‘- i -2
BEFORE EXPOSURE AFTER EXPOSURE 500X UNETCHED

500X ETCH OXALIC ACID

SPECIMEN THICKNESS BEFORE TEST---0.585¢cm.
SPECIMEN THICKNESS AFTER TEST---0.584cm.

 

 

»
-
MAX CORROSION DEPTH-=----===-=====< 0.0llem,.
CENTER OF SPECIMEN
— PER CENT COMPOSITION
[ PSR T A TN PR B L BERAS sl ~— —  SPECTROGRAPHIC————— ~VACUUM  FUSION—
Ly T'-‘.- s ..-‘. ; X .‘.'13'-".‘-_- . et i!r:s‘__ Fe Cr Ni Mn Mo Nb Al c N2 _
-‘E_.' <’A;—‘a". ’t‘_' i md p ‘-‘—;i"_ll M 1Y R
ST s R T A -5 T O E‘; = "l e92 190 102 08 - 0B84 - 0077 0023 0003 000!
By sapis 2 M (AL R X RTITEITR| omale sawne
-u.,"'i BN, rogeti- 58 S BUEC . ot J.-_-',-"- .
"'_'.. _:,:l‘.-;’:'l' o ‘_ g e b =20 e
TS BN G o, - PR Pk o § R e S e
' ST IS e e 3 e T N NI
0.010cm | B i e T At e tn N T LT s iy AR s
o™ LRSS A IR e R e
' :

 

 

500X ETCH OXALIC ACID

AFTER EXPOSURE

 

TYPICAL TEST SPECIMEN

TYPE 347

SURFACE NOT CLASSIFIED
PHOTO NO. Y-2611

 
 
 

5
- p ‘ - — 3
- ; -'.' i B I"\ r--.. %
° .\\:fi ' 9‘\ i s "N
0.010cm) ey =% g9 Q0 R g -4 0 ag
T S Rl N gt
=~ F A o‘_ ‘ ..fi“ ..'?
> T Tyl TR S
i aall B T

 

 

 

AFTER EXPOSURE 500X ETCH: ALKALINE
POT ASSIUM

FERRICYANIDE

SURFACE

    
  

0.010¢c

|

  

 

 

% 3
F ¥ & ’ . bl
W g e "'::a‘ Lok P L
3 . e = A ol TS I L -
o - v, v e el LY

AFTER EXPOSURE 500X ETCH OXALIC ACID

 

 

STAINLESS STEEL EXPOSED TO MOLTEN LITHIUM

1000 °C FOR FORTY HOURS

Fig. I3

R.J. Gray

 
¥ &
SURFACE
r \
0.005¢m < .

- - ]
| NS i
= 2

| e

o |

L

 

E‘EF(_JF.'E _r'_m.-_wr;f.,:?fs_ ICO0OX ETCH. PICRAL-HCL

 

 

g; CENTER OF SPECIMEN
| e -
\
\ —TN —
|1 e B \\" ',_’frr ‘\‘
¥ | N i ’ l
| ¥ 4
| | )
0.005cm -
|| T o
| - ‘
Y . =
! W
_\‘_- 3 b—

S

AFTER EXPOSURE  100OX ETCH. PCRAL-HCL

STAINLESS STEEL EXPOSED TO MOLTEN LITHIUM

SURFACE

0.005cm

 

 

_—

 

AFTER EXPOSURE  1000X " UNETCHED

SPECIMEN THICKNESS BEFORE TEST---0.587c¢m.
SPECIMEN THICKNESS AFTER TEST---0.586cm.
PARTIAL SOLUTION-----------0003cm be— 0.05 cm—>

PER CENT COMPOSITION =

—SPECTROGRAPHIC i ~VACUUM FUSION-—

Fe Cr Ni Mn Mo Nb Al C Nz ©O2 Ha

 

 

863 132 . : . 047 0078 0018 0003 (00Ol
SINGLE SAMPLE
LOCATION
) OF N
EXAMINATION SURFACE

AFTER EXPOSURE

 

TYPICAL TEST SPECIMEN

TYPE 405

1000 °C FOR FORTY HOURS

Fig 14

   

AFTER EXPOSURE

SURFACE NOT CLASSIFIED
PHOTO NO. Y-286I2

ETCH. PICRAL- HCL

000X

i

100X ETCH. PICRAL-HCL

R.J. Groy
69

0.0I0cm.

1

T

0.010cm.

|

 

  

e e
i ers

BEFORE EXPOSURE 500X

;.-- ‘1’-;. -I ’ .i-.--.l o .

 

ETCH: PICRAL-HCL

 

CENTER OF SPECIMEN
PR 2 < .

T e e ~ ———
- - = g ——— e - =
sy s i

= f:;’h-l. "‘l - " s b
I L Eme——— . -

 

“" R..-‘U.-F s

AFTER EXPOSURE

'lll

“{ 0.010cm.

&

 

h

AFTER EXPOSURE UNETCHED

500X

 

 

SPECIMEN THICKNESS BEFORE TEST----- 0.63l cm.
SPECIMEN THICKNESS AFTER TEST~----
PARTIAL SOLUTION----====-- -0.015¢cm.

 

PER CENT C Tl
~———————SPECTROGRAPHIC ~VACUUM  FUSION—
Fe Cr N Mn Mo Nb Al C N2 Q2
T35 255 - R - = - 0.084 0046 0.009 (000!

SINGLE SAMPLE

/ LOCATION

OF
EXAMINATION

 

500X  ETCH: PICRAL-HCL

063cm,
TYPICAL TEST SPECIMEN

TYPE 446

SURFACE

  

NOT CLASSIFIED
PHOTO NO. Y-2539

S?F ACE

 

Ak

AFTER EXPOSURE

500X ETCH: PICRAL-HCL

 

 

 

AFTER EXPOSURE 100X ETCH: PICRAL-HCL

STAINLESS STEEL EXPOSED TO MOLTEN LITHIUM

1000 °C FOR FORTY HOURS

FIG. 15

R.J. Gray
-
-

SURFACE SURFACE SURFACE NOT CLASSIFIED
PHOTO NO. Y-26I3

 

l S~ : :
0.005¢m : | wn B g |
~ = _ - ireem, - | 0.00 | o il L
" . 0.005¢cm | i Pt Mfi _

 

 

ETCH. PICRAL-HCL

 

AFTER EXPOSURE 10O0D0X

 

FTER EXPOSURE 1000 X UNETCHED

RE EXPOSURE 1000 X ETCH PICRAL-HCL

o bt

SPECIMEN THICKNESS BEFORE TEST---0.577cm

 

 

 

 

 

 

. 82 i ! SPECIMEN THICKNESS AFTER TEST---0.575cm.
CEl i OF S | N 0. —
o ENTER OF SPECIME PARTIAL SOLUTION === ===-- 0008em, ,. S —
PER CENT COMPOSITION
— -SPECTROGRAPHIC ~VACUUM FUSION— |
Fe Cr N Mn Mo NQ A,l C N2 02 H2 e

850 149 = - = = - 0080 0020 0.006 <0.00I Jriled s |

TINGLE SAMPLE <o Ve

/ DR

" B 5
-1 |

) e

LOCATION o [

OF /‘" H P |,

EXAMINATION SURFACE YR ’f;“ -

l; s;- 5 P‘ ?r
T . SNEY ... Wi . 254 L 37 A ST .M{.Hh .,J
AFTER Egh'rbgg FC'OON ETCH. PICRAL-HCL J> 2:;1{ AFTER EXPOSURE 100X ETCH. PICRAL-HCL

i .
o
0.63cm

TYPICAL TEST SPECIMEN

TYPE 430

STAINLESS STEEL EXPOSED TO MOLTEN LITHUM
1000 °C  FOR FORTY HOURS R, Groy

Fig. 16
Although this investigation is incomplete, the ferritic iron-chromium
alloys appear to be somewhat more resistant to 1000°C lithium than the austenitic
alloys on short exposures; the available evidence, however, is not sufficient
to permit recommendation of the ferritic alloys as container materials. The
austenitic grades, 304, 310, 316, and 347, were all attacked intergranularly
while the ferritic alloys showed little evidence of this type attack. The
iron-chromium alloy, 446, showed a decarburized area accompanied ;v marked
grain coarsening to a depth of 30 mils. There is evidence that a part of the
structure immediately adjacent to the attacked grain boundaries in the 304,
316, and 347 grades has undergone a phase change, i.e., upon slow cooling a
rim of ferrite surrounds the austenite grains, presumably as a result of re-
moval of an austenite-former, perhaps carbon, by diffusion and aolution in the
lithium. Magnetic tests of these areas with a colloidal solution of iron
verified the phase transformation from austenite to ferrite, Further metallo-
graphic work is in progress to check the presence of small quantities of sigma
phase in the 310 stainless steel test specimens.

In an effort to learn something about the rate of attack of stainless
steel by high temperature lithium, tests were performed at different exposure
times (4, 40, and 400 hours). Micrographs showing the effect of time on the
attack of each of the seven stainless alloys tested are presented in Figs. 17
through 21.

Special Tests. A few tests were made to determine the effect of pre-
saturating the coolant with graphite to inhibit decarburization and inter-

granular corrosion attack of the stainless alloys.

Beneficial results were noted by addition of graphite to the coolant for
40-hour exposure test of stainless alloy grades of 310 and 316 under static
conditions. Decarburization was minimized and the depth of intergranular
attack was reduced by at least a third.  Comparison of micrographs for specimens

tested with and without the graphite addition are shown in Fig. 22.-

Pure Metals in Lithium. During the past quarter, 40 and 400 hour tests
were run on some of the metallic elements and high temperature alloys in
1000°C (1832°F) lithium.  Results indicated, as did the four-hour tests com-
pleted last quarter, that iron, zirconium, and columbium show good resistance

to lithium at 1000°C. It should be noted, however, that both the zirconium

71
el

 

 

250 X ETCH: OXALIC ACID

{ FOUR HOURS|

SURFACE
-

 

 

 

 

 

l ! A
3 - g ] . b . o
0.0200m | fFr— £ kL )SA
i 1 f,' ., 1- 5 s L 52 ke o |
R Ve A
¥ p— S ', ”'r . " {;J ]
| ot . A A ¥
| e e ] r £ Y
TYPE | tin s o
3‘6 ,' - 4 ) ,T".J'n\ L L p
!_Z{ H_'.'l_ J =4 f“f i 3 -,‘\' = c-—-'j._ )
250 X

ETCH: OXALIC ACID

EFFECT OF TIME ON LITHIUM CORROSION OF
AT 1000°C

T

0.020¢cm

¥ .

 

250 X

SURFACE

o 0.020¢cm.

 

{ FORTY HOURS/{

SURFACE

 

 

 

ETCH AQUA REGIA

Fig- I7

STAINLESS STEEL

NOT CLASSIFIED
SURFACE oot NO. Y-2605

I
l

 

 

 

e ETCH: OXALIC ACID
{ FOUR HUNDRED HOURS S;Ffimt

0.020cm.

l

 

 

 

 

 

250 X ETCH. AQUA REGIA
 

    
 

SURFACE SURFACE NoT CLASSIFIED
PHOTO NO. Y-2606

 

 

 

 

 

 

" ETCH. OXALIC ACID 250 X ETCH: OXALIC ACID 250 X ETCH: OXALIC ACID

t FOUR HOURS| { FORTY HOURS{ }FOUR HUNDRED HOURS}

 

 

 

 

 

 

 

 

0.020cm.| -4 | L
L Y T LT A T L L
) 2 e, o, frid LT -
s Lo e l AR NGRS el -
".. . ‘(' Fite = ;,‘I"‘ -f o l ."."' F
§ = fr e _’dm _,_f-_'}: e
* J"’:.'H:'-" y, NSNS
17 Sl ™
~V - e B
. i~ = { -)\n, -
ETCH: OXALIC ACID 250 X ETCH: OXALIC ACID 250 X ETCH. AQUA REGIA

EFFECT OF TIME ON LITHIUM CORROSION OF STAINLESS STEEL
AT 1000°C

Fig. 18
0.005cm

LYPE
405

 

SURFACE

-

0.005cm.

 

 

000X

ETCH. PICRAL-HCL

t FOUR HOURS/

 

SURFACE

 

0.005cm.

 

 

I000X

 

 

ETCH: PICRAL-HCL

SURFACE NOT CLASSIFIED
PHOTO NO. Y-2607

 

ETCH: PICRAL-HCL

{ FORTY HOURS| $FOUR HUNDRED HOURS/

v

 

SURFACE

T

0.010cm.

|

 

 

 

 

SURFACE

 

 

 

ETCH: PICRAL-HCL 500X

SURFACE

 

 

ETCH: PICRAL-HCL

 

ETCH: PICRAL-HCL

EFFECT OF TIME ON LITHIUM CORROSION OF STAINLESS STEEL

AT 1000°GC

Fig- 19
NOT CLASSIFIED
PHOTO NO. Y-2608

 

 

 

 

 

 

 

 

 

SVIJE-'MZE SURFACE SURFACE
T lllllllllllllllllllllll . = .‘.,a:..
i | 0.005¢m v o
0.005cm | . 1-
Y
TYPE |
430 |
000X . ETCH: PICRAL-HCL 000X ETCH: PICRAL-HCL 000X ETCH: PICRAL-HCL
FORTY HOURS FOUR HUNDRED HOURS

FOUR HOURS

EFFECT OF TIME ON LITHIUM CORROSION OF STAINLESS STEEL
AT 1000°C

Fig- 20
DWG. 9925
NOT CLASSIFIED

 

023

020

INS.

DEPTH OF AJTACK
o

005

ITEIITITTII

..b',
o

 

TIME(HRS 4 40 400 4 40400 4 40400 4 40 400 4 40 400 4 40 400

 

 

 

 

 

 

 

 

 

ALL MEASUREMENTS BY METALLOGRAPHIC EXAMINATION —:
et ora]
NT INTERGRANULAR -
y I ATTACK .
[T e B PARTIAL SOLUTION Tom
St e = TotaL RemovaL 060
4840 HR.- ONE SPECIMEN g
7 400 HR.- TWO SPECIMENS 4
.oasi"oa "
2z
040
1.015
030
100
020

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TYPE b— 310 —f b— 316 —i +— 304 —{ +— 347 —| l=— 405 —| — 446 —| — 430 —

EFFECT OF TIME ON LITHIUM CORROSION OF STAINLESS STEEL
AT 1000°C RJG.

FIG. 2i

76
NOT CLASSIFIED
TYPE 310 TYPE 3l6 Y-2262

   
 

 

 
 

 

 

 

 

 

 

L

 

 

0.020
0.020
cm oM
F : Beg A LITHIUM
J& - - )
{\u— &5 *,f“,w*‘\‘“ X uuv, J. i -+ < e
o : =z .- - -
'E?. -3(‘3;3'{’\“&";34 : ‘{"? 3 ‘_: g 3 ' S o X o
fl1 '....T' *,f"‘k o "“-r I: -._‘ ‘: 'u e 1 -
X — - x - \ »
R i B b e P L R S 7 \ _
PR EE B E P s e e . _ .
e R ,u-rs...a o> -&‘ ey ',"Jl;a‘- 1 » - \ - ® I
R -l'rfigw i % “‘*.za o *f“’ i "' P e T A .
;b:fl_.-*f =) A ) . ° _' 1‘ e L - \' i - -
3 ~ iy e A4 i e L - o i Fand k 3
1M A .“3"-“,';1-_ oLy e SR g .
:?:m:‘;fi;:;—: : -:r- '1‘:‘_ "".‘-l‘ P-— L‘_ | !"' "—k b‘.“ 1—: & ;“H:-J'r:f-: !.El::-:;“ ‘:L: 1;;{' 250 x < -
ETCH AQUA REGIA
ETCH: OXALIC ACID
0.020
cm

LITHIUM
WITH
GRAPHITE

 

 

 

 

 

 

 

 

EFFECT OF GRAPHITE ADDITION ON LITHIUM CORROSION OF STAINLESS STEEL
I000°C FOR FORTY HOURS
EXPOSED SURFACE AT TOP OF PHOTOMICROGRAPH

FIG. 22

R.J. Gray
and columbium specimens gained slightly in weight due to film fofmation.; Such
films may be acting as protective coatings to prevent further corrosion. Pure
nickel, one of the major ingredients in stainless steel alloys, was severely
attacked. The uranium test which was made in a tantalum liner showed essen-

tially no solubility in 1000°C lithium,

In the case of tungsten, molybdenum, and cobalt, a mass-transfer of metal
from the test specimen to the iron capsule was noted. A similar, but more
striking mass-transfer of metal was observed in a four hour silicon test in
1000°C lithium. Such a corrosion mechanism probably precludes the use of
these elements in an environment of metallic iron in liquid metal piping
systems. Because of interest in molybdenum as a structural reactor material,
it 1s planned to determine the corrosion behavior of this metal in the absence

of iron.-

The 40- and 400- hour test results are tabulated along~with the four-hour

test results for reference in Table 10b.:

Refractories in Lithium. Beryllium oxide and several of the commercial
refractories that might find possible application in reactor construction were
corrosion tested in 1000°C lithium. Four-hour exposure tests were run on
beryllium oxide, aluminum oxide, chromite, cordierite, silicon carbide, quartz,
boron carbide, alundum, zirconium oxide, magnesium oxide, clay-graphite, and
graphite. As expected, all disintegrated except beryllium oxide which showed
a moderate weight loss. A similar 100-hour exposure test showed the hot-
pressed beryllium oxide grade to be more resistant to 1000°C lithium than the

refractory grade.-

Pure Metals in Lead. Preliminary screening tests on resistances to lead
[at 1000°C (1832°F) for 40-hr exposure] of some of the metallic elements
indicated: moderately good resistance of tungsten, zirconium, iron and

tantalum; fair resistance of molybdenum, columbium, and beryllium; and poor
resistance of titanium and nickel. Detailed results will be reported upon
completion of X-ray and metallographic examination of the test specimens now

in progress.’

Materials in Sodium Hydroxide. Several pure metals, stainless steel
alloys, and refractories were corrosion tested in 1000°C (1832°F) sodium

hydroxide with the hope of finding potential materials for use in reactor core

78
6L

TABLE 10b

Corrosion Data on Pure Metals in 1000°C Lithium

Specimen Size 1 in. X ¥ in. X % in.

Surface-Volume Ratio 0.6-1.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NUMBER TEST AVERAGE WEIGHT EQUIVALENT PENETRAT ION BATH ANALYSIS
METAL SAMPLES DURAT ION CHANGE (mils) AFTER TEST REMARKS
TESTED (Hours) (mg/cm”) (ppm)
Extruded 3 4 -17.19 ~-3.68 Severely attacked; micro shows scattered non-
Beryllium 3 40 -67.82 -14.67 Fe 140-930 continuous film and large voids. X ray iden-
Be 530-9600 tified only Be.

3 450 -100. 82 -21.78

Silicon 3 4 Completely lost All three samples apparently dissolved and
deposited on capsule wall.

3 4 +3.81 +0.33 - s ,

Titanium 3 40 -1.46 -0.13 Fe 40-310 Continuous adherent film formed identified
, Ti 43-760 by X ray as TiN.

2 400 +3.45 +0. 30

Vanad ium 1 4 _ -7.68 -0.50 Fe 37,000 Continuous nop-uniform film formed. X ray
‘ V N.D. showed VN, V,N, V, and Fe,.-
Chromium 1 4 -5.72 -0.31 Fe 8200 X ray showed only Cr.
CR N.D.
Manganese 3 4 Only thin electrolytic strip tested which
apparently dissolved.

Armco 3 4 -0.12 -0.01 - -
Iron 3 40 -0.95 -0.05 Fe 53-1100 Good resistance; micro shows no noticeable

3 400 -0.55 -0.03 attack.
Sintered 2 4 -5.00 -0.22
Cobalt 3 40 -23.35 -1.03 Fe 210-31,000 Non-uniform etching attack. X ray showed a

3 450 -120.71 -5.36 Co 150-3100 Fe.
TAM 3 4 -86.41 -3.82 Severe intercrystalline type attack; X ray
Nickel 3 400 -580-97 -25.79 showed only Ni.:

 
TABLE 10b (cCont’d)

 

 

 

 

 

 

08

 

 

 

 

 

NUMBER TEST AVERAGE WEIGHT EQUIVALENT PENETRATION BATH ANALYSIS
METAL SAMPLES | DURATION CHANGE (mils) AFTER TEST REMARKS
TESTED (Hours) (mg/cm™) (ppm)
Bureau Mines 5 4 +0.25 +0.02 Good resistance; all samples developed brassy
Zirconium 4 40 +1.09 +0.07 Fe 72-1300 tarnish identified by X ray as ZrN.
Zr ND-32
3 400 +3.04 +0.18
Sintered 3 4 -0.84 -0.04 Good resistance; micro showed continuous film
Columbium 2 40 +1.06 +0.05 Fe 2, 600, not yet identified.
Cb N.D.
Sintered 3 4 +0.23 +0.01 Good resistance on short term exposure; longer
Molybdenum 3 40 -18.02 -0.69 Fe 91-1000 exposures showed Mo deposited on Fe capsule
Mo N.D.- wall.
3 400 -26.91 -1.04
Sintered 3 4 -1.29 -0.04 Good resistance; four—hour specimen showed
Tantalum 3 40 -2.90 -0.07 Fe ND-1900, continuous thin film identified as TaC.
. 3 400 -7.04 -0.17 3300-8700 Ta
Sintered 3 4 -2.63 -0.05 Fe 900-1300 Excellent resistanpce on short exposure;
| W-N.D.- tungsten deposited on capsule wall on longer
Tungsten 3 40 -32.69 -0.67 exposures.
3 400 -175.36 -3.58
Uranium 1 4
1040 Steel 3 4 -2.08 Micro showed decarburization throughout entire
450 -7.73 % in. specimen.’ '
Haynes Ceramel 3 4 -120.81 Drastic attack; all samples exfoliated."
LT-1 ‘
Haynes Alloy 3 4 -6.78 Micro showed areas ‘cached out immediately
L605 below exposed surface.
Allegheny 3 40 -44.02 Ni 680-1100 Poor resistance.:
Alloy V-36 Fe 470-1000
Cb 260-1400
Cr 59-270
Haynes Alloy 3 450 -85.88

N-155

 

 

 

 

 

 

 

N.D.: Analysed for but not detected.:
construction. Forty-hour tests were run in the usual manner, with nickel
substituted for Armco iron as the capsule container. ‘With the exception of
the container nickel, all the metallic materials tested were drastically
attacked.- Caustic attack on the stainless alloys was characterized by an

exfoliation type attack.

The high fired grade of BeO appeared most resistant of the materials

tested. Graphite showed moderate resistance on this short exposure.

In spite of the precautions exercised in canning, numerous capsule leaks
were encountered during the heating cycle. These leaks are believed to be
partly due to excessive internal pressure developed during decomposition of
the contaminants sodium carbonate and water. The combined amount of water and
sodium carbonate present in the commercial grade of hydroxide is approximately

one percent by weight.

Materials in Sodium. Because of probable use of sodium as the coolant in
the core circuit of the proposed ARE, major activity in the liquid metal
corrosion program has been shifted to this coolant.. Tests are in progress to
investigate the resistance of various engineering materials to corrosion attack
by sodium liquid and vapor at high temperatures (up to 1000°C). Where commer-
cially available, tubing 1s used as the test capsule; otherwise, metals like
chromium, cobalt, silicon, tungsten, manganese, and vanadium, in a more readily
available form, are being contained in a relatively inert capsule during

testing.

Initial tests indicate that Armco iron will not serve as a satisfactory
capsule material at temperatures above the boiling point because of leakage of
sodium through the capsule bottom.- Metallographic examination of the capsules
after test showed the leaks apparently followed the path of stringers in the
iron. Similar tests showed that nickel can be used satisfactorily to contain
sodium at 1000°C."

Although this testing is in the preliminary stage, it appears from visual
inspection that most of the materials tested, particularly the austenitic
stainless steels, are more resistant to 1000°C sodium than to high temperature
lithium. X-ray and metallographic examination of test specimens is in progress
to obtain more exact information concerning the resistance of specific materi-

als.

81
DYNAMIC CORROSION TESTING

To avoid overlapping of effort it has been agreed that the ANP Experi-
mental Engineering Group will carry out all engineering phases of the dynamic
corrosion testing program previously planned by the Metallurgy Division’s ANP
Materials Group. The meta]lurgical contrpl and examination of all these
systems will remain the respopsibility of the Materials Group. This includes
the thermal convectjon loops (harps) and the forced circulation, or electro-

magnetic pump, corrosion test loops.

Therm#l convection Loops. The following harps have been fabricated in
the ORNL Central Shops:
- Type 316 stainless steel
- Type 316 stainless steel - extra heavy wall
Type 347 stainless steel
- Type 304 stainless steel

- Low carbon 1iron

N O~ NN D
$

- Nickel, Grade "A™
The following are being fabricated by Philadelphia Pipe Bending Company:

- Shrouded harp - V-36 alloy - low carbon steel liner

- Shrouded harp - Type 310 stainless - low carbon liner
- V-36 alloy harp

LL-605 alloy harp

. Type 316 ELC stainless steel harp

DY b b e e
¢

- Type 446 stainless harps

These will be delivered to the Experimental Engineering Group primarily
for testing in sodium when completed. Some of the harps operating in lithium
and sodium will be contained in pressure vessels which will be used as inert
gas-containing dry boxes. Three such vessels have been received from Central
Copper Works, Inc.: These dry boxes are now being filled with electrical and
thermocouple connections and vacuum and helium gas connections. The component

parts for the type 316 stainless steel lithium purifier and the type 347

82
stainless steel sodium purifier have been finished by the shops and the
purifiers are being assembled. The type 316 stainless steel purifier will be
turned over to the Experimental Engineering Group. At present it is tentatively
planned to use sodium in both purifiers.. A gas purifier, inwhich helium will
be bubbled through a packed column filled with molten lithium, is being made by
the shops. At present it is planned to make a lithium purifier similar to the
ones already completed, but using a ferritic iron-chromium stainless steel

wherever the lithium will come in contact with the metal.:

It is planned to heat the loops by passing large currents at low voltages
through the hot leg of the loop. For the first stage of the work two large low
voltage transformers now in the laboratory will be used and more suitable
transformers are being procured. A type 310 stainless steel loop filled with
lead-bismuth eutectic was heated satisfactorily in this manner. A hot leg
temperature of 1900°F was obtained and a temperature drop of 800°F was indicated
by thermocouple readings.  Due to the large temperature drop obtained on this
preliminary run and because only about a 300°F temperature drop appears to be
required, work has been stopped ona liquid metal cooling system for the harps.
The harps will be run in the dry boxes without any insulation other than radia-

tion shields on the hot leg.:

CREEP-RUPTURE AND STRESS-CORROSION TESTS

Creep-rupture tests will be run, both in and out of liquid metal environ-
ments, to obtain engineering design data, as well as to determine the effects

of liquid metals on the corrosion of stressed metals.

Two machines have been received for this work and are now being set up and

tested. Four more are on order and should be received in October.:

Negotiations are in progress to have a portion of the work performed in

available equipment at other laboratories.:

ASSEMBLY OF FUEL ELEMENTS

Essential equipment required for fuel element development'work is on order
and should be received shortly.
83
Compatibility tests to obtain information concerning the chemical stabil-
ity of binary systems of potential fuel element materials are in progress.:
Materials being studied include: molybdenum, chromium, stainless steel,
uranium oxide, uranium-beryllium compound, beryllium, beryllium oxide, and

possibly others.:

Some of this information is available from work done at other sites and

will be taken into account in carrying out the program.

In addition to the straight chemical compatibility and diffusion studies
of binary systems, more complex laminated structures more nearly representative
of the combinations to be expected in fuel elements will be fabricated and
tested for thermal and metallurgical stability and structure integrity. This
phase of the work will be carried out in our own laboratory and some phases
will be subcontracted to activities specializing in certain types of fabrica-
tion. The development of fuel elements for the ARE will represent a major

activity of the ANP Materials Group during the next few quarters.

84
RADIATION DAMAGE

D. S. Billington, Metallurgy Division

The ANP Radiation Damage Group has engaged the cooperation of the Cyclo-
tron Group at Y-12, which, it is felt, will contribute a great deal of help in
Radiation Damage problems. It is planned to begin a series of experiments to
first study the feasibility of doing damage studies inside the dees of the
86 in. cyclotron; secondly, to work out fundamental correlations on property
changes in metals; thirdly, set up "in-cyclotron™ experiments that will make
a study of the feasibility of certain fuel element assemblies proposed by the

Design Group.

Y-12 CYCLOTRON EXPERIMENTS

The first experiments to be conducted at Y-12 are on copper and 316
stainless steel. Copper was selected since it represents a good metal and
should lead to more fundamental results for use in correlating reactor damage
with cyclotron irradiation than the 316 stainless steel which was selected as
a result of its probable use in the ARE reactor. In addition to the use of
electrical resistivity as a criterion of damage before and after measurements

will also be made on the hardness and X-ray diffraction patterns.’

At the present, suitable samples have been designed to enable these
measurements to be made and their fabrication is in progress. It is planned
that these samples will be placed in a water-cooled probe with an aluminum
window and bombafded with 20 Mev protons in the Y-12 cyclotron for the accel-

erator studies. The Hanford Reactors will be used for the reactor studies.

The North American Special Research Group has done a great deal of work
on the correlation of radiation damage induced by accelerators with the,damagé
caused by reactors. They have shown by a study of the change in electrical
resistivity in graphite that an accelerator will produce the same type of
damage as a Hanford type reactor but at a rate on the average 250 times as
rapid. This is an experimental correlation and is consistent with the best

known theory.
85
It should be pointed out, however, that this correlation applies only to
graphite and only to the change in electrical resistivity. Furthermore, this
correlation considers only the damage induced by displaced atoms. Also, this
correlation is not based on "in-pile™ or "din-cyclotron' measurements but is
based on observations made after samples have been removed from the accelerator
or reactor, i.e., displacement effects are the only ones that normally would

be observed.

There is another type of radiation damage resulting from thermal spikes
that normally would be observable only during operation of the reactor or
accelerator. This phenomenon, which has not been studied‘very extensively
either theoretically or experimentally, might be expected to affect such

properties as creep and diffusion.

The point to be made as a result of the above considerations is that there
are two general effects to be studied: (1) total exposure-—displacement
damage, and (2) flux—thermal spike damage. Most of the work up to the present

has concerned itself with the first type.

NORTH AMERICAN AVIATION, INC.: LITHIUM-IRON ACCELERATOR CORROSION EXPERIMENT

There are nodata available from this experiment as yet, due to fabrication
difficulties, but it is expected that a successful cyclotron run will be com-

pleted in the near future.

It appears desirable to have the North American group study the corrosion
of stainless steel 316 by sodium at operating temperatures proposed for the

ANP program.

PURDUE UNIVERSITY

The Purdue group has been doing some preliminary studies preparatory to
studying radiation effects in molybdenum and various stainless steels. They
will make before and after measurements of hardness, electrical resistivity,

crystal structure.

A bombardment of molybdenum for two hours with a 0.3 pah beam of 10 Mev

protons resulted in no noticeable change in hardness.

86
IN-PILE CREEP

The in-pile creep apparatus described in the last ANP quarterly report
(ORNL 768) proved operable in bench tests. A test on 304 stainless steel
stressed to 21,000 pounds per square inch at 1200°F was started in the ORNL
reactor in late August. Data from this test are still being taken, but no
reliable results are expected from this first test because of temperature
control difficulties. Unfortunately, a last minute shift in hole assignments
caused the test to be placed in a stringer hole adjoining a cross core hole
through which cooling air passes at a very high velocity.  During pile shut-
down and start-up periods the temperature of the apparatus varies too fast for
the D.A.T. temperature control unit torespond. A modified apparatus embodying
a supplemental temperature control system to maintain constant the ambient
temperature of the furnace will be ready for insertion in the ORNL reactor

within two weeks.

Modification of furnace design to permit test temperatures of 1500 to
1800°F are underway. The feasibility of using a creep specimen in the form of
a cantilever beam with a loading weight within the apparatus itself is being

investigated.

The new Solid State Building is nearing completion and occupancy ought to

begin by November 1, 1950.

A new series of samples for before and after experiments is being prepared
for irradiation at Hanford. It is hoped these samples can be irradiated at
around 400°C. While this temperature is below contemplated operating tem-

peratures, it should give some preliminary data on annealing characteristics.

87
NUCLEAR MEASUREMENTS

‘MECHANICAL VELOCITY SELECTOR

A H. Snell, Physics Division

A chopper time-of-flight velocity selector for operation in the neutron
energy region up to several thousand electron volts is in process of design
and construction. The 80-channel counting equipment is being constructed.

The chopper assembly design drawings are almost ready to be turned over to the
shops.

XENON CROSS-SECTION MEASUREMENTS

A. H. Snell, Physics Division

The possibility of extending the thermal xenon cross-section vs. energy
measurements of Bernstein et al. to higher neutron energies is now under active

consideration. Two possible methods are being studied:

1. Epi-cadmium deactivation at high flux. This experiment is the
same as earlier thermal measurements of Eliot et al. and Sugarman
et al. except that the sample would be encased in cadmium. With
available flux it is not possible to detect deactivation if the
Xe resonance integral is less than 10° to 10° barns.

2. Danger coefficients in intermediate critical assemblies with
kilocurie quantities of Xel3%. It is possible with kilocurie
quantities of Xe!?® to detect a change in reactivity of an inter-
mediate assembly if the resonance integral is 10* to 10° barns.
The possibility of extracting such large amounts of Xel3% from
the MTR mock-up is_being studied in the Chehistry Division,

MOLYBDENUM CROSS—SECTION MEASUREMENTS

Columbia University

The electronic time-of-flight neutron velocity selector at . Columbia
University, under the direction of Professor W. W. Havens, is assisting the

ANP program with a measurement of the total cross section of molybdenum in

88
the intermediate range. Special samples of molybdenum blocks have been
furnished toColumbia, and the measurements are now in progress. The molybdenum
cross section is of great interest for ANP reactors since this material not
only has exceptional high temperature strength but also a relatively high
resistance to corrosion by liquid metals. If molybdenum should be found to
have a reasonably low intermediate absorption cross section; it might be
possible to employ it in the aircraft reactor even though this metal does

present great difficulties in fabrication.

89
REACTOR PRYSICS

The ANP Physics group during the past quarter has undertaken detailed
calculations of the neutron physics of the proposed ARE reactor. These include
‘exploratory calculations by bare reactor theory as well as a large amount of
multi-group calculations, including reflector effects. Since the reactor
designs now contemplated are at least partially intermediate in neutron
spectrum, the work has been forced in the direction of 13-group calculations
'similar to those carried out at KAPL. The following sections describe the
classes of calculations now underway. The first numerical results in most of
these programs should be forthcoming early in the next quarter. In addition

to this work, there is also described below an investigation by the Mathematics
Panel on the energy dependent pile equations.

ARE CALCULATIONS

- N. M. Smith, Reactor Technology Division

Multi-Group Calculations. Following the general method given in GE-RE-1
and KAPL-39, a thirteen—group numerical calculation has .been set up for a
spherical prototype reflected reactor and experience has been gained for all

persons involved.

Several exploratory calculations have been undertaken. These have re-
vealed the necessity of careful monitoring of all work in addition to the
regular numerical checks. In. the first place, some of the internal consistency
checks are insensitive, and in the second place, a large amount of work (i.e.
calculation of reactor constants and certain constant numerical factors used
~in the multi-group calculation) is done for which there exists no adequate

check other than independént calculation by another person.

Thus, through the proverbial hard way, we have learned to duplicate all
work which cannot be checked by the ordinary laws of mathematical consistency,

conservation of neutrons, etc.

A single 13-group calculation requires from 8-12 computer days, depending

on the number of = iterations required.

90
Comparisons have been made .between different reflectors with a core of
U0, fuel, BeO moderated, ha#ing stainless -steel structure and Na-cooled. The
reflectors studied were SS-347, as compared with the bare pile. Unfortunately,
errors were discovered in the reactor constants but it appeared that the
steel reflector was at least as good as the BeO reflector for thicknesses
below three inches. The errors made were such as to increase the leakage from
the core and thus enhance the effect of a reflector. Thus, the conclusion

may still be tenable.

The result with the bare core indicated that the thin reflector used
(2.5 in.) did not make sufficient change of reactivity to warrant much interest
in reflector control rods other than for power control. Again, this conclusion

must be reevaluated.

One difficulty with the GE method has been the fact that for each éroup
the ratio of the average slowing down density to that leaving the group must
be guessed. When the guess turns out to be poor new values must be used . and
the group recalculated. With the steel refleétor, in particular, the guesses
-were wild and required several tries in many groups. Two modifications of the
GE method have been suggested which will remove the necessity for this guessing.
One involves the calculation of the first moment in the lethargy scale of the
reactor cross sections and, although more accurate,‘wiil involve several times
more work per group. The other method assumes a linear change of g, the slow-
ing down density, in the group interval and has promise of eliminating the
‘guessing with but little additional work. The assumption of linearity of ¢
is implicit in the present GE method so that the latter new method should be

.at least as accurate.

IBM Calculation. Following certain suggestions of H. Hurwitz of KAPL,
the GE multi-group method has been modified for IBM calculation. Thngwork
is in the set-up stage. About two-thirds of the preparatory work--wiring of
boards, puncturing cards, etc.-~has been completed. The first calculations are
to take place soon, and after several checking tests and calculations, the

machines will be used to carry out routine 13-group calculations.

Reactor Constants. Routine methods for the averaging of reactor constants
have been devised. The macroscopic constants for the bakic reactor matérials
(fuel, moderator, structure, coolant) have been computed. All that is re-

quired then for calculation of the constants for a specific core or reflector

91
are the volume fractions of the constituents. This method has proved convenient
and time saving. It still remains necessary, however, to check by making

two independent computations of the reactor constants.

The same constants serve for either the multi-group multi-region calcu-
lations (for the same volume fractions) or for bare homogeneous reactor calcu-
lations. A set of constants requires two computer days for double evaluation.
A new set of constants obtained by changing only one of the constituents can
be obtained in about 2/3 computer days. The IBM will be made .eventually to

compute reactor constants also.

Adjoint calculations. The reactor adjoint functions, necessary for

perturbation calculations, have been computed in one case on a trial basis.

It . has been shown (Greuling) that the adjoint of the slowing down density
is the same as that of the flux. The continity and boundary conditions on
the adjoint function, including that of the thermal flux, have been worked out

and calculational procedures detailed.

Bare Reactor Calculations. It has been found expedient to make .exploratory
calculations using bare-reactor theory. Since no spatial integration is re-
quired, a numerical bare reactor calculation can be made and checked in less
than two computer hours. Thus reactivity calculations pertaining to control
considerations are quickly made. Such things as xenon reactivity, coefficients,
(both of power and of temperature) nuclear temperature coefficients of reactiv-
ity, expansion temperature coefficients, uranium mass-reactivity coefficients,
etc., are in process of calculation for the ANP Control Panel. The bare-
reactor calculations will be followed by multi-region calculations as time

permits and wherever necessary.

Multi-Group Calculations in Cylindrical Geometry. Most multi-group
calculations are made assuming the reactor is spherical. 'Detailed calculations
of lattice effects (whether a lattice of control rods or of fuel elements)

requires a calculation in cylindrical geometry.

The multi-group equations have been rewritten and reduced to difference
equations for use with cylindrical geometry. Some exploratory calculations
‘have been made. The indications are that the difference equations will be

somewhat more complicated than in the spherical case.

92
ENERGY DEPENDENT PILE EQUATIONS

‘M. L. Nelson, Mathematics Panel

The problem discussed here is that of a two-region pile with cross

sections which are strongly dependent on energy.
These equations can be written.

t

- V2 y(r,u) + Alu)y(r,u) + B(u)y(r,u)

@

= N F(u) S()g(r,v)dv. - V2 $(r,u) (1)

0

+ C(u)@(r,u) + D(u)qg(r,u) =0,

with

Y(r,u)

flux in core at point with position vector r and
at lethargy u.

¢(r,u) = flux in reflector.

A(u) 1
B(u):
F(u) L

S(u) Simple combinations of the various cross sections
' as functions of lethargy.
C(u) -

D(u)

 

 

The dot denotes differentiation with respect to u.

The boundary conditions are taken as follows:
Y, ¢ finite within their respective regions of definition

Y(I,u) = $(I,u) (a)

93
W L,w  d, (w)IF(I,u)
3 ) 2

 

dl (u) (b)

&(B,u) =0, (c)

where I is an arbitrary point on the interface between core and reflector, B
an arbitrary point on the outer boundary of the system, and the derivative in

(c) is, of course, the normal derivative.

The method of separation of variables works quite smoothly for these
equations until one tries to fit the boundary:conditions at the interface,
but it breaks down at this point. 1In particular, the bare pile equation.:can
be integrated explicitly. However, if one does this, one finds that the
solution of the above system is not unique. A closer examination of the
physical significance of the terms of the first gquation'ifiklj shows the
following. The term involving the u-derivative arises from the balance be-
tween neutrons flowing into a certain u-level from below, and those flowing
out of this level to higher values of u. The fact-fihat this term involves
only a first derivative is due to the one-way flow of neutrons in u (flow
only to.highef u, or lower energy). However, if one assumes a lowest u-value
(taken as u= 0), this term expresses only a flow out of the level, hence this
term should be absorbed into the second, or absorption, term. If one then
writes down the modified equation for u = 0, the whole system will have a
unique solution. Even with this modification, the dependence of the diffusion
coefficients d,, d, on u still offers an apparently insuperable barrier to the
application of the method of separation of variables. Hence it would seem
that some scheme of reduction to a multi-group problem is nécessary,~together
with a numerical, and presumably iterated solution of the resulting set of
equations. Various proposed methods, in particular the so-called GE method,
are being studied from the mathematical and computing standpoint.

94
DESIGN OF THE AIRCRAFT REACTOR EXPERIMENT

'FUNDAMENTAL DESIGN PRINCIPLES
C. B. Ellis, ANPDivision

During the past quarter the General Design Group of the Oak Ridge National
Laboratory has completiéd a preliminary report on the principles of the design
of the ARE. This has so far been given only local distribution, as document
No. Y-F5-15 and Supplements I to VIII. This report is written almost wholly
from the standpoint of a reactor for a supersonic aircraft. The tentative
‘conclusions listed are that the reactor core diameter should be in the 3-ft
.range, that the primary coolant should be liquid sodium, that .the fissionable
material should probably be UO,, that the moderator should probably be BeO,
that the canning material for the fuel elements should be a variety of stain-
less steel, and that the maximum fuel element surface temfiernture in contact
with the liquid metal should be not more than 1800°F, In this report it is
pointed out that the most practical type of fuel ‘element is not yet.known,
The probable merits and difficulties of eight general varieties are discussed:
loose powder pins, loose-sleeve moderator blocks, sandwich sheets, outaide-
coated coolant tubes, bonded sleeve moderator blocks, inside-coated pins,
solid-filled pins, and liquid-filled pins. No final choice of the variety
most suitable for e supersonic aircraft reactor can be made until a large
‘amount of fuel element research is finished, |

Also during this quarter, the NEPA Division of the Fairchild Engine and
Airplane Corporation has completed a preliminary study of an overall design
for a liquid-metal-cooled supersonic aircraft. This deaign is given the NEPA
distinguishing number IL-12 and is described in the preliminary report NEPA-1520.
The general appearance of the reactor shown in this report is very similiar to
‘the ORNL arrangement discussed in Y.F§-15.

During the latter part of the quarter, a decision was made to build the
ARE as a prototype for a subsonic aircraft reactor, the so-called "War MoHel,"
rather than a prototype for a supersonic power source. This change makes very
little difference in the features of the design listed above except to permit
the lowering of the desired fuel element wall temperature from 1800°F to
1500°F., Another advantage of this change, which is yet to be felt, is that
the fuel element research no longer needs to immediately result in an element
capable of ultra-high power densities.

95
Although the two reports mentioned above are essentially in agreement on
the reactor core, the NEPA design employs a divided shield whereas the ORNL
design postulates a unit shield. To develop the theoretical principles on
‘which engineering components can most advantageously be located within a unit
shield and to compare the advantages of unit and divided shieldéy a joint
ORNL-NEPA Shielding Board has been set up to begin operation September 6, 1956.
This board contains both theoretical and experimental shielding specialists
as well as mechanical engineers from ORNL and NEPA. The mission of this
Shielding Board is to produce by October 16, 1950, fully engineered designs
for both unit and divided shields in as low a grosé weight as current knowledge

will permit, together with an evaluation of the operational problems resulting
from the use of the divided shield.

Another block of fundamental design principles which has not yet been
thoroughly explored is the question of reactor controls. To investigate this
.field a similiar joint ORNL-NEPA effort, the Control Group, has been set under-
way. From the work of this joint group, there is expected to arise the first
version of a complete control system which will be adequate for the airplane
reactor and for the ARE, from both nuclear and mechanical engineering stand-

points.

Aside from the large blocks of analysis and research listed above and
described in the previous section of thisvreport,‘thé most outstanding question
0of the ARE which has not yet been attacked is probably the means to avoid .
self-welding of the moving parts of the reactor. Research on this subject
will begin shortly in the Metallurgy Division as well as will design studies
to see how,fa; the use of moving parts with contact can be eliminated from the

reactor.

DETAILED DESIGN 'AND CONSTRUCTION

'W. P. Berggren, ANP Division
R. W. Schroeder

During the past quarter an ARE group has been set up. This groupis
starting from the basis of the fundamental design principles discussed in
the preceding section.and continuing to a completed design for the ARE reactor
and all of its auxilaries and its housing in as rapid a fashion as the arrival

of research answers from the laboratory groups will permit. A preliminary

96
design of the building to house the ARE has now been drawn up. Problems of
the remote handling equipment to permit disassembly and inspection of radio-
active compbnents are being studied. Tentative detailed drawings have been
(prepared for several reactor core geometries and several possible arrangements
of equipment within a unit shield. These studies have been given to the
Experimental Engineering Group for mdck-up tests and to the Shielding Board

for theoretical analysis.,

As discussed in the preceding section, very few of the specifications of
the ARE can yet be designed with extreme firmness since the basic researth is
not yet in a finished state. However, a tentative detailed design is proceed-

ing on the assumption of the following specifications:

TABLE 11

Alrcraft Reactor Specifications

Primary coolant Sodium
Maximum fuel element surface temperature 1500°F

Average temperature of coolant emerging

from reactor core 1400°F
Active lattice shape Square cylinder, 3 ft
' diameter

Fuel element ' Pins containing U0, and BeO
powder

Fuel pin size Approximately 0.1 in. X 3 in.
with 7010 in. wall

Number of fuel pins Roughly 100,000

Alternative detailed designs are being drawn up to cover several possi-
bilities on which a decision has not yet been made. These include sodium vs.
natural lithium as a secondary’ coolant, BeO vs. stainless steel as areflector,
and a single primary pump as compared with several smaller pumps to be im-
Peddedifixthe unit shield. Figure 23 illustrates the tentative general core

arrangement now under investigation,

97
 

 

L STR LT IR END PLATE

   

N

: —ACT, vE ComE

 

 

Ve
d

 

 

   

 

S -
. ] - =~ PEFLFCTOR
: | - -
. dql | "
. -
, |
< ¢ _ S#& sHCT O~C j
N { e FCR TRIS MASNIEIATION !
< = |
< HHEC |
- o o i
[ . ' 7
N ! '
8 ’ 2

= CYLIORICAL  CONTR IR

 

36.0 —*———fiu-]io a'\__/
| J

—— T

= STRUCTURAL END PLATE

/—Mowv TING FLANGE

    

 

 

 

 

 

 

 

 

 

 

 

 

/!
LFLEXBLE DIAPKRAGH

  

COOLANT eNLET
FROM Py
@o‘m Ac/i_ADis Ve Y)
. — SURPCKT VANEST
s T R8BS 6 LvENLy SOMCED
SECTION A-A & Eveniy SPACKD

 

 

 

 

/

/L
/)

7 2

"/,

 

 

Z

4

o

= VOID FOR F S5/0n7 RPYXQD:CT
- GAS ACCUNILILATION
L~ (235 INTERNAL vOLUME)

s

  
  
 

/0

 

/ SFUFL, UG | BeO AMIXTURE

 

 

 

 

 

ZAANNN

P2
P

7

7
7

 

/4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 SN

| .
VLTS
LR N AN

3 \\:\\“\\é\o\\ N \\\\ :

 

 

Firosn C-C

L.

75—

}
i

 

—Be0 MODERATCR SESMENT

7YPE S, WD
f S.STF e CCCLA TeF

  
 
 
 
  
 

~TYPE S.STEFL SUPPIRT GRID

FUEL CCNTAINER

TYFE S.STEEL  SUPFORT o%4a
COLLAR

FI1G 23

 

ARE. TUBULA KEACTOR |
FRCPCSAL

 

SECTICN D-D

DRAWN © BREWEF 8-8-50
SCALE "
DWG. 1O. ¥y —F7-10

 

 

Pan‘c P

 
PART II, LONGER RANGE ACTIVITIES

99
'INTRODUCTION

As described in Part I, the major part of the research and design work for
the ANP program now underway at the QOak Ridge National Laboratory applies
directly to the proposed fixed-fuel:liquid-metal-cooled ARE reactor. However,
the Laboratory and its contractors are carrying on a considerable amount of
research -along alternate lines which might lead to aircraft reactors even
superior to this type, provided certain fundamental difficulties could be sur-
mounted. These studies include investigation of alternate cycles employing:
(a) vapor coolants with compressor jets, (b) circulating fuel homogeneous
reactors, (c) fixed-fuel circulating-moderator reactors, and (d) separated
lithium-7 isotopes as a primary coolant. The research and analysis underway
which attempts to solve the crucial problems of each of these potential systems

is described in the following sectionms.

100
COMPRESSOR-JET CYCLES

'C. Starr, North American Aviation, Inc.

The work accomplished by North American Aviation, Inc. during this period
included the development and application of generalized procedures relative to
‘the analysis of power cycles for possible utilization in a nuclear power air-
craft. The results of the analysis to date indicate that the compressor-jet
type power plant may be superior tothe more conventional turbojet on the basis
of 'specific thrust developed and reactor power required. However, the analysis
is not yet complete, in that not all relevant parameters have been considered
quantitatively. This is now being done in connection with the compressor-jet
cycle employing a mercury-vapor turbine to drive the air compressor. The re-
sults obtained to date are not too encouraging due to the tremendous bulk and
weight introduced by the heat exchangers and the weight of the working fluid
(mercury). However, these results are quite preliminary, and it is hoped that
further study will result in a more favorable configuration and arrangement of

power plant for this particular cycle.

Cuantitative studies similar to the mercury compressor-jet cycle are be-
ing carrled out for a compressor-jet cycle using a sodium vapor turbine as the
compressor drive, with consequent higher reactor temperature. Being of lower
priority than the mercury unit, these studies have not progressed to the point

where the results are significant.

101
CIRCULATING FUEL REACTORS

In the field of circulating fuel reactors, two efforts are underway. The
Atomic Energy Division of the H. K. Ferguson Company is continuing an analysis
of circulating fuel reactor systems in general, which will place particular
emphasis on an overall integrated design for a modified B-52 aircraft whose
homogeneous reactor operates on a suspension of uranium-bearing particles in
sodium hydroxide. On the other hand, the Chemical Research Divisions of the
Laboratory are searching experimentally for suitable high-temperature liquid
systems which will contain uranium-bearing material either in solution or sus-

pension. These two efforts are discussed in the following sectionms.

CIRCULATING FUEL BREACTOR ANALYSIS
K. Cohen, H. K. Ferguaon Co., Ato?ic Energy Diviasion

The circulating fuel reactors which are now being considered are of two

general types as follows:

1. Circulating fuel reactors in whieh the fuel has no moderating
properties. The moderator, if any, is present as a separate
liquid or solid phase.

2. Circulating fuel-moderator reactors. In this case the fuel mix-
ture has moderating properties and the interior of the reactor
has no structures other than control rods.

A number of other possibilities have been given some brief consideration,

but the two general types mentioned above appear to have the most promise.

The reason for consideration of circulating fuel reactors is the advantage
of eliminating the problemof fuel element design. This at once dispenses with
the most difficult set of problems now facing the airplane reactor designers.

Other advantages will be seen when we consider a specific design.

Of the many liquid fuel reactor designs which have been considered, the
following three possibilities have survived the initial screening. Although
the other possibilities have not been completely discarded, it is felt that a

successful airplane reactor can be developed from one of these bases.

102
A reactor is suggested with a homogeneous core consisting of a solution

or stable suspension of a uranium compound in sodium hydroxide.

heat is removed by circulating the

directly to the propulsion unit radiators.

Circulating fuel consisting of a solution of uranium in bismuth.
This system has two limitations. First the solubility of uranium
in bismuth appears to be somewhat limited except at very high
temperature. This means that either the reactor must be very
large in order to contain sufficient uranium, or the temperature
must be high enough at all points in the liquid fuel circuit to
insure the necessary solubility. In the second place, this fuel
has no moderating properties and moderation must be supplied by
structures within the reactor. Since these structures occupy a
large fraction of the reactor volume, the problem of limited
solubility of uranium in bismuth becomes even more intense. fow-
ever, the probable resistance of such a solution to radiation
damage makes this possibility interesting in spite of these
limitations,

A circulating-fuel-moderator consisting of the three compounds
UF,, NaF, BeF,. Although this system has not received very much
experimental study, there seems to be a good possibility that a
liquid mixture exists in a usable concentration range. The
chemical stability of fluorides in general is well known.

A circulating fuel-moderator consisting of a solution or stable

suspension of some uranium salt in sodium hydroxide. This sy-
stem appears to have the greatest potential advantages. Pre-
liminary work on preparation of the sodium hydroxide-uranium
mixture has given encouraging results, and experiments are now
being pushed to explore this possibility more thoroughly.

CIRCULATING FUEL REACTOR FOR SUPERSONIC PLANE

would be a sphere about 2 ft 6 in. in diameter surrounded by a 4-in. beryllium

reflector and a thermal shield to absorb leakage neutron heat. The fission

which 1s located within the shield and rather close to the core. An expansion
vessel containing helium is an integral part of the heat exchanger and provides
for expansion and contraction of the fuel under changes in temperature. The

‘heat exchanger is cooled by a stream of sodium hydroxide which is conducted

sodium hydroxide at about 1600°F to the radiators with a total heat output of
about 500 mw.

103

The core

homogeneous fluid through a heat exchanger

The reactor is designed to deliver
The concentration of uranium in the fluid will be of the order of one
percent by weight, and the total uranium hold-up in the reactor system, in-
cluding the hold-up in external lines and heat exchanger, will not exceed about
100 pounds. This liquid has physical properties rather similar to water at
atmospheric temperatures and heat transfer coefficients in the range from
5,000 to 10,000 are not difficult to obtain.

Although sodium hydroxide is a rather corrosive fluid, there is con-
siderable commercial experience in handling it up to around 1000°F. Some
satisfactory experience has been obtained at temperatures higher than this us-
ing nickel and Inconel. The development programon the circulating fuel should,
of course, include careful study to determine the proper metals for use in
construction. In view of the low vapor pressure of sodium hydroxide at the
temperatures expected in the reactor, no great problem is anticipated in the

design of the reactor container.

The rather large variation in density of the circulating fuel with changes
in temperature introduces a corresponding variation in reactivity which must
be taken up by control rods. The extent of the variable reactivity, of course,
depends on: the design range of temperature which is to be encountered in the
core. Since the temperature coefficient of reactivity is large and negative,
it exerts adesirable stabilizing effect on fluctuations in temperature. - Thus,
although the variation in reactivity either requires control of a large
fraction of reactivity or restriction to a narrow ray of temperatures in the
reactor, the control of slight fluctuations in reactivity under operating
conditions should be fairly simple. Since a fraction (approximately half) of
the delayed neutrons will be emitted in the circuit outside the core, the
amount of excess reactivity which can be present be fore reaching prompt
critical will be about .3%. This does not seem to lead to inordinate diffi-

culties in design of the control system.

A detailed report on this reactor as well as an integrated design for 1its

use in a modified B-52 airplane will be presented during the next quarter.

CHEMISTRY OF LIQUID FUEL SYSTEMS

W. R. Grimes, Materials Chemistry Division

It is recognized that a reactor with a liquid fuel which could be circu-

lated through a heat exchanger would have a number of advantages over solid

104
fuel models. The advantages would, moreover, be materially increased if the
liquid fuelwere self-moderating. It seems likely that the liquid-fuel reactor
would be among the most promising of the various designs i1f a satisfactory
fuel system were available. During the past quarter studies have been initiated
to ascertain whether uranium-bearing fused salt mixtures were of possible

value in this connection.

General Chnaracteristics of Fuel System. The projected high temperature
reactor imposes specific and stringent restrictions on the physical aud chemi-
cal properties of the fuel system. The fuel must contain a considerable
amount of uranium and must consist of elements of low neutron cross section.
The fuel should, if possible, contain a sufficient amount of a moderator ele-
ment. The system must be thermally stable to at least 1800°F and must possess
a low vapor pressure (preferably less than one atmosphere) at this temperature.
On cooling to about 1100°F in the heat exchanger, the fuel must undergoiu)bhase
changes and for reasonable ease 1n startup it should be molten at much lower
temperatures; it is hoped that melting points below 400°C can be obtained.
The system must form a melt of low viscosity, must be a suitable conductor of
heat, and must be relatively non-corrosive tosome otherwise suitable structural
material. In addition, the system must be chemically stable in the intense

radiation field.

The restrictions suggested above, while they do not necessarily eliminate
from consideration fuel systems composed of fused salts, limit severely the

number of individual salts which need consideration.

Choice of yranium Compound. No known compound of uranium has thermal
properties which approximate those required. It will, accordingly, be necessary
to employ some other salt or mixture of salts as the vehicle for the uranium.
Table 12 lists melting and boiling points of known uranium compounds. It may
be observed that only a few of these materials, notably vo, , the alkali
uranates, UF(, and the uranium trihalides, are stable at the temperatures re-
quirel. This list is further reduced by the unfavorable nuclear properties of

all the halogens except fluorine.

Choice of Vehicle for Self-moderating Systems. Beryllium or lithium
would be of some help insofar as moderation of the fuel is concerned. It
appears, however, that hydrogen-bearing systems alone will be completely self-

moderating. Examination of the literature reveals that sodium and potassium

105
hydroxides are the only hydrogen-bearing compounds which are thermally stable
at the temperatures of interest. Sodium hydroxide is to be preferred by virtue

of its nuclear properties.

Since the physical properties of this compound appear to be satisfactory,
for the purpose, a suitable solution or suspension of a uranium compound in
this material might prove of considerable value. The results of experiments

per formed to date are described below.

TABLE 12

Uranium Compounds

CONPOUND" MELTING POINT BOILING. POINT

(°c) (%c)
uo, 2700 -
uo, - dec. 650
U, 04 - dec. 1700
uFr, 1425 2300
UF, 1036 1417
UF, 65 56
UCl3 835 - 1725
UCl4 ' 590 : 787
UClS - dec.
UC16 179 277
UBr, 752 1567
UBr, 519 766
UI, 757 1427
UuI, 502 759
"UN 2600 dec.
uC 2275 dec.
Us 1800 -
U,S, - dec. 1800
Us, - dec. 1600
UoS - Probably unstable
UO,F, - dec.
UOZCl2 - dec. |
uoCl, - 750
UP,0, m.p. of porcelain vol.
Ue,F,C, -
UC,_,EO2 called "stable" (most

perborates explode)

3UO38203
U0;7S10, easily melted
Na,UO, stable —not melted at

(and polyuranates, 800°

other alkali ura-
nates)

106
Choice of Vehicle for Unmoderated Systems. If the restriction that the
system be self-moderating is removed, the number of fused salts which might
serve as solvent for the uranium is very considerably increased. Choice of
systems for study must, in light of present knowledge of the problem, be made

in a rather arbitrary manner.

The literature reveals that the alkali fluorides are compounds which are
stable in this temperature range and that these compounds form low-melting
binary eutectics with beryllium fluoride. The temperature-composition diagram
for the sodium fluoride-uranous fluoride system is also known. The liquid
state exists to about 600°C in this system. Since it seemed likely that among
the ternary systems one of low melting point could be discovered, the pre-
liminary studies have been concerned with these compounds. It is anticipated
that study of the binary UF,-alkali fluoride systems described below will be
followed systematically by examination of whichever of the ternary systems

seems most likely.

Stability of these compounds at high temperature in the radiation field
is still a matter of question. It may be stated that from the standpoints of
structural simplicity and possible rate of recombination they seem likely
choices. It must be admitted, however, that the consequences of decomposition

under irradlation are severe.

Systems Containing Sodium Hydroxide (J. D. Redman, L. G. Overholser,
D. E. Nicholson). Asnotedin the preceding discussion, sodium hydroxide 1is
in some respects the ideal choice as the vehicle for a fused salt fuel system;
therefore a considerable effort has been placed during the past quarter on
attempts to dissolve uranium compounds in this medium. Much more recently a
program of evaluation of suspensions of uranium compounds in this material has
been initiated. Work on the latter problem is still largely in the planning

and literature survey stage.

1. Solubility of Uranium in Sodium Hydroxide. 1Initial attempts to dis-
solve uranium in molten caustic were made with reagent grade caustic contained
in nickel crucibles exposed to air in an electric muffle furnace. 1In general
the caustic was heated to about 750°C to expell most of the water; some of the
carbonate was probably decomposed in this interval. The caustic was permittéd
to cool to near the melting point and the uranium compound in question was in-

troduced as a finely divided anhydrous powder. The sample was then raised to

107
the temperature desired, usually 450 to 650°C,and maintained with occasional
agitation for at least two hours. The melt was then allowed to settle, still

-at the elevated temperature and after a clear liquid was obtained a small

sample was withdrawn simply by insertion of a metal rod into the top of the

melt. The uranium content of the few milligrams of material so obtained was

determined by the standard fluorescence procedure.

By this simple technique the solubility of sodium diuranate, U,0,, UO,
and uranyl phosphate in the caustic was shown to be less than 0..03% by weighg

These studies were repeated and a number of other materials were tested
by use of the apparatus shown in Fig. 24. This assembly was fabricated from
nickel with the 1/8~in. tube silver soldered to the reactor body. The charge
of caustic and the uranium compound desired could be placed in the upper
portion of the inverted assembly and a metallic filter medium with gold or
copper gaskets fitted in place between the halves of the reactor. The assembly
was heated in a pot furnace with the charge maintained under an inert atmos-
phere if desired and after a suitable interval application of the vacuum served

to separate the caustic solution from the undissolved solid.

No filter cloth of sufficient fineness was found and the sintered stain-
less filters available reacted rapidly with the caustic. However, sections of

porous nickel supplied by the laboratory at K-25 were found to be very useful.

With this apparatus, repetition of the former studies yielded values
somewhat lower than those previously obtained. While there was some variation
even on different experiments with the same compound the values were rarely
higher than 0.005% by weight. The studies were:expanded to include uranyl
carbonate, UF4, uc, and uranium metal. In each case the solubility was shown
to be less than 0.01%.

In addition preliminary studies indicated that although sodium hexa-
me taphosphate dissolves up to 2% of UO, at elevated temperatures, mixtures cf

NaOH with this compound show virtually no solvent action.

The experimental effort to date has perhaps not been sufficient to demon-
strate that solutions of uranium in caustic cannot be prepared. The great
similarity in behavior of all the compounds tested, however, indicates that
they react with the caustic to yield.a common product, probably sodium uranate,

which 1s insoluble.

108
60l

 

DRY NITROGEN >

DWG. 9878
NOT CLASSIFIED

 

EXHAUST

 

 

N\ (7
]

 

VACUUM PUMP €«————

 

POT FURNACE
8" HIGH =  *—»

CHARGE — |

FILTER MEDIA

 

MOLTEN CAUSTIC —_|

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FIG. 24

FILTRATION APPARATUS

 

 
2. Suspensions of Urantum in Caustic. In the studies presented above
it was noted that some of the uranium compounds tested seemed to give sus-
pensions which were difficult to filter and which required a censiderable time
to settle. This was especially true of those obtained by reaction of uranium

metal with the caustic and by reaction of UF, with this material.

Late in the past period this laboratory was asked to attempt preparation
of stable slurries of this type. The preliminary experiments have been per-
formed in silver crucibles at moderate temperatures in the range 500 to 600°C

and have used U0, as the test material.

Microscopic examination of the resultant suspension after rapid cooling
has indicated that the size of the UO; particle obtained by treatment at these
temperatures is of the order of five to 10 microns in diameter. From pre-
liminary observations it appears that this diameter is largely independent of
the original size and, perhaps, of the original chemical composition of the
material added. Studies designed to establish this important point and to

disclose the chemical composition of the suspended particles are under way.

Low Melting Fluoride Systems-~Thermal Analysis (C.'J. Barton,[l.EQ Moore,
J. P. Blakely, G. J. Nessle). During the past quarter the two componént systems
of UF, with sodium, potassium, and lithium fluorides have been explored by the
me thods of thermal analysis. In addition some preliminary corrosion data have
been obtained with various metals in NaF-UF, melts at 700°C in cooperation
with the Y-12 Metallurgical group. The information so obtained will be of

value in planning experiments on radiation stability of these materials.

The uranium tetrafluoride used in thermal analysis studies was a high
purity, anhydrous material obtained from Mallinkrodt. The alkali fluorides
were anhydrous reagent grade commercial preparations. Potassium fluoride was
dried at elevated temperature and maintained and weighed in a dry box. The

other fluorides were dried and subsequently used without special precautions.

The cooling curves were obtained in the conventional manner using 1 to
1.5 moles of the salt mixtures contained in a covered crucible of high density
pure graphite. The crucible was contained in a graphite block fitted into a
'5-in. Hoskins pot furnace. The crucible assembly is shown in Fig. 25. The
handle of the graphite stirrer as shown was drilled out to serve as the thermo-

couple sleeve and extended through the furnace cover.

110
A

(e T U

.",'i.

t.‘ »

e

#
i
o
"
"

 

 

 

FIG.

2y,

CRUCIBLE

ASSEMBLY

 

 
The entire assembly is shown in Fig. 26. A slow flow of nitrogen, de-
oxidized by passage through heated copper turnings and dried by passage through
magnesium perchlorate, was directed into the crucible to minimize oxidation of
the UF,.

Chromel-alumel thermocouples with an L & N. Micromax recorder have been
used for the measurements performed todate. Platinum-platinum rhodium couples

are to be substituted as equipment is available.

All mixtures studied were heated to temperatures above 1000°C and stirred
well to insure complete fusion and homogeneity before cooling. The mixtures

were stirred occasionally during the cooling cycle.

1. NaF-UF . A phase diagram for the system NaF-UF, has previously been
reported.(!? This diagran for which a few points were checked to break in the
equipment is shown as Fig. 27. It will be noted that the system is reasonably
simple with one compound, NaUF,. crystallizing from the melt and with two
eutectics, the lower one melting near 610°C. One run in the present study used
26 mole percent UF, and obtained only one break in the cooling curve at 620°C,

essentially confirming the eutectic temperature and composition reported.

2. KF-UF,. The system KF-UF,, however, is apparently somewhat more
complex. Figure 28 indicates the points on this diagram which have been ascer-
tained with certainty. It is apparent that the compound KSUF,,-which has no
counterpart in the sodium fluoride system, is quite stable and high melting.
While there is not sufficient data above 30 mole percent UF, to justify con-
struction of the diagram, it appears that the compounds K,UF, and KUF; exist,
though the former probably possesses anincongruent melting point. This system
does not at present appear promising since the lowest melting point so far
obtained is the KF-K,UF, eutectic whose melting point is more than 100°C higher
than that obtainable in the sodium fluoride system.

3. LiF-UF,. A large number of cooling curves have been run on mixtures
of LiF and UF, in the range up to 40 mole percent UF,. The changes of slope
(freezing points) in this system are not well defined as in the other two
cases. The eutectic halts, however, are quite pronounced and, in contrast to
the behavior exhibited by the other systems studied, often show considerable
supercooling. In addition, over a portion of this range a second pronounced
halt appears at a temperature 40 to 50°C below the first. These phenomena are
not as yet satisfactorily explained. The behavior observed suggests that the

(1)iPaper by Co A. Eraus In University of Rochester Progress Report No.: 5 for April 1944, M-1515 (n.d.)

 

112
¢l

0o T
™ | r-:_\__.l

 

 

 

NOT CLASSIFIED

| 758

 
8
2

-~
o

Decrees C.

TEMPERATURE,
g

8

DWG. 9874

 

 

 

 

 

 

  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NaF -UF4

 

 

 

 

 

 

 

 

 

 

 

 

 

0 10 20 30 40 50 60 70 80 90 100
MoL Percent UF,

FIG. 27 NaoF - UF4

e
Gl
TEMPERATURE - DEGREES C

DWG. 9877

etk

1000

 

900

 

 

800

 

 

 

 

 

700

600

 

500

 

 

400

 

 

 

 

 

 

 

 

 

 

 

300

 

10 20 30 40 50 60 70 80 90 100
Mol PERCENT UF,

KF'UF4
FIG. 28
freezing point curve exhibits a eutectic point but that solid solutions are

obtained over a wide range of relative concentration.

The system is apparently deserving of considerably more study since it
appears that the eutectic point lies below 475°C. Present plans call for com-

plete investigation of this phase diagram before study of ternary systems.

4. Corrosion Studies. In a preliminary test, specimens of 18 different
metals were exposéd in a closed Hastelloy C container at 700°C for 10 days to
a melt consisting of 25 mole percent UF, and 75 mole percent NaF. The vessel
was pumped down and sealed with somewhat more than one atmosphere of helium to

prevent oxidation of the UF, during the test.

The specimens are still undergoing examination in the metallurgical labo-
ratory. Preliminary examination indicated that the high nickel alloys had not
been corroded seriously in this system. The stainless steels used were ap-

parently less satisfactory.

116 .
FIXED FUEL CIRCULATING MODERATOR REACTORS

{

K. Cohen, H. K. Ferguson Company

The homogeneous circulating fuel reactors discussed in the previous
section would, if obtainable, be free from all of the troubles inherent in a
solid fuel element at high power densities. This feature gives very great
importance to the research looking to the discovery of a suitable liquid fuel.
However, even if one must be content with a solid fuel element, there are some
advantages predicted for using a hydrogeneous coolant instead of a liquid
‘metal, now assumed for the ARE. The only hydrogeneous -bearing liquid which is
likely to be sufficiently stable at aircraft reactor temperatures is NaOH.
This liquid is known to have rather good heat transfer properties and it 1is
probable that adequate container materials can be found, although this question
is not yet settled. The most debatable point in the use of sodium hydroxide
at present seems to be the question of its resistance to radiation damage.
Also, no very thorough integrated design has yet been completed for the adap-
tation of a circulating-moderator reactor to an aircraft. As mentioned in
previous parts of this report, laboratory research on both the corrosion and

radiation damage characteristics of sodium hydroxide is now underway.

Meanwhile, the Atomic Energy Research Division of the H. K. Ferguson
Company has prepared a preliminary study of a fixed-fuel circulating-moderator
design for an alternate aircraft reactor experiment. This preliminary ARE

study is reported in HKF-106, some extracts from which are given below.

"A summary of the properties of the reactor is shown in Table 1.1 [Table
13 in present report]. The fuel rods are composed of compressed U0, (in which
the mole fraction of U-235 is 0.25) bathed in NaQH, contained in 30 mil Inconel
tube 1/2 inch in diameter. 350 of these rods are spaced in an open (20:1
volume ratio) lattice in liquid NaOH. Cooling is accomplished by circulating
the NaOH parallel to the rods. The reactor is surrounded by a NaOH reflector
........... The core diameter and height are 40 inches; the reflector is 10
inches thick, giving a reflector saving of about 5 inches........... These
dimensions were arrived at by a series of design calculations to give the
minimum weight of U-235 consistent with the amount of Inconel needed for heat
transfer. The weight of U-235 required is 120 lbs. This allows for the very
considerable self-shielding of the U-235, according to the Y3 approximation of

the spherical harmonic method..........

117
‘TABLE 13

[Table 1.1 in:HKF-106)

Chief characteristics of -NaOH-Cooled ARE Reactor

Power level
Maximum surface temperature
Maximum coolant temperature
Coolant temperature rise

Reactor dimensions
Core (right cylinder)
Reflector thickness

Fuel element
Dimensions
Composition

Number of fuel rods

"Uranium content (U-235)
Uranium-235 concentration
Coolant, moderator, and reflector

NaOH- fuel v6lume ratio in core

Rod arrangement
Spacing

Coolant flow
Path
Rate

Weight of Inconel in core
Heat transfer surface

Maximum heat flux

k,, 1500° F.

‘Temperature coefficient of k
Neutron flux (at 2200 m./sec.)

Fuel burn-up, central rod

o’

118

1000 KW
1600° F.
1500° F.

118° F.

40 in. diam. X 40 in. . high
10 1n.

40 in. long (net) 0.50 in.0.D.
UOé sintered pellets, voids filled
with NaOH, jacketed with 30 mil

Inconel.

350
120 - 1b.
25%
NaOH
20

Triangular lattice, vertical
1.90 in. (center-to-center)

Parallel to rods, upward
100 gpm

450 1b
150 ft.2
57,000 Btu/ft.2 x hr. x °F
1.42
1.2 x 10"%/°F
5.7 x 10"
1% in 300 MWD
"There 1s a very considerable reactivity change between operating temper-
ature of the NaOH (1500° F.) and the melting temperature of NaOH (600° F.) due
to the change in density. The xenon effect contributes a small amount. It is
estimated that the total change in excess k will be about 0.14. Control rods
to take 0.20 in k are provided for: these are 16 3-inch diameter Inconel
tubes coated on the inside with a layer of boron and filled with beryllium to
increase fast neutron absorption. The control rod fits into empty sleeves in

the core of the reactor and does not have to be sealed.

"The choice of Inconel is based on HKF experience in handling caustic,
which shows it to be superior to nickel from the stand point of strength and
equal to it for corrosion. The UD, tablets are based on French experience of

Zoe, which gives good reports on high density compressed and sintered powder.

"The reactor .......... is intended to operate at a maximum NaOH tempera-
ture of 1500° F. and a maximum fuel rod surface temperature of 1600° F., at
which conditions it will generate 1000 KW of heat. This heat rate was taken
as a basis to conform with the general specifications set by ANP for the A.R.E.
reactor. Higher heat rates could be achieved without much difficulty. The
hemispherical shape of the vessel was chosen to permit easy adaptation to a

homogeneous reactor ..........

"The heat is removed during operation by circulating the NaGH through an
outside loop.......... consisting of a fiump and cooler. A separate drain line
is provided to remove the NaQOH to storage after shut-down, the flow being
induced by helium pressure. In order to control the shell temperature, and
specifically to heat the shell before introducing NaOH, hot combustion gases
are circulated outside............ and to a stack. JInsulation on the outside
of the shell protects it from high temperature gradients resulting from too

rapid heating or cooling. ..........

"The heat capacities of the various parts of the reactor are sufficient
to permit emergency shut-down with safety even 1f it is impossible to continue

to circulate NaOH for heat removal.

"The control rods are operated individually by hydraulic cylinders above
the top shield (in a cold region) with emergency insertion of springs ........
....Since the NaOH must be removed from the reactor after shut-down (to prevent
excess buildup of reactivity as the temperature drops and to prevent solidi-
fication of NaOH) a helium circulating system is provided to remove the heat

generated by fission products. ............
"Since this is an experimental unit it is designed for easy dismantling
and revision of internal lattice. The reactor shell top, which contains all
openings, may be removed through the shield; the fuel rods which hang from a
heavy Inconel sheet, may be liftgd out individually and replaced or the entire
lattice may be removed as a unit and replaced by another. ............. The
only critical detail in assembly'is location of the control rod actuators
‘which are contained in a portion of the top shield. This assembly must be
lowered onto the reactor shell and centered on it in such a way that changes

in temperature will not alter the arrangement.

"Although the temperatures at which we propose to operate are beyond
usual practice in NaOH handling, we are within striking distance of present
experience. The fabrication of the fuel element is not particularly critical
‘at these low fluxes, and the bulk of the reactor is a liquid which requires no
‘fabrication. [Radiation damage to NaCH at these temperatures and fluxes should
be negligible, and fur thermore, the NaOH can be continuously replaced. From

the standpoint of short development time, the reactor is very attractive.

"The reactor is somewhat larger than one would like to see on an airplane,
but this is not a real limitation on design. The Inconel poisoning tends to
increase  the uranium weight and we weht to a larger reactor size to keep it
down. Later phases eliminate the Inconel jacketing entirely. Taking a larger
core also made the mechanical details of the control rods simpler. An extremely

small core could do the same thing by permitting control in the reflector,”

120
LITHIUM ISOTOPE SEPARATION

Particularly for supersonic aircraft reactors, eventual use of separated
lithium-7 isotopes as a liquid metal coolant appears highly attractive. Two
research groups of the Oak Ridge National Laboratory are now working on
possible methods for obtaining massive quantities at a sufficiently low cost.
The Chemistry Division at X-10 is attempting to develop ion-exchange column
methods for large-scale use, while the Materials Chemistry Division at Y-12 is
exploring the molecular distillation of the liquid-liquid extraction tech-

niques. These researches are described below.

ION-EXCHANGE SEPARATION OF LITHIUM ISOTOPES

J. H. Gross, Chemistry Division

During the past quarter a column run at 70°C with a Dowex 50 bed has been
completed. Assay results from this run indicate that enrichment is not im-
proved by hot operation of the column.

Equilibrium studies for the NH4+ - Li* exchange on a carboxylic exchange
resin have been completed. On the basis of these studies, a set of appropriate
column operating conditions has been chosen, and a pilot column run has been
conducted with the carboxylic exchanger. No isotopic analyses from this run

have been obtained at this time.

The work mentioned here is described in detail in aquantetl?F?éport;;l)t

LARGE-SCALE LITHIUM ISOTOPE SEPARATION METHODS

G. H. Clewett and W. M. Leaders

Materials Chemistry Division

During the last quarter the research effort has been concentrated on two

different methods both of which appear to hold promise of accomplishing the

separation of lithium isotopes.” One of these methods is molecular distillation

(1) C'Jheni;gry Division Quarterly Progress Report for Period Ending September 30, 1950, ORNL 870 (to be
1ssue °
121
which has been under investigation for several months. Additional data from
the two stage still have made this method appear more promising than previous

data would indicate.

The second method being considered is a chemical exchange process. This
is a recent development and already has been demonastrated to be capable of
achieving good separation in small scale equipment. The method uses lithium
amalgam and aqueous lithium hydroxide flowing countercurrently. A controlled
potential is applied across the two phases to counteract the natural tendency
of the lithium to become oxidized. - Reflux is accomplished by electrolysis at
higher potentials. Enrichment up to 94% 'Li has been achieved.:

The research designed to develop other liquid-liquid systems has been

nearly terminated. A summary of the more recent work is included.:

Work on pulse column applications has likewise been abandoned temporarily

in the favor of the amalgam system.

Lithium Amalgam-Aqueous Hydroxide System. (A. Clark, Jr., W.: T. Ward,
L. P. Twichell, and F. B. Waldrop). Original experimental work with lithium
amalgam systems in which amalgam was contacted with a solution of lithium ion
gave single stage separation factors approaching 1.05. The primary difficulty
with the early approaches to application of such a system was the excessive
reactivity of the amalgam. Numerous solvents were tried without success.:
During the quarter research was begun on an electrolysis-exchange apparatus.
The fundamental principle of the idea as conceived was to counteract the re-
action effect of the amalgam by applying a suitable potential between the
amalgam and the aqueous phase. Preliminary experiments demonstrated that
simple electrolysis apparatus could be used for the continuous preparation of
lithium amalgam from dilute (one molar) aqueous solutions with greater than
97% transfer of lithium into the amalgam. Such a technique is essential as a
reflux mechanism in any process using amalgam for purification of Li. These
initial successes made it almost mandatory to examine carefully the over-all
possibilities of this process. One of the first pieces of equipment built for
this work was a simple plastic box, one inch wide, two inches high, and twenty
inches long. This box was divided into compartments with the use of plastic
baffles every two inches.” A perforated platinum electrode was provided near

the .top of the apparatus to act as the anode in the aqueous phase. Platinum

122
wires were inserted near the bottom for contact with the amalgam which is the
cathode. Original experiments were made without agitation and sufficient
voltage was used to electrolyze nearly all of the lithium into the amalgam.
Definite enrichment of ’Li in the aqueous overflow was observed. It was also
demonstrated with this type apparatus that by proper adjustment of the voltage
applied any degree of transfer of lithium could be achieved, including zero

transfer which would be desired in an equilibrium exchange process.

An additional plastic box device was constructed similar to the one
described. The original unit was equipped with agitators in each compartment
and connected in series flow with the second unit. Mercury and aqueous lithium
hydroxide solution were fed into opposite ends of the equipment. The section
which was agitated was operated with just sufficient potential applied to
prevent any net transfer of lithium between the two phases. The reflux section
(unstirred) was operated at approximately 10 volts and virtually all of the
lithium ions were reduced into the amalgam. The reflux ratio actually attained
was approximately 50 to 1. After several attempts, a successful run was made
with this apparatus and a sample taken at the termination of the operation.

This sample assayed 93.4% 'Li in contrast to 92.5% in the feed material.

A more substantial model "electro-exchanger": was designed on the basis of
these results. It consisted of 48 compartments on one-inch centers as shown
schematically in Fig, 29, This unit was incorporated into a more elaborate
arrangement designed to minimize the labor and attention required during
operation. The over-all cycle is shown in Fig. 30 and a photograph of the
‘apparatus as currently operated is shown in Fig. 31. Mercury was pumped from
a reservoir into the reflux section and then allowed to flow by gravity through
the entire train and back to the reservoir. The lithium hydroxide feed rate
was controlled by a constant level siphon and flowed countercurrently to the
mercury under the action of gravity. At low flow rate (10-15 cc/minute) the
hydraulic heads required for this type flow are relatively small. This feature
does offer some operational problems at higher flow rates. The aqueous over-
flow from the refluxer was collected and evaporated to check on the efficiency
of the reflux section.” The material collected could also be used as assay
samples. A very elaborate mercury scrubbing system has been devised to insure
complete removal of lithium from the mercury prior to recycle into the appara-

tus. The mercury is contacted first with water, then sprayed through nitric

123
vl

STIRRERS

AQUEOUS

AMALGAM

  
     

—_—

-

POWER DRIVE

GEAR TRAIN

t

DWG. 9879

 

FIG. 29
EXCHANGE GCELLS

——» AQUEOUS

~«~—— AMALGAM
cel

 

 

 

DWG. 9873

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LiOH SOLUTION ACID
LiOH J -
REFLUXER AMALGAM EXCHANGER  (amaicam| SCRUBBER
——— — -~
He RECYCLE y
-
Y Y
H,0 (‘LiOH) WASTE ACID
(°LiNO,)
FIG. 30

COUNTER CURRENT LITHIUM EXCHANGE
Y-12 PHOTQO 10215

 

 

FIG. 3l EXCHANGE CELLS IN OPERATION

 
acid, and washed again with excess water. Qualitative tests on this scrubbed

mercury showed that the lithium content, if present at all, was very small.

Several successful runs have been made in this apparatus using low flow
rates (10-17 cc/minute each stream) and enrichments of 'Li up to 94% were
achieved. Additional work is planned in this equipment using flow rates up to

the maximum capacity of the machine to minimize the effect of back diffusion.

It has been shown that. the reflux mechanism can be included as .n inte-
gral part of the exchange apparatus by merely splitting the anode in the
aqueous phase and using a common mercury cathode for both the exchanger and
refluxer. This discovery greatly simplifies the operational problems and

eliminates the necessity of pumping the liquid between the two units.:

Two additional units are in process of design and fabrication and should

be ready for operation during the current quarter.

All of the analyses used have been made witha modified Nier mass spectrom-
eter.. The number of analyses that can be made in any given time by this
method is definitely limited. Some other method capable of handling a large
number of samples in a short period of time is definitely needed to insure the

most rapid progress of this research.:

A statistical analysis of the mass spectrometer results on a series of
nine standard samples of lithium iodide has been made. The arithmetical average
of these samples was 92.51% 7Li. The 95% confidence interval on these assays.
has been calculated and found to be as follows: 92.34; 92.68.

Oorganic-0organic Systems (J. S. Drury and D. A. Lee). During this
period, the work in the development of a two-phase liquid-liquid organic
solvent system for the chemical separation of lithium isotopes was pursued

along the following lines:
a.  Preparation of lithium compounds.
b. Solubility of lithium compounds in individual organic solvents.

c.  Distribution of lithium compounds between the phases of two im-
miscible organic solvents.

d. Search for new immiscible organic solvent pairs.

127
1. Preparation of Lithium Compounds.

In addition to the

previously reported the following compounds were prepared:

a. Anhydrous Lil

Lithium
Lithium
d. Lithium
e. Lithium
f. Lithium
g.. Lithium

Lithium
i. Lithium
j. Lithium
k. Lithium
1. Lithium
m. Lithium
n. Lithium
Lithium
Lithium

q. O-amino

dithionate
picrate
thioglycolate
acid oxalate
acid tartrate
acid phthalate
benzoate
acetyl acetonate
urate
phenoxide
salicyaldehyde
furoate
lactate
citrate

amino acetate

lithium phenoxide

compounds

2. Solubility of Lithium Compounds in Individual Organic Solvents. For

purposes of this study an arbitrary standard of solubility was adopted. One-
tenth gram of the lithium compound was added to 5 ml of the solvent and the
mixture was shaken thoroughly. Compounds failing to dissolve completely were
said to be insoluble in the solvent.  Compounds dissolving completely were

termed soluble and further studies were made of this group. 1In general, the

following solvents were used:
a. Butyl acetate
b.  Hexone
¢. Dimethyl aniline

Ethyl benzoate

128
e. Benzaldehyde
f. 2-ethyl hexanol
g. Diethyl ether
h. Nitromethane
~i. Chloroform
j. Toluene
| Lithium acid oxalate, lithium acid tartrate, lithium acetyl acetonate,
lithium urate, lithium phenoxide, lithium lactate, lithium citrate, lithium
amino acetate and o-amino lithium phenoxide were found to be insoluble in the
solvents listed. Lithium picrate proved to be soluble in butil acetate,

hexone, and benzaldehyde. It reacted with dimethyl aniline, giving a brown

product and was insoluble in the remaining solvents.

Lithium salicyladehyde, while insoluble in the listed solvents, was
soluble in ethanol. Lithium furoate was insoluble in furfural, furfuryl

alcohol and the solvents listed above.

3. Distribution of Lithium Compounds Betueen the Phases of Tmoilubisqible
Organic Solvents. Those compounds possessing suitable solubilities in im-
miscible solvents. The following immiacible solvent pairs were used:

a. Butyl acetate - ethylene glycol
b. Hexone - ethyleng glycol

¢. 2-ethyl hexanol - nitromethane
d. Benzaldehyde - ethylene glycol
e. Dimethyl aniline - ethylene glycol
f. Ethyl benzoate - ethylene glycol
g. Formamide - butyi acetate

h. Formamide i he#one

i. Formamide - 2-ethyl hexanol

j. Formamide - ethyl éther

k. Formamide - dimethyl aniline

1. Formamide - ethyl benzoate

129
Normally, the procedure used to determine the distribution of the compound
between the phases started by saturating each solvent of the pair with the
other in order to prevent a salting out effect later. Each liquid was then:
equilibrated with an excess of the compound to be studied. Equal‘voluhes of 
the saturated solution were then equilibratéd, separated, and finally anaifzed‘
for lithium.: The distribution of the compound between the phases was then

calculated.:

Favorable solubilities were found for Lil in various solvents. However,
the distribution of LiI between two immiscible phases invariably favored the
more water-like solvent of the two by a large factor.. By saturating the less
water-like phase with I,, distribution ratios of 3.5 to 10.0 in favor of the
more water-like phase were obtained. ' This significant change may be attributed
to the formation of a complex between Lil and I, in the less water-like
phase.” Probably the complex formed 1is LiP(::.j Table 14 gives the results of

these studies.

4. Search for New Immiscible Organic Solvent Pairs. Batch runs of the
systems Lil ¢ womigey ¥S- L1l * Ty ctnyt hexanoly LT (formamidey VS L3I ¥ Ipiperone )
failed to show any isotope separation effect. In these systems the lithium is
probably coordinated by the oxygen atom in each solvent.. A more favorable
condition for isotope separation would exist if the lithium were coordinated
by an element of the second solvent which is different from that which co-
ordinates the lithium in the first solvent. None of the immiscible organic
pairs in which lithium compounds are soluble fulfill this condition. Conse-
quently a search was made to find organic liquids containing N, S or P as
coordinating .elements which are immiscible with other organic solvents having
0 as the coordinating element. The solvents listed in Table 15 were se-
lected as non-oxygen coordinators while the solvents listed in Table 16
represent a standard list of organics against which the solvents in Table 14
were checked for immiscibility. The solvents of Table 16 represent oxygen-

coordinating liquids predominatly.:

In excess of sixty immiscible organic pairs have been found from the
many possible combinations made possible by comparing the solvents of Table 14

with the solvents of Table 16.- This study is being continued.

Molecular Distillation (C. C. Haws, G. B. Marrow, and M. J. Fortenberry).:
Reduction of the experimentation with the two-stage molecular still to routine
operation permitted the collection of complete data within the operating limits

of the equipment.  These data are given in Table 17. .

130
TABLE 14

Organic Solvents for Lithium Exchange

 

 

 

 

Li COMPOUND SOLVENT A SOLVENT B DISTRIBUT ION RATIO
is
Lil CH,NO, 2-ethyl hexanol Miscible
Lil CHONH, BuHc saturated with I, 6.5
Lil CHONH, BuAc Miscible
Lil (CH,0H), BuAc saturated with I, Miscible
Lil CHONH, Hexone saturated I, 8.8
Lil (CH,0H) , Hexone saturated I, Miscible
Lil CHONH, 2-ethyl hexanol sat’'d I, 3.5
Lil CHONH, Ethyl benzoate sat’'d I, 4.5
LiI (CH,0H), Ethyl benzoate sat’'d I, 4.9
Lil | (CH,0H) , Benzaldehyde sat’d I, Miscible
Lil | CHONH, Ethyl ether sat’d I, Solid phase deposited
Li picrate | CHONH, Hexone Miscible
Lil CHONH, Dimethyl aniline Reaction occurs
Lil (CH,OH), Dimethyl aniline sat’d I,| Reaction occurs

 

 

Non-oxygen Coordinator Solvents

Benzyl mercaptan

Methyl disulfide

Isoamyl sulfide
Ethyl thiocyanate

Tri-n-butylamine

W N NN s W pg

Di-n-propylamine

Isobutyl mercaptan

Ethyl isothiocyanate

TABLE 15

9. Adiponitule

10. Benzothiazole

11. Benzonitrile

12. Triphenyl phosphite

13.. Tricresyl phosphate
14. Allyl iodide

15. [ amino-n-octane

16. Phenyl isothiocyanate

131
TABLE 16

Standard List of Organic Solvents

1. Acetone 16. n-butyl ether
2. Acetyl acetone 17. Benzyl alcohol
3. Butyl acetate 18. Diethyl ether
4. Methyl isobutyl hexone 19. Diethyl cellosolve
5. Butanol 20. Dimethyl aniline
6. Benzene 21. Diacetone alcohol
7. Carbon tetrachloride 22. Triethanolamine
8. 2-ethyl hexanol 23.  Hydroxyethyl ethylene diamine
9. 2-amino - 2-methyl - l-propanol 24. Capryl alcohol
10. Isoamyl alcohol 25. Ethylene benzoate
11.- Trimethylene glycol 26. Ethylene glycol
12. Penta ether 27. Ethanol
13. Glyceral 28. Formamide
14. Furfuryl alcohol 29. Pyridine
15.- Aniline

132
TABLE 17

‘pata for Lithium Runs in the Molecular still

 

 

 

 

 

 

ASSAY RESULTS - PERCENT 'Li
o T | e o SRS LY L SR T
L (Percant)

7 4.5 475 92.8 92.6 92.5 0.3 1.02

8 11 470 92.9 92.4 92.2 0.7 1.04
10 24 475 92.9 1 92.4 92.1 0.8 1.06
11 50 475 92.9 92.7 92.5 0;4 1.03
12 10.5 500 92.9 92.5 92.5 0.4 1.03
15 24 450 92.8 92.3 92.1 0.7 1.04
17+ | 48 455 93.1 92.9
18** 21 500

19 6.5 525
22 %> 6.5 525

 

 

 

 

 

 

 

 

* Unsatisfactory or incomplete runs omitted.

*+ Results are not yet available.

Limit of error £ 0.1 percent.

133
From the data the following interesting observations and conclusions may

be drawn:

(1) A much higher separation factor than the previously reported
1.02 may be realized. This is most significant in its effect
upon product cost.

(2) A concentration difference of 0.5% exists between the lower and
middle boats of runs 8, 10, and 15. This (subject to errors in
assay) corresponds to a separation factor of 1.08, which is the
theoretical maximum for this process.

(3) After a certain minimum run time, extended operation apparently
has no effect upon enrichment (rumns 7, 8, 10, and 11).

(4) Run 10 may have reached full equilibrium (within limits of
assay errors) 1n its 24-hour eperation.

(5) Although the separation factor is known to decrease with in-
creased operating temperature;, no decrease is noticeable from
the data.

Due to the small number of stages and the relatively large probable error
involved in assay, as compared with the total enrichment involved, these data
are necessarily spotty and definite conclusions may not yet be drawn. However,
the data clearly demonstrate the feasibility of a multi-stage still, and in

so doing are of value.

Design has been finished and construction is fully 75% completed on a
15-stage molecular still of the same general design as that used in the above
described work (Figs. 32, 33, and 34). Completion of this still is expected
with sufficient time remaining for at least preliminary testing within the

present guarter.

134
KEY TO FIG. 32

Cascade or Madorsky still

Cast Copper End Heater.
Cartridge Type Heater Elements.

Condensing Surfaces.

.- Evaporating Surfaces Showing Baffles and Direction of

Lithium Vapor Movement.

Stillpot Assembly (Photograph Fig. 5).

.- Reflux Port Showing Liquid Lithium Flow.

- Calrod Type Heater Element.

Cast Copper Bottom Heater.

Caat Copper Side Heater.

 Condenser Asaembly (Photograph Fig. 6).

Evaporating Chamber for Coolant (Dowtherm or Mercury).
Anti-Splash Baffle.:

Heat Exchanger Condenser for Coolant Vapors.

- Heat Exchanger Tubes.

135
 

 

© O

 

 

DWG. 9875

FIGURE 32

i .
>

 

 

 

 

 

o @ \_ e 2 o
X eSS \Q)
N fo (
2727
@
\ o ©
@ ©%
®

MADORSKY TYPE
STILL
Y-I2 PHOTO 10214

LA

row B L . B

 
 

- T

 

 

 

2 = - . - _._‘,r

FIG. 34 EVAPORATOR FOR MADORSKY STILL
138

 

 
A second (vertical) still, which is much cheaper in construction, 1is now

being designed (Fig. 35).

With the completion of the work on the two-stage still, emphasis has
shifted to the collection of data on the rates of evaporation ©f lithium metal.
Only a few runs have been attempted thus far in this phase of the investi-
gation, the results of which were inconclusive. However, a device has been

tested which promises to perform satisfactorily in the future.

139
10.
11.
12,
13.

KEY TO FIG. 35

Vertical Still

Reflux Section.

Cértridge Type Electrical Heaters.
Outlet to Pressure Manifold.
Lithium Condensers (cooled by Dowtherm or Mercury under pressure).
Exte;nfil Heaters Brazed to Shell.

Evaporating Boat.:

Reflux Tube.

Common Header for All Internal Condenser Boiling Tubes.
Directional Baffles.

Cartridge Heaters in Boat."

Baffle to Direct Liquid into Constant Head Baffle.
Constant Head Baffle.

Pressure Manifold to External Condenser.

140
FIGURE 35
\/ACgLJUM iiii

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/ \\/
: |
) Y
i 4
/
fi % @
3 ) / J/
? /
] /
: i
@ | | AN\ X
~— R
| 5 R I @
@ - °
L © /é>/@
; Y
@ | 4 T
1 | / .e oe /
e \\\\\\\\’ \ /\\\\\\\\
/ ©
[ : b
/ \
CONDENSATE / @
; :
/
s /
IR \ @K ;\\\\\\\
NN / ';\\\\\\\
/) / @
/ /
/] 4Z
VERTICAL | /
STILL ’ N

 

7

VACUUM . =