ORNL-1227.txt

CENTRAL RESEARCH LIBRARY

DOCUMENT COLLECTION

LIBRARY LOAN COPY

PO NOT TRANSFER TO ANOTHER PERSON

se¢ {o see this document,
send in neme wilth docupiont and the {ihrary will

arrangs a foan,

 
 

 

 

 

 
i g o :o

 

<4

)

 

¥}

"

 

<)

 

ORNL-

This document consists of 214 pages.

copy {pp of 222.

, ‘ Contract No. W-7405-eng-26
AIRCRAFT NUCLEAR PROPULSION PROJECT

QUARTERLY PROGRESS REPORT

FOR PERIOD ENDING MARCH 10, 1952

R. C. Briant, Director

A. J. Miller, Assistant Director

Edited by:

W. B. Cottrell

DATE ISSUED

CMAY 7 {959

 

DAK RIDGE NATIONAL LABORATORY
. operated by
CARBIDE AND CARBON CHEMICALS COMPANY
A Division of Union Carbide and Carbon Corporation
Post Office Box P
0ak Ridge, Tennessee

1227

Series A.

 

 

 

1
|

 

 

T g o p— =

 

L cciiant oo

T T T T TR T P

 
ik e e

 

 

i1

 
 
 
  
   

 

INTERNAL DISTRIBUTION

G. M. Adamson

. J. Barton

D. S. Billington

N P. Blizard
Brasunas
CiBriant

yuce

Bug

Calljhan
Cardwell
Center j
Cisar N
Clewett %
Clifford W\
Cottrell N
Cowen ®
Eister

Ergen

Fraas
Gall A
Gr aham &
Grimes y
ollaender 4
Householdgr
Humes (Ku25)
Jones fif
Kellhqftz

    
  
 

QLM NEOOPErrE0EFO=E=UI00SOTFETE R
M ERHY RN I I TN NP ERE =0 E D

 

Its trans

Felbeck (C&CCC) X

defxned 1nt%h¢'F““

 

82-91,

  

< Enérgy Act-.
o isclosure of its
i afi?“manne to“?fi?fifine_-

person is proh:b:ted ‘

ORNL - 1227
~f3gress

. Lavers_;fill2)

 
 
   
  
  

35. W,
36. R. S. L1v1n;ston
37. R. ("
38. W. D.
39. W. B/
40. J. L.
41. ;?jj. Miller
42. SK. Z. Morgan
43¢ E. J. Murphy
fifi%l H. F. Poppendiek
745, P. M. Reyling
4 . H. W. Savage
47. R. W. Schroeder
48. E. D. Shipley
49. O. Sisman
50. A. H. Snell
51. F. L. Steahly
52. R. W. Stoughton
53. C. D. Susano
54, J. A. Swartout
55. E. H. Taylor
56. F. C. Uffelman
57. F., C. VonderLage
58. A. M. Weinberg
59. C. E. Winters
3 60. Biology Library
¥1-62. Chemistry Library
& 63. Health Physics Library
m"64. Metallurgy Library
$5. Physics Library
66-6%. Training School Library
68-77% ANP Library
78-81.% Central Files

'%Eentral Files (0.P.)

  

orized

 

7 e T TR T PR Pt 4 g

x

T e X RS I R TS e e s e

e ——

 

 
! :
_ s e .. = RN PN
ol T dFECYRE LRI
H SN ;

 

EXTERNAL DISTRIBUTION f

 
 
  

<4

 

 

    

   
     

   

    
 

 
   

       

G i

R ;
92-101% ~Argonne National Laboratory £
IOSQKHArmed Forces Special Weapons Panect (Sandia)
104-111. "Agomic Energy Commission, Wa 1ngton
112. Battelle Memorial Inst1tute
113-117. Bro;%haven National Labor ry
118, Burea§§nf Aeronautics ‘
119. Bureau of\Ships /
120-125. Carbide andsCarbon CM€micals Company (Y-12)
126, Chicago Patent Grop
127. Chief of Naval:@- Ffearch
128-132. duPont Company ¢ %\
133-157. General Electg 5' any, Oak Ridge
158-161. General Ele f;r1c Comg y, Richland
162. H. K. Ferd uson Company %
163. Hanford f-eratlons Offlc
164-167. Idaho @ferations Office
168. Toway/ gfflte College
- 169-170. Lo_f fand Area Office, AEC %
171-174. Krb?“s Atomic Power Laboratory\
- 175-177. ;bm'Alamos fii
¥ 178. /Mfssachusetts Institute of Technology (Kaufmann)
179~180_'v,ound Laboratory %
181-18é&_* ational Advisory Committee for Aerfihautlcs, Cleveland
65." National Advisory Committee for Aero%@ptlcs, Washington
186£4187. New York Operations Office .
18§ 189. 'North American Aviation, Inc.
/" 190. Patent Branch, Washington
£ 191. Savannah River Operations Office s
£192-193. University of California Radiation Laboratobhy
4 194-197. Westinghouse Electric Corporation ' 3
198-207. Wright Air Development Center
208-222. Technical Information Service, Oak Ridge
< o o a}#ifT;éu;£ €%chtg%i§§E§estr1cted Data4ga
{3
111

 

 

TR T T

 

 

 

 
 

iv

 

Reports previously issued

ORNL-528
ORNL-629
ORNL-768
ORNL- 858
ORNL-919
ANP-60
ANP-65
ORNL- 1154
ORNL-1170

Period
Period
Period
Period
Period
Period
Period

Period

Period

in this series are as follows:

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

 

 

 

 
 

Aetencenane kb b e

o i

x) *

 

TABLE OF CONTENTS

FOREWORD

1.

PART I. REACTOR THEORY AND DESIGN

'SUMMARY AND INTRODUCTION

CIRCULATING-FUEL AIRCRAFT REACTOR
Reactor with Tandem Heat Exchangers
Reactor with Annular Heat Exchangers
Reactor Shield Designs

Design procedure
Activation of the secondary coolant

Shield weights and specifications

CIRCULATING-FUEL ATRCRAFT REACTOR EXPERIMENT
Core Design

Primary coolant circuit
Secondary coolant circuit

Core temperature distribution
External Fluid Circuit

Pumps

Heat exchangérs

Primary coolant syétem
Secondary coolant system

Monitoring circuit
Preheating System

Reactor assembly

Piping

    

', co ;“ K ,.
-def1ned in  SHE A ”;“;}magjrgy Act of 1946.
,7,-~V:s“f:¢:”'*'? sclosure of its

 

”“L o any unauthorized

=D WO D ] N1 U W e

Pt et
—

13
13
15

15

16
16

16
18
18
18
18

18

19
19

 
 

 

o e B L e

vi

   

Heat exchangers

Electrical Power Circuits

Control System

Shim control

Regulating and safety rods

High-temperature fission chamber
Control console and panel

Reactor dynamic computer

Instrumentation

Building

EXPERIMENTAL REACTOR ENGINEERING

- Seals and Closures

Frozen sodium seal

Frozen fluoride seal o ,
Bellows type of face seal for ARE pump
Stuffing-box seals for molten fluorides

Lubrication of seals and shafts

Pumps

ARE centrifugal pump

Laboratory frozen-fluoride-sealed centrifugal pump
Worthite frozen-sodium-sealed pump

Modified Durco centrifugal pumps
Forced-convection-qooled sodium-sealed pumps

Canned-rotor pumps

Valves

Packing-glandiseal test eguipment

   
   
  
 

 

This document- corfalis K
defined in.-the At
Its transmittal”or’#led™
contents in g fhas

person is gf o1l

19
19
20

20
20

21
21
21

21
22

23

23

24
24
24
25
26

26

26
27
27
27
29
29

29
29
 

w2

4.“ "

 

Frozen fluorides valve
Ball check valves

Valve seat test
Heat Exchangers

Aircraft type of radiator
Sodium-to-air radiator
'Fluoride-to-fluoride heat exchanger
NaK-to-NaK heat exchanger

‘Heat transfer in circulating fluoride loops

Instrumentation

Flow measurement

Pressure measurement

Temperature meaSfirement 7 |

Level controls and level indicators B
Heatlng and Cooling of High-Temperature Systems

External heatlng systems

Induction heating

Resistance heating

Insulation testing

ARE core preheating

Technology of High-Temperature Liquids

Fluoride preparation and handling
Sampllng and analyz1ng techniques
D1ffu51V1ty of he11um through stainless steel

Cleanlng and 1nspect10n technlques

REACTOR PHYSICS

Clrculatlng-Fuel Alrcraft Beactor | |

 

30
30
30

30

30
31
31
32
32

33
33
34
34
34
35
35
36
36
36
37

37

37
39
39
39

41

41

vii

 
o B, Rl B g

. g e by

heb, o, o,

o, e R i d

Eale e ¢ o o i

Gk o

 

e el e B, i .

bl e, o S

B pan B i e e d

i i MW“aMmMmm&mmwmmMmm.mMMkm s Bt e ak

 

Oscillations »
Slow kinetic effects
Critical mass

Neutron leakage spectra
Alkali Hydroxide Moderated Aircraft.Reactor
Survey Calculations of the Circulating-Fuel ARE

ARE core design

Critical mass and total uranium investment
Reactivity coefficients |
Power distribution

Neutron flux and leakage spectra
Statics of ARE Controls

Shim control requirements
Regulator rod
Safety rods

Specific Design Problems of the Cifculating—Fuél ARE

5. CRITICAL EXPERIMENTS

- Direct-Cycle Reactor
j - ,
- Control rod calibration

’ -
/ Temperature effects

1 Reflector studies

|

»Graphite Reactor
3,»-:7‘&%’%"‘“"“*’

Circulating-Fuel Reactor

Correlation of Theory and Critical Experiments

Criticality with nonhydrogenous moderators

Criticality with hydrogenous moderatops
By

    

CT) DOATA 5 .

4 estriéte&fbaifl;aifiww;r
brgy Act of 1946,

egsure of its

viii

42
44
44
46

48
48
48
50
50

51
52

54

54
54
55

57

59

59

59
61
61

62

62

63

63
65
 

 

 

.
¥

A

{ STAWARY AND INTRODUCTION.

6

 

Foil exposures

Danger coefficients

Rod sensitivity
- Gap experiment

" PART IT. 'SHIELDING RESEARCH

'BULK SHIELDING REACTOR

 "Mockup of the Divided Shield

"Reactor Calibration

DUCT TESTS
Theoretical Treatment of Duct Transmission
Measurement of Air-Filled Ducfis in Watér

Straight ducts

Ducts with bends

Comparison with theory
TOWER SHIELDING FACILITY

Tower Facility Design

Experimental Program

NUCLEAR MEASUREMENTS

.Measurements with theIS-Mev Van de Graaff Accelerator

Total cross section of iron

Inelastic scattering levels in iron
Time-of-Flight Neutron Spectrometer

Background measurements

  
  
    
    

Indium resonances

OPRexgy Act of 1946,
losure of its
‘ unauthorized

defined in Lhe” AtQuiN
- Its transsise€al”.
" contents; WM y~ha
" person’ isTprG
R

65
66
66
68
69

71

73

73
76

79
79
80

80
80
80
89

89
89

93
93
93
93

03

93
94

1X

 
 

 

e e Sk s e

 

kg kb

SECUR

 

PART I1I. MATERIALS RESEARCH

SUMMARY AND INTRODUCTION

- 10.

1.

 

CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS

| Low-Melting-Fluoride Fuel Systems

NaF-Ber-UF4
KF-Ber-UF4
BbF-Ber—UF4
LiF-NaF-Ber-UF4

Fuel containing zirconium fluoride

Simulated Fuel Mixture for Cold Critical Experiment
~Ionic Species in Fused Fluorides

Preparation of Standard Fuel Sahples
Preparation of Pure Hydroxides

Coolant Development

LiF-ZrF4

| NaF-ZrF,

KF-—_ZI'F4
RbF-ZrF,
NaF-KF-ZrF4
NaF-RbF—ZrF4
NaF-_KF-LiF-ZrF4

Preparation of Pure Fluorides

Fuel preparation equipment

Fuel handling equipment

CORROSION RESEARCH

Static Corrosion by Fluorides . - o

    
  

 
  
   

RRBE = trlcted Date as

kgy Act of 1946.
gt losure of its
ally unauthorized

This document Q)
defined in the A%
Its transmitgel-g

95
97

99
99

100
101
101
101
101

102
102
103

104
104

105
105
105
105
105
105
106

106

106
107

1109

110
 

 
  

SECURILY_INTORMATION

Effect of pretreatment of fuel
Corrosion of structural metals
Corrosion by fluorides with various additives

Melting point of fluorides after corrosion tests
Static Corrosion by Sodium Hydroxide

Corrosion of special alloys
Corrosion by sodium hydroxide with various additives

Corrosion of refractory materials
Static Corrosion by L1qu1d Metals

Corrosion by low meltlng point alloys

Corrosion by sodium-lead alloy

-

Facilities for Dynamic Corrosion Testing

Thermal convection loops

Seesaw corrosion tests
Differential temperature tests
Rotating dynamic corrosion tests

Forced convection loops
Dynamic Corrosion by Fluorides

Corrosion by fluorides in thermal convection loops

wh

Corrosion by fluorides in a seesaw furnace

Standpipe tests of fluoride corrosion
Dynamic Corrosion by Hydroxides

Corrosion by hydroxides in thermal convection loops
Corrosion by sodium hydroxide in seesaw tests

Standplpe tests of sod1um hydrox1de corrosion
Fundamental Corros1on Besearch |

Poss1b1e equ111br1a among_fluorldes and metals

   
   
 
  
  

ted Data as

| Thls docume.n " "Rt
' 4b~ 3 *‘rgy Act of 1946.

- defined in Je-
. “-Its trans} isclosure of its

unauthorized

 

110
110
112
113

114
114

115

116
116

116
118

119

119
120
120
120
122

122

122
124
124

127

127
129
129

129
132

X1

o

 
 

  

 

]
1
,’ |
é i FMF measurements in fused fluorides . , _ 135
; | Electrode potentials in fused sodium hydroxide | 135
Polarography of sodium hydrox1de in silver and plat1num ' 136
, Magnetic susceptibility of stainless steel exposed to fluorides 136
12. METALLURGY AND CERAMICS ’ - - 139
§ Fabrica;ion of Reactor Elements - . - -”f:'139
§ ~ YCold drawing of tubular solid fuel elements - =7l 7140
: ARE control rod | 140
Cone-Arc Welding 142
| Equipment} | % 142
_ Principle of operation ' | 142
i ,'t - Experimental procedure | _ 142
| Brazing 143
: Flow tests . | 144
z Corrosion of brazing alloys 144
; Mechanical Testing of Materials | 147
Inconel creep and stress data | 147
; Tube-burst tests | ' 148
§ Operation of creep and stress-rupture equipment 148
| Ceramics Laboratory ' ‘ ' ' 151
] Ceramic appliéations to reactors | 151
Coatings for the radiator - 151
é Ceramic laboratory equipment 152
% Microscopic examination of fluorides ‘ . _ 152
13. HEAT TRANSFER AND PHYSICAI PROPERTIES RESEARCH - ._ - 153

- -

Viscosity of Fluoride Mix3 153

 
      
   
 

~

._?ustrlcted Data as
0 kpergy Act of 1946.
shendesclosure of its

Y any unauthorlzed

'"“Thls document _,.
'w;deflned in th*“

A : _r.:
contentsf7;;jnV]:¥
person - 1s*propt

 

   
 

 

 

T Viscosity of NaF-KF-LiF
Viscosity of NaF-KF-LiF-UF,

Modifications of viscosity apparatus

Thermal Conductivity of Liquids and Solids

L3

Heat Capacities

Vapor Pressure of Liquid Fuels

Physical Property Data

Natural Con?ection in Confined Spaces with Internal Heat Generation
Analysis_of Heat Transfer in a Circulating-Fuel System

Heat Transfer Coefficients

Heat transfer in molten lithium,
Heat transfer to fused salts and hydroxides
Entrance region heat transfer in a sodium system

Heat Transfer of Boiling Liquid Metals

14. RADIATION DAMAGE
Irradiation of Fused Materials

Pile irradiation of fuel

Cyclotron irradiation of fuel

i

" Inpile Circulating Loops
Creep Under Irradiation

Radiation Effects on Thermal Conductivity

PART 1V. APPENDIXES

| .SUMMARY AND INTRODUCTION

%/15. THE' SUPERCRITICAL-WATER REAC

st

  
  
  
       

  

':Thxs.document , "f ; '~fӣicted Data as

- |  defined in E AT MR Act of 1946.
: ' Its™ trans”;_“y'wg@a; - losure of its
conternts- A ;ie “man -; unauthorized

)

153
154
154

154
155
155
156
156
159
161

161

161
161
162

163
163

164
165

165
165
166

167
169

171

X111

 
 

 

 

 

i b i i,

 

Al dinatia i i

 

 

 

Description of Reactor
Conclusion of the NDA Study
Recommendations of the NDA Study

16. ANALYTICAL CHEMISTRY

.Studies‘of Diatomaceous Earth
Analytical Studies of Fluoride Eutectics

Uranium

Beryllium

Total alkali metals
Total fluoride
Nickel

Chromium

Manganese

Slllcon
Solublllty of Borlc ‘Acid in Water
| Clarlty of Borated Water.

Studles of Alkali and Alkaline Earth Hydroxide Coolants

Analytical Services
17. LIST OF REPORTS ISSUED

18. DIRECTORY OF ACTIVE ANP RESEARCH PROJECTS AT ORNL
Reactor and Component Design

Shleldlng Research

Materlals Research

   

Technical Administration . -~ . . . 2,

This document zk.:=;-- 2 “iété&5bfiiafaq_

"+ . defined 1n_tr-

X1V

171
171
172
175
175
175
176
176
176
176
177
177

177
177

1T

178
178

179

183
183

185

186
191

© oy
 

 

FIGURE

O O A N ot 8 W BN

st
o

11

12
13
14
15

- 16

 

LIST OF FIGURES

TITLE

Tandem Heat Exchanger Arrangement for Circulating-Fuel Reactor
Annular Heat Exchanger Arrangement for Circulating-Fuel Reactor
Circulating-Fuel ARE Core Design

Arrangement of External Fluid Circuit Equipfient

Frozen Flfioride'Seal Tester “

ARE Centrifugal Pump

Ball Check Valve

Schematic Diagram of Circulating-?uel‘Aircraft Reactor

Leakage Spectrum Through the Reflector of the Circulating- Fuel
.Aircraft Reactor

Leakage Spectrum Around the Fuel Pipes of the Circulating-Fuel
Aircraft Reactor : :

Critical Mass vs. Core Diameter for Hydroxide Reactors with
Thick Reflectors of the Same Comp051t10n

Power Distribution in the Core of the Circulating-Fuel ARE
Leakage Spectrum from the Reflector of the Circulating-Fuel ARE
Leakage Spectrum from the Open Ends of the Circulating-Fuel ARE

Transverse Neutron Flux Spectrum for Three Sections Through the
Core of the Circulating-Fuel ARE ‘

Radial Neutron Flux’ Dlstrlbutlon in the Core of the Clrculatlng-
Fuel ABE '

Loadlng Chart of Crltlcal Assembly of Direct-Cycle Reactor

:Beact1v1ty as a Funct1on of Control Rod Position

Reactivity vs. Temperetfiref'“""“,

  
 
 
  
  

. This document co,;& >3 f;;fi“1cted Data as
: def1ned in "the AR Sger gy - At of 1946.
e e Le'fiJ“ B{sclosure of its -

contents 13 Q)Y unauthor1zed

o person is T prohthi

PAGE

10
14
17
25
28
30
45

46
47
49
52

53
53

54

55
60
61
61

XV

 
B
3

4
A
o

 

 

 

e i B i

 

 

 

i . R o e i . b, s s

FIGURE

20

91

22
23

24
25

26

“a1

.28

29

30
31
39

33
34

35
36

XVi

 

TITLE

Relétive Position of Reactof, Divided Shield, and Gamma-Ray
Spectrometer in Bulk Shielding Facility

| Gamma-BRay Spectra at 130Acm from the Water-Reflected Reactor

Fuel Assembly Arrangement of Bulk Shielding Facility Reactor

Center Line Measurements of Neutron Transmission Through
Cylindrical Ducts in Water

Traverse Measurements of Neutrons in Water Beyond Cyllndrlcal
Ducts

Center Line Measurements of Neutron Transmission in Water

Through Cylindrical Ducts with Variable Bends

Traverse Measurements of Neutrons in Water Beyond Cylindrical
Ducts with Variable Bends

Comparlson of Relatlve Source Strengths (n ) from Varlous Ducts

Comparlson of Relative Source Strengths (n ) from Various
Geometries of the 3- 1n.Duct

'Proposed 300-ft Tower Shleldlng Facility

X- 6 Aircraft Shield Mockup on Tower Shielding Fac1l1ty
Total Cross Section of Iron

Corrosion of Inconel in a Fluorlde Fuel [(NaF-KF- UF ) + 2% Zrl
for 100 hr at 816°C

'Intergranular Attack of Type-310 Stainless Steel Tested in 44%

Lead-56% Bismuth Alloy for 100 hr at 1500°F

Alloying Attack of Inconel Tested in 43% Tin-57% Bismuth Alloy

for 100 hr at 1500°F

Seesaw Apparatus for Dynamic Corrosion Tests

Sections of Hot and Cold Legs of an Inconel Convection Loop After

Circulating the Fluoride Fuel (NaF-KF-LiF-UF,) for 500 Hours

  
 

Thiezdoeomen\,_.--f )
defined -in thea:‘;‘~

 
 

Act of 1946,
gsure of its

 

PAGE
74
75
77
81
82

83

84
87

88
90
91
94

113
117

117

121

125
 

 

i

T

FIGURE
37
38

39

- 45

46
47
48

49
50

 

TITLE

Sections of Hot and Cold Legs of a Type-316 Stainless ‘Steel

Convection Loop After: Circulating the Fluoride Fuel
(NaF-KF-LiF- UF ) for 123 Hours

Nickel Thermal Convection Loop Operated for 117 hr with Sodium
‘Hydroxide Under an Air Atmosphere

Nickel Thermal Convection Loop Operated.fdr 296 hr with Sodium
Hydroxide Under a Hydrogen Atmosphere

Inconel Thermal Convection Loop Operated for 135 hr with
Potassium Hydroxide Under a Hydrogen Atmosphere

Sectional View of L-Nickel Specimen after 117 hr (14,000 cycles)

- with Molten Sodium Hydroxide in Seesaw Apparatus

Transverse SectionsAThrough %-in. OD, Cold-Drawn, Tubular,
Solid Fuel Elements_ o

Corrosion Test of Nlcrobrazed Inconel Tube-to-Header Specimens
Exposed to NaOH for 100 hr at 1500°F

Corrosion Test of Nicrobrazed Inconel Tube-to-Header Joint

Exposed to NaF-KF-LiF-UF, for 100 hr at 1500°F

Corrosion of Inconel Tube-to-Beader Joint Brazed with a 60%
Manganese-40% Nickel Alloy after 100 hr at 1500°F in NaOH

Creep and Stress-Rupture Data for Fine-Grained Inconel Sheet
Creep and Stress-Rupture Data for Coarse-Grained Inconel Sheet

Dimensionless, Wall-Mixed Mean Temperature Difference as a
Function of Reynold’s and Prandtl’s Moduli for a Heat Transfer
System with Insulated Pipe Walls

Heat Transfer Coefficients for Boiling Mercury

Comparison of Bench and Inpile Creep Rates of Nickel

  
 
  

     
  

      
 

 

. \ ':elbi )
Thi§~dobumg,v)wfifigfi-fliflRéstrlcted Data as
defined ingth '“'?L RN gy Act of 1946.
Its trap#h Y aclosure of its
conten Q™ unauthorized
person 1

 

PAGE

126
128
130
131
132
141
145

146

147
149

150

160
162
166

XVvii

 
 

 

 

 

 

 

Y Y N T e

 

T g

T

 

 

T

 

e e

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

ety s i A -

 

'y

TABLE

- b
D ke O

13
14

15

16

17

18

O © N N B W B

 

LIST OF TABLES

TITLE

Shields for Circulating-Fluoride-Fuel Reactors

Fuel Temperature After Each Pass Through Core

Performance of Fluoride-to-Fluoride Heat Exchanger

Analysis of Heater Section (Model A-1)

Cofiposition of Vérious Fluoridé Fuels and Coolants‘

Volume Fractions of the Circulating-Fuel ARE Core and Reflector
Uranium Requirements of the ARE

Reactivity Coefficients

Shim Control Requirements

CompariSon.of Control fioé Calibfafiions
Beacti#ity'ChangelIntroduced by Substituting Beryllium for Air

Reactivity Change Effected by Substituting Plastic and StalnleSa
Steel for Beryllium in the Reflector

Experimental Data on Critical Assemblies 1 and 4
Calculated Results for Critical Assemblies 1 and 4

Compafison of Experimental and Calculated Values of the Ages of
Thermal Neutrons in Reryllium and Graphite

Expéfiméntal and Calculated Values for the Cadmium and Cadmium-
- Indium Ratios '

Expérimental and Calculated Values for the Totai Loss in keff
upon Introduction of Various Materials into Assembly 4

Comparison of Calculated Effectlve Source Strength for Stralght
Cylindrical Ducts : ’/(_

 
    

 This document c;rh
defined in thf-; ol

  

PAGE

12
16
32
33
38
50
50
51
56
61
62

62
64
64

64
66

67

85

X1X

 

 
#
%

i g

 

 

 

4
o
3

 

 

   

ORMATION

TABLE TITLE PAGE
19 Comparison of Calculated Effective Source Strength for Bent,
o Cylindrical Ducts 86
20 Improvement of Count-to-Background Ratio w1th Aluminum and
' Berylllum Filters L 94
21 Comparison of Break Temperatures from Heat1ng and Coollng
Curves of the System NaF-BeF, UF , 100
22 Break Temperatures from Coollng Cur§es of the System KF;BeF -UF, 101
23 . Break Temperatures from Coollng Curves of the System BbF BeF -UF,101
24 Break Temperatures from Heating and Cooling Curves of Zir-
: conium-Bearing Fuel Mixtures 102
25 Disposition of Standard Fuel Samples | 103
26 Static Cofros1on of Various Meterlals in the Untreated
Fluoride Fuel (NaF-KF-LiF- UF ) in 100 hr at 1500°F 111
27 - Static Corrosion of Inconel and Type- 309 Stalnless Steel by
- the Fluoride Fuel (NaF-KF-LiF-UF,) with Magnesium and Zir-
~conium Additives in 100 hr at 1500 F _ 7 112
28 Melting Points of Samples of Fluoride from Corrosion Tests 114
29 " Static Corrosion of Structural Metals by Sodium Hydroxide '
with Various Additives in 100 hr at 1500°F , 115
30 Melting Points of Various Alloys ' 116 |
31 Corrosion of Types-310 and -317 Stainless Steel and Inconel
by Various Low Melting Point Alloys in 100 hr at 1500°F 118
32 Corrosion Data from Inconel and Stainless Steel Thermal Con-
' vection Loops Operated with Various Fluorides 123
33 Free Energies and Equilibrifim Constants for Reactions of
Metals with Alkali Fluorides 133
34 Free Energies and Equilibrium Constants for Reduction of |
Uranlum Tetrafluoride with Metals 134

   
   
  

DATA

e I D ;?fstrlcted Data as

deflned in the’il-g;‘w gy Act  of 1946.
a1 6T sNosure of its

contents runauthorized

person j&

:-'$ -
e 1 (\i \MU

(_} e A ,
o SECURIY
 

TABLE
35
36

37

38

39

 

TITLE

Viscosity of NaF-KF-LiF-UF, Mixtures as a Function of Uranium
Tetrafluoride Concentration

Physical Properties of Fluoride Salts
Physical Properties of Miscellaneous Materials
LITR Tests on Fused Fluoride Fuels in Inconel at 1500°F

Summary of Service Analyses

   

‘qfigtr1cted Data as
N Act of 1946,

   

T

o

§o
T ey,

PAGE

154
157
158
164
178

XX1

 
 

 

 

 

L 3 * .
. *
" ¥ 3
! . i A a v ' ‘
. e
. a v *
&
% v
g
- . "
e
. .
<
=
« .o .
'
- .
. - . < .
- . '
-~ .
»
A
W TR T I ” ey e Y ~ e A ey v

 
 
 
 
 
 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

FOREWORD

This is the quarterly progress report of the Aircraft Nuclear Propulsion
Project at the Oak Ridge National Laboratory and summarizes the technical
_ progress on the project during the period covered. It includes not only the
work of the Laboratory under its own contract, W-7405-eng-26, but also the
research for the national ANP program performed by Laboratory personnel. The
reportis divided into four parts: I. Reactor Theory and Design; II. Shielding
Research; III. Materials Research; and IV. Appendixes. Each part may be
regarded as a separate entity and has a separate "Summary and Introduction.™

  
    
     

.~ This doéument',fif;?”"-“éstr1cted Data as

" defined in the XIS Edergy Act of 1946.
Its trang ihsclosure of its
conten i unauthorized
persoil 4

oo e
i

 

 
 

 

 

pdcadide . S o o e s i b

e

 

 
 

 

DA,

vzt

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 
 

kb s

 

|
!
\

 

e gt

 

N

¥

)

1

-y

SUMMARY AND

Analysis of the circulating-fuel
aircraft reactor has been extended
to systems incorporating intermediate
heat exchangers, various secondary

coolants, liquid moderators, and the

use of heavier reactor shielding
(sec. 1). All these systems utilize
the fundamental advantage of the bi-
functional fuel-coolant, and appear to
be capable of supersonic nuclear
propulsion. The location of the heat
exchangers around the reactor results
in lower shield weight, even with a
larger shielded-volume diameter, than
a tandem reactor and heat exchanger
arrangement. In order to perform a
limited amount of aircraft maintenance
without special shielding, various

modifications of the minimum divided

shield specifications have been in-
vestigated.

Studies of the performance and
design of the circulating-fuel air-
craft reactor are sufficiently en-
couraging that the first Aircraft
Reactor Experiment (ARE) to be con-
structed by the Oak Ridge National
Laboratory will be of this type (sec.
2). The reactor core, as designed for
the ARE, consists of a beryllium oxide
moderator with a multipass fuel-coolant
system. The core and a surrounding
beryllium oxide reflector are contained
in an Inconel pressure shell. Design
of the reactor, fluid circuits, build-
ing, and assoclated equipment are
essentially complete. The reactor is
expected to be in operation early in

1953, ‘

The developmental work in reactor
plumbing and associated hardware has

been primarily concerned with the

technology of high-temperature fluoride
mixtures, and a secondary effort has

been the study of liquid metals (sec.

INTRODUCTION

3). The techniques of the prepara-
tion, purification, and handling of
the fluoride mixtures have been
developed so that 100-1b batches of
the treated fluoride may be prepared
and loaded 1in adequately cleaned test
equipment. Techniques of pumping,
sealing, and controlling the fluoride
coolants and lubricating moving parts
of the systems have been demonstrated
at temperatures above 1300°F, and it
is considered that these techniques
are adequate for ARE application.
A centrifugal-flow fluoride pump has
operated for weeks with neither me-
chanical failure nor leakage. Liquid
sodium technology appears to be well
in hand, since continued success has
been experienced in the operation of
sodium (or NaK) pumps, seals, and heat
exchangers. The NaK-to-NaK heat ex-
changer loop has now operated for 2300
hr with a maximum temperature of 1500°F.
Gross heat transfer studies indicate
that space-economical systems and
components can be built to handle
copious quantities of heat, as required
by fluoride systems, at temperatures

between 1200 and 1800°F.

The reactor physics calculations,
which have further defined the statics
of the circulating-fuel ARE, have led
to some general observations regarding
the kinetics of both the circulating-
fuel ARE and ANP reactor (sec. 4).
Although the thrombosis effect is an
important concern in the control of
these reactors, the loss of the delayed
neutrons may not be i1f the circulation
of the fuel itself 1s as good a damping
mechanism as now indicated. These
kinetic difficulties are of less con-

" cern to the ARE than to the ANP, since

the circulation rate in the ARE is so
slow that the control rods can cope
with the thrombosis effect and a large
fraction of the delayed neutrons are

T

 

 
g s

 

e e b et

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

emitted into the active volume. The
current ARE design has a critical mass
of 22.3 1b, a total uranium investment
of 74 1lb, 71% thermal fissions, and a
leakage-to-absorption ratio of about
1 to 3. Brief studies of hydroxide
moderated reactors (including KOH,
LLiOH, NaOH, RbOH, and SrOH) show that,
except for KOH, the hydroxide moderated
reactors require low critical masses
and small core volumes for minimum
critical mass.

Measurements on the critical ex-
periment of the simulated General

Electric direct-cycle reactor have
been completed and the simulated cir-
culating-fuel reactor is now being
assembled (sec. 5). Evaluations, in
terms of contributions to reactivity,
have been made of several reflector
modifications of the direct-cycle
assembly. In addition, the data from
the earlier graphite reactor assembly
have been correlated with the data
from theoretical calculations of the
assembly. The correlation lacks
precision but gives results that are
at least consistent with the experi-
mental facts.

 

 

 

 
 

 

 

L]

n

%

-y

FOR PERIOD ENDING MARCH 10, 1952

1. CIRCULATING-FUEL AIRCRAFT REACTOR

A. P. Fraas, ANP Division

A circulating-fuel aircraft reactor
system in which the fluid fuel circu-
lates directly through the turbojet
radiators was described in the last
report.{!) Other circulating-fuel
aircraft reactor systems, 1ncorporating

such features as 1intermediate heat

exchangers, various secondary coolants,
the use of liquid moderators, and
heavier reactor shielding, have been
considered in an attempt to determine
the most practical system for a func-
tional supersonic aircraft, ~All these
systems utilize the fundamental ad-
vantage of the bifunctional fuel-
coolant — the elimination of a heat
transfer stage within the reactor core.

~ Two series of shielded, full-size,
circulating-fuel aircraft designs have
been studied; one involves an annular,
or "wrap-around,"”" type of heat ex-
changer, and the other is a tandem
arrangement with the heat exchanger
behind the reactor. Analyses show
that the annular arrangement gives
the lower shield weight. These design
studies brought forth the important
technique of lacing the heat exchanger
matrix with about 8 vol % B4C to
keep the radiation from sodium or
NaK in the secondary circuit to
tolerable values even with something

~approaching a unit shield.

It also appears from these studies

that 1t would be advantageous to use
a liquid moderator such as water or
fused hydroxides. Not only would the
problem of cooling the moderator and
reflector be greatly simplified, but
by using perhaps a 12-in.-thick re-
flector, the problem of heating of the

 

()R, W. Schroeder, *“Circulating-Fuel Aircraft
Reactor,” Aircraft Nuclear Propulsion Project
Quarterly Progress Report for Period Ending
December 10, 1951, ORNL-1170, p. 7.

pressure shell by gammas, neutron
captures, and inelastic scattering
would be greatly reduced. More de-

‘tailed studies of the practicability

of design for cooling the structure
with a liquid moderator have been
initiated.

REACTOR WITH TANDEM HEAT EXCHANGERS

The first basic configuration con-
sidered was one in which the reactor
core and the intermediate heat ex-
changer were placed in tandem. A
longitudinal section of a tandem re-
actor and heat exchanger arrangement
is shown in Fig. 1. A basic premise
was that by interposing the heat ex-
changer between the pumps and the re-
actor core, a uniform flowdistribution
among the fuel tubes would be assured.
The design required that the liquid
fluoride fuel flow through the tubes
in the reactor core at 11 ft/sec for
operation at 400,000 kw with a tempera-
ture rise through the reactor of 400°F.
The basic layout is best adapted to a
liquid moderator (i.e., H,0 or NaOH)
from the standpoint of fabrication and
of means for cooling the moderator
and reflector.

In this design, the fuel enters the
reactor at the top rear, makes a com-
plete loop through the fuel tubes in

.the core, and discharges to the heat

exchanger. The fuel tubes are of
stainless steel with 1 1/2-in. ID and
0.015-1n. wall thickness. If water
were the moderator, a double-walled
construction could be used. It may be
noted that if the inner tube (con-
taining circulating fuel) ruptured or
cracked, the only fuel lost would be
the amount that filled up the space
between tubes before it froze in
contact with the cold outer tube.

e T T T R p T e e o ™

Gl

 
=“SECREF
DWG, E-Y-F7-47T1R{

   
 
    
   
 
   

APPROX. 42 in. H0 ARCUND HEAT
EXCHANGER AND PUMP HOUSINGS

Y, -in. STEEL PLATE FOR REACTOR AND
SHIELD SUPPORT, 2 REQUIRED

H,0 MODERATOR HEAT EXCHANGER

INLET PIPE

   

T .

ACTIVE ‘
LATTICE

 
   

H,0 MODERATOR
PiPE OUTLET

   
 
   
 
    
 
   
 
  

  
    
   

SECONDARY
COOLANT
LINES

 

FUEL PUMP
HOUSING

  

i-in. THICK SHELL

CIRCULATING AIR AND
THERMAL EXPANSION GAP

 
 

Y4 -in. THICK

 

FUEL EXPANSION
TANK, 4 REQ'D

  
 
   

  

¥ -in. THICK ALUMINUM

'/Z-in. MINIMUM H,0 LAYER AT 300-350 psi AROUND
ENTIRE REACTOR AND HEAT EXCHANGER CASTINGS

Fig. 1. Tandem Heat Exchanger Arrangement for Circulating-Fuel Reactor.

e

LH0da¥ SSTY90Ud ATALIVND IDAf0dd ANV
 

 

The water (or hydroxide) moderator
enters the active lattice around the

periphery of the rear of the reactor .

and flows toward the outlet at the
forward end of the reactor. As the
moderator flows in it is distributed
by a baffle sheet that is orificed

"to give the flow distribution required

by the power distribution within the
active lattice.

The control system for this proposed
circulating-fuel reactor design con-
sists of two curtains of cadmium rods
mounted on two endless tracks. The
curtains are moved from the reflector
into the active lattice by an endless

~chain-sprocket type of mechanism.

Each cadmium rod is mounted upon two
steel shafts with steel rollers attached
to the ends of the shafts. These
rollers are linked together to form
an endless chain. The sprocket 1is

driven by a worm-gear drive, which in

turn is driven by a hydraulic pump
mounted within the pressure shell.

Each cadmium rod is a cylinder ap--

proximately 1/2 in. OD by 1/4 in, ID
by 12 in. long, and there are about 50
rods in each control curtain.

REACTOR WITH‘A‘NNULAR’ HEAT EXCHANGERS

The second basic configuration
considered was a reactor with an
annular, or "wrap-around," type of
heat exchanger such as shown sche-
matically in Fig, 2. Radial webs are
used both to separate the annular heat
exchanger into sectors and to support
the shell around the outside of the
reflector. To permit assembly, the
reactor pressure shell would have to
be split axially, probably on the
center line. The pressure shell and
other structures are jacketed and

cooled by NaOH at around 1200°F, A

heat exchanger to cool the NaOH is
provided in the region where the NaK
enters the pressure shell. Although
not an essential element in this general

FOR PERIOD ENDING MARCH 10, 1952

!type of design, the concentric helical

coil carrying the NaOH moderator
through the reactor core is interest-
ing because it eliminates a header
problem. By using perhaps ten tube
fittings, welds at the ends of these
coiled tubes could be avoided so that
a material such as molybdenum might
be used.

REACTOR SHIELD DESIGNS

It is important to know the effects
of various parameters on the weight
of the aircraft reactor and shield,
but it 1s difficult to determine these
effects quantitatively since the pres-
ent knowledge of shielding does not
permit weight estimates closer than
+5%, at best. However, engineering
designs have been completed for the
shields for the preceding reactor
arrangements, and real weight differences
that result in lighter weight shields
for the annular arrangement than for
the tandem arrangement have been found
to exist. In the shield weights cal-
culated the use of B,C in the heat
exchanger and between the reactor and

" heat exchanger has been found to sub-

stantially reduce the activity of the
secondary coolant.

Design Procedure. The primary
objective in the design of these
shields was to minimize the activation
of the secondary coolant by delayed
neutrons from the fuel. This has been
effected by (1) attenuating the neutron
leakage from the active lattice to a
level below that of the delayed neu-

‘trons in the heat exchanger, and

(2) lacing the heat exchanger with
enough neutron absorbing material
(B,C) so that relatively few neutrons
would be absorbed in the secondary
coolant. The header sheet and a 5-in.
B,C layer between core and heat ex-
changer was sufficient to reduce the
neutron leakage flux to the heat ex-
changer, whereas the use of B,C in

U

 
 

 

 

   
 
  
  
  
   
 

 

 

  

 

 

 

 

 

    
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

sy L i i o i s o e i s i ok i ER T - N ‘.
\ .
.-—I
o
SECREE
: DWG. E-Y-F7-2i3aR1
- NOK ]NLET SIS NN o VNN SN AN NQK OUTLET
Y» B4C CANNED WITH 0080 STEEL .
/ Yo NaOH I/*/2 STEEL R
NaOH -NoK HEAT EXCHANGER { ] S ; HEAT EXCHANGER
%8 OD TUBES ON %» ¢ - ¢ % —_ \ OUTLET MANIFOLD
SQ PITCH HELICAL GOIL GOOLANT TUBES ,,,,’\
1.250 OD TUBING \ N Y STEEL
27.000 | \ Al
60 ST | | g\‘ 5 B4C
o 0080 STEEL =1t} %\\ }y STEEL
£ 0 0O00BOI 6\ Y% B,C CANNED IN
g d “ 2 Bg
D 0 q DD = B 0.060 STEEL
T O ‘\
1 A g ?\
N b\ b b‘q ¢ 2 PRESSURE SHELL
b 3?*95? oL 9@9@3 o 1 Lok
NAOH PUMP E \ Y SN
L,
4q9qqdqaqadag SN
O ¢
QaQqqoaag S AN
Z ’\\
9 - AN Y% B,C CANNED IN
Q O 0.080 STEEL
Py '/2 NaQH
% STEEL

 

 

NeOH OUTLET TO
NoOH-NaK HEAT EXCHANGER

 

 

 
    

10% CIRCULATING
NaOH REFLECTOR

O S
R~
= RN
o = W,
s
WY = )
TR LY = R
\ Y T~ \‘\
WALy B3 ey
3 v, =
\ \
N B
MY g
\\\'\\\\ = =
Y

S S
B = 6 ANNULAR TYPE HEAT

 

 

 

HEAT EXCHANGER
INLET MANIFOLD

 

HEAT EXCHANGER OUTLET

 

 

 

— LINE, 6 REQ'D
3 O\ = EXCHANGER Y% 0D TUBES =
o B 5 _ = X >
NaK INLET \” =M—--= o /32; & 3 PcH = %y NaK OUTLET

 

NOTE:
DIMENSIONS ARE IN INCHES

Fig. 2. Annular Heat Exchanger Arrangement for Circulating-Fuel Reactor.

 

 

140d3¥ SSIY90uUd ATHALYVAO 1Daf 0d4d dNV
 

 

every ninth tube in the heat exchanger
served to supress the activity of the
circulating fuel.

The balance of the shield design

work was straight-forward and followed
the procedures outlined by the Shield-
ing Board.(?? Both Lid Tank and Bulk
Shielding Facility test data were used.
The fission-product-decay gamma rays
for the equilibrium fission-product
concentration at full power were taken
as being equivalent to three 1.0-Mev
gamma rays and one-half of a 3.0-Mev
gamma ray per fission.

Activation of the Secondary Coolant.
The degree of activation of the second-
ary coolant is, of course, a function
of the materials 1in that coolant.
Since NaK (56% Na and 44% K) has many
advantages as a high-temperature heat
transfer medium, and since 1t would
probably be about as bad from the
activation standpoint as any coolant
that could be used, an estimate of its
activation was made with the heat ex-
changer shielded and laced with B,C as
outlined above. It was found that 1f

- 25% of the NaK in the system were

considered as concentrated at a point
and the self-absorption of the sodium-
decay gamma rays released in the air

radiator or other similar component

 

(Z)Report of the ANP Shielding Board for the Air-
craft Nuclear Propulsion Program, NEPA-ORNL, ANP-353,
October 16, 1950.

FOR PERIOD ENDING MARCH 10, 1952

were assumed to be 50%, the dose at
1.5 meters (5 ft) from the center of
the source would be 200 r/hr after
long periods of operation at a reactor
heat output of 400,000 kw.

The activation of a quite different
coolant, a fused salt consisting of
approximately 60% LiCl, 30% MgCl,,
and 10% KCl, was also estimated. Al-
though with no B,C in the heat ex-
changer the activity of this fused
salt was only 5% of that of the Nak,
the activity for the case with the B,C
in the heat exchanger was 35% of that
of the NaK. Thus, it appears that
there is little to be gained from the
shielding standpoint in the use of
this molten chloride in place of Nak.

Shield Weights and Specifications.
Several engineering designs of shields
have been completed for both the tandem
and annular heat exchanger arrange-
ments. For either reactor arrangement
the divided shield resulted in the
lowest shield weight. However, the
desire to be able to carry out a
limited amount of airplane maintenance
work without special shielding argues
in favor of more shielding around the
reactor. The various shield designs
for the two reactor and heat exchanger
arrangements are given 1in Table 1.
For comparable shields the annular
heat exchanger and reactor assembly
gives appreciably lower weight than
the tandem arrangement.

11

 
¢l

 

T G s i g

TABLE 1

-Shields;for,Circuiating~Fluoridé-Fuel'Reactors

 

 

 

- TANDEM HEAT EXCHANGER PARTIALLY DIVIDED

ANNULAR HEAT EXCHANGER
PARTIALLY DIVIDED

 

 

1 2 3 4 1 2

Details in report Y-F15-10 Y-F15-10 Y-F15-10 Y-F15-10 Y-F15-10 Y-FI5-10
Date of design Jan. 1952 Jan. 1952 Jan. 1952 Jan. 1952 Feb. 1952 Feb. 1952
Reactor shield diameter (in.) 150 150 121 121 148 118
Crew shield weight (1b) 5,000 11,000 36,000 14,000 5,000 36,000
Weight of reactor, intermedi- 151,000 130,000 75,000 75,000 123, 000 62,000

ate heat exchanger, and

reactor shield (lb
Total weight of reactor, in- 156,000 141,000 111,000 89,000 128,000 98, 000

termediate heat exchanger, :

and all shielding (including

crew shield) (lb% ' _
Reactor power (kw) 400, 000 400,000 400,000 400;000 400, 000 400, 000
Diameter of reactor core (in.) 32 32 32 32 32

Liquids in primary and second-
ary circuits

Temperature loss in ingermedi-
ate heat exchanger (°F

Pressure loss in intermediate
heat exchanger (psi)

Crew shield size (ft)

Beactor-crew separation
distance (ft)

Radiation inside crew com-
partment (r/hr)

Radiation 5 ft from center
of reactor {(r/hr)

Radiation 50 ft from center
of reactor (r/hr)

. Radiation 300 ft from center

of reactor (r/hr)

 

Fluoride-NaK
100

100
6% X T% x 12%

50

300

 

32
Fluoride-NaK
100

100
6% X T4 x 12%

50

2,400

36

 

Fluoride-NaK
100

100
6% X % x 12%°

50

1
350,000 |

5,600

156

 

Fluoride -NaK
100

100
5 X5 x 12%

120

1
380, 000
5,600

156

 

.Fluoride-NaK

100

50
6% X T% x 12%

50

300

 

Fluoride-NaK
100

30
6% X Th x 12%

50

380, 000
5,600

156

 

1Y0d3¥ SSAYO0uUd X THIALMVND lDHrOHd dNV
 

 

'FOR PERIOD ENDING MARCH 10, 1952

2. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT

R. W. Schroeder, ANP Division
E. S. Bettis, Reactor Projects Division

- The studies of the circulating-fuel
aircraft reactor using fused fluoride
have culminated in the design of two
aircraft reactor systems (one with and
one without an intermediate heat ex-
changer), both of which are potential
power plant systems for supersonic
nuclear propulsion, Furthermore, these
aircraft reactor systems appear to
permit higher performance than the
sodium-cooled reactor, primarily be-
cause of the bifunctional capacities
of the primary coolant, which is the
circulating fuel. Consequently, the
~ ARE to be constructed by the Oak Ridge
National Laboratory at Qak Ridge will
be a circulating-fuel type of reactor,

The circulating fuel of the reactor
will be a mixture of molten alkali
metal fluorides plus uranium fluoride,
- The structural metal of the core and
container shell will be Inconel, since
this metal has been on order for the
ligquid-metal ARE and currently appears
at least as corrosion resistant to the
circulating fluorides as any of the
stainless steelsor The reactor c¢an
advantageously be moderated with
beryllium oxide, which is also cur-
rently available for the previously
contemplated liguid-metal ARE,

The power is generated within the

circulating fuel as it is passed
through the beryllium oxide moderated
core. The core is provided with a
beryllium oxide side reflector, which
is housed with the core within an
Inconel pressure shell. The reflector
and pressure shell are cooled by a
separate circuit using a mixture of
nonuranium-bearing fluorides, Power
is abstracted from the fuel by means
of four fuel-to-helium heat exchangers
through which the fuel is circulated.

The heliumis cooled by passage through

four helium-to-water heat exchangers
and the water - the ultimate heat
source — is discharged.

Control of the reactor is provided
in three forms, (1) shim control, (2)
regulating rods, and (3) safety rods,
Shim control is achieved by varying
the uranium concentration in the circu-
lating fuel. One boron rod, which
passes through the core center lineand
effects approximately 0.75% Ak/k,

serves as the regulating rod. The
boron safety rods, equally spaced
about the.center of the core, have

approximately 5% Ak/k per rod.

CORE DESIGN

The core design is illustrated by
Fig. 3. It may be noted that the
over-all assembly includes an Inconel
pressure shell in which beryllium oxide
moderator and reflector blocks are
stacked and through which fuel tubes,
reflector coolant tubes, and control
assemblies pass. The innermost region
of the lattice is the core, which is a
cylinder 3 ft in diameter and 3 ft
long. The core is divided into six
60-degree sectors, each of which in-
cludes 13 vertical stacks of hexagonal
beryllium oxide blocks. FEach sector
includes one serpentine, fuel-tube
coil, which passes through the 13
moderator stacks in series, as 1l-
lustrated., The six serpentine coils
are connected in parallel by means of
external manifolds.

A reflector with a nominal thick-
ness of 6 in. 1is located between the
core and the pressure shell on the
cylindrical surface only. The re-
flector consists of beryllium oxide

i’ "‘:’ 13

Wl g

Y e
& :6}' £y e
PENE

..

 
EhE s gdst: gl b mpldh Bk So ot e podt e G LSl SER R G i ESArr iR i e L byl e e e s s e g il oo i o s i b R

 

 
  
 
 
  
 
 
  
  
   
   
 
 

‘__. . fiSREfl‘
= _ DWG. E-Y-F7-202 Ri
WELD BACK-UP RING
BOTTOM
g?SE c%Fr,aFéORT : CORE AND REFECTOR COOLANT
INLET (SHOWN ROTATED 30 DEGREES -
FROM TRUE POSITION )
REINFORGING PLATES
e FUEL OUTLET
s MANIFOLD
o ASSEMBLY
~
£

CONTROL ROD

‘ BLY
ASSEM THERMAL INSULATION

CORE AND REFECTOR
COOLANT QUTLET PIPE
{ ROTATED 30 DEGREES
FROM TRUE POSITION

FUEL INLET
MANIFOLD

1Yg FUEL TUBE WELDING JOINT

  
 
   
  

FUEL TUBES

1.235 0D, 0.060 WALL
BOTTOM TUBE .

SHEET, 4 THICK TOP HEAD

TOP TUBE SHEET

   

CORE AND REFLECTOR
CAN, Y1g THICK

   

COOLANT ANNULUS,

REFLECTOR COOLANT TUBE
g THICK

0500 0D, 0.020 WALL

    
 

NOTE:
ALL DIMENSIONS
IN INCHES

35.250 ACTIVE LATTICE

2 7.500 34.500 - 1
| 48.500

 
  

 

 

 

 

 

Fig. 3. (Circulating-Fuel ARE Core Design.

Riin e S ol oo o i e sl ity bERE { ;o

140434 SSAYI0Nd ATHALUVND IDAC0Ud dNV
 

 

wd

.}

8

blocks similar to the moderator blocks
but with}-in, holes for the passage of
reflector coolant, The reflector
coolant enters at one end of the
pressure shell, passes through the
reflector, bathes the pressure. shell
and fills the moderator interstices,
and exits at the other end of the
pressure shell,

Primary Coolant Circuit, Hydro-

dynamic and thermodynamic considera-

tions have influenced the design to
such an extent that a discussion of
these problems appears to be desirable
prior to description of the primary
coolant circuit itself, In a circu-
lating-fuel type of reactor, heat 1is
generated in the flowing fuel. The
particle heat generation is dependent
on the particle residence time within
the reactor., As the flow proceeds
through the tubes, the velocities
adjacent to the walls are lower than
the bulk mean velocities, so that the
fuel particles adjacent to the walls
are subjected to higher temperature
rises., 1Tlhis effect is mitigated by
convection within the fuel stream, but
a temperature difference will exist
between the wall and the stream
center line, Studies have indicated

~that this temperature difference is

directly proportional to the fuel
power density, tube diameter, Prandtl’s

~number, and Reynold’s number. Appli-

cation of these studies to various
postulated ARE configurations have
established that excessive temperature
differences are encountered in the

laminar flow region (Reynold’s number

below 5000) but that small temperature
differences are encountered in the
turbulent flow region. Consequently,
the reactor core was designed for a
Reynold’'s number of 10,000, thereby
allowing for some deviation of flow
rate or fluid physical properties.

The design for a straight-through
flow core arrangement was revised
because the calculated Reynold’'s

FOR PERIOD ENDING MARCH 10, 1952

number was less than 1000, Con-
sideration of the parameters for in-
creasing Reynold’s number indicated
that the desired correction would
entail increasing tube diameter, in-
creasing the number of tubes in series,
increasing the volumetric flow rate,
or decreasing the volume of fuel with-
in the core, or combinations thereof.
Investigation of these variables led
to the observation that a few series
passes would be required even with
maximum feasible exploitation of the
other variables. Since this was the
case, and since the design problems
involved in the use of many series
passes appeared to be similar to those
associated with the use of a lesser
number of series passes, it was decided
to use 13 series passes and avoid
compromising the volumetric flow rate
or the core fuel volume.

Secondary Coolant Circuit. The use
of a liquid circuit other than the
fuel circuit within the pressure shell
is desired for the following reasons:
(1) to maintain the pressure shell
essentially isothermal at a tempera-
ture of approximately 1150°F; (2) to
cool the beryllium oxide reflector
with a fluid other thanm circulating
fuel:; and (3) to fill the beryllium

"oxide moderator interstices and other

internal voids with a liquid maintained
at a pressure higher than the fuel
pressure to prevent a fuel tube leak
from adding reactive material to the
core, The ligquid to be used has to be
compatible with the structural material
under dynamic conditions and should
cause no undesirable reaction with
the fuel in the eveht of a leak in the
separating wall. At the present time
it is believed that nonuranium-bearing
fluorides, perhaps the fuel carrier
without the UF,, will best meet these
specifications,

The possibility of cooling the

moderator by positive flow of this
fluid through the moderator interstices

15

 
 

 

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

has been studied. Yt has been found,
however, that the mass flow through a
gapof constant periphery varies as the
third power of the gap width (in the
laminar flow region, with fixed
pressure drop), Temperature analyses
- with postulated gap width distributions
indicated that gross temperature
inequalities could exist throughout
. the moderator because of tolerance
accumulations affecting interstice
widths. Accordingly, it has been
decided to obstruct flow through the
moderator so as to render it virtually
stagnant. The moderator heat then is
conducted to the fuel stream, and thus
a relatively accurate analytical
determination of moderator tempera-
tures can be obtained.

Core Temperature Distribution. The
circulating fuel in the ARE core flows
in parallel through six serpentine
“tubes, each tube traversing the core
13 times. In passing through the core

13 times, the fuel is heated from
1150 to 1500°F., The calculated mixed

mean fuel temperatures at various

stations in the reactor are given in

Table 2. | '

| TABLE 2

o Fuei'Temperaturé after each Pass
Through Core

_ MIXED FUEL
STATION TEMPERATURE (°F)

Entering reactor 1150

Leaving pass No. 1 1171
| ' 2 1191
3 1211

4 1233
-5 1257
6 1281

7 1305

8 1335

9 : 1365

10 1392
11 1426
12 1462
13 1500

16

There are two sources of heat for
the fuel circulating through the
reactor: 1nternal heat generation
(from fission and gamma-ray absorption)
and heat transferred to the fuel from
the remainder of the core. The heat
transferred from the remainder of the
core 1s produced in the following
manner. Heat is generated in the
moderator and the parasitic core
material as a result of gamma-ray
absorption, and additional heat 1is
generated in the moderator by neutron
slowing down. All this heat is trans-
ferred to the circulating fuel and
results in a temperature gradient
across the fuel, moderator, and
parasitic material,

EXTERNAL FLUID CIRCUIYT

The ARE fluid circuit, intended to
handle toxic and corrosive fluids at
1500°F, requires a considerable amount
of design and developmental effort.

‘Where possible, commercially available

components are used, but pumps, heat
exchangers, and certain other com-
ponents are being constructed es-
pecially for the ARE. The location of
the major system components, including
the reactor pits, the heat exchanger
room, the reactor, and the two heat
disposal loops, is 1llustrated in
Fig. 4. It may be noted that the heat
exchanger room is shielded from the
reactor pits to permit servicing after
fuel drainage and flushing. The

"shield was designed to minimize

activation of fluid circuit structure
by reactor neutrons during power
operation and to attenuate reactor
post- shutdown gamma rays to a level
permitting access to fluid circuit
components after shutdown.,

Pumps. The pumps to be used in the
fuel and moderator coolant circuits
are vertical-shaft, tangential-dis-

"charge, centrifugal pumps in which a

 

B T

 

T

 

 
LT

 

UNCL ASSIFIED
DWG. D-Y-F7207R4

 
         

; ‘.:".‘ .:. :. - : Y .... 5 K . . i :'_ it ._.. S _., .: ° T

W

 

    
 
 

 

INERT SALT
COOLING
SYSTEM NO. 4

   

   
  
   

  
 

  

He TO H,0 EXCHANGER

PRIMARY COOLENT

SYSTEM NO. 1

  

  
 
 

SOURCE

   
 
 
   
   

——
- — . — g —— S—

  
  
  
 

  
 
  
 

   
     

   

EXCH. »

 

TEST PIT | } } £
<

HEAT =

He TO H,0 EXCHANGER l w

ROOM 4

o

{
!
i
|
|
it
|
s
I
| |
|

  

COOLANT TO He EXCHANGER

  
 
 

 

PRIMARY COOLENT
SYSTEM NO. 2

 
 

LEGEND
— e o e COOL ANT FROM SOURCE
COOLANT TO SOQURCE

  
   
 
 
 

—— ———— 1

 

 

   
 

 

 

     

 

 

 

 

 

 

— //== INERT SALT FROM SOURCE o
% /=== INERT SALT TC SOURCE S
00 00— HELIUM FROM RODS S
o O == HELIUM TO RODS @
— W ——— WATER IN a
——wo wo ~— WATER OUT v
—_——— OIL N S
—— — — — 0L OuT o

   

ROD COOLING
BLCWERS

 
 

   

INERT SALT
COOLING
SYSTEM NO. 2

 

 
 

 

 
       

 

-

  

Fig., 4. Arrangement of External Fluid Circuit Equipment.

‘0T HOYVI HNIGNA doIydd Hod

cS61

 

T N P T R T R T Ty T T TR ET R
 

 

-

ANP PROJECT QUARTERLY PROGRESS REPORT

gas seal 1s used and the liquid level
is maintained. The gas seal is formed
by a floating graphite ring between

 the stationary nose on a bellows seal

and the hardened rotating nose on the
shaft. The primary pump seal carries
no appreciable pressure differential
and, consequently, requires that the
oil-circulating system above the seal
be pressurized to balance that of the
circulating system below the seal. The

‘circulating oil serves to cool the

shaft, bearing, and lower seal. In
addition, internal cooling of the
structure is obtained by the circu-
lation of helium. Another feature of
the pump is that the entire rotating
assembly can be put together outside
the pumping casing, and in event of

a failuqe it can be installed with
comparative ease,

fieat Exbhangérs. The fuel-to-

‘helium heat exchangers are on order

with the Griscom-Russell Company; they
were designed by Griscom-Russell to
ORNL specifications. The heat ex-
changers are of the cross-flow type
with the fuel flowing through 1-in.-0D,
0.109-in. wall Inconel tubing. The
tubing is finned on the gas side with

'stainless steel, helically wound
" strips. Each heat exchanger includes

five parallel tube banks; each bank
consists of seven to nine horizontal
tubes in series with integral return

~bends. The fuel inlet and outlet

headers are constructed of 4-in. pipe

to which the five tube banks are
Weldedf

- The helium-to-water heat exchangers
also are finned-tube, cross-flow ex-
changers, but steel tubes with copper
fins are used.

Primary Coolant System. Fuel flows
from the six core tubes to a common
header, and a common trunk line conveys
the fuel into the heat exchanger room.
The fuel then divides into two com-
pletely independent loops, each

18

capable of dissipating 1500°F kw,
Within each loop the fuel divides into
two parallel 750-kw fuel-to-helium
heat exchangers and then passes through

a centrifugal pump to a common return

trunk (Fig. 4).

In an individual loop, helium flows
from a centrifugal blower through a
fuel-to-helium exchanger in which the
temperature of the helium is increased
from 250 to 750°F, The helium is then
cooled to 250°F in a helium-to-water
exchanger from which it passes through
another heating and cooling cycle and
is returned to the blower. This so-
called "double-sandwich" arrangement
permits two helium heat transports per
cycle and thereby halves the helium
flow rate required through the blower
and ducting for a specified power and
temperature rise, ' '

Secondary Coolant System. The
reflector and pressure shell coolant
system, as 1in the case of the fuel
system, is divided into two heat
disposal loops, each capable of
handling the load associated with
1500-kw reactor power, Heat 1is
conveyed from the primary coolant to
helium and then to the water sink,

Monitoring Circuit. All lines and
components containing fuel or reflector
cotlant are double-jacketed, with
helium passing through the annulus,
The helium pumping head 1is maintained
by drawing helium from the system,
cooling 1t, and admitting it to rotary-
type compressors at various poilnts in
the system as indicated. Monitoring
for fluoride leaks into the helium is
achieved by passing helium samples
through halogen detectors,

PREHEATING SYSTEM

The relatively high melting points
of the circulating fuel and secondary
coolant (around 752 to 932°F, depending

T I T TR

T o

 

 

At S o

TR T
 

 

[y

 

F L)

N

”

-1y

upon the particular fuel composition)
require that all equipment within
which these coolants are to be main-
tained be preheated to permit loading
and unloading,

"Reactor Assembly. The insulation

heat leakage at operating temperature
has been calculated to be approximately
15 kw, and the heat capacity of the
assembly has been established as
approximately 4000 Btu per °F. After
having studied various preheating
methods, including pumping hot fluids
through the fuel passages or through
the moderator interstices, it was
decided that the simplest method
entailed the application of electrical
resistance heaters to the outside of
the pressure shell, withheat transport
to the inner regions by conduction.
With a constant heat input of 14 kw,
in addition to the insulation losses,
it is estimated that the assembly can
be heated from room temperature to
1200°F in approximately four days.
With this heat input rate, certain
calculated temperature gradients are
as follows:

1. Across pressure shell, 20°F,

2. Across helium gap between shell
inside diameter and reflector outside

- diameter, approximately 100°F,

3. From reflector outside diameter
to core center line, approximately

250°F.

Ppiping. The heat capacity of the

~ piping is small relative to the piping
~insulation leakage. Consequently, the
power required to preheat the piping
"is substantially the same as the power
required to maintain it at the ultimate

temperature. If the liquid-bearing
pipes are empty, as is the case during
preheating, the pipe resistance to
axial heat flowis high, and the appli-
cation of heat to specific points on

FOR PERIOD ENDING MARCH 10, 1952

the pipe surface tends to cause large
temperature valleys to exist between
the points of heat application. By
attaching electrical resistance heaters
to the outer pipe of the double-

‘walled piping, however, the flowing

helium acts as a thermal conveyor and
facilitates axial heat transport,
Accordingly, substantially isothermal
helium contacts the inner pipe, and it
is thus possible to heat the pipe
uniformly.

Heat Exchangers. The fuel-to-
helium heat exchangers require pre-
heating prior to filling and the
addition of heat to compensate for
thermal leakage during filling or at
any time after filling when the fuel
pumps are inoperative. These heat
exchangers have a total of approxi-
mately 40 ft? of exposed free-flow
area (both upstream and downstream),
which, if allowed to radiate to the
adjacent cold structure, would dissi-
pate approximately 40 kw when the
tubes are at 1325°F. Accordingly, it
appears necessary to include radiation
barriers in the form of gates that can
be lowered during warm-up or zero- or
low-power operation, When these
radiation barriers are in position,
the fuel-to-helium heat exchangers are
enclosed and may be preheated by
means of hot helium supplied by the
external piping annuli.

ELECTRICAL POWER CIRCUITS

Because of the extreme incon-
veniences associated with forced
reactor shutdown, fuel drainage, re-
filling, and restarting, the objective
has been to design the system so that
no one failure will necessitate a
forced shutdown. Critical pumps,
blowers, etc. are duplicated, and each
set is served by an independent d-c
circuit that includes independent

19

T

 
 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

buses, switches, and a-c-d-c motor-
generator sets, The use of direct cur-
rent for these circuits permits con-
venient speed control where required and
the use of battery sets floating on the
line to safeguard against outside
power failures. Alternating-current
instruments are fed from the d-c
circuits via d-c-a-c motor-generator
sets so that these instruments will
derive their power from the batteries
in an emergency.

- The heat capacity of the system 1is
large relative to the heat leakage
rate, so that no large temperature
loss will be incurred for about 6 hr
in an emergency. Consequently, it 1is
not necessary to use battery power for
system heat addition. System heaters,
therefore, are connected directly to
the 440-v a-c¢ circuit. Similarly, the
building crane and certain other items
will not need to be operative during
any short period in which it is neces-
saTy to use battery power, so these
components are supplied by the 440-v
a-c system. ‘ '

CONTROL SYSTEM

‘The ARE is controlled by three
essentially separate systems: (1)
regulating, (2) shim, and (3) safety.
A mechanical regulating rod is pro-
vided, since the time-constants of the
self-stabilizing effects of the fuel
expansion 1in the ARE are too long to
provide stiff control of the reactor.
Shim control is conveniently effected
by the addition of higher-uranium
concentration fuel to the circulating-
fuel volume.

Specifications for the control
console and panel have been released
with a specified delivery date of
October 1, 1952, The high-temperature
fission chamber has operated at 1292°F,
and an MIR-type servo for the regulat-
ing rod has been ordered., The kinetic

20

behavior is currently being analyzed
on the analogue computor.

Shim control.. Fuel enrichment will
be accomplished by adding enriched
molten fuel at the surge tank. The
molten concentrate will be introduced
through a small (%-in.) line, whach,
by means of a video pickup device, can
be observed from the control room
while fuel additions are being made,
The fuel-enrichment mechanism has been
laid out in elementary form. The
principles of operation have been
established, but the tanks, weighing
devices, valves, heaters, etc. have
not been detailed. This system 1is
much simpler than was first thought
possible and it makes possible a safe
and relatively easy technique of
bringing the reactor to critical after
the initial loading.

Regulating and Safety Rods. Because
of the low power density of the circu-
lating-fuel ARE, the temperature
coefficient of the fuel must be
supplemented by a servo-actuated
regulating rod if transient conditions
are to be controlled, This servo will
operate in exactly the same manner as
the fuel temperature coefficient. The
error signal that actuates the servo
will be a mixed signal of temperature
and neutron flux; it 1s given by the
equation '

€= (8, -6, ;) *58.5(p-py,

T

where
€ = error signal,
O, -6, ¢~ reactor inlet error in
OF,
(p - pO) = reactor power error in

megawatts (the factor
58.5 is in °F per mega-
watt).' :

 

 

 

 
 

 

i)

When € < 0 the servo withdraws the rod,
and when € > 0 the servo inserts the
rod.” This servo equation is different
from that previously reported as a
result of a change in reactor design.

A change has been made in the
location of the drive mechanisms for
the regulating and safety rods. They
will not be located over the reactor

pit, and they will have straight

linkages to the rods. This change
will allow considerable simplification
of the over-all mechanical control
system, |

An MTR-type servo has been ordered
for the regulating rod. The regulat-
ing rod and safety rod designs have
been determined and the fabrication of

" these rods has been turned over to the

Metallurgy Division. After fabri-
cation, the rods will be run in the
"cold" critical assembly to determine
their worth, Calculations indicate
that the safety rods are worth approxi-
mately 5% Ak/k, and the regulating
rod will be loaded so as to be worth
about 0.75% Ak/k. One safety rod
operating mechanism has been finished
by the machine shop. This assembly
has not been tested, but the design
seems to be satisfactory.

High-Temperature Fission Chamber.
The fission chamber has been operated
at 700°F but no insulator material has
been found that will withstand higher
temperatures. A design is in progress
that eliminates the insulator in the
high-temperature region. This chamber
has not been tested and there is no
certainty that 1t will function
properly. Because of the uncertainty
of high-temperature operation of the
fission chambers, sufficient cooling
is being provided to maintain the
chambers at 400°F, where it 1is known
they will function.

FOR PERIOD ENDING MARCH 10, 1952

Control Console and Panel. Complete
specifications for the components of
the control room have been released,
including detailed construction draw-
ings and fabricational and material
specifications., Sets of these specifi-
cations have been mailed to ten pros-
pective bidders, and the bids are to
received by March 1. The instructions
to bidders specified that the complete
order is to be filled by October 1,
1952,

The items covered in these drawings
and specifications include an operating
console, instrument racks, relay racks,
recorders for nuclear measurements,
amplifiers, power supplies, and
assorted mounting hardware, Items not
covered in this outside order are
nuclear chambers, some preamplifiers,
servos, and process instruments. The
order takes care of from 90 to 95% of
the electronic equipment needed for

the ARE.

Reactor Dynamic Computer. The
entire circuit of the ARE has been
put on the analogue computer, and
the kinetic behavior of the system 1is
being analyzed., The analysis is not
yet complete, but the work 1is pro-
gressing satisfactorily and indi-
cations are that a fairly compre-
hensive study will result from this
work, This computer analysis has
received the full attention of two
engineers for the past eight months.,

INSTRUMENTATION

A basic purpose of the ARE is the
acquisition of experimental data, so
the importance of complete and reliable
instrumentation cannot be over-
emphasized, Most ARE process in-
strumentation is intended to observe
and record rotational speeds, flow
rates, temperatures, pressures, oOr
ligquid levels. 1In the low-temperature

21

-~ e e g

T

 
 

 

 ANP PROJECT QUARTERLY PROGRESS REPORT

loops, this equipment is sufficiently

conventional to obviate the need for

detailed descriptions., The high-
temperature loops involve special
sensory problems, however, which may
merit some discussion, In some
instances several alternate sensory
principles are under development con-
currently, No attempt will be made to
describe each of these alternate
methods in detail in this report;
“however, the various technigues that
are being used or developed are out-
lined in the section on "Experimental
Reactor Engineering" (sec., 3).

22

BUILDING

The building to house the ARE is
proceeding on schedule., Additional
contracts have been negotiated with
the contractor to complete work not
specified in the original contract,
and this new work is to be completed
so that the building can be released
to OBNL by June 1, 1952. The auxiliary
power specifications have been com-
pleted and orders for the equipment
are being released. An elementary
electrical diagram for this equipment
has been drawn,

 

 

 
 

 

Y

FOR PERIOD ENDING MARCH 10, 1952

3. EXPERIMENTAL REACTOR ENGINEERING

H. W. Savage, ANP Division

Liquid metals, hydroxides, and
mixed fluorides are being investigated
as heat transfer media and fuels for
aircraft reactor experimentation at
temperatures of 1200 to 1800°F. Some
developmental effort on the use of
sodium and sodium-potassium alloy
continues, although the effort on the
technology of molten fluoride mixtures,
as required by the circulating-fuel
reactor, predominates. The develop-
ment of fluoride systems 1s limited
by the corrosion problem (see sec. 11
"Corrosion Research”). Chemical and

_physical treatment of fluoride com-

ponents to eliminate contaminants
and improving handling, storing, and
transfer techniques to avoid reintro-
duction of contaminants are being
studied as means of limiting corrosion.

The applicability of known methods
of pumping, sealing, controlling, and
measuring properties and quantities
of these high-temperature coolants and
fuels is being investigated. The high-
temperature reactor systems, however,
place unique restrictions on materials,
lubricants, leakage, and performance
of mechanical devices and associated
equipment. Pumping has been accom-
plished with conventional hydraulic
designs, but alleviation of thermal
distortions and stresses, cooling of
bearings, and the development of
liquid- and gas-tight seals have been
required. Alleviation of thermal dis-

tortions and the development of liquid-
-tight seals have also been required

for valves. 1In addition, the valves
must contain seal materials that will
not interdiffuse in the presence of
the coolant at high temperatures.
Lubrication of the moving parts of
these devices at high temperatures
has been accomplished. The cumbersome,
and at times massive, geometry of

these devices has made it necessary
to provide auxiliary heating to avoid
freezing of the coolant.

Heating and cooling of liquid
metals, hydroxides, and mixed fluorides
are being investigated from room
temperature to 1800°F. The high
melting temperatures of hydroxides
and fluorides have introduced pre-
heating, insulation, and operational
complexities, since 1t is desirable
to avolid freezing of the coolant and
possible bursting of containers upon
remelting. Heat transfer studies at
temperatures at which the materials
are molten thus far appear to be
straightforward, and the technology
advances at the rate at which the
controlling physical properties are
defined.

Equipment performance 1s improving
markedly, and a number of mechanical
and other devices for fluorides have
been operated for periods exceeding
1000 hr in some cases without visible
equipment damage and without leakage.
Heat exchangers are being designed
for aircraft and laboratory applica-
tion, and the NaK-to-NaK heat exchanger
has now operated for over 2000 hr at
1500°F. Flow, volume, pressure, and
temperature control appear to be
straightforward 1f care 1s exercised
at the high temperatures involved.

SEALS AND CLOSURES

Frozen-sodium-sealed pumps have
operated over extended periods, and
frozen fluorides that have been used
for sealing shafts show considerable
promise of giving satisfactory service.
The metallic braid with self-contained
packing lubricant appears to be highly

23

 
 

T

i g

st

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

satisfactory for use with molten
fluorides. Controlled isolation of
two sections of a fused salt system
has been accomplished by the use of a
"freeze valve." Welded or metal-
gasketed joints are used in plping
systems. Improved weld designs and

welding techniques that allow full

penetration of weld metal have made
possible the operation of equipment
without failure for extended periods.
Flanged joint seals with oval-ring
gaskets have proved satisfactory for
operation at temperatures to 1300°F.
Frozen Sodium Seal. The frozen-
sodium-sealed pump regorted in the
last quarterly report‘!’ has operated

"approximately 1500 hr during this

quarter without failure of the frozen
sodium seal. A modified Durco cen-

trifugal pump has been constructed

that has a finned sleeve for forming
a frozen sodium seal by means of con-
vective cooling.

Frozen Fluoride Seal (W. B. McDonald
and P. W. Smith, ANP Division). An
additional 160 hr of testing was ac-
complished during this period with the
frozen fluoride seal previously re-
ported(?? (Fig. 5). 1Initial tests
were conducted to determine whether
a frozen fluoride seal 1s feasible,

and additional tests were conducted

to determine (1) the effect of fluorides
on the Stellite-coated shaft around
which the seal is formed; (2) the
maximum pressure that can be sealed
without leakage; (3) the problems
encountered in starting the shaft after
it has been stopped and the fluorides
permitted to freeze around the shaft;
and (4) the determination of design
and operational parameters with this

 

(1)H. W. Savage, ‘“Experimental Reactor
Engineering,”” Aircraft Nuclear Propulsion Project
Quarterly Progress Report for Period Ending
December 10, 1951, ORNL-1170, p. 42.

(2)W. B. McDeonald, ¢‘‘Seal Testsfi; op. cit.,
ORNL-1170, p. 43. '

24

type of seal. These tests 1indicate
that greater clearance is desirable
between shaft and cooling sleeve than
was used in the frozen sodium seal.
A build-up of magnetic, metallic
material, from 1 to 2 mils thick, was
found on the shaft and cooling sleeve.
Several shallow surface scratches were
found on the shaft; however, these
were no more severe than those normally
encountered with stuffing-box packing
materials. A significant difference
was found between a frozen fluoride
seal and a frozen sodium seal; when
rotation 1s stopped and fluorides are
permitted to freeze around the shaft,
it is necessary to apply heat to the
shaft in order to free it for continued
operation. Maximum pressure limits
for this seal were not determined;
however, the seal was operated agalnst
60-ps1i pressure, which 1is satisfactory
for the operation of the ARE.

Bellows Type of Face Seal for ARE
Pump. Sealing of the ARE pump 1is
accomplished by two bellows type of
face seals installed in series. The
seal below the bearing space consists
of a graphite ring floating between
two hardened, metallic sealing faces,
one on the nose of the bellows and the
other attached to and rotating with
the shaft.

Sealing against the atmosphere 1is
accomplished by the conventional lapped
face seal above the bearings. This
arrangement permits the lower seal,
which must operate at higher tempera-
tures, to operate with no pressure
differential across it and thus
lengthens the life of the seal. This
arrangement would also permit complete
isolation of the system from the

" ambient atmosphere in the pump room

if leakage should occur in the primary
seal.

The temperature of the upper high-
pressure seal can be easily controlled

 

 

 

 

 
 

 

 

&7

A

1)

-

and should present no problem.  Any
leakage of bearing lubricant into the
system would be trapped immediately
below the seal and could be easily
removed from the system.

SHAFT (316 STAINLESS STEEL, STELLITE COATED)

FOR PERIOD ENDING MARCH 10, 1952

Stuffing-Box Seals for Molten
Fluorides (H. R. Johnson, ANP Divi-
sion). Thin-walled bellows have thus
far been unsatisfactory for sealing
against high-temperature flucrides

UNCLASSIFIED
DWG.14424

SUPPORTED BY OUTBOARD BEARINGS AND DRIVEN

BY V-BELT

 
 
 
 
   
  
 
   
  
    
     
    
 

THERMOGOUPLE
THERMOGOUPLE

THERMOCOUPLE

THERMOCOUPLE —
LEVEL

INDICATOR

¥4

 
  

THERMOCOUPLE N — — —
U WELL —

0.060-in. CLEARANGE

FINNED SLEEVE, TYPE-346
STAINLESS STEEL

SPACER

GAS PRESSURE

THERMOCOUPLE

GAS PRESSURE APPLIED
TO THIS SURFACE

POT, TYPE—316
STAINLESS STEEL

" Fig. 5. Frozen Fluoride Seal Tester.

25

T T T ST T WY

 
 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

owing to metal and weld failures. A
stuffing-box seal for a valve stem
has been developed that will furnish
a positive seal against fluorides at
1500°F. Such a seal has operated
satisfactorily and has sealed against
fluid pressures to 50 psig. The seal
consists of a conventional type of
stuffing box in which successive layers

of Inconel braid, graphite, nickel -
‘powder, and another layer of Inconel

braid are packed under compression
around a shaft. A conventional stain-
less steel globe valve with the stem
packed in this manner has operated
during this period for more than 700
hr at 1500°F and 30-psi pressure and
is still in operation,

" The molten fluorides pump described
in the following is sealed with a con-
ventional stuffing-box arrangement by
using four rings of commercial-graphite-

impregnated asbestos packing separated
by Teflon washers. The washers are

machined from Teflon bar stock, since

it has been determined that sheet

Teflon formed by extrusion or rolling
processes has a "memory" for its pre-

extruded shape and at elevated temper-

atures tends to return to this shape.

~ The stuffing box is surrounded by a

cooling coil through which refrigerated
ethylene glycol is circulated.

Lubrication of Seals and Shafts
(H. R. Johnson, ANP Division). Tests
indicate that when the valve stem seal
described above is lubricated periodi-
cally (a few drops each week) with
tricresyl phosphate (a compound used
for lubricating wire-drawing dies),
friction is reduced to the point where
this high-temperature valve can be
operated with as little friction as a
standard valve operated at room tem-
perature.

Further tests show that when such a
gland is packed with Inconel braid and
finely powdered UF, is substituted for
the nickel and graphite powders, a

26

shaft can be rotated at 1000 rpm with
the entire assembly heated to 1500°F
with no damage to the shaft.

According to the Celanese Corpora-

tion of America, extensive tests indi-
cate that tricresyl phosphate has
approximately six times the film
strength of a lubricating oil of ap-
proximately the same viscosity. Con-
tinuous heating at well over 200°F in
in the presence of air is required to
bring about decomposition. "The lubri-
cating effect of tricresyl phosphate
on iron is believed to be due to forma-
tion of iron phosphide Fe,P."(37 A
valve stem lubricated periodically (a
few drops once a week) with tricresyl

.phosphate has operated 660 hr, and it

remalns very easy to turn.

PUMPS

Development of a mechanical pump
has progressed to the point where
reliable operation can be expected in
forced-circulation systems containing
either liquid metals or molten fluo-
rides with fluid temperatures in the
pump as high as 1300°F. Designs have
been completed, and fabrication of
pumps, which are expected to operate
reliably at 1500°F or above, is under
way. All pumps tested to date are of
laboratory size; however, pumps de-
signed to meet ARE flow and head re-

- quirements have been designed and are

to be fabricated.

ARE Centrifugal Pump (W. G. Cobb,
G. F. Wislicenus, J. F. Haines, and
A. G.Grindell). The ARE fluid circuit
consists of two pumps in the circu-
lating-fuel system and two pumps in
the moderator coolant system., The
head and flow requirements for each of

 

(3)3. J. Bikerman, Surface Chemistry for
Industrial Research, p. 369, Academic Press,

1948.

 

 
 

 

.

»

these pumps are 1dentical. Each must
deliver approximately 45 gpm with a
head of approximately 45 ft of fluid

‘pumped. Two models have been proposed

and designed for ARE service.

The ARE centrifugal pump, Model D,
requires a gas seal, whereas ARE cen-
trifugal pump, Model F, has a frozen
fluoride seal. The design features
of Model D were described in the last
quarterly report. The assembly of this
pump is shown in Fig. 6. The Y-12
shops are presently in the process of
fabricating three Model D pumps from
type-316 stainless steel. The Model
F frozen-fluoride-sealed pump is almost
identical to the Model D pump, with
the exception of the inclusion of a
gas-cooled section in which the fluo-
rides surrounding the shaft will be
frozen. The principle of the frozen
fluoride seal has been described in
the section on "Seals and Closures" as
well as in the two previous quarterly

~ reports.

Laboratory Frozen-Fluoride-Sealed
Centrifugal Pump (W, G. Cobb and P, W.

‘Taylor). The laboratory size cen-

trifugal pump described previously(*’

‘has been used to circulate molten

fluorides in an isothermal loop op-
erating at temperatures up to 1300°F
for more than 500 hours. Although some
shutdowns have occurred because of
leaks in the plumbing system, no indi-
cation of failure has been encountered
in the pump itself. This pump has
been modified to incorporate a Graphi-
tar, rotary face seal in place of the
graphite-impregnated-asbestos stuffing-
box seal, and the parting faces of the
pump casing have been moved to a point
above the liquid level in the pump.
These modifications are expected to
make it possible for this pump to

 

(4.

AircraftNuclear Propulsion Project Quarterly
Progress Report for Period Ending June 10, 1951,
ANP-65, p. 168.

- FOR PERIOD ENDING MARCH 10, 1952

circulate fluids at temperatures 1in

excess of 1500°F.

Worthite Frozen-Sodium-Sealed Pump
(W. B. McDonald and W. R. Huntley).
The Worthite centrifugal pump modified
to incorporate the frozen-sodium seal,
testing of which was described in the
last report,‘'’ has been transferred
from an isothermal loop to operation
in a figure-8 loop. This pump has
operated for a total of approximately
1500 hr with no seal failure. Occa-
sional shutdowns for repair have been

caused by the mass transfer of copper

from the gasket used to seal the part-
ing faces of the pump to the shaft in
the cold section of the sealing annulus.
This resulted in a build-up on the
sleeve, which filled the annulus and
froze the shaft. The copper gasket
has been replaced by a soft nickel
gasket and the mass transfer condition
has apparently been remedied. Maxi-
mum sodium temperature in this loop
is 1500°F and the sodium temperature
at the pump is between 1000 and 1100°F.

During the total operating time no
difficulty has been experienced 1in
starting the cold pump from stand-by
conditions. As soon as molten sodium
is lifted from the sump to the pump
the shaft rotates freely. Some bear-
ing failures have been encountered
owing to the proximity of the bearing
housing to the high-temperature pump
casing. Further modifications have
been designed for moving the bearing
housing away from the pump and also
for furnishing better lubrication to
the bearings. Such a pump 1s expected
to operate at 1500°F.

Modified Durco Centrifugal Pumps.
Three Durco centrifugal pumps are
currently being modified for the forced
circulation of fluorides or sodium.
The sodium pump will 1ncorporate a
frozen sodium seal, and the two fluo-
ride pumps will be modified, one to a
frozen fluoride seal and the other to
a stuffing-box seal.

27

T

 

o e

 
 

8¢

 

 

Shaft, type-316 stainless steel.
Shaft liner spacer, any 300-series stain-
less steel.

"O" ring, material to be recommended by
seal manufacturer.
Impeller, type-316 corrosion-resistant

steel casting. _

Pressure sleeve, any 300-series stainless
steel.

Seal face ring, Ontario steel.

Bearing housing, type-316 corrosion-re-
sistant steel.

Fig. 6. ARE

10.
11.

12.
13.
14.
15.

UNCLASSIFIED
DWG. ESK—15329aR

Shaft liner, any 300-series stainless steel.
Single-row radial ball bearings.
Slinger, type-316 stainless steel.

Special graphite ring, alternate with seal.

arrangement.

Gasket, copper.

Pump body assembly.
Impeller casing assembly.
Bellows shaft seal. '

Centrifugal Pump.

 

 

190434 SSAY90Ud ATHALYVAO 1JAf0ud dNV

 
 

 

 

 

 

-y

“

-l

Construction is under way ona type-
316 stainless steel Durco centrifugal
pump having a conventional stuffing
box packed with Inconel braid and a
lubricant similar to that described
in the first part of this section.
The portion of the shaft passing
through the stuffing box is hard-faced
with a layer of Stellite 1/16-1in,
thick.

A Durco centrifugal pump (Model H

'34MDVX-80) made of type-316 stainless

steel has been modified to incorporate

a frozen sodium seal. The seal is
cooled by forced convection. The
lubricating oil for the bearings 1is
circulated through a cooling radiator.
This pump will deliver 45 gpm with a
head of 45 ft of fluid and is expected
to operate over a temperature range of
1000 to 1300°F. The modified pump is
being assembled in an isothermal loop
that has temperature measurement,
flow measurement, and control equip-
ment.

Forced-Convection-Cooled Sodium-
Sealed Pump. A type-316 stainless
steel Durco pump similar to that
described above has been modified to
incorporate a copper-finned-sleeve

frozen sodium seal that will be cooled

by forced-air convection. This pump
1s designed to be used for supplying
sodium at approximately 1500°F to a
fluid-to-air radiator for the opera-
tion of a Boeing turbojet engine. The
pump will operate at 3000 rpm and
deliver 50 gpm of sodium at 20 psig.
The temperature of the sodium at the
pump will be approximately 1100°F,

Cénnédéfidtdr.Pumfi (M. Richardson,

Reactor Experimental Engineering Divi-
sion). The new, 3/4-hp, canned-rotor

"pump for the NaK loop has been in-

stalled and operated for about 8 hours.
At present the pump motors. are of two
different types - a standard wound
motor and a motor with G-E Class H

insulation for 500°F operation. The

FOR PERIOD ENDING MARCH 10, 1952

1000°F, Bentley-Harris, insulated
wire has arrived and is in the shop
to be wound. It appears that the No.
22 wire will fit in the existing stator
slots without enlarging the slots.

The loop was started with con-
siderable difficulty because of a
partial plug in the suction-line valve
seat. However, the plug was partly
cleared and the loop checked to tem-
peratures of 500 to 650°F by using the
newly installed cooler that feeds cold
metal back through the motors. Only a
slight rise above normal operating
temperatures for the motors was noted.
The pump was shut down when a leak
occurred at the discharge flange.

VALVES

In the construction of a plumbing
system for the forced circulation of
high-temperature fluids such as will
be encountered in the ARE, 1t will be
necessary at several points to 1in-
corporate valves that will operate
reliably in any emergency. Some of
these valves may be normally open or
normally closed and required to open
only once, whereas others may be re-
quired to open at more frequent inter-
vals. It 1is an absolute requirement,
however, that each valve incorporated
in the system perform its designated
function quickly and reliably at any
desired time during the entire Jlife of
the system. During this period empha-
sis has been placed on the development
of valves for such application.

Packing-Gland Seal Test Equipment.
Equipment has been designed and 1s
now being constructed for extensive
testing of the packing-gland type of
seal described above (see "Seals and
Closures™). Although initial tests
have produced a valve that appears
to be reliable, the optimum design of
such a seal, as well as the most

29

 

TR

nor o

T TR e S

 
i

 

 

satisfactory packing materials and
lubricants, should be determined.

. Frozen Fluorides Valve (W. G. Cobb).
It has been determined in actual tests
that two sections of a high-tempera-
ture forced-circulation fluoride sys-
tem can be isolated from each other by
means of freezing a short section of
the connecting pipe. This frozen
section can be melted and operation
resumed at any desired time. Such a
valve, however, 1s not quick-acting.
Several minutes are required to either
freeze or melt the fluoride. It 1is
thought that a valve of this type used
in series with the conventional valve
described above would provide ade-
quate protection against leakage past
a valve seat. Such an application
would be the isolation of the dump
tanks from the ARE system. '

Ball Check Valves (D. R. Ward).
Specially designed ball check valves
(Fig. 7) have been operated to trans-
fer fluoride at 1400°F. The balls
were constructed of type-440 (passi-
vated) stainless steel and the housing
andlfittings were of type-316 stainless
steel. There was some slight evidence
of the balls sticking in the open
position, but in all cases increased
pressure or slight tapping of the
valve corrected this condition.

Valve Seat Test (A. P. Fraas and
G. Petersen). Zirconium and molybdenum

‘were recommended and tested as valve

seat materials because neither element
diffuses readily at high temperatures,
and zirconium does not readily alloy
with other metals. Both materials

" operated satisfactorily as valve seats

for 175 hr in sodium at temperatures
up to 1500°F. Metallurgical examina-
tion following this test indicated
that the zirconium was slightly at-
tacked but there was no evidence of
attack on the molybdenum.

30

ANP PROJECT QUARTERLY PROGRESS REPORT

UNCLASSIFIED
ASK DWG.15287 Rf

 
   
  
   
  
   

SWAGELOK FITTING TYPE-
300-1, TYPE- 316 STAINLESS
STEEL, MODIFIED BY DRILL-
ING CLEAR THROUGH FOR _~
/15710.-0D TUBING -~

~ g

NOTCH TO PERMIT LIQUID
FLOW UPWARD AFTER

BALL HAS WSEN\\\\\\\\

"4-in. DIA BALL STAINLESS 7
STEEL {400 SERIES)~___ N5

 

 

Yg-in. IPS PIPE COUPLING,
4-in. LONG, TYPE - 316
STAINLESS STEEL

END OF TUBING IS FLARED  DRgZ.
WITH FLARING TOOL AFTER /b

TUBING 1S PASSED THR-
OUGH SWAGELOK FITTING

 

 

FIRST WELD

"SWAGELOK FITTING TYPE-
300-1, TYPE-316 STAIN-
LESS STEEL, MODIFIED

BY DRILLING CLEAR THR-
OUGH FOR 3/g-in.- 0D TUBING

Fig. 7. Ball Check valve.

HEAT EXCHANGERS

Heat transfer studies of the fluo-
rides and liquid metals indicate that
essentially conventional types of
heat exchangers will be applicable to
these systems once the corrosion prob-
lems with these fluids are surmounted.

The NaK-to-NaK heat exchanger has now’

operated for well over 2000 hr with no
apparent loss in performance. Other
heat exchangers have been designed for
aircraft application as well as for
use 1in other laboratory tests.

Aircraft Type of Radiator (J. F.
Bailey and W. C. Tunnell, ANP Divi-
sion). A literature survey of the
present fluid-to-gas heat exchangers

 

 

 

T T e T,

 

 
 

 

)

uty

indicated that some aircraft and loco-
motive radiators with similar require-
ments for space, weight, and efficiency
have been operated with the gas passing
through the tubes and the fluid flowing
through the interstices between the
closely stacked tubes. This has re-
sulted in a design proposed for a
fluoride-to-air radiator.

Preliminary calculations on the
basis of arbitrarily chosen dimensions
indicate that 1/2-in.-0D by 0.010-in. -
wall Inconel tubes flattened to a
rectangular cross section and stacked
in close contact with each other may
result in a desirable heat exchanger
configuration. Air will flow through
the flattened tubes and fluorides
around the tubes in the interstices.

. The advantages of the design appear to

be: (1) parallel (essentially) coun-

“ter-flowing fluids, (2) no serious

thermal expansion problems, (3) low
air pressure drop across the exchanger,
(4) low specific volume (in.® per
megawatt of heat transferred) of the
heat exchanger, (5) simplified fabri-
cation because all welds are exposed
for easy inspection and repair, and

(6) versatile tube spacing.

Sodium-to-Air Radiator (A. P. Fraas
and G. D. Whitman). The testing of the
high-performance, air-radiator, core
element described in the last report(®’
has not yet been accomplished because
of delays encountered in Nicrobrazing
the parts. The contractor who is
performing the Nicrobrazing operation
has promised delivery by March 10.

The forced-circulation sodium loop
for testing the performance of this
heat exchanger has been completed. The

 

‘S)A. P. Fraas and M. E. LaVerne, “Heat

Exchanger Tests,”” op. cit., ORNL-1170, p. 47.

FOR PERIOD ENDING MARCH 10, 1952

loop is heated by I?R losses in the
piping. An air flow test of the core
element, which was made prior to the
Nicrobrazing operation, indicates that
the air pressure drop is approximately
30% lower than had been estimated for
the "cold" condition.

Fluoride-to-Fluoride Heat Exchanger
(D. Salmon). The performance of a
small shell-and-tube heat exchanger
operating with fluoride mixtures 1in
pure counterflow in a single-fluid
figure-8 system was investigated.
Tubing chosen for the study was 0.10-
in.-ID by 0.025-in.-wall Inconel, and
the parameters btaken were flow rate,
number of tubes, and length of ex-
changer. The results of the investi-
gation were plotted over a range of

- flows from 1 to 1000 1b/hr/tube and

for lengths from 1 to 10 feet.

Considering the limitations of

available heat sources and pumps, a
heat exchanger containing 52 tubes and
having an effective length of 1.5 ft
was analyzed using the above data. Its
performance 1s shown in Table 3 for
a constant heat input to the system

of 70,000 Btu/hr.

For obtaining heat transfer it
would be desirable to have the Reynold’s
number range extend well into the
turbulent region, and it is seen from
the tabulated data that the turbulent
region would be barely penetrated at
the highest flow for the heat exchanger
analyzed.

The main value of such a heat ex-
changer system would be for endurance
testing to determine the effect of
high-temperature dynamic corrosion and
possible deterioration of heat transfer
performance for a heat exchanger 1in
which small tubes and delicate fabri-
cating techniques are used.

31

T B ST

 
 

ANP PROJECT QUARTERLY PROGRESS REPORT

- TABLE 3

Performance of Fluoride-to-Fluoride Heat Exchanger

 

 

 

SYSTEM HEAT TRANSFER |PRESSURE DROP | REYNOLDS NO. INLET COLD TEMP.
FLOW RATE | IN EXCHANGER | IN EXCHANGER IN FOR INLET HOT TEMP.
(gpm) (Btu/hr) (psi) EXCHANGER OF 1500°F (°F)
- 0.52 37,500 0.027 104 970
- 1.56 30,000 0.26 312 1336

5.2 23,500 2.5 1040 1456
10.4 20,000 10.0 2080 1478
15.6 18,500 21.0 3120 1486

 

 

 

 

 

NaK-to-NaK Heat Exchanger (A, P.
Fraas). The previously reported(®’
NaK-to-NaK heat exchanger, core ele-
ment consisting of 52 tubes 1/8 in. OD
has continued in operation. The total
running time has reached 2230 hr, over
2000 hr of which have been with a heat
exchanger inlet temperature of 1500°F
- and an outlet temperature of 1000°F for
the hot stream. The cold stream 1is
about 50°F lower at inlet and outlet.

- No apparent losses in heat transfer
- performance or increase in pressure
drop have occurred. The only diffi-
culty experienced after the 1initial
"shake-down" of the loop has been that
the accumulation of oxide in the header
tank became great enough after 2000 hr
to interfere with the operation of the
level control. The header tank and
filter have been removed, cleaned out,
reinstalled, and operation resumed.

Heat Transfer in Circulating Fluo-
ride Loops. A series of loops are
being constructed for circulating

 

(6)rbid., p. 45.

32

fluorides at various rates to obtain
thermodynamic, hydraulic, and other
engineering data. The first loop uses
convection circulation and has been in
operation for 500 hours. The loops
are made of 1/2-in.-ips Inconel pipe,
heated by electric tube furnace ele-
ments, and cooled by natural convective
air over a coiled pipe section. By
means of a heat balance on the heater
section, flow rates, velocities, and
heat transfer coefficients are cal-
culated. Data derived from the first
loop, designated Model A-1, are given
in Table 4.

A natural circulation loop (desig-
nated Model B) has been designed to
maintain 600°F temperature difference
between hot and cold legs. The di-
ameter of the tubing between the hot
and cold legs 1s reduced to achieve
velocities of the order of 5 ft per
second. This loop, fabricated from
l1-in.,-ips Inconel pipe with a finned-
tube cooling section, will be heated
by electric tube furnace elements.
The loop is expected to operate with a
maximum hot-leg temperature of 1500
to 1600°F and a cold-leg temperature
of 900 to 1000°F. o

¥ o

 

 

 

 

 

iR
 

 

¥

»

Ah

FOR PERIOD ENDING MARCH 10, 1952

TABLE 4

Analysis of Heater Section (Model A-1)

 

 

 

 

 

 

 

 

 

RUN NUMBER
1 2 3 4

Heat input (Btu/hr) 36,460 36,660 33,600 29,600
Over-all heat transfer

coefficient (Btu/hr+ft?:°F) 43.0 34.0 34.8 37.5
Flow rate (lb/sec) 0.09 0.08 0.07 0.03
Velocity (ft/sec) 0.35 0.30 0.29 0.13
Inside heat transfer _

coefficient (Btu/hr-ft?:°F) 132 82 108 129
Heater outlet temperature {(°F) 1,450 1,275 1,343 1,580
Heater inlet temperature (°F) 1,172 950 1,029 9717
Flow rate from hydraulic

analysis (1lb/sec) 0.03

INSTRUMENTATION extended periods. The degree of ac-

Operation of a high-temperature
forced-circulation system with either
liquid metals or molten fluorides
necessitates maintenance of level in-
dication and control, accurate de-
termination of flow rates, and reliable
pressure measurements at several points
in the system. Since the use of most
commercially available equipment for
performing these functions is limited
to temperatures considerably lower
than the minimum operating tempera-
ture, an extensive developmental pro-
gram has been carried out to either
produce new instruments or remove the
temperature limitations placed on
existing instrumentation. Development
has now progressed to the point where
levels can be controlled in both high-
temperature fluoride and sodium systems
with great reliability for periods
extending more than 2000 hours. Al-
though equipment 1is still lacking for
the calibration of flow measuring
devices, means are now available for

curacy eventually attainable will
depend upon the procurement of ac-
curate calibration equipment. Pres-
sures are reliably measured at many
points in fluoride and sodium forced-
circulation systems with both com-
mercial and locally developed instru-
mentation.

Flow Measurement. Electromagnetic
flowmeters have proved to be very
reliable when used with systems cir-
culating sodium or other liquid metals
having good electrical conductivity.
With liquid metals it is possible to
measure toas high a degree of accuracy
as 1s permitted by the calibrating
equipment. In the case of fluorides,
which have very low electrical con-
ductivity, the prospects of using an
electromagnetic flowmeter are nil. A
venturi section with associated pres-
sure sensing instruments appears to be
the most attractive device for measur-
ing flow in a high-temperature fluo-
ride system because of the low head

33

T T - —rve

 
 

 

loss. The extrapolation of water-
calibrated venturies to fluorides has
been sufficiently accurate.

Pressure Measurement. No pressure-
sensitive instrument has been developed
at ORNL or found commercially availa-
ble that will operate at the tempera-
tures encountered in the forced-circu-
lation systems. The highest tempera-
ture rating of any known commercial
instrument is 450°F. Consequently,
all pressure measuring devices have
used trapped gas, which results in the
pressure at the gas-liquid interface
being measured instead of that of the
liguid stream. However, this error in
pressure measurement can be minimized
by calibrating each set of instruments

-and associated gas traps with accurate

Bourdon tube pressure gages are

‘sufficiently reliable for rough pres-
- sure measurements but are not desirable

if a high degree of accuracy is to be
obtained. Bourdon tube gages, which
indicate_the”differential pressures
directly, are commercially available.

. Th? null balanéé typé of preésure
gages indicate differential pressure
on a single gage to the nearest hun-

- dredth of a pound per square inch.

Although use of this instrument 1is
limited by temperature, it was connected
to the hot liquid stream by gas lines
and was operated successfully with the
gage at room temperature.

Manometers have also been used to

‘measure differential pressfire. Al-

though they provide a reasonably ac-
curate measurement of the differential
pressure, these devices are more dif-
ficult to seal than the other systems
described and are therefore not recom-
mended for use with a trapped gas sys-

‘tem.

Temperature Measurement.- Tempera-
tures up to 2000°F are measured by

34

ANP PROJECT QUARTERLY PROGRESS REPORT

means of Chromel-Alumel thermocouples

and are recorded on multipoint Brown
temperature recorders. It has been
found that the accuracy of temperature
measurement depends to a great extent
on the fabrication and installation

~of the thermocouples. Two methods of

thermocouple attachment are being used:
(1) pulse-welding each thermocouple
wire to the outside surface of pipes
or containers and (2) forming a beaded
junction of the two thermocouple wires.
In each case the thermocouple is given
additional rigidity by tying it down
with nichrome wire. With either of
these methods a temperature error 1is
introduced if the thermocouple is not
insulated from ambient air or other
adjacent material. Borax has been
found to form an airtight coating that
prevents further oxidation.

Tests conducted with thermocouples
placed in deep wells centered in the
flowing stream and other couples
directly opposite on the external
surface of the pipe have revealed
that there is no more than a 15°F

temperature drop between the internal

thermocouple and that placed on the
outside. Thermocouples made by beading
the wires are accurate within +1°F.

The Brown temperature recorders,
with careful adjustment, will record
temperatures accurately within *1°F
over the full range from 0 to 2000°F.
They do however show a tendency to
drift after extended periods of opera-
tion, and cause errors of the order of
5°F. This necessitates frequent cali-
bration of the instruments; therefore
a program for routine inspection has
been set up, which should reveal any
instrumentation errors soon after
their occurrence.

Level Controls and Level Indicators
(A. L. Southern, ANP Division). Four

level indicating devices have proved

 

 

 

 
 

 

ry

¥

A

satisfactory for use 1in high-tempera-
ture liquids. The conventional probe-
type of level control operates a relay
or solenoid when the level of the con-
ducting fluid rises to short out the
probe. Such a level control has proved
most satisfactory both with liquid
metals and molten fluorides, although
in the case of liquid metals some
difficulty was encountered because
condensation on the spark plug in-
sulation shorted out the plug.

Another level indicator using the
principle of a resonant cavity has
been designed, and simulated experi-
ments indicate that this instrument
may give satisfactory performance.

The variable inductance level in-
dicator consists of a small coil wound
on the outside of the tank in which
the level is controlled and a tapered
iron core mounted on a float 1i1nside
the tank. The change in level raises
and lowers the tapered core inside

the coil and alters the inductance.

Tests made with the iron core at room
temperature demonstrate that it gives
a linear response that can be directly
correlated to a level inside the
system.

The fourth level indicator de-
veloped consists of a small tube ex-
tending to the bottom of the tank in
which the level is to be controlled.
Gas flowing through this tube bubbles
through the liquid. The pressure re-
quired to maintain a constant flow of
gas through the ligquid is directly
related to the height of the liquid

level above the end of the tube and is
measured by a manometer connected

between the gas tube and the gas space
above the free surface of the liquid.
Such an instrument can be used for

indicating and controlling liquid

levels.

FOR PERIOD ENDING MARCH 10, 1952

HEATING AND COOLING OF HIGH-
TEMPERATURE SYSTEMS

Heating and cooling of liquid
metals, hydroxides, and mixed fluo-
rides are being investigated from room
temperature to 1800°F. Laboratory
methods of heating, some of which are
applicable to reactor preheating,
have 1included a variety of electrical
means, and gas heating ovens are being
installed. Cooling has been accom-
plished by liquid-to-liquid and liquid-
to-air heat exchange. Knowledge of
heat transfer properties, film coef-
ficients, and the effects of turbulence
is increasing.

Initial attempts at preheating
systems to be used in handling fused
fluoride mixtures melting between 800

‘and 1000°F resulted in an epidenic of

heater failures. Heating methods used
with sodium were studied and improved
to avoid overheating of heater elements,
and the insulation, which had been
chosen for inertness in contact with
sodium, was improved to reduce heat
losses. The heating techniques de-
scribed below, which were developed
for use with fluorides, are equally
applicable to hydroxides and liquid
metals.

External Heating Systems. Several
varieties of external heater units
have been tested in fluoride systems.
The different heat units included an
integral heater-insulation assembly
that can be prefabricated in the de-
sired shape, a flexible heater cable
that can be wrapped around the object
to be heated, and calrod heaters that
are specially constructed to have
greater contact with the surface to
which they are attached. The limiting
temperature of these heaters is de-

termined by the heater elements and

varies from 1250 to 1800°F, depending
upon the desired lifetime.

35

LiEtolid akid oo

 
 

 

 ‘ment temperature of 1250°F.

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

The integral heater-insulation
assembly, which can be prefabricated
in a variety of sizes, consists of a
conventional clamshell tube furnace
element inserted into and attached to
preformed insulation by means of
clamps. This assembly has resulted in
a substantial saving of installation
time as well as in reduced heat loss
to the ambient air. The heaters have
operated more than 500 hr with

" filament temperatures at 1800°F without
becoming detached from the insulation.

Heaters of this type are preferred
when assembly time is at a premium.

A fléxiblévheater cable that can
be tightly wound on surfaces of com-

.plex geometry has been used to obtain

contact heating and reduction of
heat-up time. The heater, consisting
of nichrome wire covered with a double

. layer of fiberglass tubular insulation,

has been used successfully with fila-
Nichrome
wire commercially covered with this
insulation has been procured and 1is
being tested. This thin insulation is

. superior to other types of insulation,
‘such as ceramic beads, since the wire

can be more closely wound on the surface
to be heated and the temperature drop
from heater wire to heated surface is
considerably lower.

Dies have been constructed for
flattening the tubular calrod heaters
to give greater contact with the pipe
to which they are attached. As modi-
fied, one side of the heater 1s shaped
to conform to the curvature of the

pipe for nearly the full length of
. the heater.

4 Care must be exercised in
such shaping since it decreases the
volume of the heater that contains the
magnesium oxide insulation and results
in heater filament relocation and
distortion and subsequent hot spots
in operation.

36

Induction Heating (A. L. Southern,
ANP Division). High-frequency heating
is advantageous and availlable as a
means for rapid heating of small com-
ponents. Megatherms producing 20 kw
of power at 220 kc have been modified
with water-cooled coi1ls to obtain
temperatures over 1500°C. This method
of heating is uniform when the ma-
terial to which heat is applied is of
symmetrical geometry. On complex
surfaces, however, that section closest
to the coil is heated as much as 200°C
higher than the rest of the material.
The power producing equipment 1is

readily adaptable toheating components

containable in a space up to 24 1in.
in diameter and length.

Resistance Heating (R. G. Affel,
ANP Division). The generally appli-
cable method for rapid heating of
equipment appears to be direct resist-
ance heating. This type of heating is
superior to other types, in cases
where applicable, since the heat 1is
generated internally at the point
where it is desired and heat conduc-
tion by means of air or other gas 1is
eliminated. Direct resistance heating
has been used successfully for heating
sections of pipe where freezing would
be critical and for preheating thermal-
convection loops 1n preparation for
the filling operation. Sections of
l-in.-ips pipe 6 ft long have been
heated to 1500°F and convection loops
to 2000°F, for the purpose of cleaning
with hydrogen-firing. Heating rates
of the order of 100°F/min may be ob-
tained. The uniform heating obtained
and the elimination of heater failures
make this method of heat application
attractive,.

Insulation Testing (R. G. Affel,
ANP Division). The most satisfactory
insulation for use with sodium and
sodium-potassium alloy has proved to

 

 

 

 
 

 

T}

i

be the lead-mill slag-wool type be-
cause of its relative inertness when
in contact with these liquids at high
temperatures. For fluorides, however,
the chemicals themselves are rela-
tively inert and the insulation 1is
only required to be nonoxidizing and
nonexplosive when suddenly placed in
intimate contact with the high-tempera-
ture liquid. However, with fluorides
it was found necessary to reduce heat
losses through insulation to alleviate
overheating of heater elements, and
Johns-Manville "Superex" insulation,
rated for 1900°F service, was used
since it 1is available in preformed
shapes to cover standard pipe sizes.
This new insulation has reduced the
insulation surface temperature on

1500°F thermal-convection and forced-

circulation loops from 400°F for the
lead-mill slag to an average of 200°F
for the Johns-Manville "Superex" in-
sulation.

ARE Core Preheating (D. F. Salmon,
ANP Division). The effectiveness of
various methods of preheating the ARE
core 1s being examined analytically
and by test. The most promising pre-
heating methods are believed to be:

1. Electric resistance elements ap-
plied directly to the outside of
the pressure shell,

2. Circulation of a high-temperature
gas or liquid through the coolant
tubes or core 1interstices,

3. Cyclic charging and dumping of the
actual reactor coolant, heating
the fluid between cycles, and
utilizing a portion of the tempera-
ture excess above its freezing
point. '

The test vehicle wofild be a fhll-éize

60-deg 7 segment of the reactor. The
methods of preheating may be tried

FOR PERIOD ENDING MARCH 10, 1952

singly or in combination and core tem-
perature vs. time for varying quanti-
ties of heat input recorded, in order
to determine actual requirements and
to galn operating experience.

In the course of the investigation
a graphical solution of the transient
heating of a hollow cylinder of beryl-
lium oxide with a constant flow of
helium at 1500°F inside the cylinder
was made. A thermal diffusivity con-
stant of 0.186 ft?/hr for beryllium
oxide was assumed, and an effective
internal heat transfer coefficient of
25 Btu/hr ft?+°F was used. The solu-
tion indicated that the time required

to reach an outer beryllium oxide

surface temperature of 1200°F with the
above conditions is approximately 40
minutes. The low heat capacity of
helium gas would obviously be the con-
trolling factor for preheating time
and would greatly increase the time
required.

TECHNOLOGY OF HIGH-TEMPERATURE LIQUIDS

The successful operation of equip-
ment containing liquid metals, fluo-
rides, and hydroxides at high tempera-

tures 1s largely dependent upon purity

of the fluid sample and cleanliness of
the ultimate container. Attainment of
the desired purity of the fluid will
require close monitoring of each fluid
batch from the time it 1is prepared
until 1t 1s sealed in its container
and the development of special equip-
ment for the preparation, storage,and
transfer of the various fuel mixtures.
Similar precautions must be taken to
ascertain the cleanliness of fluid
containers, and degreasing, pickling,
hydrogen-firing, and electrolytic
cleaning may be involved.

Fluoride Preparation and Handling

V(B. C. MacPherson, H. P. Kackenmester,

.. A. Mann, J. C. White, E. Wischhusen,
ANP Division). Pilot-plant-scale

37

T YT I T T e

 
Lo -

o g

s e b, i

 

s i it i i

ki

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

preparation and handling of several
eutectic fluoride mixtures, Table 5,

‘have been accomplished by using tech-
"niques specified by the Materials

Chemistry Division. For most of these
batches, the mixed components were
liquified in an 1nert atmosphere in
the presence of stainless steel and
Inconel chips at 1800°F to remove some
of the objectionable contaminants.
They were subsequently stored solid
or transferred for use in the molten
state under helium or argon gas pres-

simply forcing the fluid through a
frangible diaphragm located between
the two chambers. The pretreatment
consists of heating the fluorides to
1800°F for several hours in the presence
of stainless steel and Inconel chips
before filtering.

Storage and handling require that
the fluorides are not recontaminated
and that they are contained in simple
equipment in suitable physical condi-
tion for handy use. These require-

 

 

 

 

sure. ments are somewhat contradictory, and
TABLE 5
- Composition of Various Fluoride Fuels and Coolants
7 FLINAK FULINAK* FUNAK FUBENA
~ * Mixture Number - 12 14 2 17
 :C0m£dsiti0h'(wt %)
UF, 7.8 71.4 12.6
NaF 11.7 10.8 16.1 39.5
KF _ , 59.1 54.5 12.5
LiF 29.2 26.9
BeF, " 47.9
‘Cohpoéifion (mole %)
UF, | 1.1 27.5 2.0
~"NaF | 11.5 10.9 46.5 47.0
KF | 42.0 43.5 26.0
LiF | 46.5 44.5
BeF, | 51.0

 

 

 

 

*Made by adding UF, to Flinak

The equipment for fluoride melting
and treatment has been redesigned and
simplified in construction and opera-
tion. The mixing chamber, storage
chamber, filter, and transfer lines
can be enclosed within a single furnace
to assure more adequate heating, and

filtration can be accomplished by

38

for the present all fluorides are

stored in the solid state under puri-
fied inert gas when received from the
melting and pretreatment equipment.
Handling is accomplished by connecting
the storage containers to test equip-
ment and making a pressure transfer
after remelting the fluorides.

g T T T

 

 

 

 
 

U

 

 

SR . i

)

»

wi

Division). H _
ferred for an inert gas blanket be-

Since fluoride fuels in other than
solid and liquid form may be needed
ultimately, preliminary experiments,
which were quite successful, have
been carried out on pelleting the
eutectic salts.

Sampling and Analyzing Techniques
(J. P. Blakely, Materials Chemistry
Division). A molten fuel sampler has
been designed, constructed, and used
successfully to obtain a sample of
molten fuel during the filling of a
convection loop. The sampler con-
sists of a heavy-walled, 100-cc,
graphite crucible held between two
metal plates with necessary transfer
and gas lines attached. The sampler
can be used either independently of
or in conjunction with the actual
filling of a loop.

Another sampler has been designed
that will make possible repeated

“sampling of a system that is in opera-

tion. When in use, molten fuel will
continuously drip through the sampler
and the sample container can be moved
as desired to intercept the drops.

Analysis of the individual fluo-
rides and eutectic mixtures of NaF-
KF-LiF-UF, (taken in this manner) have
shown traces of oxygen, hydrochloric
acid, sulphur dioxide, silicon, silicon
tetrafluoride, sodium, potassium,
boron trifluoride, and lead chloride.
Lithium and uranium have not been ob-

served.(7>

. Diffusivity of Helium Through

Stainless Steel (E. Wischhusen, ANP
Helium is generally pre-

cause o6f its low density and relative
mobility. Recent tests have demon-
strated that helium, under 54-psi
constant pressure, may be contained

 

. (7)Letter from C. R. Baldock to R. C. Briant,
Y-B16-3, Jan. 15, 1952.

FOR‘PERIOD ENDING MARCH 10, 1952

in stainless steel and/or Inconel
seamless tubing of 0.030-in. wall
thickness at 1500°F+ for 150 hr with-
out detectable diffusion through the
walls. A Westinghouse mass spectro-
graph type of helium leak detector,
sensitive to 1 part in 3.5 million,
was employed in this test.(3)

Cleaning and Inspection Technigques
(L. A. Mann and D. R. Ward, ANP Divi-
sion). Metal components have been
receilved 1in various conditions of
cleanliness and soundness; therefore
inspection and cleaning specifications
for components and assemblies have
been established to eliminate struc-
tural defects such as crevices and
pits and to remove oxides and other
contaminants from metal surfaces. The
cleaning techniques include degreas-
ing, pickling, hydrogen-firing, and
possibly electrolytic cleaning. In
addition to these techniques, the use
of additional flushing solutions 1s
required when cleaning previously used
equipment. The cleanliness of internal
surfaces may be determined with a
boroscope.

The process used for degreasing
metal parts depends upon the type of
grease encountered, the completeness of
degreasing required, and the tolerable
amount of film or deposit that can be
left onor in the degreased part. Tet-
trachloroethylene appears to be the
most promising of all degreasers for

~these purposes and trichloroethylene

the next.

Following the degreasing operation,
pickling with HNO, has been established
as the most satisfactory general clean-
ing process fornickel, Inconel, Monel,
and stainless steel equipment. This
type of pickling leaves the metal

 

(S)F. Wischhusen, Containment of Helium in
Stainless Steel and Inconel at the 1500 F+ Range,
ANP-72, Oct. 16, 1951.

39

 

Ty et O

 

 
 

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

covered with what is thought to be
essentially a monomolecular layer of
protective (dgainst many corrodents)
oxide; otherwise, 1t 1s as clean as
possible. Oxides may then be removed
from the surface of these metals by
contacting the metal (oxide) surface
with very dry oxygen-free hydrogen
gas at temperatures of from about
1500°F for nickel to above 1800°F
for chromium. The hydrogen 1s dried
to a -40°F dew point by first passing
it over a palladium (or other) catalyst
at room temperature and then passing
it through activated alumina.

‘Electrolytic cleaning has also been
examined as a possible technique,
although inherently it has the diffi-
" culty of requiring properly shaped and
placed internal electrodes. Of the
various electrolytic procedures ex-
amined, only the anodic treatment in
phosphoric acid solution evidenced
the high degree of cleaning desired.

" In used equipment most of the fluo-

-~ rides may be drained out while molten.
However, a coating will remain that

40

is dangerous 1f radiocactive or toxic
fluorides are present. It has been
determined that some molten salts and
hydroxides will remove the fluorides
by flushing at temperatures above the
melting point of the fluorides, but
such a treatment is difficult and
could easily mask the surface condi-
tion by additional corrosion. Of the
several noncorrosive cleaning solu-
tions tested, tap water and water con-
taining 10 to 50% H,0, gave the best
results, the latter being perhaps
slightly more effective. However,
the results obtained amounted only to
softening the fluorides so that they
could be easily removed mechanically
by brushing or with high-velocity
water. :

Prior to assembly all metal parts
are visually inspected for surface
defects and cleanliness. Internal
surfaces are examined by use of a
boroscope that will reveal major in-
ternal flaws, scale, slag, etc., but
its effectiveness is limited by the
patience and thoroughness of each
individual inspector. ’

 

 

 

 
o o i

 

 

»y

&)

[ 3

a4

. FOR PERIOD ENDING MARCH 10, 1952

4. REACTOR PHYSICS

W. K. Ergen, ANP Division

. The statics and uranium investment
of the 350-megawatt circulating-fuel

aircraft reactor were described in the-

last report¢!? and at present are not
a major concern. However, the kinetic
behavior of a circulating-fuel reactor
introduces several new considerations
into a somewhat obscure kinetic
picture. Malfunctions that brought
an increase as large as 3% of uranium
into the reacting zone would cause an
average temperature rise of ~100°C if
a thermal expansion coefficient of
3 X 10°* per °C is assumed for the
fuel. (An example of this type of
mal function 1s the thrombosis effect,
in which a precipitation of uranium
accumulates outside the active lattice
and is suddenly transported into the
reactor.) In addition to the thermal
expansion of the fuel, direct nuclear
effects could yield a temperature
coefficient of reactivity whose sign
is not now known. The loss of delayed
neutrons from the active lattice,
caused by the circulation of the fuel,
has been previously viewed with con-
cern because of the damping effect
associated with the presence of these
neutrons in static-fuel reactors. It
now appears plausible, although not
yet proved, that the circulation of

‘the fuel itself acts as a damping

mechanism, and possibly as a powerful
one.

The kinetic difficulties are not
very serious for the ARE. The flow
velocities of the fuel are so low
that the control rods can keep up
with the possible entrance of excess
fuel into the reacting zone or with

the exit of poison from this zone.

 

(I)N. M. Smith, Jr., “Reactor Physics,”
Aircraft Nuclear Propulsion Project Quarterly
Progress Report for Period Ending December 10,
ORNL-1170, p. 13.

Consequently, the negative temperature
coefficient of reactivity will not be
called upon to compensate for large
reactivity changes. The residence
time of the fuel in the reactor 1is
comparable with the longest delayed-
neutron periods; therefore a con-
siderable fraction of the delayed
neutrons 1s given off in the reacting
volume and is available for the damping
of oscillations. Since 1t can reason-
ably be expected that the uranium will
be returned to the U?3% stockpile,
the uranium investment 1is, within
limits, a minor problem, The work
on the ARE was mainly concerned with
specific design problems and culminated
in a fairly detailed study of the
specific design selected for the
feasibility study. The critical
mass of this design was 22.3 1b,
which led to a total uranium i1nvestment
in the system of 74 pounds.

CIRCULATING-FUEL AIRCRAFT REACTOR

The kinetics of the circulating-
fuel aircraft reactors are currently
under intense investigation. The
problems are the thrombosis effect,
direct nuclear effects, and fast
oscillations.

The considerable amount of fuel
outside the reactor could cause
mal functions to occur and bring an
excesslive amount of fuel into (or out
of) the reacting zone.. As presently
conceived, the main mechanism that
compensates for these reactivity
changes before any control rods can
act 1s the thermal expansion of the
liquid fuel. For thermal volume
expansion coefficients of 3 x 10°*
per °C, an increase by 3% of the
fissionable material in the reactor

41

 
 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

3

would call for temperature rise of
100°C. Actually, the temperature
rise would be somewhat larger because
at short times after a sudden increase
of fissionable material in the reactor
a temperature overswing occurs, and
at longer times the effective fuel
expansion coefficient i1s decreased by
the expansion of the fuel tubes.
Furthermore, the temperature rise of
100°C refersto the average temperature,
~and local temperatures may increase
somewhat more. All this limits the
allowable change in fuel concentration
"in the reactor to about 3% - a very
stringent requirement.

In addition to the thermal ex-
pansion of the ligquid fuel, direct
nuclear effects, such as adaptation
of the neutron temperature to the

fuel temperature or Doppler broadening

of resonance lines, could yield a
. temperaturé coefficient of reactivity.
~Should it develop that such a temper-
~ature coefficient is positive and of
' the same order of magnitude as the
negative temperature coefficient
resulting from fuel expansion, then
the circulating-fuel ANP reactor
would be almost impractical. However,
it is hoped that the direct nuclear
effect can be made to yield a negative
temperature coefficient of reactivity
once these effects are well under-
stood. A negative temperature coef-
ficient would aid the fuel expansion
effect. A strenuous effort is being
made to investigate the temperature
coefficient of reactivity resulting
from direct nuclear effects. This is
a very difficult problem and in its
full extent is somewhat novel. No
reportable results have been obtained
thus far,

In conventional reactors, any
oscillations of periods that were
short compared with control rod action
times would be very effectively damped
out by delayed neutrons. The circu-
lating-fuel reactor loses a large

42

fraction of 1ts delayed neutrons in

the volume exterior to the reacting
zone, and at the beginning of this
quarter considerable concern existed
as to the possibility of undamped or
increasing oscillations. This concern
has been greatly alleviated by the
reasoning given below, in which 1t 1is
shown that 1t is plausible that the
circulation of the fuel 1tself will
act as a damping mechanism, and
possibly as a powerful one,

'The uranium investment of circu-
lating-fuel reactors is at present not
a major concern, However, an in-
vestigation of the use of low-assay
uranium (say 10% U%2%) is planned,
but no work has been completed thus
far. '

O0scillations. Work in reactor

"kinetics was concentrated on the

effect of the circulation of the fuel.
It appears likely that the circulation
of the fuel itself acts as a damping
mechanism. It should be emphasized
that the following réasoning does not
constitute a rigid proof and that
there is not yet any quantitative
measure of the damping effect of the
fuel circulation. Furthermore,
antidamping as a result of coupling
between mechanical and nuclear oscil-
lations is still a possibility. The
following simplifying assumptions
were made for this analysis of the
circulating-fuel reactor kinetics:

1. All particles of the fuel
spend the same time, £, in the reactor.

2. The flux and power distri-
butions are constant over the reactor
so that the temperature rise of a
fuel particle during a time interval
dt is proportional to the total reactor
power P(t) and the proportionality
constant € 1s independent of spatial
coordinates, and so that a temperature
changeof an element of fuel influences
the reactivity to an extent that is

 

 

 

 
 

 

Iy

FL S

$

 

"

4

independent of the position of the
element.

3. Delayed neutrons are neglected.
4. A change AT of the average fuel
temperature, 1, causes an instantaneous

change ~oAT in the reactivity.

5. The reactor inlet temperature
i1s constant.

The equations used are

. a -
P =-—PpPT _ (1)
T
. €
T = €p “ P(o)do . (2)
t-6

In addition to the symbols already
specified, 7 is the prompt generation
time, P and T are functions of the
time t; and T is the deviation of the
average fuel temperature from the
value that gives zero excess re-
activity.

With these simplifying assumptions,
Egq. 1 1s equivalent to the first of
Egs. 1 in an HRP quarterly report.¢?’
Equation 2 can be understood by
considering that T changes during a
time element dt for two reasons: 1t
increases by €Pdt as a result of the
power input and it decreases because
hot fuel is expelled from the reactor.
(This explanation would be very simple
1f T were defined as the excess of
the temperature over the constant
reactor inlet temperature. Such a

‘definition of T would differ from the

one used here only by an additive
constant that disappears 1f the time

 

(2)T. A. Welton, “Damping Produced by Delayed
Neutrons,”' Homogeneous Reactor Project Quarterly
Progress Report for Period Ending August 15, 1951,
ORNL-1121, p. 99.

. traction.

FOR PERIOD ENDING MARCH 10, 1952

derivative T is formed.) The decrease
is proportional to the fuel outlet
temperature, which i1s € times the
integral of the power over the time
the expelled fuel spentin the reactor.
The factor 1/6 is used because the
temperature of the expelled fuel
influences the average temperature
T only to the extent of the ratio of
the volume of the expelled fuel to
the total volume of fuel in the re-
actor; for the fuel expelled during
dt this ratio is dt/6.

By dividing Eq. 1 by P, differ-
entiating twice with respect to ¢,
substituting T from Eq. 2, and differ-
entiating again with respect .to t,
the following equation 1s obtained:

;3

5 log P+ =P =~ [P(t) ~P(: ~6)]=0. (3)

at

 

This third-order, non-linear,
differential-difference equation has,
evidently, a simple solution if P(t)
happens to be periodic with the period
&/n, where n is a whole number. This
special case occurs, of course, only
if € has a specific relation to the
parameters determining the period of
osciliation (a€/7, and the amplitude
and average value of P). 1Ia the
special case the solution turns out
to be 1dentical with the well-known
solution for constant power ex-
(3) This is also physically
easy to visualize: 1f each particle
sees, on 1ts transit through the
reactor, a whole number of cycles,
then each particle attains the same
temperature by the time 1t reaches
the reactor outlet. With constant
outlet temperature, the power ex-

~ traction 1s constant.

 

(3)“Effect of Temperature Coefficient,”
Homogeneous Reactor Experiment Report for the

Quarter Ending February 28, 1950, ORNL-630, p. 23.

43

T T

 

 
e

 

ANP PROJECT QUARTERLY PROGRESS REPORT

If the period is not &/n, the case
of small oscillations can be con-
sidered, that 1s, the linearized
form of Eq. 3. The Nyquist theorem,
or its equivalent, then shows that
there are no antidamped oscillations.

In the nonlinear case, the following
reasoning may be used to make it
plausible that all oscillations are
damped: |

Divide Eq. 1 by P, and mulfiiply by
Eq. 2:

ST - P(t)ft P(o)d
" = ¢ 7 o, Los 8 o) do
. t -

Integrate from « to a *+ p, using
partial integration on the last term
on the right. a and ¢ + p are, so
far, arbitrary times.

times, this relation cannot be ful-
filled for all particles, and the
right side is really negative 1f the
oscillations are undamped. Apparently
this discrepancy between the vanishing
left side and the nonvanishing,
negative right side disappears 1f the
oscilllations are damped, and hence
P(o) 1s depressed as compared with

Ploc - 0).

Slow Kinetic Effects. (%’ Slow
kinetic effects are those that occur
in times of 1 sec or longer. These
effects are not a major worry because
of the possibility of compensating
for them by control rods and servo
mechanisms. Hence, the slow effects
have not been investigated systemati -
cally for the aircraft reactor.
However, a study was made of the net
positive reactivity coefficient re-
sulting from the reduction 1n the
Xxenon absorption cross section with

t=atp
t atp
a € €
- T? + eP - — log P(t)f P(o)do = "Ef log P(o) [P(o)-P(c=6)]do.
7 t-0 a
t=aq
" Now assume that fiheréis an undamped increased moderator temperature. The

periodic oscillation, and identify
p with the period. Then the left
side of the equation vanishes and the
right side is, by a known theorem(*)
€0. Presumably the right side 1is
equal to zero only if the transit time
€ is in a certain relation to the
other parameters of the oscillation.
Since, 1in reality, different particles

" of the fuel have»different transit

 

(4)This follows readily from theorem 378 of
Inequalities by G. H. Hardy, J. E. Littlewood,
and G. Polya, University Press, Cambridge, Eng.,
1934, p. 278. In order to apply theorem 378
directly, log P has to be nonnegative, which
always can be accomplished by choosing the units
of power sufficiently small. The units of power
are normally arbitrary.

44

study shows that for a 200-megawatt
reactor, an introduction of an excess
reactivity of 0.0091 causes a temper-
ature increase of only 39°F in the
first second. (For further specifi-
cations of the assumptions, see ref. 6.)

Critical Mass. Instead of the
NaF-UF, coolant used in the corre-
sponding calculations of the last
quarterly report,(®) a NaF -BeF, -UF,

 

(S)C. B. Mills, A Flux Transient Due to a
Positive Reactivity Coefficient, Y-F10-63,
Jan., 14, 1952.

(G)C. B. Mills, **Circulating-Fuel Reactors,”
ep. cit,, ORNL-1170, p. 14.

—

 

e e I e T Lo o L L

 
 

1952

FOR PERIOD ENDING MARCH 10,

DWG. 14425

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REFLECTOR

 

PRESSURE SHELL

'Schematic Diagram of Circulating-Fuel Aircraft Reactor.

Fig. 8.

45

 
 

i
]
4
g--‘
i

 

 

 

 

 

'O

ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20 18

Fig. 9. Leakage Spectrum Through the

craft Reactor.

coolant was used.(’) Furthermore,
the Inconel structure was assumed to
occupy 2% of the reactor volume
(instead of 1.5% as in the last re-
port). The net effect of these two
changes was a small decrease in
critical mass. Using beryllium metal
instead of beryllium oxide as a

‘moderator would reduce the critical

mass by only 6.5%. The fuel self-
shielding corresponds to changes
Ak of 2.2 and 5.8% for 1-in.-ID and
2-in.-ID tubing, respectively.

 

(T)C. B. Mills, The Circulating Fuel ANP

Reactor with NaF-BeFQ—UFé Fuel-Coolant, Y-F10-83,

Feb. 5, 1952.

46

(-
oo

-
£
i
..?.-

SFERET
DWG. 14426
| ] *
COMPOSITION BY VOLUME FRACTION —
CORE o REFLECTOR
FUEL-COOLANT ( NoF-BeF,-UF, ) 0.3407 -
STRUCTURE (INCONEL) 0.0197 STRUCTURE 0.05
MODERATOR (BeO) 0.6076 MODERATOR 095 _|
INERT SALT 0.0320
>..
£ 2 0.005
g O
EE
S 0.004 ‘
oW 0.04350 '—HJ-~L_1_
THERMAL
S 0003 ESCAPE L
D ; AT
¢ = 0002 y=18.6
T
o
Q
W > 0.00 lej;____
2
x
~7
< a0
— 6 14 12

o 8 6 4 2 0
LETHARGY (v) |

Reflector of the Circulating-Fuel Air-

+

Neutron Leakage Spectra. As has

" been described,¢®’ leakage spectra for

3
}.....-.-.

an aircraft reactor were computed.
The reactor is shown schematically in
Fig. 8. It has two-pass fuel flow,
and the core, which is aright cylinder
with conical ends, 1s surrounded on
all but the fuel entrance sides by a
thick beryllium oxide reflector.
Leakages through the reflected sides
and end, through both the thin, un-
reflected annulus around the fuel-
coolant pipe, and through the pipe
itself (cut off 6 in. from the core
surface), are shown in Figs. 9 and 10.

(S)C. B. Mills, Circulating Fuel-Coolant of the
ARE of February 12, 1952, Y-F10-992 (to be issued).

 
 

 

FOR PERIOD ENDING MARCH 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SECRET
DWG. 14427
] l l | | 1 I I l
5.040 COMPOSITION BY VOLUME FRACTION -
‘ CORE REFLECTOR
. 0035 FUEL-COOLANT (NaF-BeF,-UF, ) 0.3407  FUEL 0.84 _
§ o "~ STRUCTURE (INCONEL) 0.0197 STRUCTURE 0.16 ||
: 9 MODERATOR (BeO) 0.6076
B £ 0030 = inert saT 0.0320 T
- W@
£ 0.025 —
z 4 0.00195
< THERMAL
5 £ 0020 ESCAPE o
o= AT
lé(’-! S 0 015 v =i86 ———h} ]
g« | |—J
Y a 0.010 : ]
I |__r"J .
0.005 i 7 .
i :
dl
(a) 20 18 16 14 12 10 8 6 4 2 0
LETHARGY (v)
.040 1 i T 1 1 T T T
0 COMPOSITION BY VOLUME FRACTION
, 0.035 CORE REFLECTOR
§ @ FUEL-COOLANT (NaF-BeF,-UF,) 0.3407 FUEL .00
I 8 STRUCTURE ( INCONEL) 0.0197
E & 0030 |
b - MODERATOR ( BeO) 0.6076
- INERT SALT 0.0320
= 0.025 _
Zz J 0.00293
- g THERMAL M4
e > 0.020 ESCAPE ]
- T 1 Jlles - l'l
8 = 0.015 ' ; — _
<
Y @ 1 .JJ
5 & | | J_[
- 0.010 - | r__l_r'— ]
= | |
0.005 L:"Hrr ]
| L _

 

 

 

 

 

 

 

 

 

 

 

 

(b) 20 18 16 14 12 10 8 6 4 2 0O
- LETHARGY (¢)

Fig. 10. Leakage Spe'ctrum Around the Fuel Pipes of the Circulating-Fuel Air- u
craft Reactor. (a) Through the annular area around the fuel pipes. (b) Through
the fuel pipes. '

-y 47

 

 
 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

ALKALI HYDROXIDE MODERATED REACTOR(?)

In continuation of the work on
circulating-moderator reactors re-
ported in the last quarterly report, (%)
the critical masses plotted in Fig. 11
as functions of core diameter were
computed by the bare-reactor multigroup
method. A thick hydroxide reflector
was assumed, the cross section of
uranium and the hydroxides only were
included (thus neglecting the poisoning
by structures, etc.), and the following
density values (in g/cc) were used:
[LiOH = 1.21, NaCH = 1.61, KOH = 1.59,
hbOH = 2.9, Sr(OH), = 3.4. All curves
would show minimums 1f extended to
lower reactor volumes. The large
critical masses for potassium hy-
droxide result from the large ab-
sorption and small scattering cross
section of potassium, and these cross
sections also explain the moving of
the minimum to large core volumes,
Except with potassium hydroxide, the

hydrogen moderation results in small

critical masses.

SURVEY CALCULATIONS OF TiHE
'CIRCULATING-FUEL ARE

C.B. Mills, ANP Division

At the beginning of this quarter, a
survey calculation was carried out(!??

on the circulating-fuel ARE. The main
variable was the vbyume fraction
occupied by the fuel-coolant. Although

this fraction varied from 6.85 to
22.4%, the total uranium investment
(assuming 6.8 ft® of fuel external to
the reactor) varies only from 41 to
43 1b and goes through a flat minimum
of 40 1b around a value of 15% for

 

(Q)C. B. Mills, Critical Masses of Sone
Alkeli-Hydroxide Moderated Reactors, Y-Fl10-8¢,
Jan. 21, 1952,

(10)c. B. Mills, The Circulating Fuel ARE Core
Series of December 13, 1951, Y-F10-76.

48

this fraction. The per cent thermal
fission decreases from 83 to 67, as
the fuel volume fraction increases
over the above range. The most recent
version of the ARE has a small (7.60)
volume per cent of fuel-coolant circu-
lating in a series-parallel piping
arrangement. Critical mass, reactivity
coefficients, and other numbers of
design interest for the circulating-
fuel ABE are presented.

A specificARE design, now obsolete,
was investigated('!) prior to the work
on the ARE design presented here.

ARE Core Design. The cross section
of the circulating-fuel ARE core
normal to the axis consists of 82
hexagons of beryllium oxide with
central holes for 78 fuel tubes and
4 control rods. The fuel tubes are
1.235 OD and have 0.060-in. wall
thickness. The core length 1s 35.25
in., and the ends of the core are bare.

The reflector i1is composed of
hexagonal blocks of beryllium oxide
with 57 central cooling holes con-
structed to build out the core plus
reflector to a right cylinder with a
diameter of 47.750 inches. A 2-in.-
thick Inconel pressure shell with 48
in. ID, 52 in. OD, and 48.50 in.
in length encloses the core and re-
reflector assemblies.

The fuel-coolant volume 1in the
heat exchanger and the plumbing outside
the reactor is 4.4 ft3 (does not
include 0.2 ft® in tube bends at the
reactor ends). The total core volume
is 20.42 ft®, so there is 1.55 ft® of
fuel in the core. The ratio of fuel
in the core to total fuel is 0.26.
Volume fractions of the core and
reflector, as used in the design
calculations, are given in Table 6.

 

(11)c, B, Mills, The ARE with Circulating Fuel-
Coolant, Y-F10-82, Jan. 11, 1952.
 

 

'FOR PERIOD ENDING MARCH 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\/
SECRET
DWG. 14428
CRITICAL MASS DETERMINED BY
BARE REACTOR METHOD, USING A
REFLECTOR SAVINGS VALUE OF L
i '_ (THE DIFFUSION LENGTH)
= |
= 40 ‘ K OH SrOH 100
n RbOH
wn
<{
§ .
4 30 75
g | NaOH.
=
-
o
5 20 50
>
&
> LiOH
10 ' : , 25
-
-0 { 2 | 3 4
. ' CORE DIAMETER (ft)
fiig. 'li; 'C'r'i'tiéa'll‘mlass vs. Core Diameter for Hydroxide Reactors with Thick
Reflectors of the Same Comp_o"s_ition._ , ‘ u

[
S

LS
. 49

 
 

 

 

 

 

 

& B R R i

ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 6

Volume Fractions of the Circulating-Fuel

ARE Core and Reflector*

 

 

 

 

 

 

CORE REFLECTOR
(vol %) (vol %)
Fuel-coolant 7.6
Beryllium oxide 80.7 92.23
Inconel 1.96 0.69
Inert salt 8.26 6.34
Void 1. 40 0. 64

*Communicatedby R. W. Schroeder, February 8, 1952,

Crltlcal Mass and Total Uranium
Investment. The critical mass for

the circulating-fuel ARE is obtained

from bare(!?) and reflected(!®) reactor
calculations of equivalent spherlcal
core with the volume fractions as
given in the preceding section. The
critical mass is arrived at by a
series of approximations in which the
keff corresponding to an arbitrary
uranium mass 1s calculated until the
mass corresponding to the required
keff 1s found. The %, from the
bare and reflected calculations are
weighted when determining the actual

eff of the ARE according to the

'percentage surface area of the core,

which 1s assumed to be either bare or
reflected.

The résulting'critical.mass of the
ARE reactor described in the preceding

‘section 1is 22.3 1lb and will provide a

 

(IZ)M. J. Nielsen, Bare Pile Adjoint Solution,
Y-F10-18, Oct. 27, 1950.

(13)D. K. Holmes, The Multigroup Method as

Used by the ANP Physics Group, ANP-58, Feb, 15,

1951.

50

"to be bare.

maximum k_.. of 1.034, which is dis-
tributed among the various reactivity
effects as follows:

criticality 1.000
fiéSion-product override 0.005
excess for experiment 0.024
excess for delayed neutron loss 0.0054

The uranium 1nvestment 1n the
circulating-fuel ARE is summarized in
Table 7.

TABLE 7

Uranium Requirements of the ARE

Uranium reacting voiume‘ | 22,3 1b
Tbtal.uranium inventory 85 lb'n>
Assumed éfror rafiée ;riO% to 20% 94 to 68 1b
A “ best’’ guess allowing.-4.5% for

critical experiment correlation 74 1b

There will be a small reduction in
the uranium requirement as the result
of reflection and fissioning in the
end reflector region, which was assumed
The reflected reactor has
70.73% thermal fissions and a leakage-
to-absorption ratio of about 1 to 3.

_ Reactivity Coefficients. The
values of the reactivity coefficients,
summarized in Table 8, were obtained
by bare reactor calculation meth-
ods.(12:13) These methods appear
justified, since the corresponding
bare reactor calculation for critical
mass gave keff 1.054 for 15 1b of
uranium, which 1s close to the value
of keff = 1.0398 for the reflected
reactor. However, the reflected
reactor calculations for Ak/k per °F
gave -0.64 x 10-%, so -that the actual
reactor (whose surface 1s 13.92/41.39
bare and 27.47/41.39 reflected) would
have a Ak/k per °F of -1.95 x 10°°
 

H

For the change Ak/k corresponding to
a temperature change from 68 to 1283°F,
the value calculated for a reflected
reactor is very nearly equal to the
above value of -0.0318 calculated for
a bare reactor.

"' FOR PERIOD ENDING MARCH 10, 1952

is that of fission fragments in the
fuel-coolant. Figure 12 presents the
results of the three separated power
densities. The power density 1in the
fuel-coolant is shown both as fis-
sions/cc/sec and normalized to an

TABLE 8

Reactivity Coefficients

 

 

 

 

RANGE OF VARIABLE _
WITH CHANGE OF OF COLUMN 1 SYMBOL VALUE
o A'k/k -7
Thermal base {reactor temperature) 1283 to 1672°F o -5.76 x 10
68 to 1283°F Dk /k -0.03180*
| Lk/k |
Uranium mass 11.75 to 15 1b — 0.404
(QM/m)U
D/ k
Coolant density 90 to 100% of ——— 0.0153
quoted density (Ap/p)coolant
Dk/k
Moderator density 95 to 100% of —_ 1 0.305
quoted density (AW/p)moderator
_ . Ak/k
Density of structure (Inconel) 100 to 140% of —_— | «0.174
quoted density (D;o/p)structure
) D/ k
Core radius 100 to 101% of . _— 0.438
: o gquoted radius (AR/R)COre

 

 

 

 

*Total change over range of column 2.

Power Distribution. The power

distribution in the ARE fuel-coolant
has been evaluated by separating the
total fission energy into three parts:
fission fragment energy absorbed in
the fuel-coolant; fission-neutron
energy absorbed in the moderator; and
gamma-ray energy from direct fission,
fission products, and (n,?%¥) ab-
sorptions. The large energy density

average of 1 fission/cc/sec in the

reactor core. lhe factor converting
this to watts/cc from fission fragments
in the fuel-coolant solution is given

" on the graph, and to this must be

added the gamma-ray heating, which is
given on the graph directly 1n watts
per cubic centimeter. Power density
in the moderator as a result of
heating by gamma rays 1s also given

. e
. Fo 1‘%} AT

™y
Laadfp LoD 51

 
 

i
|
4
|
i

 

s

 

B cn et b oo S Ao M - bbb s 5

 

 

C

POWER DENSITY IN THE CORE {NORMALIZED TO 1 IN THE RADIAL DIRECTION)

on the graph in watts per cubic
centimeter. The assumption 1is made

‘that gamma-ray heating is the same

for moderator and coolant. A first
approximation to the total power
density in watts/cc in the moderator
is obtained by adding gamma-ray heating
to neutron heating. A refined calcu-
lation will be made when the design
is fixed. :

ANP PROJECT QUARTERLY PROGRESS REPORT

Neutron Flux and Leakage Spectra. -
Neutron leakage from the surface of
the reactor is shown in Figs. 13 and
14. Leakage {from the reflector
surface in neutrons/cm?/sec and leakage
from the ends are also presented in
Figs. 13 and 14, and the relative
importance of open ends on neutron
flux out of the reactor core may be

noted.

. SFCREF
DWG. 14429

 

—
REACTOR

6.15%
1.00%

 

 

CYLINDRICAL , BARE -END REACTOR, 35.25 in. IN LENGTH, 35.75 in. IN DIAMETER
CORE VOLUME, 20.4 ft°, 7.6% FUEL -
FISSIONS Zcc/sec = 1.61 x10'' (AVERAGE)
POWER DISTRIBUTION
90.0% OF POWER IN FUEL, FROM FISSION FRAGMENTS
2.5% OF POWER IN MODERATOR, FROM NEUTRONS
OF POWER IN CORE, FROM GAMMA-RAY HEATING
OF POWER IN REFLECTOR, FROM GAMMA-RAY HEATING
035% OF POWER LEAKING FROM ENDS AS GAMMA RAYS

 

 

 

 

 

 
   

 

 
 

 

 

THESE ARE APPROXIMATE FIGURES TO BE NORMALIZED TO 3 MEGAWATTS
NOTE:
1.6 FOR GAMMA AND NEUTRON HEATING USE SCALE DIRECTLY (watts/ec). ——
\ FOR FISSION FRAGMENTS MULTIPLY SCALE BY 92 FOR watts/cc IN
1.4
1.2
TRADIAL POWER
1.0 —=——=—GAMMA-RAY HEATING ™
T (APPROX. wofls/cc)\
0.8 ~— AN :
S —-\-szAL POWER
0.6 ™~

 

~

\\\

 

 

+=——CORE SURFACE

 

0.4 -
ON HEATING \\

'
t
!
{
1

 

 

 

 

 

 

 

 

 

o NEUTR
A e -(fi__ (APPROX . watts /cc) ,
02 e T = 1 o
AT T vREACTOR END
i l B fi\ _ REFLECTOR SURFACE
0 10 20 30 40 50 60

DISTANCE FROM CENTER OF REACTOR (cm)

Fig. 12. Power Distribution in the Core of the Circulating-Fuel ARE.

52 ' | G s
 

 

FOR PERIOD ENDING MARCH 10, 1952

 SEERET
DWG.14430

 

0.009

 

 

0.008

 

0.007

0.006 -
S

0.005 -

 

 

 

0.004

 

 

0.003

LEAKAGE IN NEUTRONS /cm?2/sec
FOR 1 FISSION /cc/sec IN CORE

 

 

 

0.002

. 0.0431 M
0.00t THERMAL i

LEAKAGE 'Ll-
v=186

8 16 14 12 10 8 6 4 2 0O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LETHARGY (¢)

Fig. 13. Leakage Spectrum from the Reflector of the Circulating-Fuel ARE.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

~SEERT
DWG. 14434
0.4 -
§§ 0.12
O
ggg OJO.
2§
& 0.08
E 8
o~ 0.1046
%g 0.06 THERMAL
Q LEAKAGE
=
=% u=186 ! -
WT 0.04 ' =
gt | | L
< & |
Wi 0.02 L. |_

 

 

 

 

 

 

 

 

 

 

 

 

18 16 14 12 10 8 6 4 2 0
LETHARGY (v)

Fig. 14. Leakage Spectrum from the Open Ends of the Circulating-Fuel ARE.

{ ’"”g .
E ;; .5, ! LTy 53

ULy

 
 

trons/sec/cm?.

 

kil i i, o b . s,

 

o it

ANP PROJECT QUARTERLY PROGRESS REPORT

Neutron flux normalized to 1
fission/cc/sec was computed to be
364 for the monthermal neutrons and
80 for the thermal neutrons in this
reactor. At full power, 3 megawatts,
the total integrated flux in the
center of the reactor is 16.7 x 10'3
neutrons/sec/cm?. The thermal flux

is 3.0 x 10'% neutrons/sec/cm?, and

the fast flux 1is 13.7 x 10!'% neu-

at three points in the reactor are
given in Fig. 15. The important
difference 1s that of amplitude.
Note that high-energy neutron flux
is relatively somewhat higher toward
the reactor center, but the difference
is small. Figure 16 shows the corre-

.sponding plots of flux vs. radius for

four energies. The importance of
moderation by the reflector 1s quite
apparent.

TP
DWS.14432

THERMAL VALUES
AT 186

@
AT CORE CENTER

AT '/, RADIUS
I8
AT CORE
BOUNDARY

NEUTRON FLUX (NEUTRON /7 cm2 / sec
. FOR AVERAGE OF 1 FISSION / cc / sec IN CORE

AT CORE CENTER

AT %, CORE RADIU

AT CORE

 

0 2 4 6 8 10 2 14 ‘e . 18 20
LETHARGY (¢)

_Fig; 15.. Transverse Neutron Flux
Spectrum for Three Sections Through
the Core of the Circulating-Fuel ARE.

54 ool U

Neutron flux spectra

~reduces k

STATICS OF ARE CONTROLS

Static calculations of the various
mechanisms for the regulation, safety,
and shim control of the circulating-
fuel ARE are given in terms of changes
in reactivity. The one boron carbide
regul ator rod, fully inserted, effects
a net change 1in reéctivity_of 0.0075.
The three safety rods each effect a
net change in reactivity of 0.053.
Details of the operation of these
control mechanisms are given in sec. 2.

Shim Control Regquirements, Esti-
mated reactivity changes, Ak/k, f
from room temperature to 1000°F andfto
an assumed controlled reactor temper-
ature of 1283°F are given in Table 9.

The effect on reactivity of xenon
at full power 1s to reduce the kegf
Y

0of the clean reactor calculation

Maximum transient xenon
further by Ak = 0.0082.
The temperature coefficient as a
result of the xenon at full power is

Ak = 0.0031.

Ak/k ;
— =+ 6.4 x 10-7 per °F
AT |

Xe

in the vicinity of 1400°F.

Regulator Rod. In the poison
control system, there is one axial
boron carbide regulator rod lying
along the longitudinal axis of the
cylindrical reactor. The permanent
reactivity effect of the regulator
structural material 1s included 1in
the core volume fractions. Detailed
calculations have been made for the
sodium-cooled ARE reactor for this
reactor control system design. The
only important reactor characteristic
that affects the control rod effective-
ness 1s the neutron spectrum, and
this spectrum is quite similar for
ek o S

 

 

a o

ha Y

NEUTRON FLUX PER UNIT LETHARGY FOR 1 FISSION /cc/sec AVERAGE

FOR PERIOD ENDING MARCH 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TSHECRET—
DWG. 14433
60
55
50 | ~THERMAL NEUTRONS
. \“
45 I~
| \\ .
40 ‘ ~.
\\
35 ™~
30
25
20
5 /l mev INEUTRONS
e
500 ev NEUTRONS =
10 T E—— | e
04evNflflmfi§:1~4=-~~-l__
5
0

 

 

 

SPACE POINT »n (r=2062n1n cm)

Fig. 16. Radial Neutron?Flux Pistributionin the Core of the Circulating-Fuel

ARE.

the sodium-cooled reactor and  the
present ARE Hence, the control rod
effect on react1v1ty 1s as sumed to
be the ‘same.

The net change in react1v1ty when
the regulator rod is fully 1nserted

is to be made 0 0075

, Safety Rods.4 Three safety rods are
equally spaced on a 15-in. circle
around the reactor axis. The spacing
is sufficiently large to reduce the
shadowing effect of each rod on the
others to a relatively small value,
and this effect is not considered.

Ty

§ ¥
Wl

These rods are 2-in. diameter cylinders

" uranium requirement.

of boron carbide with wall thickness
of 0.335 in., and a beryllium oxide
rod is attached to the end of each
rod. The normal position of the
safety rods is out, so that beryllium
oxide will be added to core material
with a resulting decrease in the
| The net change
in k, for each rod is thus the sum
of two reactivity contributions:

(1) from the removal of beryllium oxide
moderating material and (2) from the
insertion of a 2-in, -diameter boron
carbide neutron-absorbing rod. Each
boron carbide rod is worth -5.3% 1in

S

: ey x
g,
e

 

 
o
N

Y

 

TABLE 9

Shim Control Requirements

 

 

 

Dk/k ;¢ FOR CHANGE

 

 

 

 

 

| FROM 100 | FROM 1000| OF 1°F AT
EFFECT ASSUMPTION EQUATION FOR (Lk/k ¢ () /AT T0 1000°F | TO 1283°F|OPER. TEMP.
Expansion of liquid fuel |Volume expansion coefficients -0.0554 -0.0174 -6.16 x 107°
o -4 7
per F; fuel, 1.67 x 10 AR/ & Ne/k
(melting ignored), Inconel, 1.47 x 1074
0.20 x 10°%; difference, Bn/m)y B/ P gorant |
1.47 x 1074
Dimensional expansion Radial expansion determined Ne/ & 0.0024 0.00076 2.67 x 10°°
with constant material by Inconel; axial, by BeO. 6.1 x 107° —————
(AR/B)core
Linear expansion coefficients -0.0083 -0.0026 |-0.92 x 1075
per °F; Inconel, 6.7 x 10'6,
BeO, 4.9 x 107%, weighted Dk k
average expansion coefficient:| 18.3 x 10°°
linear 6.1 x 10'6, volume (Ao/p)moderator
18.3 x 10°% per °F
Change of density of Inconel volume expansion coef- Ne/k +0.0031 +.00098 |[+3.5 x 10°°
Inconel in core ficient per °F, 20 x 10~ 20 x 10”6
(Ao/p)structure
Change of density of Same as metal-cooled ARE -0.0054 -0.0017 -5.7 x 1078
reflector BeQ '
Change of cross sections -0.0271 -0.0047 -2 x 107°
(except Xe) with reactor D/ k
temperature °F
Change of Xe cross section| 0.26 of decay products in the 4.6 x 1079
with reactor temperature | core, remainder in external
Total Me/k  [-0.0907 | -0.0247 |-7.2 x 1075

 

 

 

 

 

 

LY0d3d SSAYI0Ud ATHALYVNO 1DoAfoud dNV _‘
 

 

 

LR

"Ak/k, and each beryllium oxide rod

in the same position is worth +0.16%.
The Inconel around the boron carbide
rods is not an effective poison when
the rods are inserted and results in
a decrease in poison rod effect of
0.21%. The net effect of insertion
of the three safety rods is thus
15.9% 1in k, This value is larger
than that quoted for the metal -

cooled reactor because there 1is no

poison NaK to displace. The maximum
effect of control rod motion is thus
16% in k_ .., corresponding to about
6 lb of uranium in the core.

SPECIFIC DESIGN PROBLEMS OF THE
CIRCULATING-FUEL ARE

Dumping all the ARE fuel into a
single tank does not result in a
critical mass in the tank. The
assumptions pertinent to this con-
clusion are given in Ref. 14; the
most i1mportant assumption 1is the
absence of good moderators in or near
the tank. '

Two ARE designs with volume frac-
tions of fuel in the core of 6.85 and
22.5% were compared as to the sensi-
tivity of their critical mass to the
addition of potassium (which has a
large absorption cross section) to
the fuel mixture.¢ %) Of importance

 

(14)C. B. Mills, A Note on Fuel Dumping fronm
ARE No. 5, Y-F10-84, Jan. 18, 1952.

(IS)C. B. Mills, Effect of Potassium in the
Fuel-Cocolant Solution in Two ARE Reactors,
Y-F10-85, Jan. 21, 1952.

FOR PERIOD ENDING MARCH 10, 1952

during the discussion was whether
the existing beryllium oxide blocks
should be cut to accommodate the
larger fuel percentage. For the
6.85% design, increaseof the potassium
fluoride mole fraction in the fuel
from 0 to 60% increased the critical
mass by 55%. For the 22.5% design,
the corresponding increase was 105%.
Since the use of potassium may be
necessary, the large sensitivity of
the 22.5% design to potassium works
to the disadvantage of this design.

The power generation in a boron
carbide curtain on one end of a
3 -megawatt ARE reactor was found to
be 0.53 watts/cm?, 20% of which is
developed in the first 0.05 centi-
meter.¢1%)

The problems of fuel-tube-wall
corrosion may not be completely
solved by the time the ARE is being
built, and for this reason, as well
as for reasons of fabrication, 1t 1s
highly desirable to use thick tube
walls. Under assumptions specified
in the reference,{!”’? increasing the
tube wall thickness from 20 to 40 or
60 mils increases the critical mass

by 18.5 and 43%, respectively.

 

(IG)C. B. Mills, The Power Generation in a
346 Curtein ARE on One End of the ARE No. 1
Reactor, Y-F10-87, Jan. 21, 1952.

(17)C. B. Mills, Effect of Structure on
Criticality of the ARE of Jaeruary 22, 1852,
Y-F10-89.

R . ~y
st 8 W b e

57

 
 

 

T ™ TN T "
i . T ™ " TR T T > " T T T T R TP SO Y TY R T R Ty oy ” o

 
 

 

O

L

"reactor have been made.

uranium,

5. CRITICAL EXPERIMENTS

A, D, Callihan,

The group responsible for studies
of preliminary reactor assemblies has
continued investigations during the
past quarter with the mockup of the
G-E, direct-cycle reactor described in
the preceding report.{(1) Relative
evaluations, in terms of contributions
to reactivity, have been made of
several reflector modifications., Data
obtained several months ago from a
critical assembly of uranium and
graphite have been analyzed, and
preliminary plans for experiments on
the liquid-fuel-coolant aircraft
The data from
the graphite assembly have been corre-
lated with the theoretical calculations
of the assembly. The correlation
lacks precision but gives results that
are at least consistent with experi-
mental fact. |

" DIRECT-CYCLE REACTOR(?2)

E. V. Haake and D. V. P. Williams
Physics Division
, W. G. Kennedy
Pratt and Whitney Aircraft Division

Dunlap Scott
‘ANP Division

The preceding quarterly report(1)
described briefly an assembly of
' beryllium, methacrylate
plastic, and stainless steel that was
designed to yield information of
value to the General Electric Company

 

(I)A. D, Callikan, “Critical Experiments,’”’
Aircreft Nuclear Propulsion Project Quarterly
Progress Report for Period Ending December 10,
1951, ORNL-1170, p. 35.

(Z)The critical assembly of this reactor will
be discussed in greater detail in a report that
is now being written to the General Electric
Company.

Physics Division

in the development of the direct-
cycle nuclear reactor for aircraft
propulsion, A loading chart of this
assembly is shown in Fig, 17. During
this quarter temperature effects on
reactivity have been studied, and some
comparisons have been made of the
effect on reactivity of the beryllium
reflector and of varying relative
thicknesses of stainless steel and
plastic substituted for the beryllium.
A control rod calibration was made,
and although the data obtained are
unique to the direct-cycle reactor,
they are of value for other types of
reactors, since they indicate the
method and precision of evaluating
reactivity changes introduced by
changes in reflector, structural
elements, and other components,

Control Rod Calibration. The
change in reactivity introduced by
displacement of each control rod has
been determined by two methods. In the
first, the "period" method, a measure-
ment 1s made of the period of the
supercritical system resulting from
the insertion of a control rod, and
the corresponding change in reactivity
is determined. In the second, the
"rod-drop" method, a safety rod is
evaluated from the transient occurring
in the flux as the safety rod is
rapidly removed from the assembly.
Since each control rod is practically
coaxial with a safety rod (Fig. 17),
the over-all value of corresponding
rods is taken to be equal.

A comparison of the total rod
calibrations obtained from the rod-
drop method with the integrated value
from the period measurements 1is given
in Table 10. Changes in reactivity
are expressed in "cents," where 100
cents ("one dollar™) is equivalent to

oo TR
fifi“ufié i
'\l/, Tt Le I W 59

 
 

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

SECRET
- DWG.144324

ABCDEFGHI JKLMNOPQRSTUVW X

0w ~N O D W .

 

 

- FE5] Be REFLECTOR, JACKET 18 in. LONG, EACH HALF BACKED BY 6in. OF

kel GRAPHITE (END).

FUEL ELEMENT, EACH {8in. LONG, EACH HALF BACKED BY 6in. OF
GRAPHITE. TOP: 6 LAYERS OF STAINLESS STEEL, WITH 5 TO 40 - mil
BY 3in. URANIUM DISKS, HORIZONTAL. BOTTOM: 1in. OF PLEXIGLAS,

*STAINLESS STEEL, AIR, FUEL

PLEXIGLAS

LETTERS REPRESENT CONTROL ROD POSITIONS NUMBERS,
SAFETY ROD POSITIONS.

 

 

 

Fig. 17. Loading Chart of Critical AsSemny of Direct-Cyclé Reactor.

60 © o gsh
 

a

<

§

0.00730 Ak/k. Tt is believed that the
precision of the rod-drop method is no
greater than 2 cents, so the agree-
ment 1n some cases 1s fortuitous.

TABLE 10

Comparison of Control Rod Calibrations

 

 

REACTIVITY CHANGE

 

 

o : : PERIOD,
CONTROL | CORRESPONDING | INTEGRATED | ROD DROP
ROD SAFETY ROD (cents) (cents)
A 5 18.2 16
B 6 15.9 16
C 1 - 15.9 18
D 3 18.3 18

 

 

 

 

Y

Data obtained in the incremental

"calibration of control rod A by the

period method are shown in Fig. 18,
which gives the change in reactivity
occurring when the rod is withdrawn
from the reactor. '

FECRER
DWG.14435

Ol.00131
Q.00141/
000102
0.000876

4
000073 &

5

0.00584

ROD VALVE (cents)

Q m b o O

0.000438
0000292
0.000146 .
o

 

0 4 8 12 16 20 24
2 ROD POSITION FROM MiD-PLANE {in.)

Fig. 18. ljkéébffifiifi'ashfiLFfifiction'

of Control Rod Position. ’

Temperature Effects. During the
operation of reactor assemblies an
irreproducibility of varying degree
has been observed occasionally in
day-to-day locations of control rods

FOR PERIOD ENDING MARCH 10, 1952

required for criticality under con-
stant loading conditions. In the work
reported here these variations exceeded
the sensitivity required to detect the
reactivity differences produced by
some structural changes under study.
Investigations have indicated the
probable cause to be ambient tempera-
ture changes. The concomitant re-
activity differences, from the re-
activity at 72.9°F, are plotted as
functions of temperature in Fig., 19,

 

 

 

 

 

 

 

 

 

 

 

 

 

SECRET
DWG. 14436
10 |—— ALL VALUES REFERRED TO| // 0.00073 -
= O cents AT 72.9°F: S
t 100cents = 0.0073 <& )
g K .
>
I_
S ////
'_..
g 0 /7 °
!
o °
z
wl
o
z /
I
G40 Pa: ' -000073
63°F 729°F 80°F
1
0.9 1.0 14 12 1.3 1.4
MILLIVOLTS [IRON-CONSTANTAN)
Fig. 19. Reactivity vs. Temperature.

Reflector Studies. An experiment
was initiated before the reactor was
in final form to compare the reactivity
value of a reflector element with that
of a fuel element to ascertain the
most satisfactory location of the
safety rod. A section of the beryllium
reflector 22 in., long (one-half of the
44-in. total reflector jacket length)
was removed from positions M-21, V-11,
and W-11 (Fig. 17), and the resulting
change in reactivity was ascertained
by a calibrated control rod. The
resulting data (Table 11) give the
reactivity change (with an estimated
error of 1 cent) when a beryllium-
filled tube is substituted for an
air-filled tube,.

Sk 075

 

 
S o il

e

ki

e

 

logiees sk cpilda Lo o

sl el o

 

 

TABLE 11

Reactifiity Change Introduced by
' Substituting Beryllium for Air

 

 

 

. REACTIVITY CHANGE
TEST CELL (cents)
M-21 4.4
 V 11 10.9
V-11 and W-11 27.5

 

 

In a second series of experiments,
with final reactor geometry, the

~beryllium in a section 9 by 36 by 6 1in.

thick(3®) was replaced by a type- 310
stainless steel section 9 by 36 by 3

~in, thick., This change was made
‘separately at the side of the reactor,
“at the top, and at the bottom. Results
‘giving the change in reactivity
‘effected by the substitution and the
“changes in location of the reflector

alterations are listed in Table 12

‘(designations are the same as in Fig,

17). The system was more reactive

‘with the beryllium reflector.

ANP PROJECT QUARTERLY PROGRESS REPORT

GRAPHITE REACTOR

E. L. Zimmerman, Physics Division.

Limited effort has been directed to
the analysis of data obtained from a
uranium-graphite reactor and to cor-
relation of the data with theory.
This material will be presented when
the analysis is complete. Such results
as are now available are discussed 1in
a subsequent paragraph on the "Corre-

lation of Theory and Critical Experi-
ments."

CIRCULATINGfFUEL REACTOR

Dunlap Scott, ANP Division ‘

A preliminary assembly is to be
made of the projected aircraft pro-
pulsion reactor by using beryllium
oxide as a moderator and reflector and
molten UF, -BeF,-NaF as a circulating
fuel and coolant. This assembly 1s to
be operated at room temperature and
egsfintially zero power and will utilize

TABLE 12

React1v1ty Change Effected by Substituting Plastic and Stainless Steel
for Beryllium in the Reflector

 

 

 

 

~ LOCATION FINAL POSITION REACTIVITY CHANGE
IN REACTOR | INITIAL POSITION PLASTIC IN STAINLESS STEEL IN (cents)
Side V,W-11,12,13 v-11,12,13 W-11,12,13 ~56. 5
Top L,M,N-2,3 L,M,N-3 L,M,N-2 -41.3
Bottom L,M,N-21, 22 L,M,N-21 L,M,N-22 -33.0

 

 

 

 

 

 

v(3)1t must be noted that since the outside
dimensions of the aluminum tubing are 3 by 3 in.,
the cross section of the units of the reflector
(and the fuel) is 2 7/8 by 2 7/8 in. , and the
remainder is aluminum (wall) and air-filled
space, For convenience in discussion, the 3-in.
dimension, and multiples thereof, will be used to
deslgnate-the reflector alterations. In the
quantitive presentation of the data actual thick-
nesses of the material will be stated, with the
understanding that the voids are present.

62

beryllium oxide and packed-powder fuel;
beryllium fluoride will be omitted from
the fuel for convenience in preparation,
One possible experimental variable will
be the uranium density of the fuel.
In the first experiment the density
will be that calculated for the molten
fluoride mixture for the ARE. Also
 

the powder mixture will have the same
uranium-sodium atomic ratio as the
design fuel. However, it will probably
not be possible to achieve the over-
all fuel density of the design fuel in
this mockup. Design of fuel con-
tainers and control and safety devices
is under way.

CORRELATION OF THEORY AND
CRITICAL EXPERIMENTS

D. K. Holmes, Physics Division

An attempt has been made to de-
termine how closely theory can be made
to check with results from critical
experiments., The several types of
results that are subject to theoretical
analysis include (1) criticality with
a nonhydrogenous moderator and with a
hydrogenous moderator, (2) foil ex-
posures, (3) danger coefficients, (4)
"rod sensitivity, and (5) the gap
experiment, ’

Although criticality calculations
for both hydrogenous and nonhydrogenous
moderated reactors are acceptably
consistent, the results could be
improved in the case of a beryllium
moderator by the existence of better
cross-section data, Multigroup calcu-
lations of hydrogen moderated reactors
yield high multiplication constants.
Danger coefficient data provide a
further check of the multigroup method,
and the cross sections upon which these
calculations are based agree reasonably
well, at least in the cases of iron
and nickel. Foil measurements also
show appreciable correlation with
predicted values. However, elementary
theory on the loss of reactivity 1in
the gap experiment is consistently

about 50% low.

Criticality with Nonhydrogenous
Moderators. Since considerable effort
has been devoted to setting up the
"multigroup" technique for numerical

FOR PERIOD ENDING MARCH 10, 1952

integration of the age-diffusion
equation,(*) a very definite attempt
has been made to determine whether
these calculations can predict experi-
mental results. Two separate bare
assemblies have been studied, one with
beryllium as moderator and the other
with graphite. Two assemblies are
described in Table 13,

The results of the multigroup
calculations on these two reactcrs are
given in Table 14,

- The values of k_ listed in Table
14 encourage the belief that the method
and the cross sections on which the
calculation 1s based are accurate;
however, these results are somewhat
offset by the calculations of the ages
of thermal neutrons in beryllium and
in graphite. Since both these calcu-
lations involve the same multigroup
method and the same cross sections,
the correlation with experimental data
would be expected to be similar to
that for the multiplication constants.
Table 15 shows that whereas the calcu-

lated value for graphite may be
acceptable, the calculated age for
beryllium is not acceptable, The

following possible explanations for
this discrepancy have been examined:

1. Incorrect experimental value
for the age,

2., Incorrect values for the beryl-
lium scattering cross section,

3. Neglect of p-scattering 1in
beryllium.

The measurements of the age in beryl-
lium seem to be in good agreement.
Changes in either the total scattering
cross section or the p-scattering

give the
age, when

contribution such as to
correct value for the

 

(4)M. J. Nielsen, Bare Pile Adjoint Solution,
Y-F106-18, Oct. 27, 1950. '

oo .
Lo Ui 63

 
 

 

 

., i s, i

o S g e, sl MG

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

applied to the criticality calculations,

lead to values for keff of around

0.90.

An additional complication is
introduced by the possibility of the
existence of a significant cross
section for the (n,2n) reaction in

beryllium.¢(5) Various experimental
estimates for this cross section may
lead to increases of as much as 20%
in the calculated k for assembly 1.
At present the best approach to match-
ing the calculations with experimental
data is probably that of allowing some
p-scattering to raise the calculated

TABLE 13

Experimental Data on Critical Assemblies 1 and 4

 

 

ASSEMBLY 1 ASSEMBLY 4

 

Critical dimgnsions
Volume fractions bf constituent materials
Fuel (93.4% U?3%)
Béryl]ium
Carbon
Stainless steel

Aluminum

 

Total hass 93. 4% U235

21 x 21 x 23.22 in. 51 x 51 x 44.111 in.

 

0.00658 0.0015032
0.90202

0.88916
0.000826 0.0003127
0.03004 0.06073
19.36 kg 57.092"

 

*This figure is the total mass of fuel which would be present if the loading had been carried uniformly
to the outside of the assembly. Actually, this was not the case, but the difference is considered to be
negligible for purposes of calculatjon since the affected region is only 3 in. wide (as compared with
over-all dimension of 51 in.) and is at the very outside where the fuel has its lowest importance.

TABLE 14

Calculatéd Results for Critical
Assemblies 1 and4

 

 

 

ASSEMBLY 1 | ASSEMBLY 4
koL, . L9912+
eff 0.98 0.9912
Median energy for 1.1 ev 0.15
fission
Fraction of fissions | 0.107 0.274
. that are thermal

 

 

 

*These figures are based on uranium cross
sections corrected for self-shielding owing to
fuel lumping (10-mil disks), The removal of the
self-shielding factors (calculations may be open
to question) made only a I to 2% difference in

eff’

64

TABLE 15
Comparison of Experimental and
Calculated vValues ot the Ages of

Thermal Neutrons in Beryllium
and Graphite

 

 

 

BERYLLIUM GRAPHITE
(cm?) (em?2)
Measured age - 93 350
Calculated age 69 391

 

 

 

 

(S)W. K. Ergen, On the (n,2n) Reaction in
Beryllium with Neutrons of a Polonium-Beryllium
Source, Y-F20-12, Apr. 30, 1951.
 

 

 

age and also some (n,2n) cross section
to bring the k back to unity. In
the absence of reliable experimental
data on both points, such calculations
would be purely speculative. In any
case, 1t seems fair to conclude that
the method of calculation gives results
that are at least consistent with
experimental fact. '

Criticality with Hydrogenous
Moderators. A modified form of the
multigroup method has been used to
calculate the multiplication con-
stants for four critical assemblies
consisting of either bare or reflected
cores of solutions of uranium hexa-
fluoride in water. An extra term is
added to the usual age-diffusion
equation to take account of the
relatively large energy losses of the
neutrons because of scattering in
hydrogenous materials. Since the
reflectors, when used, were es-
sentially infinite and of water, a
simple, reflector-saving correction to
the bare calculation was used. The
results{(®) gave effective multipli-
cation constants ranging from 0.97 to
0.988 for the critical assemblies,
Since the reactors ranged from 41.5 to
94% thermal and a simple age-diffusion,
multigroup calculation gave effective
multiplication constants about 20% too
high, it may be expected that the
method of calculation will give
acceptable results, in general, for
hydrogenous reactors.,

Foil Exposures. It is possible to
calculate the activity of an aluminum
catcher foil placed against a fuel
disk, either bare, cadmium covered, or
cadmium-indium covered 1f the neutron
flux in the reglon and the uranium and
cadmium capture cross sections are
known as functions of energy. By
denoting the activations, bare cadmium
‘covered, and cadmium-indium covered by

 

(6)C. B. Mills, Water Mederated Reactors,
Y-F10-78, Jan. 7, 1952.

' comparison,
- used,

FOR PERIOD ENDING MARCH 10, 1952

AB, Ac, Agp, rTespectively, the follow-
ing proportionalities are obtained:

: 7
0
E

th

Ef(E) H(E)dE

fc(E) L(E) ¢(E)dE

ferE) Ly(E) $(E)dE

th
where
E0 = some high neutron energy,
say 10 Mev
E,, = thermal energy,
zf(E) = macroscopic fission cross
section of the fuel disks,
H(E) = neutron flux as a function
of energy,
and fC(E) and fCI(E) are the frac-

tions of the incident neutrons of
energy E. that reach the fuel disk.
From these expressions it may be seen
that the extent to which a calculation
of foil activities checks experimental
results 1s some measure of the accuracy
of the calculation of the flux as a
function of energy. For purposes of
‘ ratios of activations are
The results are given in Table
16 (both calculation and experiment
refer to assembly 4, the graphite
assembly), The variation of the
ratios over the reactor must be
attributed to the loading in assembly
4 not continuing uniformly to the
edge of the assembly so that the outer
3 in, gave a reflector-like effect;
thus a higher fraction of the neutron
flux 1s thermal near the edge of the
reactor. The calculation represents

Tl S
bod 073
65

O

 
e bR e

 

 

 

e R

i bt s - bl A b, | i i a .

bk,

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

an average over the entire reactor.
The calculated values reported in
Table 16 do not take into account the
relative depression of the low energy
end of the flux spectrum near the foil
upon introduction of the cadmium.
Including this effect would tend to
raise both calculated values.

TABLE 16

Experimental and Calculated vValues for
"the Cadmiumand Cadmium-Indium Ratios

 

 

EXPERIMENTAL

OUTSIDE
EDGE OF
ASSEMBLY | CALCULATED

 

CENTER OF
ASSEMBLY

 

Cadmium ratio

Ap
= — 2.11 2.60 2.72
AC
Cadmium-indium
AB
ratio = —— 3.40 3.90 2.98
’ ACI

 

 

 

 

~ _Danger Coefficients. A further
check of the multigroup method and the
cross sections on which calculations

~are based is available in the danger

coefficient measurements, Experi-
mentally, the measurement consists of
determining the loss in reactivity
when a block of some material 1is
placed in the center of the assembly.
(The comparison is with a void of the
same size and at the same position as
the block of material; the experiments
were performed on assembly 4.) There
are two methods of calculation that
may be used to check these experiments,
In the "difference method," the re-
activityis recomputed for the assembly
using the multigroup method with the
proper additional amount of the
particular material added to the
assembly., [Since the material 1is, by
this procedure, essentially spread

o

66 , Q-

‘relatively small,

uniformly over the volume of the
reactor, whereas in the experiment the
material is concentrated at the center
of the assembly, the calculated
results are multiplied by (7/2)3 to
properly weight the importance of the
center of the reactor.] The new k

. . eff
is then compared with the old one
obtained before the introduction of
new material, Since the volume
fractions of the added materials are
this method involves
taking small differences of large
numbers; however, it is felt that the
numerical methods used are guite
adequate for this case, and the method
has the advantage of including the
scattering cross section and the & for
the added material as well as 1its
absorption cross section. The second
method is a perturbation technique
that involves the calculation of an
"importance" function(!) (of energy)
for the assembly. The perturbation
method allows the material to be
placed directly at the center 6f the
reactor but, as used for the danger
coefficient calculations, takes into
account only the absorption cross
section of the added material; thus
any gain 1in moderation over the void
is not included in the perturbation
method., 1In Table 17 the total loss
in k, upon introduction of a block
of the material of the size listed
into assembly 4 1s given.

An additional effect owing to the
lumping of the added material, which
would reduce the effective absorption
cross section, was not included in the
calculation; such an effect would
reduce the magnitudes of the numbers
calculated.

Rod Sensitivity. The control rods
for assembly 4 were typical sections
about 3 by 3 in. 1n cross section and
extended from the center to the edge
of the assembly., An experiment ,which
is of interest froma theoretical point
 

FOR PERIOD ENDING MARCH 10, 1952

 

 

 

 

 

TABLE 17
Experimental and Calculated Values for the Total Loss in keff upon Introduction
of vVarious Materials into Assembly 4
Ne X
eff 10
SIZE OF BLOCK CALCULATED
MATERIAL (in.) EXPERIMENTAL DIFFERENCE PERTURBATION
Sodium 3 x3 x1 0.066 0.002
Iron I x 3 x1 1. 44 1.42 2.23
Iron 3 x3 xY4 0.450 0.368 ¢.560
Nickel 3 x 3 x4 0.657 0.700
Mol ybdenum 3x3xY 1.31 0.600 0.960

 

 

 

 

 

of view, is the measurement of the
incremental sensitivity of such a rod
as a function of position in the
assembly, 1i.e.,

1 Ak

“erf

eff
Ax

 

evaluated when the control rod 1is
withdrawn (from the center) a distance
x. In the relatively simple case of a
very thin poison rod (which leaves
behind a negligible void as it 1is
withdrawn) the data for assembly 4
fitted the theoretically expected
importance function, i.e.,

1 Dk, gg

 

 

~ cos? yx

thin poison rod

where x is measured from the center of

the reactor, and v = /L, where L is
the length of the reactor. However,
the "typical-element™ control rod,
which does leave a 3 by 3 in. void

channel behind as it is withdrawn,
showed an entirely different behavior.
The sensitivity was fairly constant as
a function of x until the rod had been
withdrawn somewhat over three-quarters
of the total rod length, at which
point the sensitivity rose to a maximum
(about 20% above its value at the

center of the assembly) and then fell

rapidly as the edge of the reactor was

approached.

It seems fairly certain that the
removal of a block of moderator of
width Ax at a point x in the reactor
(leaving a thin void section behind)
should give a change in k_ 7 that
varies again as cos? Yx; tflus the

“peculiar behavior of the rod sensitivity

seems most likely to be the result of
the void channel, which is of length

(L/2) - x when the rod has been with-.

drawn a distance x. Thus 1t 1s
possible that a term provided by the
transport of neutrons along the channel
from near the center of the reactor to
a region of much lower importance near

the outside might account for the

~ T ey g 67
God 081

\J

 

 
i s, Al 4 e okl

 

where

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

experimental result. A typical form

- ‘ -’ -
for the theoretical estimate of this
effect is:

L Ok ¢y

 

k
eff void

where P(x’ - x") is the probability
that a neutron will make a flight
through the void, leave the channel
walls at x', and re-enter at x”. This
term will depend upon the flux at x’
and the solid angle subtended at x'
by a small area at x". I(x' — z") is
the change in importance of a neutron
while being transported from x' to x”.
Since the integral is over all values
of x' and x”, account is taken of the
gain in importance owing to transport
toward the center of the reactor, but
this will be smaller than the loss of
importance since the flux falls off
from the center as cos yx. Numerical
evaluation of the integrals shown
indicates that the void contribution
to the rod sensitivity rises from zero

at the center of the assembly to a

maximum near the three-quarter point
and then falls off rapidly. It is
possible to write

k Ax total

_ A[l Ak] . B{l Ak}
| k Ax moderator k Ax void

] = cos? yx
moderator ’

68

has the values for the void effect
given by the integration above, and to
choose A and B so that an acceptable

 

d X x
Ax ~ Ejo“/ov I(x' = x") P(x! = x") dx' dx”

 

fit to the experimental data 1is
obtained,

 Gap Experiment. With calibrated
control rods it 1is possible to measure
the loss in reactivity upon separating
the two halves of the reactor at the
center plane, Such an experiment was
performed on assembly 4 with sepa-
rations up to 0.3 in. and losses in
reactivity up to about 0.005. A calcu-
lation has been made by using the
mul tigroup method and allowing an
apparent absorption cross section at
each lethargy corresponding to the
probability for loss of neutrons from
the gap [with a weighting factor of
(7/2)% to account for the fact that
the losses are actually from the
center of the reactor]. The leakage
losses from the gap were calculated by
using the results given in CP-3443.(7)
The calculated losses in k& are
uniformly about 50% lower than the
experimental results. However, since
the maximum separation of halves 1is
only 0.3 in., whereas the dimension of
a face of the reactor is 51 in., the
entire experiment is in the range of
"small gap" for which the results of
CP-3443 are known to underestimate the
leakage from the gap. An improvement
of the method of CP-3443 has been made,
and the losses are now being recalcu-
lated on the new basis.

 

(T)M. G. Goldberger, M. L. Goldberger, and
J. E. Wilkins, Jr., The Effect of Gaps on Pile
Reactivity, CP-3443, Feb. 20, 1946.
 

Rt e

 

RN

 

 

 

 

 

[ P PP

 

 

T R AT TR i) R N N -l o R Yo

 

 

 
a ¢ . » o . oo w

- M
\ s q w
@ n « 0
“ '
’ * \
-
'
“
.U
e
o T— il B I B b e e i o b i st Rt e R e i T T T T R T et o " Ty R ki | R

 

 

 

 
 

 

i s

SUMMARY AND INTRODUCTION

E. P, Blizard, Physics Division

The mockup of the divided shield is
now being measured in the Bulk Shield-
ing Facility (sec. 6). The angular
and energy-dependent gamma-ray measure-
ments that have been obtained are
gratifyingly detailed., It is not
certain that the neutron spectral and
angular distributions will be as
amenable to measurements; the in-
struments for these measurements are
sti]ll being developed,

Research on ducts has included
detailed measurement of the effect of
duct geometry on neutron transmission,
as well as the experimental corrobo-
ration of a simplified theory of
neutron transmission in ducts (sec. 7).
The agreement between theory and
experiment for duct transmission is
within a factor of 2 for attenuations
as high as 10°%., Duct parameters in-
vestigated include diameter, length,
and angles of a single bend,

A comprehensive design of the Tower
Shielding Facility, which will make

possible full-scale (but not full-
intensity) measurements of divided
shields has been completed (sec. 8).
The resulting configuration, basically
a 300-ft tower with a 100-ft cross
member for the reactor and crew shield,
meets all requirements regarding
freedom from spuriously scattered
radiation and flexibility. Tt 1s
estimated that this facility will
cost about two million dollars and
that it will be completed 1in the
middle of 1953.

Additional cross-section measure-
ments have been obtained on the 5-Mev
Van de Graaff and the time-of-flight
neutron spectrometer for use in reactor
cross-section measurements has been
completed (sec., 9). The spectrometer
has been installed and the counting
rates optimized and tested on the LITR.

Measurements of the total cross section
of 1ron on the 5-Mev accelerator extend

from 0.7 to 3.6 Mev.

71

 
a ¢ . » o . oo w

- M
\ s q w
@ n « 0
“ '
’ * \
-
'
“
.U
e
o T— il B I B b e e i o b i st Rt e R e i T T T T R T et o " Ty R ki | R

 

 

 

 
 

s oo v

6. BULK SHIELDING REACTOR

. Meen

» Cochran
Haydon
. Henry

. Holland

mRET G
W= "TvoH

H. E. Hfingerford
E. B. Johnson

J. K. Leslie

F. C. Maienschein

G. M. McCammon

T. N. Roseberry

Physics

The divided shield mockup, supplied
by the General Electric Company, has
been installed in the Bulk Shielding
Facility, For these measurements the
reactor was reloaded to completely
fill the lattice of the reactor and
to minimize the effect of the borated
water., The gamma-ray spectroscopy 1is
well under way, but instrumentation

required for the neutron spectroscopy-

will not be completed until this
summer. ‘

MOCKUP OF THE DIVIDED SHIELD

The divided shield mockup consists
of a tank cylindrical on the sides and
roughly hemispherical in front (Fig.
20). A vertical slot that barely
allows clearance for the reactor is
cut along the length of the cylindrical
section so that the reactor and its
supporting bridge may be moved back
out of the shield. Cylindrical air
voids on the sides provide a region of
no attenuation for neutrons and gamma
rays. This has the effect of simulat-
ing a reactor 4 ft in diameter and
extending out to the walls of the air
voids. Mounting brackets are provided
to hold two, large, roughly hemispheri-

cal, lead dishes that can be installed

to mockup the lead shadow shield.
These lead dishes are not being used
for the present experiments. '

For the éxperiments now under way
the shield has been filled with borated
water (0.4 wt % boron), and all

Division

measurements are being taken along the
center line out from the front face of
the reactor, as follows:

1. Center line measurements of
thermal-neutron flux, fast-neutron
dosage, and gamma-ray dosage, such as
were made on the unit-shield mockup,

2. Energy and angular distribution
of gamma rays,

3. Energy and angular distribution
of neutrons, '

Unit-shield measurements will pro-
vide a temporary estimate of measure-
ment 1, and measurement 2, the gamma-
ray spectroscopy, is well under way.
Assuming 140 cm to be a typical shield
thickness, measurements have been made
with the spectrometer(!’ at various
angles with respect to the center line
(see Fig. 20). The results have been
described by Maienschein(?2?) and are
shown in Fig, 21. These data supersede
the preliminary spectradata previously
reported. (3)

 

(g, . Maienschein, Hultiple-Crystai Gamma-
Ray Spectrometer, ORNL-1142 (in press).

(Z)F. C. Maienschein; Gamma-Ray Spectral
Measurements with the Divided Shield Mock-up,
Part I, ORNL CF-52-3-1, Mar. 3, 1952.

(3)Figure 5.1, ““Preliminary Gamma-Ray Spectrum
at 130 cm from the Water-Reflected Reactor,”
Aircraft Nuclear Propulsion Project Quarterly
Progress Report for Period Ending September 10,
1951, ORNL-1154, p. 85.

coa (85 73

 

 

 
122

i e il i e, s b, i Kot B i S i il G B L R e e e 22 i S5 A B 322 s MR, a5 it . s s R, e it i, e

~BEGRET
DWG. 14237

    
    
 
 
   
 
 
  

STEEL SHELL

 

STEEL SHELL -1

BORATED /‘fi

AIR
PACE

 

REACTOR

REACTOR ¢ _

 

 

 

 

 

 

 

214.5¢cm

 

 

 

 

 

 

AIR SPACE

 

PROPOSED LEAD
SHADOW SHIELDS

 

 

 

 

 

BORATED
WATER

SECTION A—A DIVIDED SHIELD /
PLAN VIEW d

o —————————— e ———————————————————

Fig. 20. Relative Position of Reactor, Divided Shield, and Gamma-Ray Spectrometer in Bulk Shielding
Facility.

140434 SSTU90Ud ATHALYVAD 1D03A[0¥d dNV
 

ke

 

Lk bl iEe

- GAMMA—-RAY FLUX, I (y*s/_cmzléec/Mev/wafi/sterodicn)
‘ o
no

104

N

10

Fig.

FOR PERIOD ENDING MARCH 10, 1952

v

SEGREF—
- DWG. 14235

 

 

-0 2 4 6 8 10 {2 14

GAMMA—RAY ENERGY (Mev) | ( | j |

21. Gamma-Ray Spectra at 130 cm from the Water-Reflected Reactor.

s i
\ CER QT 15
S SHUo
ke o s bl i

 

ANP PROJECT QUARTERLY PROGRESS REPORT

Measurements are now being made at
various other distances from the
reactor. The reactor will then be
moved out of the shield into the open
water, and a few of the measurements
will be repeated for comparison with
the spectra i1in the borated-water
shield., Finally, an attempt will be
made to measure the spectrum of gamma
rays emerging from the face of the
reactor,

The third measurement, the energy
and angular distribution of the
neutrons, will be undertaken as soon
as the necessary instruments have been
developed, which now appears likely to
be some time this summer.

REACTOR CALIBRATION

For the divided-shield experiments
the reactor was loaded so as to com-
pletely fill the lattice when rolled
into the shield as shown in Fig. 20.
Furthermore, since the shield con-
tained borated water and duplicate
measurements were to be made with the

~reactor in the open water behind the

shield, the reactor was loaded so as
to minimize the effect of the borated

o)
£
v

76

water and was surrounded on four sides
with beryllium oxide reflector as shown
in Fig, 22. Fuel elements were added
in the interior of the lattice until

" criticality was reached with 3.1 kg of

U235, The complete interior could not
be filled, but two positions (44 and
46) had extra beryllium oxide elements
and two positions (43 and 47) were
left filled with water. Repeating the
critical experiment in the open water
necessitated a slight rearrangement of
the extra beryllium oxide elements,

~which amounted to exchanging the

positions of the extra beryllium oxide
elements (44 and 46 above) with that
of the water (43 and 47). The final
assembly required about 30 g less of
fuel. .

The power distribution of the two
critical assemblies is being measured
with gold foils, following the method
of Meem and Johnson.(*)> All data on
the divided-shield measurements will
be normalized to 1 watt using these
power calibrations, '

 

(4)J. L. Meem and E. B. Johnson, Determination
of the Pover of the Shield-Testing Reactor -~ JI.
Neutron Flux Measurements in the Water-Reflected
Reactor, QORNL-1027, Aug. 13, 1951,
 

 

FOR PERIOD ENDING MARCH 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

} ’ T
- REACTOR GRID PLATE—\ STANDARD FUEL ASSEMBLYW
%V R
D)) ) s%wm
bV oDIDIBD
W
@fi ST 1% ol

 

 

 
s ,
o bl b ¥ o 3 A ir ¥ * W e )
* ia
i I
i t w W
' . v ‘s
¥ I
v e
" x
i
N
.f
b ;
¢ i
!
e - " T e w ™ R i e Al o kol RL R e s b

 

 

 

 

 
 
 

7. DUCT TESTS

C. E. Clifford
M. Hullings

F. Muckenthaler
.. Abbott

A. Simon

~Physics Division

Duct work has been directed toward
obtaining further experimental cor-
roboration of the simplified theory
of neutron transmission through
cylindrical, air-filled ducts in
water. The effects of further vari-
ations in duct diameters, lengths, and
angles of a single bend (with straight
sections of equal length) on neutron
transmission through the duct have
been measured in the Thermal Column
Facility. In addition to the above
variations, measurements have also
been made to predict the effective
source area that contributes to the
radiation reaching the exit of a
single duct. An increase of a factor
of 6 in the calculated source strength
resul ted when the source was increased
from an area equal to that of the
duct mouth to a much larger area.
This was measured for only one duct
size (3 in. ID) and is reported to
indicate the order of magnitude of
the effect. This is, of course, very
important for extrapolating the
single-duct data to a reactor design
in which an array is used. '

The agreement with the theory has
remained good in that the dose can
be predicted for a given duct geometry
within a factor of 2 when the geo-
metrical attenuation of the duct is
as high as 10°%. A report of all
the experimental and theoretical work
on ducts 1s essentially complete and
will be issued during the next quarter.

THEORETICAL TREATMENT OF DUCT
TRANSMISSION

The generalized equation used to
predict a measured dose at the end of

a duct can be written as follows:

 

b Fflaz { a? Aa?
- n .
o [ 4] (81,2) 81,% sin 6,
Aa?
8l ,, sin &,
where
D = relative dose {(thermal-neutron
flux),
n, = relative source strength per
unit area,
a = diameter of duct,
!, = length of mth straight section,
A = experimentally determined
constant, including neutron
albedo, etc.,
9m = angle between mth and mth + 1
straight section.
For the.sfiécial case of equal-
length straight sections with bends

of equal angles,the equation becomes:

B 2
“ "4 ) [8l2) (812 sin 6]

 

"where m 1s the number of straight

sections. This case gives minimum

transmission for a given center line

length, It should be noted that the
equations are not valid for angles
of bend that areso small that neutrons
can pass directly through the duct

Ty o~
o e S g
S Y i d

79

&Hi

 
 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

‘without a wall collision. The formula
becomes increasingly more accurate as
the angle of bend approaches 90 degrees.

MEASUREMENT OF AIR-FILLED DUCTS
IN WATER

For correlation with the above
theory, measurements were made with
various configurations of air-filled,
cylindrical ducts 1n the Thermal
Column water tank, including straight
ducts and ducts with single bends up
to 90 degrees. The source areas were
also varied. The measured results
have been compared with the preceding
theory and are within a factor of 2
- for a given duct geometry.

Straight Ducts, One end of the
duct was placed against a fission
source {uranium slugs) in the Thermal
Column water tank, and the counter
was located in the various slots of
‘the counter holder at the other end
of the duct. (These ducts had 1/4-in.
lucite walls and were closed at each
‘end with a 1/4-in. lucite disk.) The
fission-source box included a movable
cadmium shutter to cover the source
‘for background measurements.

Measurements were taken on all
straight-duct configurations with the
full source, which consisted of 24
natural uranium slugs, 1 1/8 by 4 1/8
in., in a rectangular array (9.3 by
12.4 in.). Additional measurements
were made on the 6-in. -ID ducts using
a 6-in.-~diameter circular source by
placing a cadmium shutter with a
6-in.-diameter hole beneath the slugs.
Typical curves for the neutron trans-
mission, as measured along the center
line of the 6-in.-ID ducts, are given
-1n Fig. 23, and traverse measurements
are shown in Fig. 24.

Ducts with Bends., The transmission
of neutrons through a flexible rubber

80

6

hose (60 in. in length, 1/4-in. walls)
with single bends of 0 to 90 degrees
was measured with both the full
rectangular source and a 3-in,-diameter
circular source, The ducts were
sealed with 1 1/2-in. rubber plugs to
give an actual air column of 57 inches.
The two straight sections shortened
as the angle of bend increased, since
the hose was always bent with a 12-in,
radius of curvature, For each angle
the straight sections were of equal
length.

The smallest angle of bend was
equivalent to a displacement of the
center of the duct by one diameter
to prevent neutrons from reaching
the counter without having made at
leastone collisionor having penetrated
the surrounding water. Center line
and traverse measurements for the full
source are shown in Figs. 25 and 26.
Center line measurements were also

made on a 4 1/4-in.-diameter duct

with aluminum walls.

Comparison with Theory. 1In order
to compare the measured results with
the predictions of Eq. 1 or Eq. 2, an
effective source strength must be
determined. To simplify the calcu-
lations and also to give some 1dea
of the change in duct transmission
with energy, the effective source
strengths have been arbitrarily defined
as the relative thermal flux (the
response of a specified counter at
10, 20, and 30 ¢m of water from the
end of the duct) multiplied by the
geometrical attenuation of the duct,
that is, D/n, from Eq. 1 or Eq. 2.
It was assumed that the counter was
large enough to give a reading pro-
portional to the integral of the flux
leaving the duct. This can be shown
to be nearly correct by a comparison
of integrals under the traverses with
the center line readings for various
ducts. If the formula 1s correct,
n, should be a function only of the

0
distance from the duct end, provided,

-, L
¢ r

,.m-,
[N
¥

i
t [N

. T
-}s— lex et

L
 

'FOR PERIOD ENDING MARCH 10, 1952

SSECHRET—
DWG. 14438

DUCTS: 6 in. ID, PLASTIC WALLS
6 L=24in.

4 A L=36in,

® L=48In.

2 B PURE WATER FOR L =24 in.

SOURCE: RECTANGULAR,9.3 X 12.4 in.
COUNTER AT x = ¢

 

 

 

 

—

P O®S

 

™

 

—

LOW-PRESSURE, 25 in.,BF3 COUNTER RESPONSE { counts /min)
O OO

 

0 to 20 30 40 50
z,COUNTER DISTANCE FROM END OF DUCT (cm)

Fig. 23. Center Line Measurements of Neutron Trans-
mission Through Cylindrical Ducts in Water.

L=h nor
&w% L8e

81

O

 

 
ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

~SEoRET
DWG. 14439
3
>
105
8
6

 

4 \—

+ DUCTS: 6in. 1D, PLASTIC WALLS _ \ |

@ L =24In.
A L= 36in.
O L=48in
V PURE WATER FOR L = 24in.

SOURCE : RECTANGULAR, 9.3 X 12.4 in. —\1
COUNTERAT z=6.3 cm  —

A O OO

 

 

 

 

 

w

 

 

N
|

N

LOW-PRESSURE, 25-in.,BF5 COUNTER RESPONSE (counts/min)
H o O | n

 

10 ' ‘ '
-30 -20 -10 0 10 20 30 40

‘3 ¥, COUNTER DISTANCE FROM DUCT CENTER LINE (cm)

 

Fig. 24. Traverse Measurements of Neutrons in Watef Beyqnd Cylindrical
Ducts. o o

82

 
 

12Y, - in.,BF3 COUNTER RESPONSE ( counts/min )

FOR PERIOD ENDING MARCH 10, 1952

S ECRET-
_ DWG. 14440
105
8 ' ,
8 DUCT: 3-in.IDx 60in.; RUBBER WALLS;
- 1Y -in. RUBBER PLUG IN EACH END.
4 STRAIGHT
N 12,5 deg. BEND
2 G 33.25 deg. BEND
46.5 deg. BEND
104 60.6 deg. BEND
8 ‘ - X 90 deg. BEND
6 SOURCE : RECTANGULAR,9.3 x12.4 in,
: COUNTER AT x= ¢
4
2
103
8
6

    

0 t0 20 30 40 50 . 80
z, COUNTER DISTANCE FROM END OF DUCT (cm)

Fig. 25. Center Line Measurements of Neutron Trans-

mission in water Through Cylindrical Ducts with Variable
Bends. ' ' ‘ *

83

 

 

 
i

 

 

B

i d g

i b e i 50 S s Ml . £ ey oA o i n R,

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

84

12Y, -in.,BF5 COUNTER RESPONSE (counts/min.)

SECEER-

DWG. 14441

n

o ©®O
[4Y]

D

Mo

» mooa
o

DUCT: 3in.ID, X 60 in.,
RUBBER WALLS 1Y, -in.
2 RUBBER PLUGIN EACH END

® STRAIGHT

A 13.5 deg.BEND

{ 33.25 deg. BEND
B 46.5 deg. BEND

¥ 60.6 deg. BEND

E:

RECTANGULAR, 9.3 X 12.4 in.
COUNTER AT z=88 cm

SOURC

-30 -20 -10 0 10 20
- " x,COUNTER DISTANCE FROM DUCT CENTERLINE (cm)

Fig. 26. Traverse Measurements of Neutrons in Wafer
Cylindrical Ducts with variable Bendsh

 

30

Beyond
of course, that the source i1s either
the same size as the duct mouth or es-
sentially infinite in extent, That
the formula is accurate to within a
factor of 2 may be seen in Table 18,
for straight ducts, and Table 19, for
ducts with bends. These tables
summarize the calculations of n, for
each duct geometry measured to date.
The measurements will continue on a
series of 4 1/4-in. cylindrical ducts.

The effective source strength, n,,
is plotted as a function of the
equivalent centimeters of water between
the counter and the end of the duct

TABLE

- FOR PERIOD ENDING MARCH 10, 1952

in Fig., 27. The lower group applies
to the 3-in. duct with the 1 1/2-in.
rubber plug adjacent to the source,
whereas the upper group represents
data from ducts that did not have
rubber plugs. The difference is not
simply ascribable to the attenuation
of the plug since this correction had
already been applied. No satisfactory
explanation has yet been obtained.
Further work is being done to clarify
the above enigma. The increase 1in
n, upon increasing the source size
beyond the area of the duct mouth 1is
shown in Fig. 28, which also shows
angular correlation.

18

Comparison of Calculated Effective Source Strength for Straight,
Cylindrical Ducts

Low~pressure, 25-in., BF; counter

Calculated from counter responseé at 10, 20, and 30 cm of water from end of duct

 

 

 

g g iR

ol RELELL Ll Skl i

 

 

DUCT - EFFECTIVE SOURCE STRENGTH
DIAMETER LENGTH n, (neutrons/cmz/sec)
(in. ) (in.) SOURCE 10 ¢m 20 cm 30 em
6 24 Full(®) 1.05 x 103 1.29 x 102 1.58 x 10
36 Full 2.64 x 102 1.14 x 102 1.51 x 10
48 Full 7.925 x 102 1.08 x 102 1.59 x 10
6 24 6 in.(©) 4.1 x 102 5.62 x 10 7.96
36 6 in. 3.43 x 107 5.28 x 10 7.92
48 6 in. 3.04 x 102 4.91 x 10 7.96
8 924 Full 7.4 x 1072 9.99 x 10 1.33 x 10
36 Full 7.99 x 102 1.04 x 102 1.33 x 10
48 Full 7.1 x 102 9.62 x 10 1.33 x 10

 

 

 

 

 

 

 

(a)

Includes plastic ends.

{8)

(¢)

6-in,~diameter circular source.

o
£
o5

~,<
{4t
v

{ P
e

24 uranium slugs, 1 1/8 by 4 1/8 in., in rectangular array, 9.3 by 12.4 inches.

85

 
 

ANP PROJECT QUARTERLY PROGRESS REPORT
TABLE 19

| Comparisohflof Calculafed Effective Source Stréngtfi for Bent,

 

 

3

Cylindriqal Ducts

Low-pressure, 25-in., BF; counter

Caiculated from counter responses at 10, 20, and 30 cm of water from end of duct®

 

 

 

 

 

 

 

 

 

 

DUCT
DUCT CENTER LINE ONE-BEND EFFECTIVE SOURCE STRENGTH
. DIAMETER LENGTH ANGLE ny (neutrons/em”/sec)
(in.) (in.) SOURCE (deg.) 10 ¢cm 20 em 30 cm
3 57 Full(® 0 6.7 x 102 9.18 x 10 | 1.29 x 10
Full 13.5 1.32 x 10° 1.87 x 10%| 3.57 x 10
Full 33.25 1.37 x 103 1.22 x 102| 1.68 x 10
Full 46.5 1.26 x 103 9.1 x 10 9.0
Full 60.6 1.45 x 10° 9.25 x 10 9.2
' Full 90 6.6 x 10 6.1 x 10
3 57 3 in.(€) 0 1.1 x 102 1.64 x 10 4.22
3 in 13.5 1.71 x 102 2.60 x 10 6.05
3 in. 33.25 1.57 x 102 2.14 x 10 3.29
4.25 40 3 in. 0 2.68 x 102 4.4 x 10 6.95
: 3 in. 90 3.99 x 102 3.92 x 10 4,31

 

{a)
(b)

- (0)3

86

-in.-diameter circular source.

Includes rubber plugs on all 3-in. ducts.

24 uranium slugs, 1 1/8 by 4 1/8 in., in rectangul ar array, 9.3 by 12.4 in,
 

 

'FOR PERIOD ENDING MARCH 10, 1952

v

 

 

 

 

 

 

 

 

 

 

 

 

 

. ~SEERET—
. DWG. 14442
104 I I I 1
8 6-in.-1D DUCT, 6-in.-dia. CIRCULAR SOURCE _
6 O 24 in, —
A 36 in. -
4 0 48 in. T
3.in.-1D x 57-in. DUCT, 3-in.~ dia. CIRCULAR SOURCE —— |
X STRAIGHT
2 V 13.5 deg. BEND —
¥ 33.25 deg.BEND
03 4Y4-in.-1D x 40-in. DUCT, 3-in.- dia. CIRCULAR N |
SOURCE - |
8 , © STRAIGHT - |
6 0o o 90 deg. BEND -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RELATIVE SOURCE STRENGTH ( neutrons/cm2.sec)
a0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 10 20 30
EQUIVALENT CENTIMETERS OF WATER FROM END OF DUCT

Fig. 27. Comparison of Relative Source Strengths (no) from
Various Ducts. Source area equal to duct mouth area.

ol 098 "
o BoF 3 o
s

Sk

b i R o ol i3

 

i ki, e, 2, ;

 

 

ANAP PROJECT QUARTERLY PROGRESS REPORT

88

RELATIVE SOURCE STRENGTH (neutrons/cm2-sec)
m

SR

DWG. 14443
104
8
6
DUCT: 3-in.-IDX 57 in.

4 FULL SOURCE :

® STRAIGHT

A 13.5 deg. BEND
2 B 33.25 deg.BEND

V 46.5 deg. BEND

§ 60.6 deg. BEND
103 CIRCULAR SOURCE : 3-in.- dia.

8 ¥ STRAIGHT

 

0 10 20 ' 30
EQUIVALENT CENTIMETERS OF WATER FROM END OF DUCT

*-Fig.‘28. Comparison of Relative Source Strengths (n,) from
Various Geometries of the 3-in. Duct. ;

Co L e
o 0T
. T * ot
- .
¢ i
v o d
T s . -
PRI
Po%e
o
 

 

 

]

uow

b

8. TOWER SHIELDING FACILITY

E. P. Blizard

C. E. Clifford

Physics Division

The preliminary proposal(1) for the
Tower Shielding Facility has been con-
siderably revised, and a firm proposal

"will soon be submitted to the AEC.
‘Calculations of spurious scattering

from the ground and structure indicate
that the shield components (reactor
and crew shields) should be at least
200 ft from the ground and free of
scattering material in the immediate
vicinity of either component. As a
consequence, the original configuration
has been discarded in favor of the
present design, which calculations
indicate will satisfy the requirements.

A new cost estimate for this Tower
Shielding Facility, based on much more
complete analysis, indicates that the
total capital expenditure will be
about two million dollars.

TOWER FACILITY DES TGN

The tower will consist of two
vertical members 300 ft high connected
at the top by a 200-ft bridge., At the
center of the bridge is a 100-ft cross
member from the ends of which will be
suspended a shielded reactor on one
side and a crew shield on the other
(Fig. 29). The distance from the
reactor or crew shield to the ground
will be variable up to about 250 ft;
above 250 ft the scattering from the
cross member would interfere. The
normal operating altitude will be
about 200 feet, The variable height
will make 21t possible to identify and

 

(g, p. Blizard, C. E. Clifford, H. L. F.
Enlund, J. L. Meem, and A. Simon, “Tower Shielding
Facility Proposal,?’ Aircraft Nuclear Propulsion
Project Quarterly Progress Report for Period
Ending December 10, 1951, ORNL-1170, p. T73.

£y

o
oo
P

vy
gm;z

3

3

o

J. L. Meem
A. Simon

measure the ground scattering. This
is a very welcome potentiality since
the calculations of this effect may
not be accurate and it can thus be assured
that the divided-shield measurements
do not incorporate an unknown spurious
component. In addition, 1t will be
possible to estimate the added crew
exposure from ground scattering that
will be incurred on take-off and
landing.

It will be possible to mount large
pieces of structural material (e.g.,
aluminum) at various locations around
the reactor or crew shields to simulate
airplane structure. Thus the radiation
scattering from the airplane itself
can be estimated on the basis of
direct experimental evidence,

EXPERIMENTAL PROGRAM

The program for the Tower Shielding
Facility as presently conceived will
consist of four parts:

1. Calibration of the facility and
determination of background from
spurious scatterings,

2. Measurement of a shield mockup for
the X-6 airplane (Fig. 30},

3. A general study of divided shields,
including optimizationwith respect
to weight and parametric studies
of reactor-crew separation, weight
distribution, diameter of reactor
shield, size of crew compartment,
etc,,

~

4, A study of the scattering to be
expected from aircraft structure,

engines, and ground.

89

L

 
i e . 5.

A

e

ik

 

 

90

—RESFRICTED~
DWG. Ct2107

N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

#

A

 

 

 

& " o W
W - M

‘# 1"/4

W

Fig. 29. Proposed 300-ft Tower Shield Facility.
sk iAo\ e e S D A o o s 1. i i i, MG, ) i, Ll s i e Ml

N €« owo - i . . - R B 2 ' o

SECRET— :
DWG. C-9394 Rt

 

 

 

I
A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REACTOR =N

CONTROLS ™ T
WATER” ~ z

LEAD —t =

 

SHADOW DISKS —
DUCTS
REACTOR

 

 

CREW SHIELD

REACTOR SHIELD

 

 

 

 

 

 

 

N

TOWER STRUCTURE—”y

16

Fig. 30. X-6 Aircraft Shield Mockup on Tower Shielding Facility.

‘0T HOUVIN ONIONA aorydd H¥od

cS61

 
 

 

 

 

 

 

 

   

 

 

 

 

 

 
 

 

*

9, NUCLEAR MFEASUREMENTS

A. H. Snell, Physics Division

The total cross section of iron has
been measured on the 5-Mev Van de
Graaff from 0.7 to 3.6 Mev with a
resolution of approximately 35 kev.
Measurements of the inelastic scatter-

ing levels in iron are now under way

with the accelerator. The time-of-
flight neutron spectrometer has been
installed in the LITR and used (after
optimization of the background-to-
count ratio) to scan the transmission
of indium at a resolution of 1.25 usec
per meter. |

MEASUREMENTS WITH THE 5-Mev VAN de
GRAAFF ACCELERATOR

H. B. Willard C. H. Johnson
J. K. Baair J. D. Kingdon

Physics Division

During the last quarter the 5-Mev
Van de Graaff accelerator has been
used to obtain detailed information on
the total cross section of iron and
the inelastic scattering levels
excited in iron by fast neutrons. In
addition, equipment has been assembled
and preliminary measurements have been
made on the total cross sections of
Li® and Li7 and the fission cross
U235 U238u

sections of and

Total Cross Section of Iron. The
total cross- sectlon curve for iron was

measured w1th a resolution of 35 kev
The dotted curve of this
~ figure indicates the real variation

(Fig. 31).
(outside the statistical error of 3%)

of cross sectlon with energy and shows
many unresolved resonances.,

Inelastic Scattering Levels in

- Iron. Measurement of 1nelastic

scattering levels in iron by fast

neutrons places the first level
excited by this process at 0.83 Mev
above the ground state. No absolute
cross sections have been obtained.

TIME-OF-FLIGHT NEUTRON SPECTROMETER

G. S. Pawlicki, ORINS
E. C. Smith, Physics Division

The time-of-flight neutron spec-
trometer for operation up to several
thousand electron volts has been in-
stalled in the LITR and checked with
indium. When the hard background
level was attenuated by a 3-in.-thick
beryllium filter, the transmission of
indium was satisfactorily scanned at a
resolution of 1.25 psec per meter.

Background Measurements. The
instrument for background measure-
ments, as originally installed in
hole HB-1 of the LITR, looked directly
at fuel elements, and the transmission
of the shutter was not up to design
expectations., A spectrum check made
with the instrument indicated a surplus
of neutrons above a 1/E distribution
at energies greater than 1 kev. A
large fraction of these high-energy
neutrons was subsequently shown to
have an energy above 0.1 Mev. As a
temporary measure an aluminum filter
was first used that did improve the
background-to-count ratio at some
sacrifice in counting rate and defi-
nitely confirmed the suspected reason
for the poor background observed.

By reloading the lattice to provide

3 in., of beryllium in the beam, the

background was reduced to design
expectation without loss of counting
rate. The tabulation in Table 20 shows
the counting rates and backgrounds

93

 

e T

T T T

TN T W

TR T T

 
e

o B B il &

i g

UNCLASSIFIED

 

 

 

 

 

 

_ DWG.14444
5 ‘
111" 1711 T 1T 17 1T°"7T 1 1 1 1 |
5 .
4 — —
T —
5
— 3 l— _
3
4°— Q
b T |
|
|
2 — b! —
‘.} ~——— MEASURED CURVE
| ! AVERAGED CURVE ]
¢ — _ |
o o e
0 1.0 2.0 3.0 4.0

NEUTRON ENERGY (Mev)

Fig. 31. 'Total Cross Section of Iron.

TABLE 20

" Improvement of Count-to-Background

Ratio with Aluminum and Beryllium
“Filters

 

 

COUNTS BACKGROUND
{No./min) {(No./min}

 

Original loading 209 38
Aluminum filter 149 6
3-in. beryllium filter 256 12

 

 

 

 

94

observed for 20 wsec channels at a
resolution of 6.8 usec per meter,

Indium Resonances. With the re-
loaded lattice the transmission of
indium has been surveyed at a resolution
of 1.25 pusec per meter, The three
known resonances at 1.43, 3.9, and
9,5 ev were scanned, In addition,
resonances were observedat 10.6, 15.4,
23.8, 48, 78, and 90 ev., The region
of 3.9 ev and higher will be repeated
with separated isotope samples of known
purity and enrichment,

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 
A L e W

 

 

x

4

SUMMARY AND

The research on high-temperature
liquids has been almost entirely con-
cerned with the development of a fuel-
coolant for the circulating-fuel
reactor, and the effort on liquid
moderators and coolants has been
correspondingly reduced (sec., 10).
The proposed reactor loading technique
requires that the circulating-fuel
solution permit widely varying uranium
concentration with near uniform
melting point. However, for the added
requirement of low (< 10 cp) viscosity,
the previously proposed systemn,
NaF-BeF, -UF,, would provide a suitable
fuel. Of the other available fluorides,
however, several compositions in the
system NaF-KF-ZrF, -UF, appear promising
on the basis of the limited data
available., In addition to this system,
numerous other systems containing
zirconium fluoride are being examined.

The effort on corrosion research
was divided among static and dynamic
tests of fluoride and hydroxide
corrosion and static tests of liquid
metal corrosion (sec., 11). The severe
corrosion experienced in dynamic tests
with fluorides at 1500°F -~ despite
relative inertness 1in static tests -
has redirected attention to fluoride
corrosion phenonema. Although there
is attack of up to 14 mils, the 0.622-
in,~-ID Inconel convection loops will
circulate the fuel, NaF-KF-LiF-UF4,
for the duration of 500- and 1000-hr
tests without plugging. On the other
hand, the 300-series stainless steel
loops have a life expectancy (before
plugging) of 150 hr and the 400-series
stainless steel loops plugin less than
50 hours, It is known that the
uranium-free fluorides are less cor-
rosive than uranium-bearing fluorides
and that in tests with the latter
uranium dioxide crystals form in the
cold zone, The effect of various
additives on both fluoride and hy-
droxide corrosion is now being evaluated

INTRODUCTION

and dynamic corrosion of hydroxides 1is
being determined in various environ-
ments. In one such experiment with
potassium hydroxide in Inconel under
hydrogen, there was no corrosion or
mass transfer after 135 hr at 715°C.

The metallurgical processes involved
in the construction and assembly of a
high-temperature reactor core, includ-
ing fabrication of control rods,
welding and brazing of core structure,
and creep and stress-rupture of metals,
are being investigated (sec. 12)., The
cone-arc welding method may be adapted
for tube-to-header welds in the core,
but the complexity of the heat ex-
changers suggests the use of a brazing
technique. Several boron-free brazing
alloys are being tested to determine
their gqualities as brazing material
and their corrosion resistance. ARE
safety rods will probably be con-
structed of type-430 stainless steel
because of its compatibility with
boron carbide, Stress-rupture and
creep data have been obtained on
fine- and coarse-grained Inconel sheet.

Heat transfer research and physical
properties measurements have been
largely directed to the immediate
needs of circulating-fuel reactor
studies for these data. Measurements
of viscosity, thermal conductivity,
heat capacity and vapor pressure of
one or more of the various fluoride
mixtures are being obtained. The
theoretical analysis of heat transfer
in a circulating-fuel reactor has
established the design parameters for
the core fuel circuit (assuming no
wall heat transfer). Mathematical
natural convections for ligquid fuel
elements have been developed for the
case of turbulent flow. Some data
have been obtained on the heat transfer
of boiling mercury, and data should
soon be available on the heat transfer
of fused salts, hydroxides, and lithium.

97

 

.
'L
;
k
r
L

- o T e e

e

 

s
i
&
k
k-

 
Kbk piadi,

 

bbb e i

i, S e

 

ANP PROJECT QUARTERLY PROGRESS REPORT

Radiation damage studies included

irradiated fuel capsules and inpile
liquid loops and measurements of the
effect of irradiation on creep and
thermal conductivity. In several of the
inpile experiments with fluoride mix-
tures in Inconel capsules, the rate of
attack on the container material has
been considerably higher than that
observed i1n the out-of-pile controls,

98

Radioactive decay curves were taken
while sodium was being circulated in
the inpile loop for 165 hours. Partial
confirmation of the previously observed
decrease in the thermal conductivity
of Inconel at 1500°F under irradiation
has been obtained. The creep of
nickel under irradiation has a higher
rate than 1ts bench counterpart after
about 115 hours.

 

e

 
n

 

w

 

FOR PERIOD ENDING MARCH 10, 1952

10. CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS

Warren Grimes, Materials Chemistry Division

The research on high-temperature
ligquids has been concerned almost
entirely with their development for
use as fuels and/or moderators for an
aircraft reactor. The properties
required of these liquids along with
periodic descriptions of progress in
development of such materials have
been the subject of previous re-
ports.(!} During the past guarter,
however, effort on reactor coolants
and on liguid moderator materials has
been markedly reduced, in spite of the
potential long-range importance of the
latter, to increase the effort on

“various aspects of the fluoride

program.

The interest in a circulating-fuel

- ARE has emphasized a need forliquids

of very low melting point at low
uranium concentration., In addition,
it is very desirable to choose the
fuel system so that a dilute (sub-
critical) fuel solution may be intro-
duced to fill the core, and small
increments of a concentrated solution
of uranium in the same solvent may be
added to bring the system to criti-
cality. This requires that solutions
widely varying in uranium content and
melting at temperatures well below the
operating range be available and that
no high-melting compounds be formed at
intermediate concentrations. '

The qualifications seem to be met
adequately by the system NaF-Ber-UF4.

 

(I)W. R. Grimes, **Chemistry of Liquid Fuels,”
Aircraft Nuclear Propuision Project Quarterly
Progress Report for Period Ending March 10, 1951,

. ANP-60, p. 127; W. Grimes, *“Chemistry of Liquid

Fuels,” Aircraft Nucleaer Propulsion Project

- Quarterly Progress Report for Period Ending June

10, 1951, ANP-65, p. 84; W. Grimes, “Chemistry of
High-Temperature Liquids,” Aircraft Nuclear
Propulsion Project Quarterly Progress Report for
Period Ending September 10, 1951, ORNL-1154,
p. 154; W. Grimes, ¢*“Chemistry of High-Temperature
Liquids,” Aircraft Nuclear Propulsion Project
Quarterly Progress Report for Period Ending

Decenber 10, 1951, ORNL-1170, p. 79.

However, 1t appears that the viscosity
of this liquid in the lowest melting
region of the system is too high., Of
the other available materials, several
compositions in the system NaF-KF«ZrF4-
UF, appear promising on the basis of the
limited data available. Numerous other
systems containing zirvconium fluoride
are being examined.

Studies of the 1onic species 1in
molten fluorides by determination of
transference numbers have not proved
capable of unique interpretation.
General planning for production of the
simulated fuel for the "cold" critical
experiment has been completed and
experimental preparation of fuel and
fuel assemblies is under way.

LOW-MELTING-FLUORIDE FUEL SYSTEMS

J. P. Blakely R. E. Traber, Jr.
L. M. Bratcher C. J. Barton

Materials Chemistry Division

The previous report{?) presented a
preliminary list of five promising
fluoride fuel systems of low uranium
concentration, of which four contained
lithium fluoride as a component,
Since lithium fluoride would be
attractive as a component only if the
heavier isotope of lithium were to
become available inhigh concentration,
only the NaF-BeF,-UF, system appeared
promising for immediate use, However,
the toxicity of beryllium fluoride and
the preliminary indications of high
viscosity of beryllium fluoride-bearing
liquids have prompted efforts to find
other systems for immediate use.
Promising results have been obtained
recently with mixtures containing
zirconium fluoride,

 

(2)J. P. Blakely, L.. M. Bratcher, and C. J.
Barton, “lLow Melting-Fluoride Fuel Systems,”
op. c¢tt,, ORNL-1170, p. 79.

99

 

T T n T

 
 

ANP PROJECT QUARTERLY PROGRESS REPORT

It has been necessary to study the
binary and ternary systems of zir-
conium fluoride with the alkali
fluorides, Work on these uranium-free
materials is reported under "Coolant
Development™ in the following. Ad-
dition of small amounts (1 to 5 mole %)
of uranium fluoride to several of the
systems has been shown to affect the
melting points only slightly, and it
appears likely that any of the low-
melting coolants obtained from the
preliminary studies can be readily
converted to fuels., The available
data on zirconium fuels are reported
under appropriate headings in the
following, along with additional data
on several systems containing beryllium
fluoride.

- NaF-BeF,-UF,. Difficulty in es-

tablishing reproducible melting points
by thermal analysis of mixtures of
high beryllium fluoride concentration
has been discussed,(%?)> 1In an effort
to overcome this difficulty, heating
curves have been compared with cooling
curves for a number of compositions in
this system. Selected mixtures were
heated to 850 to 900°C and cooled,
with stirring, by techniques previ-
ously described. The solidified
materials were then reheated and the
heating curve, measured with chromel-
alumel thermocouples, was recorded in
the manner previously described.

The data shown in Table 21 afford
a comparison between break temperatures

TABLE 21

CompariSbn of Break Temperatures from Heating and Cooling Curves
of the System NaF-BeF,-UF, '

 

 

COMPOSITION (mole %)

BREAK TEMPERATURE (°C)

 

 

 

 

 

NaF BeF, UF, ON HEATING ON COOLING
48 52 0 345, 375, 380 332""
s 51.5 1 330, 365, 385 350"

o 51 2 365, 385 340""

6.5 50. 5 | 3 245, 315, 369, 385 355" "
46 “ 50 4 295, 325, 385 3427"

55 49.5 5 270, 360, 395 352"

'40' | 60 0 230, 300, 330, 365 250, 390
‘39.5“ 59. 4 1 290, 330, 355 , 270, 415

39.2 58. 8 2 325, 355 295, 445
a8 58. 2 3 325, 370, 400, 405 315, %" 430
38.4 57.6 4 325, 365 285, 290, 315

38.0 57.0 5 330, 360 317

 

 

e : §
Highest break temperature considered most reliable indication of melting point.

*-,*. . . - < - ¥ .
Supercooling definitely indicated.

100

 

 

S e e e W e v on et

 
 

#

 

taken from heating curves and cooling
curves, which were made with the same
materials and equipment. The highest
temperature at which a break is shown
in the heating curve is considered the
most reliable indication of melting
point available. As the data indicate,
melting points for this system are
Jittle affected by uranium concen-
tration in the range studied. It
appears that mixtures of useful
uranium concentration may be prepared
with melting points below 400°C.
KF-BeF,-UF,. A previous document( 3)
presented some preliminary data on the
system KF-BeF, -UF,. Results of ad-
ditional studies with these materials
are shown in Table 22. All the

‘mixtures studied show melting points

considerably higher than those avail-
able in the NaF-BeF,-UF, system,

TABLE 22

Break Temperatures from'Cooling Curves
of the Systen KF-BeFZ-UF4

 

 

 

 

FOR PERIOD ENDING MARCH 10,

Break Temperatures from Cooling Curves
of the System RbF-Ber—UF4

TABLE 23

1952

 

 

 

 

COMPOSITION (mole %) BREAK TEMPERATURES™
RbF BeF, | UF, (°c)

40 60 0 375, 420

39.4 59.1 1.5 380, 440

38.2 58. 2 3 325, 370, 445
38.0 57.0 5 365, 442

36.0 54.0 10 430, 544

COMPOSITION (mole %) BREAK TEMPERATURES”
KF | BeF, UF, (°0)

52 48 0 400, 475, 520

51.2 | 47.3 1.5 397, 514, 533, 563
50.5 | 46.5 3 390, 472, 525

49.5 | 45.5 5 385, 510, 600
44.5 | 40.5 10 385, 525, 607

‘Table 23.

»

 

 

 

 

*
Highest break temperature considered most
reliable indication of melting point.

*3 ‘
. Temperature uncertain,

 

 

 

 

*
Highest break temperature considered most
reliable indication of melting point.

than those for NaF-Ber-UF4 mixtures,
the RbF-BeF,-UF, system seems to show
promise, especially below 5 mole %
uranium tetrafluoride. The physical
properties of this system are probably
very similar to those of analagous

‘sodium fluoride-bearing systems.

LiF-NaF-BeF,-UF,. The above-
mentioned heating technique has been
applied to various mixtures in the
system LiF-NaF-BeF,-UF, in an attempt
to verify the very low melting points
tentatively reported(?’) for these
materials. Although it is possible
that mixtures melting below 325°C may
be prepared from these materials, the
data previously reported seem to be
much too low. Additional studies of
this system will probably be postponed
in favor of more immediate problems.

Fuels Containing Zirconium Fluoride,
Data froma number of studies of alkala
fluoride—zirconium fluoride mixtures

..is reported in a following section on

- BbE-BeF,-UF,. Additional data on

mixtures of low uranium content in the
system RbF-BeF,-UF, are presented in
Although it appears that
the melting points are somewhat higher

 

(3)J. P. Blakely, L. M. Bratcher, and C. J.
Barton, “Low-Melting Fluoride Systems,” op. cit,,

ORNL-1154, p. 155,

"Coolant Development." Only a limited
number of experiments on mixtures con-
taining uranium tetrafluoride have
been completed and the results are
summarized in Table 24. It seems
likely from this data and that pre-
sented under "Coolant Development™
that the NaF-KF-ZrF,-UF, system may be

101

 

 
 

e o i

it I s, e b bt s

PR RN

Sk 5

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

of definite value as a fuel. Measure-
ments of viscosity have not yet been
made for this material, but microscopic
examination of the solidified melt has
indicated a complete absence of
glasses. It 1s likely that the
viscosity of the liquids will resemble
those of typical fused salt mixtures
rather than glasses. Additional
experiments with this and similar
systems are under way.

TABLE 24
Break Temperatures from Heating and

Cooling Curves of Zirconium-Bearing
' Fuel Mixtures

 

 

COMPOSITION (mole %) |  prpay TEMPERATURES®

 

 

NaF | KF | ZrF, | UF, (°c)
37.5 | 20 |42.5| 0 395, 465, 475
37.1 | 19.8 | 42.1] 1 263, 380, 460
36.7 | 19.6 | 41.7| 2 395, 410, 450
36.3 | 19.4| 41.3| 3 390, 410, 440
35.9 | 19.2| 40.9| 4 375, 451, 505 (?)
35.5 | 19.0 | 40.5| 5 375, 400, 455

5 52 | 43 0 305, 405, 425

4.9 | 51 | 42.1| o0 360, 380, 425

 

 

 

 

 

X
Highest break temperature considered most re-
liable indication of melting point.

SIMULATED FUEL MIXTURE FOR COLD
| CRITICAL EXPERIMENT

D; R. Cuneo L. G. Overholser
Materials Chemistry Division

The cold critical experiment is

‘designed to yield data at room temper-

ature that will be of value in calcu-
lating the critical mass of the final
high-temperature ARE. The primary
requirement of the simulated fuel is
that the density of uranium in the tubes
of fuel mixture be uniform and equal to
the design density of uranium in the
ARE fuel at operating temperatures,

102

In addition, it is desirable to have
the sodium-to-uranium ratio equal to
that in the real ARE fuel and, if
possible, to have the over-all density
of the simulated and real fuels the
same. The hydrogen-to-uranium ratio
must be considerably less than 1; this
will be possible i1f the water content

of the materials is less than 0.01%."

It is presently anticipated that
the ARE fuels will be simulated for
this experiment by packing realistic
fuel tubes with a mixture of uranium
tetrafluoride and sodium fluoride
powders. The Y-12 Production Division
will be responsible for preparing the
enriched fuel and for filling the fuel
tubes. The ANP chemistry group will
render assistance in formulation of
the mixture and experimental study of
tube filling and packing conditions,

Preliminary experiments with uranium
tetrafluoride and sodium fluoride
powders designed to yield information
as to effect of particle size, mixture
composition, water content, and pack-
ing technique on uranium density,
over-all density, and ease of segre-
gation of the mixture are under way.

IONIC SPECIES IN FUSED FLUORIDES
M. T. Robinson

Materials Chemistry Division

Determination of the electrolytic
transference numbers in fused fluoride
mixtures containing uranium tetra-
fluoride was attempted in an effort to
ascertain the nature of the ionic
species in these systems. The equip-
ment and general technique used in the
study of electrolysis of fused fluorides
have been discussed in a previous
report. (%) Uncertainties owing to the
use of an electrically conducting

 

(4)M. T. Robinson, ““Jonic Species in Fused
Fluorides,” op. c¢it., ORNL-1170, p. 82.

 

 

 
 

 

i

L3

diaphragm in the cell and, in part,
to the lack of high accuracy in the
analytical methods available limit the
usefulness of the data obtained.
Consequently, no further studies of
electrolytic transference are now
anticipated, Before these experi-
ments can be improved, an electrical
insulator capable of holding the
salts without reaction and con-
siderably better analytical methods
than now exist will be needed. (An
excellent review of such experiments
has been given by Baimakov and
Samusenko.(%)) The following experi-
mental results from the recent study
are, however, of interest,

The number of electrons necessary
to reduce one uranium(IV) ion to metal
decreases from about four in the more
dilute solutions to less than two 1in
the more concentrated ones. The
changes in the ion fraction of uranium
in the two compartments of the cell
are not altogether consistent. The
anolyte concentration of uranium does
not differ significantly from that
found before electrolysis, whereas the
change in the catholyte concentration
is little more than the experimental
error. On the other hand, the alkali
metal concentration rises sharply in
the catholyte and slightly in the
anolyte. | :

The apparent deviations from
Faraday’s law may represent a case of
stenolysis, i.e., electrode processes

-occurring in the pores of the diaphragm

separating the anolyte and catholyte.
If all current passing through the
cell did so by a stenolytic mechanism,
two electrons would apparently suffice
to reduce one uranium(IV) ion to
metal, whereas complete absence of
stenolysis would require four electrons
per reduced uranium(IV). Tt is
possible that the diaphragm used was

 

(S)Y. V. Baimakov and S. P. Samusenko, “Determi-
nation of Transference Numbers of Ions in Fused
Salts,” Trans. Leningrad Ind. Inst., 3-26 (1938).

‘FOR PERIOD ENDING MARCH 10, 1952

not completely permeable touranium(IV)
and that the deviations from Faraday’s
law result froman increase of stenoly-
sis with increasing uranium tetra-
fluoride concentration. A photo-
micrograph of the surface of the
diaphragm, however, seemed to indicate
a pore size sufficient to pass any of
the 1ons probably present,

The stenolytic behavior, coupled
with the observed concentration
changes, indicates that the quadrai-
valent uranium is probably not present
as Ut4., The small radius of this ion
and its large charge indicate a high
mobility. That this is not observed
is evidence for the presence of some
more complicated species.

PREPARATION OF STANDARD FUEL SAMPLES

G. J. Nessle J. E. Eorgan
V. C. Love Jack Truitt
C. J. Barton

Materials Chemistry Division

A total of 68 batches of fluoride
fuels was prepared during the past
quarter, including two large (3 to
4 kg) batches of pretreated and filtered
material for experiments in dynamic
corrosion in thermal convection loops.
The materials were used to fill equip-
ment furnished by other ANP groups as
indicated in Table 25.

TABLE 25

Disposition of Standard Fuel Samples

NUMBER OF
TYPE OF EXPERIMENT CONTAINERS FILLED
Static corrosion tests 243
Cyc}dtron bombardment 26
Radiation'damage _ 15
Viscosity measurement 8
Density measurement 4
Loop test : ‘ 2

103

 

 
e bl o i i el i, o

R ki

 

 

e e, i, B e, R,

b b Rl b bl ol

 

 

PREPARATION OF PURE HYDROXIDES

D. R. Cuneo D. E. Nicholson
E. E. Ketchen L. G. Overholser
Materials Chemistry Division

Development of moderator-coolants
has been limited during the past
quarter to preparation of pure alkal:
and alkaline earth hydroxides for
corrosion studies. Barium, strontium,
and sodium hydroxide have been prepared
with acceptable purity, that is, con-
taining less than 0.1, 0.06, and 0.1%,
respectively, of their carbonates,
Purification of the potassium hydroxide
has not been so successful.

Barium hydroxide has been prepared,
in a quantity sufficient for use in
dynamic corrosion experiments, by a
procedure(%) that yields 1-1b batches
of material containing less than 0.1
wt % barium carbonate. Strontium

~ hydroxide prepared by the same process

contained 0.06%of strontium carbonate,
and spectrographic analysis showed
about 0.1% barium hydroxide. Strontium
hydroxide may be dehydrated without
difficulty but is not yet available in
large quantity.

Additional batches of sodium hy-
droxide have been purified by removing
sodium carbonate from a 50% aqueous
solution of sodium hydroxide and
dehydrating under vacuum at 425°C.

- These batches have assayed 100.0%

-sodium.hydroxide, and the sodium
carbonate content has been less than

0.1% in'all_qases.

Potassium carbonate can be removed
from potassium hydroxide by dissolution
of the latter compound in isopropyl
alcohol, but it has not proved possible
to remove the alcohol from potassium
hydroxide without carbonization and
contamination of the potassium hy-

~droxide. Potassium hydroxide prepared

 

(6-)1.. G. Overholser, D. E. Nicholson, E. E.

Ketchen, and D. R. Cuneo, *“*Preparation of Pure

-Hydroxides,” op. cit., ORNL-1170, p. 84.

104

 ANP PROJECT QUARTERLY PROGRESS REPOKRT

by dehydration of an aqueous solution
containing 5 wt % potassium hydroxide,
which had been treated with barium
hydroxide, still contained about 0.3%
potassium carbonate, The potassium
carbonate content of higher concen-
tration potassium hydroxide solutiomns
decreased, however. There was 0.1%
potassium carbonate in the 38 wt %
aqueous solution of potassium hy-
droxide. Recent experimental determi-
nations have shown that barium carbonate
is appreciably soluble in potassium
hydroxide. Preliminary studies of the
feasibility of using calcium hydroxide
for this carbonate removal are 1in
progress. In addition, equipment is
under construction for preparing
potassium hydroxide by reaction of
pure water with metallic potassium.

COOLANT DEVELOPMENT

J. P. Blakely R. C, Traber, Jr.
L. M. Bratcher C. J. Barton

Materials Chemistry Division

Investigation of low-melting
fluoride mixtures for possible appli-
cation as coolants has been continued
during the past quarter. Imn addition
to possible value as coolants, any
low-melting compositions discovered
are of potential value as solvents for
uranium tetrafluoride in the prepara-
tion of fuels of low uranium content,

Recent interest in fuel systems
containing zirconium fluoride has led
to a considerable program of study in
phase equilibria in alkali fluoride-
zirconium fluoride systems. Although
several of these systems show promise,
in no case are the data sufficient for
construction of the phase diagram, In
many of these mixtures well-defined
thermal breaks occur well below the
ligquidus temperature, It has not been
definitely established whether these
effects are the result of solad
transitions or the presence of some

 

 

 
 

 

3

impurity such as ZrOF,, which is known

" to be present in the only available

zirconium fluoride.

LiF-ZrF,. The system'LiF-ZrF4,
which is of possible future interest,
is under study in order to obtain a
complete picture of the alkali fluoride-
zirconium fluoride systems. Cooling
curves have been run on nine mixtures
ranging between 15 and 70 mole % zir-
conium fluoride. There appears to be
a eutectic at about 39 mole % zir-
conium fluoride melting at 560 * 10°C.
No congruently melting compounds were
observed, but there 1s some evidence
for an incongruently melting compound
that is probably Li,ZrF . Breaks other
than melting points or the eutectic
halt were observed at 590, 515, and
470°C in some mixtures, This study
will be continued in greater detail
at some future date. ‘

NaF-ZrF,. The study of the system
NaF-ZrF, is still in progress; mixtures
between 10 and 50 mole % zirconium
fluoride have been studied. It appears
likely that there is a eutectic at
approximately 35 mole % zirconium
fluoride, melting at 595+ 10°C, Breaks
other than melting points or eutectic
halts were observed at approximately
735, 530, and 495°C with some mixtures.

 KF-ZrF,., Although cooling curves
have been run on a rather large number
of mixtures in the system KF-ZrF,,
covering the range from 5 to 95 mole %
zirconium fluoride, the melting point
of some compositions has not been
determined with certainty. The
existence of a eutectic at 13 mole %

~zirconium fluoride, melting at 765 *

10°C, was established and a compound
K,ZrF,, melting congruently at 910 *
10°C, has been demonstrated. Mixtures
in the range 30 to 60 mole % zir-
conium fluoride show well-defined

breaks or halts at 415 to 435 and 485°C

for some compositions.

FOR PERIOD ENDING MARCH 10, 1952

RbF-ZrF,, The system RbF-ZrF, has
been studied in the range 5 to 60
mole % zirconium fluoride and appears
to be quite similar to the KF-ZrF,
system., There 1s a eutectic at about
6 mole % zirconium fluoride, melting
at 720 + 10°C, and the compound Rb,ZrF,
melts congruently at 870 %+ 10°C, It
appears likely that there is a eutectic
between 30 and 40 mole % zirconium
fluoride, melting at 575 + 10°C, but
there are also well-marked breaks on
the cooling curves for mixtures in this
range at about 405°C.

NaF-KF-Zr¥,. Cooling curves have
been run on more than 50 mixtures in
the ternary system NaF-KF-ZrF,, and it
should be possible to draw contours
for this system when the corresponding
binary phase diagrams are completed.
The lowest melting composition that
has been found in this system contained
43 mole % zirconium fluoride, 5 mole %

 sodium fluoride, and 52 mole % potassium

fluoride. This mixture exhibited no
break in the cooling curve above 405°C
and stirred down to this temperature.
When heated, i1t started stirring at
405°C and showed no further break up
to 500°C, at which temperature the
experiments were terminated, Other
similar systems showed considerably
higher melting points., The 405°C
melting point therefore seems to
require. further verification. Com-
positions with about the same amount
of zirconium fluoride as the above-
mentioned mixture but with more sodium
fluoride and less potassium fluoride
show well-marked breaks in the cooling
curves at approximately 450°C, which

may indicate a eutectic melting at

this temperature; the composition of

‘this eutectic has not been established.

NaF-RbF-ZrF,. The NaF-RbF-ZrF,
system has been investigated about as
thoroughly as the NaF-KF-ZrF, system.
The lowest melting point recorded was
460°C for the mixture containing 52

105

 
 

ANP PROJECT QUARTERLY PROGRESS REPORT

mole % zirconium fluoride, 11 mole %
‘rubidium fluoride, and 37 mole %
sodium fluoride. Other breaks were
observed in the cooling curve for this
and other compositions at temperatures
ranging from 410 to 440°C. These
could be caused by a solid transition
or by a eutectic of unknown compo-
s1t1on. :

7 NaF-KF-LiF- ZrF . Onlyzafew mixtures
dwere prepared 1n the four component
system NaF KF LiF- ZrF, Increasing
amounts of zirconium fluoride were
added to the ternary NaF-KF-LiF
eutectic (42 mole % KF, 46.5 mole %
LiF, 11.5 mole % NaF), Wthh was con-
sidered to be one component of a
pseudoblnary mixture. The cooling
curves that were run after each ad-
dition of zirconium fluoride showed
that the melting p01nt climbs rather
rapldly, reaches a maximum of 810°C at
20 mole % zirconium fluoride, drops
sharply to 450°C at 30 mole % zir-
¢onium fluoride, and then rises again.

. PREPARATION OF PURE FLUORIDES

CG. J. Nessle ~ F. F. Blankenship
“J. E. Eorgan  W. R. Grimes

: ,Matérials Chemistry Division

- The lack of successful operation of
the thermal loops with fluoride fuel
material as presently supplied and the
definite plans for an ARE in the
relatively near future have emphasized
the need for research on equipment for
preparation of pure, homogeneous,

fluoride solutions and handling of the
llqulds so as to keep them free from
contamination until they are used.

The mass-transfer phenomenon, which
tfansfers metal from hot to cold
regions of the thermal loops, is still
not well understood. However, all the
plugs so far observed have contained
considerable quantities of oxide
sludge admixed with the metal particles.
It has not been established that the
oxides are primarily responsible for

106

corrosion, but freedom of the system
from oxides would certainly help to
clarify the picture. In additiomn, it
is essential for reactor operation that
the fuel be free from sludge formation.

The raw materlals for fuel manu -
facture are, in general, quite hygro-
scopic; 1t 1is difficult to prevent
some pickup of water during the un-
avoidable handling of the powders.
Furthermore, any water picked up by
the fluorides will hydrolyze uranium
tetrafluoride to insoluble uranium
dioxide when the temperature is raised.
Uranium tetrafluoride also oxidizes
readily at elevated temperatures to
yield volatile uranium hexafluoride
and soluble and corrosive U0,F, if any
oxXygen 1s present. It seems evident
therefore that improvement of the
corrosion picture would result if
satisfactory means existed for pre-
paring and handling fluorides so that
oxidation or hydrolization could not
occur. (onsequently, equipment de-
31gned to prevent these reactions
during fuel preparation and transfer
are being developed

Fuel Preparatlon Eoulpment The
fuel preparat1on equipment w111 be a
reactor in which all lines in contact
with the molten fluor1des are of
nickel so that they can readlly be
reduced with hydrogen to remove oxide
films and scale. The reactor will be
charged with dry, pure fluorides,
sealed by the gasketed flanges to the
gas manifold, and evacuated to remove
as much adsorbed oxygen and water as
possible at low temperatures. The
mixture will then be melted and
treated at 500 to 600°C with dry "
oxygen-free hydrogen to reduce any
UO,F, or UF, formed by oxidation or
present in the uranium tetrafluoride
used. The water will react with the
uranium tetrafluoride at this tempera-
ture and yield hydrogen fluoride, which
will be swept and pumped out, and

 

  

 

 

 

 

 
 

o

 

 

uranium dioxide, which will be sus-
pended in the melt, By treatment of
the molten liquid at 500 to 600°C with
dry hydrogen fluoride, the suspended
oxides can be reconverted to fluorides.
The excess hydrogen fluoride will be
pumped off, the liquid swept with pure
helium, and the pure liguid transferred
through a nickel filter to a clean

"receiver by use of helium pressure.

The pure material will be used for
capsule and dynamic loop corrosion
studies, as well as for all physical
property measurements, It is antici-
pated that a number of preparations
will be made under varying conditions
of temperature, exposure time, liquid
composition, etc, to define the optimum

FOR PERIOD ENDING MARCH 10, 1952

conditions for operation before the
material is used for systematic
testing,

Fuel Handling Equipment. TIf the
fuel preparation is to be conducted
under such stringent conditions, 1t
will obviously be undesirable to have
it contaminated before testing. Conse-
quently, a capsule-fillingrig is being
developed in which corrosion capsules
can be treated with hydrogen, hydrogen
fluoride, fluorine, or any desired
reagent, filled with an inert gas,
connected to a source of pure fuel,
filled, sealed, and welded without
exposure to alr or to an uncontrolled
atmosphere.

107

 
 

e Rl L e

 

TS Y

T Y TR I I Y

L A g

 
 

 

3

11. CORROSION RESEARCH

W. R. Grimes, Materials Chemistry Division

W. D. Manly, Metallurgy Division
H. W. Savage, ANP Division

The major emphasis in the corrosion
testing program returned, during the
past quarter, to the procblem of con-
tainer materials for the fused fluoride
mixtures. This effort was extended
considerably followingthe difficulties
encountered during circulation of the
fluorides by forced convection 1in
closed metal systems. Empirical
testing of corrosion of fluoride
mixtures by the static capsule tech-
nique previously described has been
continued, and dynamic testing of
fluoride corrosion 1s being intensively
investigated by several technigques,
for example, thermal convection loops,
seesaw tests, and standpipes. In
addition, a large program of a more
fundamental nature dealing with
possible reactions of metals with the
fluoride melts has been initiated.

Static corrosion testing under
generally isothermal conditions has
been continued with several high-
temperature liquids and structural
metals by using slight variations of
the capsule technique previously
described. Other materials have been
tested in the molten fluorides to
find possible bearing materials and
valve seat materials that would not
be seriously attacked by the fluoride
mixture, Stellite, which 1s a good
valve seat material, is apparently
unattacked in the fluoride mixture

"~ when tested statically.' Additions of

zirconium and magnesium as possible

.corrosion inhibitors were made to the

fluorides prior to running corrosion
tests. It was found that magnesium
increased the corrosion attack,
whereas additions of zirconium de-
creased the attack of the fluorides
on 300-series stainless steels.

Static corrosion tests have also
been run on low melting point alloys
for secondary coolants, such as
sodium-lead alloys and the hydroxides.
Low melting point alloys containing
mixtures of lead, tin, cadmium, and
bismuth have been tested to find
suitable container material. It has
been found that lead-cadmium and
bismuth-cadmium alloys can be stati-
cally contained in type-317 stainless
steel. Corrosion testing in sodium-
lead alloyshas shown that the addition
0of sodium to lead lessens the attack
on stainless steels and Inconel.
Additions of zirconium, sodium, sodium
carbonate, sodium cyanide, and sodium
hydride were made to sodium hydroxide
prior to running statlc corrosion
tests, and the effects of these
additions have been tabulated.

Al though the static capsule tech-
nigque has proved satisfactory for
routine screening of materials, 1t
lacks sufficient sensitivity to
anticipate the results of the more
difficult, large-scale, dynamic tests.
When routine dynamic tests that can
be conducted in large numbers are
developed, the static capsule method
will be discarded. Thermal convection
loops, which have hitherto been used
almost exclusively for dynamic cor-
rosion tests, have now been supple-
mented by a seesaw capsule technique
and standpipe tests.

Dynamic corrosion of the fluorides
and hydroxides has been studied with
these devices. Dynamic corrosion
testing reveals the corrosion behavior
of different alloyswith the fluorides,
the effect of inhibitor additions, and
provides a means for determining the

109

 
 

 

 

Sl ¢ By b e R B A

i e s s

 

 

 

 

effect on the reaction of various

atmospheres, such as argon, hydrogen,
oxygen, and vacuum, in contact with

the bath., One dynamic experiment has

shown that potassium hydroxide can be
successfully contained in Inconel
without corrosion or mass transfer,
For this experiment, the loop was
operated under a hydrogen atmosphere
for 135 hr with a hot-leg temperature
of 715°C and a cold-leg temperature of

- 440°C., This effort on container

materials for molten hydroxides 1is
supplemental to a fundamental approach
to the corrosion problem through
reactionsof hydroxides with structural
metals at high temperatures.

STATIC CORROSIONBY FLUORIDES

F. Kertesz
Materials Chemistry Division

D. C. Vreeland
Metallurgy Division

During the past quarter testing
techniques have been improved and
corrosion ‘tests of a wide variety
of materials in various fluorides
have been completed. Even so, weight
change and penetration data from the
recent tests with dilute fuels do not
justify a conclusion as to which
metal among Inconel and stainless
steels types-316, -321, -347 and -316
ELC is superior. The latest static
tests of Inconel and 300-series
stainless steels in carefully prepared

~fluoride mixtures support the earlier

conclusion that there is little
corrosive reaction at 1500°F.

A series of screening tests in the
untreated fluoride fuel (NaF-KF-LiF-UE,)
for 100 hr at 1500°F showed that the
refractory metals and Stellite were
unaffected and that the stainless
steels were attacked to a depth no
greater than 1 mil. Analyses of the
fluorides after corrosion tests have

110

. ANP PROJECT QUARTERLY PROGRESS REPORT

indicated that no major reaction
capable of changing the physical
properties of the fluoride melt had

occurred. The effects of several
additives to the fluoride have been
determined. Zirconium and uranium

dioxide apparently decrease the cor-
rosiveness of the fluoride, whereas
magnesium and uranium oxide increase
the corrosion. -

Effect of Pretreatment of Fuel
(H. J. Buttram, N. V. Smith, R. E.
Meadows, C. R. Croft, Materials
Chemistry Division). The original

tests with fuels of high uranium

content (120 to 140 1b of uranium per
ft®) showed extensive corrosion (4 to
8 mils) if the materials, as received,
were melted under inert atmospheres
and used in the tests. After treatment
with stainless steel and Inconel,
however, these fuels showed con-
siderably reduced corrosion (0.5 to
2 mils) in 100 hr at 800°C. The
more dilute fuels (5 to 10 1b of
uranium per ft®) containing lithium
fluoride and beryllium fluoride,
which are under study at present, are
not appreciably improved by the
treatment procedure; they show 1 to
3 mils of penetration whether treated
or not treated.

Corrosion of Structural Metals

(D. C. Vreeland, E. E. Hoffman, R. B.
Day, L. D. Dyer, Metallurgy Division).
In addition to a series of screening
tests of various materials in an
untreated fluoride fuel for 100 hr
at 1500°F, some data have been ob-
tained on the corrosion of the treated
fuel on molybdenum, cold-worked
Inconel, and stainless steel. There
is no discernable temperature effect
between 850 and 1000°C on the cor-
rosion of Inconel and several stainless
steels,

The screening tests of various
materials in fluoride fuel (43.5 mole
% KF, 44.5 mole % LiF, 10.9 mole % NaF,

T W

 

 
y

 

 

and 1.1 mole % UF4)-have been com-
pleted. These tests were run with
dehydrated, untreated, fluoride
mixtures for 100 hr at 816°C under
vacuum. Molybdenum, columbium, Monel,
and Stellite 25 were apparently un-
attacked. All the 300-series stain-
less steels tested (types-310, -317,
-321, -347, and -304 ELC) were attacked
to a depth of 1 mil or less, except
type-304 ELC, which was attacked to a

depth of 2 mils. Stainless steels of
TABLE

FOR PERIOD ENDING MARCH 10, 1952

the 400 series were not attackedover
1 mil. Z-nickel, which was the most
severely attacked material of those
tested, was affected to a depth of 5
mils. The attack on Inconel was 3
mils. Details of these tests are
presented in Table 26.

Molybdenum, Timken Alloy 6 (16%
Cr, 26% Ni, 6% Mo, balance Fe), Timken
Alloy 3 (16% Cr, 13% Ni, 3% Mo,
balance Fe), and a 74% nickel-26%

26

Static Corrosion of Various Materials in the Untreated Fluoride Fuel

(NaF-KF-LiF-UF,)* in 100 hr at 1500°F

 

 

 

 

DEPTH OF METAL
MATERTAL AFFECTED (mils) METALLOGRAPHIC NOTES
Globe iron 2 Surface of specimen very rough
Molybdenum ' Surface of specimen roughened
Type-310 stainless steel 1 Subsurface voids
Type-317 stainless steel 1 Intergranular penetration
Type-321 stainless steel 0 to 1/2 Intergranular penetration
Type-347 stainless steel 1 Subsurface voids
Type-430 stainless steel 1 Slight intergranular penetration and
decarburization
Type-446 stainless steel 1 Subsurface voids
Hastelloy B 0 to 1/4 Subsurface voids
Hastelloy C 2 Subsurface voids
.Ihébfiéixxfii 7 1 Subsurface voids
 Stellite 25 (L-605) No visible attack
Z-Nickel Voids along grain boundaries
Tantalum 1 Surface of specimen roughened
Columbium Surface of specimen roughened
Monel No apparent attack
Nichrome V | 3 Subsurface voids, some following grain
: , - R boundaries
Inconel 3 Subsurface voids, some following grain
o e boundaries
Type-~304 ELC staifiieés.steei 2 Subsurface voids, some following grain
: boundaries, attack somewhat irregular

 

 

*Composition in mole %: NaF, 43.5; KF, 44.5; LiF, 10.9; UF,, 1.1

111

T P T

 
 

 

e

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

molybdenum alloy have been tested in
a pretreated fluoride fuel {(compo-
sition in mole %: NaF, 46.5; KF, 26.0;
UF,, 27.5) for 100 hr at 816°C.
Molybdenum was not attacked, and the
Timken alloys and molybdenum-nickel
alloy were attacked only slightly
(1 mil or less).

Static tests have also been run on

specimens of as-received and ap-
proximately 20% cold-worked Inconel
and 20% cold-worked types-316 and -310
stainless steel in the above-mentioned
fluoride fuel at 816°C for 100 hours.
This amount of cold working appeared
to have no significant effect on the
corrosion properties of these materials
under the testing conditions employed.
Tests have also been run at 850, 900,
and 1000°C for 100 hours. The extent
of attack varied from 2 to 5 mils,

‘but no definite correlation between

temperature of test and amount of

attack could be established. Ap-
parently these materials are in-
sensitive to test temperatures within
this range. '

Corrosion by Fluorides with Various
Additives. A short series of tests
on Inconel and type-309 stainless
steel have been run using fluoride
fuel (NaF-KF-LiF-UF,) with zirconium
and magnesium added to the tests in
the form of turnings. The results are
summarized in Table 27,

The additions of magnesium ap-
parently increased the corrosion of
both type-309 stainless steel and

Inconel. Additions of zirconium
appeared to lessen the attack of the
fluoride on these metals. In the

tests with the zirconium additions 1t
was noted that built-up surface layers
of 1 1/2 to 2 mils and 1/2 mil in

thickness were present on the Inconel

TABLE 27

Static Corrosion of Inconel and Type~-309 Stainless Steel
by the Fluoride Fuel (NaF-KF-LiF-UF,)* with MagneSium
and Zirconium Additivesin 100 hr at 1500°F

 

 

 

' DEPTH OF METAL
METAL BATH ADDITIVE AFFECTED (mils) METALLOGRAPHIC NOTES

Inconel None 2 ' Subsurface voids

Inconel 2% magnesium 5 Large voids

Inconel ' 2% zirconium 1 Surface layer 1 1/2 to 2 mils
thick, 1/2 to 1 mil of at-
t ack beneath the surface
Layer

Type-309 stainiesé steel None 2 Subsurface voids

Type-309 stainless steel 2% magnesium 5 Subsurface voids, irregular
attack

Type-309 stainleés.steel 2% zirconium No visible attack, specimen
gad surface layer 1/2 mil

eep

 

 

 

 

*Composition in mole %: NaK, 10.9; KF, 43.5%; LiF, 44.5; UF,, 1.1

112

 

 

 
 

and type-309 stainless steel, re-
spectively. A photomicrograph of the
Inconel corrosion specimen is shown
in Fig, 32. An attempt is being made
to identify these layers by spectro-
graphic methods.

Preliminary experiments have in-
dicated that addition of uranium
dioxide to the fuels does not increase
the corrosion i1n static experiments.
Additionof uranium oxide has, however,
greatly increased the corrosion ob-
served even in the static tests., It
is likely that this is true of any
hexavalent uranium compound including
UO,F,, which is the expected product
of oxidation of uranium tetrafluoride.
This is discussed in a separate
section below.

T T T
.

e .
b S

@

aw

 

- FOR PERIOD ENDING MARCH 10, 1952

Melting Point of Fluorides After
Corrosion Tests (J. M. Didlake, G. J.
Nessle, C. J. Barton, Materials
Chemistry Division). Samples from
26 static corrosion capsules and
dynamic loop experiments were re-
covered to establish whether any
changes in the melting point of the
material had occurred as a consequence
of the corrosion reactions. As in-
dicated by the data in Table 28, the
melting point of these samples is not
affected by the corrosion. In some
cases removal of the liquid from the
capsule left some very high melting
point materials behind. Although
analyses of these materials shows
them to be wet with the molten fluoride,
they seemed to consist of oxides of
uranium and the structural metals and

 

re

 

U

Fig. 32. Corrosion of Inconel in a Fluoride Fuel [(NaF-KF-UF4i t 2% ZrI for gfl?
100 hr at 816°C. 200X. _

113

P
£
£
b
f2
?-»3

 
fhge ER e oy E Rt L : b

i il S o i iz e i

 

 

A s ik G . e, A B, v . B R 1 30 M A, . i

4

!

4

1
I8

?

 

ANP PROJECT QUARTERLY PROGRESS REPORT

of metallic particles from the con-

tainer walls.

melt had occurred.

'STATIC CORROSION BY SODIUM HYDROXIDE

D. C. Vreeland
. R. D. Day

E. E. Hoffman
L. D. Dyer

Metallurgy Division

It was clear that no
major reaction capable of changing the
physical properties of the fluoride

The static corrosion of sodium
hydroxide has been determined with a
variety of specimens and the addition

of presumed inhibitors. None of a
number of special alloys that were
tested was severely corroded after
100 hr at 1500°F, although the attack
was erratic. The corrosion of struc-
tural metals in sodium hydroxide with
various additives was in no case
reduced to insignificance., Of several
refractory materials exposed to the
test conditions only vitrified be-
ryllium oxide survived without crum-.
bling, and it lost considerable

~ weight.

Corrosion of Special AlIdys. A
few tests of special alloys have been

TABLE 28

Melting Points of Samples of Fluoride from Corrosion Tests

 

 

 

 

 

 

 

MELTING POINT (°C)
NO. OF SAMPLES FINAL* ORIGINAL "
.- CAPSULE MATERIAL
 Type-316 stainless steel 9 527 530
Inconel ) 1 526 530
Type-316 stainless steel 10 450 452
Inconel 1 450 452
Type-316 ELC stainless steel 2 450 452
Type-321 stainless steel 1 456 452
Type-347 stainless steel 1 450 452
Type-316 stainless steel 2 458 460
Type-316 stainless steel 4 338 345
LOOP SECTION MATERIAL
Type-316 stainless steel (hot) 1 455 452
Type-316 stainless steel (cold) 1 455 452
Type-410 stainless steel (bottom) 1 453 452
Type-410 stainless steel (hot) 1 453 452
Type-410 stainless steel (cold) 1 453 452

 

*Mean value if more than one

114

sample was

used.

e £

-~

7 e b
Vo by
L‘ o

Y l.‘
i

Gyl

3 C,&_
 

7

run in sodium hydroxide for 100 hr at
1500°F. A 26% molybdenum-74% nickel
alloy showed light attack to a depth
of 2 mils, whereas a 25% nickel-75%
iron alloy was attacked to a depth
of only 1 mil. Other alloys composed
of equal parts of iron and nickel;
iron, nickel, and cobalt; and nickel
and cobalt have also been tested.
Al though one of these alloys was
severely corroded, in general, the
attack was quite erratic with deep
local penetrations. Since the capsules
and specimens were machined from as-
cast bars without prior working or
annealing, it is believed that the
cast structure of these materials may
have caused this type of attack. It

FOR PERIOD ENDING MARCH 10, 1952

is planned to test materials of
simil ar compositions that have been
hot-worked and annealed so as to have
a more uniform structure.

Corrosion by Sodium Hydroxide with
Various Additives. A series of tests
were run with sodium hydroxide and
various addition agents, 1including
zirconium, sodium, sodium carbonate,
sodium cyanide, and sodium hydride.
The addition of approximately 50% of
sodium cyanide appeared to lessen
attack somewhat on type-316 stainless
steel and Inconel. None of the agents
tested appeared to reduce corrosion to
an insignificant level., Details of
these tests are presented in Table 29.

TABLE 29

Static Corrosion of Structural Metals by Sodium Hydroxide with
Various Additives in 100 hr at 15060°F

 

 

 

WEIGHT
- CHANGE DEPTH OF METAL
MATERTAL ADDITIVE (mg/in.z) AFFECTED (mils)|  METALLOGRAPHIC NOTES
Type-304 stainless steel 2 : 5-mil oxide layer
Type-304 stainless steel | 10% Zr 2 1/2 6-mil oxide layer
Type-304 stainless steel | 30% Na,COj 3 6-mil oxide layer
A-Nickel ' 1% Zr No attack
Inconel 12% Zr 3 11-mil oxide layer
Type-316 stainless steel | 48% NaCN -286.6 2 Intergranular attack
Inconel 49% NaCN -207.3 7 Intergranular attack
Inconel T.4% Na Complete Specimen converted
penetration completely to corrosion
‘ . product
Inconel 50% Na Complete Specimen converted
penetration completely to corrosion
. product
A-Nickel 14% Na +51.1 No corresion product,
‘ but edge of specimen
rough
A-Nickel 54% Na -2.5 No attack
Inconel 31% NaH Complete Specimen converted
penetration completely to corrosion
product
Type-304 stainless steel | 31% NaH -24.9 4 1/2 Oxide lavyer

 

 

 

 

 

115

 
 

ol ik G, B e v

 

 

Lead-sodium alloys

ANP PROJECT QUARTERLY PROGRESS REPORT

Corrosion of Refractory Materials,
A number of refractory materials were
tested in sodium hydroxide in capsules
of nickel at 800°C for 100 hours.
Carbides of tungsten, tantalum,
columbium, titanium, and silicon and
zirconium nitride, all in the hot-
pressed condition, disintegrated
during the test. Vitrified beryllium

oxide lost considerable weight but

survived without crumbling.

STATIC CORROSIONBY LIQUID METALS

D. C. Vreeland E. E. Héffmén
R. B. Day L. D. Dyer

Metallurgy Division

Liquid metal corrosion studies
during the past quarter have been
limited to the testing of various
low melting point alloys of potential
use as a secondary coolant and a
number of lead-sodium alloys. Inconel

~and types-310 and -317 stainless steel

were tested in the low melting. alloys
for 100 hr at 1500° F. The type-317

stainless steel was Gnattacked by the

82% lead-18% columbium alloy and only
slightly attacked by the 60% bismuth-40%
cadmium alloy, whereas Inconel and the
type-310 stainless steel were generally
subject to somewhat greater attack.
containing as
little as 5% sodium have been suc-
cessfully contained in Inconel and a
number of 300- and 400-series stain-
less steels with no greater than

1/2-mil attack after 100 hr at 1500°F.

Corrosion by Low Melting Point
Alloys. A low melting point alloy is
being considered as a coolant in a
secondary heat exchanger, so several
eutectic compositions of low melting
point alloys have been employed as
corroding mediums in static corrosion
tests with types-310 and -317 stainless
steel and Inconel. The compositions

116

and melting points of the eutectic

alloys used are listed in Table 30.

TABLE 30

Melting Points of Various Alloys

COMPOSITION MELTING POINT
BY WEIGHT (°c)
44% Pb-56% Bi 124
43% Sn-57% Bi 138.5
60% Bi-40% Cd 144
. 68% Sa-32% Cd 176
38% Pb-62% Sn 183
82% Pb-18% Cd 248

In these experiments, which were
run for 100 hr at 816°C under vacuum,
all the low melting point alloys
conta1n1ng tin (except Inconel in’

w68% tin232% cadmium> alloy) proved to.

be extremely vigorous in their attack
on the metals tested. It is believed
that the type of attack in these
mediums can be classified either as
intergranular, as in Fig. 33 that
shows the characteristic penetration
along the grain boundaries sometimes
accompanied by voids, or as an alloying
type of attack in which the attacked
surface actually alloys with the
molten coolant being tested, as shown
in Fig. 34.

Inconel did not show much promise
in the tests. Type-317 stainless
steel was unattacked by the 82%
lead-18% cadmium alloy and only
slightly attacked by the60% bismuth-40%
cadmium alloy. The results of these
tests are summarized in Table 31,
It is planned to test other structural
materials in the low melting point
alloys that proved least corrosive in

the initial tests.
 

 

 

 

W

il

»

i

FOR PERIOD ENDING MARCH 10, 1952

[ UNCLASSIFIED
40 5

ol 4t
Lo

‘;jl ,\‘;t N F
Sea e BYGS

 

Fig. 33. Intergranular Attack of Type-310 Stainless Steel Tested in
*44% Lead-56% Bismuth Alloy for 100 hr at 1500°F. 200X,

 

UNCLASSIFIED
_ Y-5328

Fig. 34. Alloying Attack of Inconel Tested in 43% Tin-57% Bismuth Alloy for
100 hr at 1500°F. 100X.

i

2

(T @ - 117

&
5

PO
£ 74

 
Sl des i,

i o SR s it R D s o

 

 

©

'ANP‘PROJECT QUARTERLY PROGRESS REPORT

TABLE 31

Corr051cn of Types-310 and -317 Stainless Steel and Inconel by Varxouscfi

Low Melting Point Alloys in 100 hr at 1500 F

 

 

 

 

 

 

 

A series of static corrosion tests of
various materials in three different
sodium-lead mixtures (80%. sodium-20%

118

s
,('f“

T R
e Ly N
Vv b

ALLOY .
COMPOSITION DEPTH OF METAL
BY WEIGHT MATERIAL AFFECTED (mils) METALLOGRAPHIC NOTES ..
43%75n-57% Bi Type-310 stainless steel 10 Many large voids
43% Sn-57% Bi Type-317 stainless steel 11 Irregular attack, many voids .
43% Sn-5T% Bi Inconel 8 5-mils of a uniform layer on
surface with an underlying
layer of voids, 3-mils in
| thickness )
- 60% Bi-40% Cd Type-310 stainless steel 15 Attack very irregular, varying
s ' 0 to 15 mils, both grains and
grain boundaries attacked in
some cases
- 60% Bi-40% Cd Type-317 stainless steel 2 Intergranular penetration and
' voids in a few areas )
-60% Bi-40% Cd Inconel Complete pene- Voids throughout entire speci-
tration of men, tube attacked to a depth
_ specimens of 20 mils
' 38% Pb-62% Sn Type-310 stainless steel | Complete pene- Erratic attack
S : ‘ tration of
. ; B specimens
38% Pb-62% Sn Type-317 stainless steel Complete pene- Erratic attack, only 2 mils .
. _ tration of affected 1n places -
7 \, specimens
. 38% Pb-62% Sn Inconel 8 Erratic attack
68% Sn-32% Cd Type-310 stainless steel | Complete pene- Erratic attack, tube failed by
: tration of penetration of wall
. specimen
68% Sn-32% Cd Type-317 stainless steel Complete pene- Erratic attack
: tration of
specimen
68% Sn-32% Cd | Inconel 2 Uniform attack .
82% Pb-18% Cd Type-310 stainless steel 2 Intergranular attack
82% Pb-18% Cd Type-317 stainless steel 0 No attack on specimen or tube
82% Pb-18% Cd Inconel 10 Intergranular attack
44% Pb-56% Bi Type-310 stainless steel 6 Intergranular attack varying
from 2 to 6 mils
44% Pb~56% Bi Type-317 stainless steel 15 Intergranular attack :
44% Pb-56% Bi Inconel 4 Intergranular attack varying .
' from 1 to 4 mils
Corrosion by Sodium-Lead Alloy. lead; 50% sodium-50% lead; and 5%

sodium-95% lead) has been completed.
The tests were run at 816°C for 100 hr

under vacuum.

Only in the case of
 

3
e

Inconel in the 5% sodium—95% lead

"mixture was the depth of attack

greater than 1/2 mil; Inconel was
affected to a maximum depth of 3 mils,
The various test combinations included
Inconel and types-317 and -346 stain-
less steel in all the alloys, type-316
stainless steel in the 50-50 alloy,
and type-304 stainless steel in the
5% sodium-95% lead alloy. Except for
the aforementioned case of Inconel,
the 5% sodium-95% lead alloys show no
attack whatsoever. All specimens
were ductile on bending 180 degrees.

It was noted that some of the
stainless steel specimens that had
been tested in the 80% sodium-20% lead
mixture had apparently become carbu-
rized.
observed before in testing similar
materials in sodium.(!’ No carbu-
rization of any of the stainless
steels was noted with the 50% sodium-50%
lead or 5% sodium-95% lead mixtures,
which indicates that occurrence of
carburization is enhanced by the
presence of higher percentages of
sodium in the bath metal. Future
plans include running static cor-
rosion tests with lower percentages
of sodium in the bath metal.

FACILITIES_FOR 'DYNAMIC CORROSION
TESTING

Static corrosion tests are satis-
factory'for the routine screening
of corrosion resistant materials, but
they lack the sensitivity to antici-
pate the results of dynamic testing.
The mass transfer phenomenon and the

deep penetration in the hot zone

that are characteristic of all the
dynamlc thermal convection loops are
not observable in static isothermal
capsule studies. These latter ex-
periments are, however, easy to

 

(I)F. N. Lyon (ed.), Ligquid Metals Handbook,
NAVEXOS P-733, p. 90, June 1, 1950.

OF% 107
Gt Cf ,.,Ld i

This phenomenon has been

FOR PERIOD ENDING MARCH 10, 1952

perform in large numbers and under
conditions of better reproducibility.
Dynamic corrosion testing 1n the
thermal convection loops in which the
bulk of the testing has been per-
formed has now been supplemented by a
number of techniques that promise to
duplicate the phenomena shown in the
large loops with much smaller and
more economical equipment. One of
these technigues is the standpipe
test 1n which controlled thermal
gradients may be introduced. The
most promising, however, 1s the seesaw
technique with which fluid motion as
well as thermal gradients are readily
attained.

Thermal Convection Loops. The
large thermal convection loops, which
have been in operation for almost
two vears, have been described in
previous reports.(2'3'%)  The loops
are usually fabricated of 1/2-in.
pipe and are shaped roughly in the
form of a rectangle 1 ft wide and 3 ft
high. When the bottom and an adjacent
side of the rectangle are heated,
convection forces in the contained
fluid establish a flow that attains
a velocity of up to 8 ft/min de-
pendlngupon the temperature difference
across the hot and ¢old sides of the
rectangle. &

Recently, smaller convection
loops have been used in these ex-
periments. These loops afford the
simplest and most direct means of
studying mass transfer, since the

 

(2)E. M. Lees, J. L. Gregg, and R. B. Day,
“PDynamic Corrosion Tests in Thermal-Convection
Loops,” Aircraft Nuclear Propulsion Project
Quarterly Progress Report for Period Ending
September 10, 1951, ORNL-1154, p. 124.

(B)E. M. Lees, *“Thermal-Convection Loops,”
Aircraft Nuclear Propulsion Project Quarterly
Progress Report for Period Ending June 10, 1951,
ANP-65, p. 150.

E. M. Lees, “Thermal Convection Loops,”
Aircraft Nuclear Propulsion Project Quarterly
Progress Report for Period Ending Marchk [0, 1951,
ANP-60, p. 203.

119

 
B B R i, B W

& ki L

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

characteristics are such as to ac-
centuate the effects leading to mass
transfer. The loops can be adapted
to allow the exclusion of oxygen,
water vapor - and to some extent
metallic oxides — from the system.
The loops are formed of 1/2-in.-ID
tubing with a nominal wall thickness
of 0.035 in. into the form of a
rectangle, 7 by 17 in., with a loading
tube extending above the hot leg of
the loop. A dehydration pot is
connected with the loop by standard
1/4-in. tubing. Glass tubing is
attached at the tops of the loading legs
of both the loop and the dehydration
pot by means of Kovar seals for
connect1ng ‘vacuum and hydrogen lines.
The furnaces for heating the loops

_are constructed from heating coils
“‘mounted in split, cylindrical ceramic
forms fitted around the sections of
;the_loop._

Seesaw Corrosion Tests. The seesaw

tests require a sealed metal tube
approximately one-third full of molten
“material.
“'a furnace capableof heating to 1700°F

;and up to 18 sealed corr081on capsules

The capsules are placed in

To create
a temperature differentr1al, the
capsules may be placed so that one end
emerges from the furnace. The tubes
are gently rocked on a central fulcrum
similar to a seesaw so that the molten
medium flows by grav1ty to both ends
of each tube The apparatus is shown
in Fig. 35. After several thousand
cYcles,'the tubes are sectioned for
examination. It is anticipated that
in this manner the mass transfer and
enhanced corrosion phenomena can be
obtained and a variety of condltlons
for 1mprovement studled

D1fferent1a1 Temperature Tests. A |

type of test frequently referred to
as the standpipe test involves capsules
similar to those used in the seesaw
tests, except that they are almost
completely filled with the fluid and

120

with desired pretreatment,

are held stationary in a vertical
position. This is not a true dynamic
test because the fluid is essentially
immobile. With this type of test,
however, the temperature gradient may
be controlled so that it is hottest
at the top, bottom, or center. Normal
convection provides the only means of
circulation. After a suitable time,

the tubes are sectioned and examined.

Another design with which a large
thermal gradient may be obtained
employs a U-shaped tube. In this
apparatus there is a large temperature
differential between the two legs,
but there is the smallest possible AT
from top to bottom of each leg (to
minimize local convection currents).
The object of such tests with this
apparatus is to determine the contri-
bution of temperature difference alone
to corrosion and/or metal transport.

A third type of apparatus will
permit one portion of a metal specimen
to be water-cooled internally. One
end of the metal will be immersed in a
crucible containing stirred or un-
stirred molten test fuel at aspecified
temperature. This equipment will be
contained in a 5-1 pot furnace.

Rotating Dynamic Corrosion Tests.
Another tester incorporates means for
preparing the li1iquid to be used
and for inserting a rotating specimen
that acts as a small pump for agitating
the liquid. Basic components of the
apparatus are a lined pot for the
preparation of the eutectic and a
lined receiver with a screen between
the pot and receiver. The sequence of
operations includes preparing the
eutectic from components in the pot,
| and then
up-ending the device to allow the

molten fluoride to run into the

receiver. The pots are then separated
in a dry box and the receiver fastened
onto the stuffing-box assembly that
includes a weighed andmeasured rotating

s By

T b L
 

 

 

¢t

Fig.

 

 

35.

 

    

Seesaw Apparatus for Dynamic Corrosion Tests.

‘0T HOUVIN ONIANI d0oldidd 404

cS6T

 
B2l Lo caig D L g Do Rgsio i cBb g -

e e R

 

phenomena,

‘materials.

 

 

ANP PBOJECT QUARTERLY PROGRESS REPORT

test specimen. After a test at any
desired temperature the specimen can
be completely examined and the eutectic
analyzed. '

Forced Convection Loeps. Forced
circulation loops are not routinely
used for corrosion research. Several
pump loops, heat exchanger loops, etc.
are operating with both sodium and the
fluorides, but these are primarily for
the development of the mechanical
equipment (sec. 3). The flow rates
attainable in thermal convection loops
have been adequate to expose corrosion
When corrosion tests with
higher flow rates are necessary, they
will be performed in forced convection
loops.

‘-DYNAMIC'CORROSION BY FLUORIDES

A program for‘dynamic testing of

various fluoride mixturesin containing

materials is being conducted in the
large (1/2 in. ID) thermal convection

loops. Duringthis period the emphasis
‘has been on screening tests to deter-

mine the more feasible container
At present none of the
stainless steels or Inconel will
satisfactorily contain the fluoride

mixtures under dynamic conditions at

1500°F, In general, the uranium

‘bearlng fluoride (NaF-KF-LiF- UF,)

ls more corrosive ‘than the eutectlc
mixture NaF KF-LiF. Furthermore, with
the above fluoride fuel mixture, no

_ 400-series stalnless steel loop has
~operated more than 15 hr before

plugging, whereas no 300-series stain-

‘less steel loop has operated for more
~than 150 hr before plugging.
‘loops, on the other. hand, generally
‘run to scheduled termination at 500 hr

‘Inconel

without plugging, but the loop material
is usually attackedto 10 mils or more.
Preliminary results of fluoride

corrosion in seesaw tests have sub-
stantiated previous results, which

122

indicated that uranium-free fluorides
were less corrosive than similar
systems containing uranium tetra-
fluoride. In these seesaw tests, as

~well as similar standpipe tests, the

formation of uranium dioxide crystals
in the cold zone was observed.

Corrosion by Fluorides in Thermal
Convection Loops (G. M. Adamson,
K. W. Reber, Metallurgy Division).
It is of particular interest that no
system or mechanical device other
than convection loops has ceased
operatlon because of plugging, parent
metal pipe failure, or liquid leakage
(except at poorly designed welds).
It is equally true, however, that some
of the other systems would not have
endured had the walls and parts been
thin enough to permit full penetration
of 1ntergranular corrosion., All
loops, with one exception, have been
of the same design and have been
fabricated from 1/2-in,-ips, sched-
ule-40 pipe.

Table 32 is a summary of the cor-
rosion data from all loops, including
those still in operation, in which
tests have been made with fluorides.
It is evident that considerable cor-
rosion occurred in all loops examined
so far and that the presence of uranium
in the coolant accelerates plugging.
The two stainless steel loops that
contain fluoride mixtures without
uranium show no signs of plugging
after 350 hr of operation, whereas
all stainless steel loops containing
the uranium-bearing fluorides have
plugged.

The 300-series stainless steels
with uranium-bearing fluorides usually
plugin from 100 to 200 hr in convective
systems, and the life appears to be

"on the order of 10 hr/mil (0.001 in.)

of thickness. The 400-series stainless
steels usually plug in 50 hr or less.
Inconel loops, on the other hand, have
not plugged in any of the 500- and
 

w 4 B w

 

 

 

 

¥ ak -'.“‘ - . N w» * - .y
. .
oy 1 4o o “w o - *®
. " .
-y (]
1 @
Corrosion Data from Inconel and Stainless Steel Thermal Convection
Loops Operated with Various Fluorides
HOT MENIMUM MINIMUM
. TIME OF REASON LEG TEMPERATURE TEMPERATURE METALLOGRAPHIC. NOTES
. CIRCULATION FOR TEMPERATURE (starr) (terminatien) BAP!
LOOP NO. MATEREAL COOLANT (hr) TERMINATION (°F) (°F) X-RAY STUDIES HOT LEG COLD LEG
78 Inconel Flinsk 12¢% 1000 Scheduled 1500 1220 1220 Intermittent layer of pits, § mils, {No attack, ne mass transfer
) : intergranular attack, 10 mils
111 Type-316 stainless steel Flinak 12 173 Leak 1525 1210 1150 Very cvarse-grained metal, inter- Very coarse-grained metal,
granular attack 10 mils, some continuous layer, 1 mil,
) ‘ reduction of wall no attack
Pt .
8.7 A 112 Type-316 stainless steel] Fulinak 14¢9) 82 Plug 1500 1210 960 Second phase in hot leg, possibly intergranular attack up to 8 mils Intermittent mass transfer
Q‘T"i ~ in cold leg i layer
$ 113 Type-316 stainless steelf Fulinak 14 ' 123 Plug 1600¢ D 1250 ? Intergranular attack 8 mils, some |Thin mass transfer layer
. . pitting
210 Inconel Fulinak 14 - 500 Scheduled 15G0 1245 1245 Layer of pits 10 mils, with maxi- Slight surface roughening
v mum of 13 mils; layer on surface
Pl _ : 2 mils
€_§,35 211 Enconel ' Ffilinak 14 506¢ Scheduled 1500 1250 1245 $Second phase in hot horizontal Layer of pits, 4 to 8 mils, vough |Thin continucus layer on
: - end cold legs surface surface, surface rough
212 Inconel Fulinak 14 : 38 Leak 1500 1230 1230 Intermittent layer of pits, 5 mils
40 Type-410 stainless steel|Fulinak 14 9 Plug 1564 1246 1115 Second phase in hot horizontal
and cold legs, possibly in hot leg
43 Type-410 stainless steel| Fulinak 14 12 Plug 1500 1230 1166 Second phase in both parts of hot
leg
48 Type-43¢ stainless steelf Fulinak 14 8 Plug 1500 1250 ? Second phase found in all sectioms
49 Type-430 stainless steel] Fulinak 14 9 Leak 1500 1250 1150
275 Type-347 stainless steel| Fulinak 14 39 Leak 1500 1250 1100 Second phase in hot horizontal leg {Intergranular attack, 8 te 13 mils,|Layer of 1 mil with nonme-
grains 50 times larger than in cold} tallic particles on and in
leg it
251 Type-310 stainless steel| Fulinak 14 . -73 Plug 1500 1260 1140 Second phase in hot horizontal leg,
possibly in cold leg
104 Nickel Fulinak 14 500 Scheduled 1500 1260 1185 Second phase in all parts of hot
leg, possibly in cold leg
118 |Type-316 stainless Fulinsk 14 147 Plugtf?
steell )
214 Inconel Flinak 12
+ NaK 500 Scheduledt)? L
116 Type-316 stainless steel{ Flinak 12 350 (s
119 Type-316 stainless steelf Flinak 12
+ NaK 350 5
213 Inconelif ®} Fulinak 14 200 o
216 Inconel Fubena 17¢%? 225 e
211 Inconel Fubena 17 375 8
365 Nimonieft? Fulinak 14 225 (g
276 Type-347 stainless steelf Fulinak 14 19 (g

 

 

 

 

 

 

 

 

 

 

 

{@)aat recorded temperature, usually 1 to 2 hr before termination.
€8 Flinak: 11§ mole % NaF, 42.0 mole % KF, 46.5 mole % LiF.
(Pylinak: 10,9 mole % NaF, 43.5 mole % KF, 44.5 mole % LiF, 1.1 mole % UF,.

(d)This loop held at 1500°F for 70 hr and then turned up to 1600°F for the remsining time.

()ifydrogen-fired with H, atmosphere.

YA

(’)Lcmps recently terminated, other data not available.

(S)Laop still operating.
(A)Fybena:

COMhis loop two-thirds usual size.

51 mole % BeF,, 47 mole X NaF, 2 mole % UF,.

‘0T HOMVW ONIONA QOI¥dd H0d

cS61

 
 

i R,

i e L ..

o L

 

Rl W, o Ao -G e

e, i i PG

 

el L i

 

 

 

ANP PROJECT QUARTERLY PROGRESSJREPORT

100-hr testé, and corroéion &ata
indicate a life on the order of 100
hr/mll of thlckness

Chemical analyses of hot- and
cold-leg materials after operation
indicate an increase in concentration
of heavier components in the cold leg
and of lighter components in the hot
leg. The tube walls of the hot and
coldlegs from an Inconel loop (No. 210)
that ran for 500 hr with the fluoride
fuel (NaF-KF-LiF- UF,) are shown in
Fig. 36. The attack shown 1in the
hot leg is more than can be tolerated
if a thin-walled tube is to be used.
Very little, if any, attack is found
in the cold-leg section. The hot and
cold legs of a type-316 stainless

" steel loop (No. 113) that plugged

after 123 hr of circulating the
fluoride fuel are shown in Fig. 37.
The mechanism and nature of the
plugging in the stainless steel loops
have notyet been determined. Metallic
dendrites can be found in these loops,

“but as yet it is not certain that
Kthey are present in sufficient quanti-
~ties to cause the plugging.

Corrosion by Fluorides in a Seesaw
Furnace (A. D. Brasunas and L. S.
Rlchardson, Metallurgy Division).
Two seesaw tests were made with the
fluoride m1xture_(10.9 mole % NaF,

3.5 mole % KF, 44.5 mole % LiF,
1.1 mole % UF,) contained in Inconel
tubes in which the tubes were "rocked”
for 162,000 cycles (450 hr). The
hot-zone temperature was approximately
800°C and temperature drops of 120 and
180°C were maintained. Manytiny,
metal dendrites were observed em-

‘bedded in the fluoride at the hot and

cold ends of the tubes. They were

 presumably attached to the wall prior
‘to the solidification of the fluoride

and have been identified as face-
centered-cubic metgl with a lattice
parameter of 3.541 A. An analysis of
these crystals indicated the following

124

~and 81.5% nickel.

‘uranium tetrafluoride.

13.6% iron, 4.9% chromium,
The fluoride mixture
was also analyzed for metal con-
stituents, and chromium was found to
be the major metallic impurity. This
explains the change in composition of
the crystals and tube surface from
that of the original Inconel (14% Cr,
6.5% Fe, 79.5% Ni).

composition:

The cold zone also contained
appreciable quantities of black
crystalsof high melting point embedded
in the fluoride. These crystals were

‘identified by x-ray diffraction as

uranium dioxide. The oxygen may have
come from impurities added to the
fluoride during preparationor loading,
The roughened surfaces of the hot
zones of the tubes indicated that some

‘crystal deposition may have occurred

throughout the tube.

Preliminary experiments employing
the seesaw capsules indicated that the
uranium-free fluoride systems do not
show the mass transfer phenomenon and
that they are generally less corrosive
than similar systems containing
The mass
transfer phenomenon 1s not yet under-
stood, but it is probably caused by
the reaction of higher valence com-
pounds of uranium,. o

Standpipe Tests of Fluoride Cor-
rosion (A. D. Brasunas and L. S.
Richardson, Metallurgy Division).
Results have been obtained from tests
with Inconel capsules containing the
fluoride fuel (10.9 mole % NaF, 43.5
mole % KF, 44.5 mole % LiF, 1.1 mole
% UF,). In duplicate tests, with a
maximum temperature of 815°C and a
temperature difference of 165°C, one
capsule showed very slight mass
transfer and the other none. In both
these tests, however, separation of a
second, reddish-black phase occurred,
which was assumed to be uranlum
dioxide.
 

1952

FOR PERIOD ENDING MARCH 10,

. 0“ '

 

 

f Hot and Cold Legs of an Inconel Convection Loop After

F1008 o0

Sect
lating the Fluoride Fuel (NAF

.

36

Fig.
ircu

&an i

QO o™

— —t

=

©

[

o~

a3

Nt

w0

bt

= %

C =

=
o

=]

0'

3 [y,
o

- O

c o

=
pnd

—

g -

B o

=]

[

e

i )

- O

vow

E
oo

' Q maw
oyt 3
0 o]
o T
H .
~
- 123
~ Ly

250X.

1250°F.

-

section,

C
 

e

.

 

e b

§ R ORI TP ST

4

 

 

k. i

i e, ¢

 

 

 

 

 

e
& %é‘gi?;;m
L

o e
Ehaie %gss AR SR \\w@&@%\
e e

=

SR
Lt
SR

et

T«617

s
\§?~:§&\%&§

%

: SRR
e . : SRk 5
e e quf‘
BT Em e 2T i & ; R
W%’%‘S; 2 5 e 5 SRS AR
e 2 G

2
R

R
e
N

S
R

S o
o
PN e

 

Fig. 37. Sections of Hot and Cold Legs of a Type-316 Stainless Steel Convec-
tion Loop After Circulating the Fluoride Fuel (NaF-KF-LiF-UF,) for 123 Hours.
(a) Cold leg section, <1250°F. 250X. (b) Hot leg section, 70 hr at 1500°F,
63 hr at 1600°F. 250X. '

126 | £

 

-
.
L E-
< -
-
.
=
-
:
M
-
=
I
- ¥
"
o
.
_
 

 

DYNAMIC CORROSIONBY HYDROXIDES

The dynamic corrosion of sodium
hydroxide in nickel convection loops
and of potassium hydroxide in Inconel
loops has been investigated., Both
these systems have been operated in a
carefully maintained hydrogen atmos-
phere, and comparable runs have been
"obtained with the sodium hydrox-
ide-nickel system in air and in
vacuum. All these loops were carefully
cleaned and hydrogen-fired. The
hydrogen itself was first dehydrated
and then loaded and maintained under
the desired pressure.

Appreciable mass transfer occurred
in the three sodium hydroxide-nickel
thermal convection loops that have
been operated. The presence of oxygen
definitely increased the oxidative
corrosion in the loop; this type of
corrosion can possibly be completely
eliminated if the loop 1s operated
under an atmosphere of hydrogen.

One Inconel-potassium hydroxide
loop operated for 135 hr showed neither
mass transfer nor appreciable cor-
rosion. The maximum hot-leg tempera-
ture was only 715°C.

Preliminary tests in the seesaw
furnace and standpipe have been
conducted with sodium hydroxide.
Standpipe tests under hydrogen atmos-
phere have shown neither mass transfer
nor oxidation with hot-zone tempera-
tures of 740°C., Both seesaw and
standpipe tests under vacuum have
shown the now-familiar crystal for-
mation (predominately nickel oxide)
and mass transfer.

Corrosidh“By'HydrOXides in Thermal

Convection Loeps (G. P. Smith, J. V.
Cathcart, W. H. Bridges, Metallurgy
Division). Five Inconel loops con-
taining potassium hydroxide and four
nickel loops containing sodium hy-
droxide have been operated. These

ST R (T g e

G i M

" FOR PERIOD ENDING MARCH 10, 1952

loops areof the second type previously
described under "Thermal Convection
Loops.”™ The care taken in these
experiments is exemplified by the
following procedure:

1. The loop and hydroxide (in the
attached loading pot) were
hydrogen-fired at temperatures

up to 150°C,

2. The hydroxide was dehydrated by
maintaining the system under

vacuum for 48 hr at 500°C.

3. The system was refired with
hydrogen.

4. The hydroxide was loaded into
the loop under hydrogen pressure,

5. The hydroxide in the loop was
maintained under hydrogen, air,
or vacuum, as the case may have
been.

Operation of one of the sodium
hydroxide-nickel systems ended prema-
turely. The remaining three loops
were operated in air, vacuum, or
hydrogen, and in each loop a con-
siderable amount of mass transfer
occurred between the hot and cold
legs. In the loop operated under
vacuum for 317 hr, there were moder-
ately heavy deposits of nickel in the
cold leg both in the form of dendritic
needles and as a dense, compact layer
on the loop walls.

The loop run under an oxygen atmos-
phere for 117 hr (Fig. 38) showed no
polished regions even in the hot leg.
However, the wall surfaces of the

_entire inside of the loop were covered

with a heavy, black powder that was
found to contain nickelous oxide,
metallic nickel, and an unidentified
constituent. Crystals of nickel
could be distinguished in the cold
leg, and there was an extremely heavy
deposit of nickel crystals in the

127

U

 
 

 

1 . ANP PROJECT QUARTERLY PROGRESS REPORT

  

 

 

 

 

 

 

 

X >
i *
: , iy - s
n"’:_{;:‘z(:
3
- A
: -~
5 - .,
.
4 5
€ "\ .
i
3 -
&
4
o .
A .
3
%
1
4
.
3
{
g .
*
. . ]
: w v
. =
s
-4 . -

 

PR Fig. 38. Nickel Thermal Convection Loop Operated for
S 117 hr with Sodium Hydroxide Under an Air Atmosphere.

o
:

Y
.

    

L

T

   
 

 

 

o »

top crosspiece of the loop. There
was no evidence of intergranular

- attack.

The sodium hydroxide-nickel system,
which operated under hydrogen for
296 hr (Fig. 39), showed considerable
mass transfer between the hot and
cold leg. The hot region showed a
high surface polish that extended over
the entire high-temperature region.
It was significant that no oxidation
occurred inside the loop, as evidenced
by a complete absence of black powder.
Metallographic examination of this
loop has not been completed, and
hence no information is available
concerning the extent of removal of
metal from the hot-leg walls,

Operation of four of the five
Inconel loops with potassium hydroxide
ended prematurely, but the fifth loop
was operated for 135 hr with hot- and
cold-leg temperatures of 715 and

440°C, respectively (Fig. 40). The

run had tobe stopped at 135 hr because
a failureof the temperature controller
allowed the cold leg to freeze. No
mass transfer occurred in this loop.
Virtually no change could be noted in
the wall thickness of either the hot
or the cold legs. A slight smoothing
out of the irregularities in the loop

wall was noted in both the hot’ and -
cold legs, but otherwise there was,

little ev1dence of corr031on

 

Corros1on byVSodium Hydrox1de 1n1

Seesaw Tests (A. D. Brasunas and L. S
Richardson, Metallurgy D1v131on)

Several tests were made in the seesaw

dynamic corrosion apparatus by using
ASTM-grade nickel in conjunctlon with
molten sodium hydrox1de in a vacuum.

An abundance of crystals was formed'
in the cold zones of the tube afterm'
117 hr, and the usual surface polishing

was noted in the hot zones (Fig. 41).
Similar tests with nickel and sodium
hydroxide produced, in addition to
the usual metal crystals, appreciable

FOR PERIOD ENDING MARCH 10, 1952

quantities of nonmetallic crystals.
Both the black, comb-like crystals
and the hexagonal, green platelets

~were positively identified as nickel

oxide.

Standpipe Tests of Sodium Hydroxide
Corrosien (A. D. Brasunas and L. S.
Richardson, Metallurgy Division). Two
standpipe tests with molten caustic
have been completed. One of the tests
with nickel and sodium hydroxide was
operated under vacuum and the other
under a hydrogen atmosphere. For the
test in vacuum, the maximum hot-zone
temperature was 820°C and a thermal
gradient of 140°C was maintained over
the length of the pipe. A moderate
amount of metal crystal formation
occurred, and some oxidation was
noted. The test under hydrogen, with
a maximum hot-zone temperature of
740°C and a temperature gradient of
140°C, showed neither mass transfer
nor oxidation,

FUNDAMENTAL CORROSION RESEARCH

Several fundamental approaches to
an understandingof corrosion phenomena
are being pursued. It i1s still too
early to say how successful these

studles'will be; however, simple,
free-energy and equilibrium-constant
‘ ~data have confirmed some of the ex-
‘J{perlmental results, for example,
" chromium, iron, andnlckel are attacked
by uranium tetrafluoride. Chromium

was the most severely attacked and
nickel the least. Measurements of the

potential differences between cells of

various fluorides are being undertaken.
Electrode potentials i1n sodium hy-

“droxide will be measured as a means
of ascertalnlng the purity or thermal
history of nickel, nickel oxide, and

sodium hydroxide. Polarographic
studies of the sodium hydroxide-nickel
oxide system will be undertaken if
preliminary tests are promising.

129

 
 

 

 

 

 

 

 

 

 

 

 

 

130

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B IR O PP UNCLASSIFIED "

 

LI

wi

%,
€.

e
i

Fig. 39. Nickel Thermal Convection Loop
Operated for 296 hrwith Sodium Hydroxide Under
a Hydrogen Atmosphere.

 
 

 

~FOR PERIOD ENBING MARCH 10, 1952

UNCL ASSIFIED
Y-5417

 

 

 

 

 

 

kS
i
-
P
. . [
i B
[N
= L
B
e
D
. [
—
3 P
r .
B B
& e

->

 

I

 

 

Fig. 40. Inconel Thermal Convection Loap Operated for 135 hr
with Potassium Hydroxide Under a Hydrogen Atmosphere.

o b

s

 

 

131

 

 

 
 

 

 

ANP PROJECT QUARTERLY PROGRESS REPORT

 

Fig. 41.

with Molten Sodium Hydroxide in Seesaw Apparatus

| UNCLASSIFIED |
Y- 5936 :

rSéctional View of L-Nickel Specimen’After 117 hf.(14,000 cycles)

Note polishing at hot zone

and abundant crystal formatlon at cold zone.

Possible Efifiilibria'Among Fluorides

and Metals (L. E. Topol and L. G.
Overholser, Materials Chemistry
Division). The values of the free
energy and equilibrium constant of
the following reactlon were computed
at 1000 and 1500°K from known thermo-
dynamic data (5)

M+ xAF = MF, + x4,

where M is iron, chromium, or nickel,
and A 1s potassium or sodium.

The activities were calculated
assumingM and AF tolnaof unit activity
and

- = 1/(x+i)

 

(5) L. Qu1ll (ed.), The Chemistry and
Metallnrgy of Hlsc‘ellaneous Materials: Thermo-=
dynamics, McGraw-Hill, New York, 1930.

132

Table 33 lists the values of AF and
K and also approximate estimates of
the activity of the metallic fluoride.

These results indicate: sodium
fluoride is less corrosive than
potassium fluoride and the reaction
increases with temperature; chromium
is the most soluble of the metals
considered, iron is second, and nickel
last; and the trivalent cations are
slightly less reactive than the
divalent (however, from free energy
considerations the trivalent 1ons are
the more stable at the temperatures
considered).

Converting the values of a to parts
per million, a concentration of 10-1
ppm is found for Fe*? at 1000°K. This
figure is much less than the ex-
perimentally determined value (similar
results are found for the others), ‘and
if the activity of AF 1s corrected
(since mixtures of alkali fluorides
 

are used), the metallic fluoride

(MF) concentrations are further de-

creased. Thus;it'seems impossible to
explain the cofroSibn of métals by
alkali fluorldes at hlgh temperatures
by postulatlng the above-mentioned

'equlllbrlum‘reactlon to be the chief
effect.

FOR PERIOD ENDING MARCH 10, 1952

where M is iron, chromium, or nickel,
and the corresponding equilibrium
constant is '
B (UF3)* (MF,)
(M) (UF,)*

in which the activities of the molecular

species are denoted by parentheses.

TABLE 33

Free Energles and Equxllbrlum Constants for Reactlons of_
: Metals with Alka11 Fluorldes ' ‘

 

 

 

 

 

M AF T (°K) AF (cal) K | e
Fe*? KF - 1000 +84, 800 10-18.6 10- 6
1500 80, 000 10- 12 10-4
Fet? KF 1000 143,700 10-31.4 10-7-5
1500 138,200 10202 10°°
Crt? KF 1000 69, 800 10°15.3 10°%
1500 62,000 10°9-1 10-3
Cr*s KF 1000 111,700 10244 10°%
1500 101, 000 10-14-7 10°3%-7
Nit?2 KF 1000 93,800
1500 90,500
Fe*? NaF 1000 91,000
1500 86,500

 

 

 

- o _

The available free-energy data(®)
and several assumptions were used to
evaluate the reaction of iron, nickel,
and chromium with uranium tetrafluoride
at 1000°K and the act1v1ty of the
metalllc fluorlde

The reaction involved may be
written

- M+ xUF, = £UF; + MF,,

The activity of ¥ = 1, and assuming

Cypg = yp, = G then

x x+1

from which

K 1/(xz+1)
a = [___] . (UF )x/(x+1).
x* 4

133

 
 

 

S, Sl R L

 

 

If uranium tetrafluoride is the sole
reactant present its activity 1is also
1, and yields

1/(x+1y
K )
a = -

1 xx

If the uranium tetrafluoride is one of
several components of a system, its
activity will be approximated by its
mole fraction, and therefore the
metallic fluoride activity will be
decreased.

Table 34 gives the values found
for free energy, AF, equilibrium
constant, K, and metallic fluoride

act}V}tlgs, a, (if Qyp, = 1), a, glf
equilibrium concentration of UF, = in-
- itial concentration = 1 mole %), and
a, (1f egquilibrium concentration of

UF, = 0.1 mole %).

From these results it is seen that
al though the trivalent forms of iron

‘and chromium are more stable, the

divalent i1ons will be preferentially

ANP PROJECT QUARTERLY PROGRESS REPORT

will readily be reduced by the pure
metal. Of the metals considered
chromium i1s the most readily attacked,
iron is next, and nickel last. 1In
addition, a comparison of these results
with those found for the alkali
fluorides indicates that uranium
tetrafluoride is the much better
oxidizing agent of the two. (It
should be stated that the accuracy of
the data used 1s not certain and even
qualitative conclusions may be in
error. )

There 1s some experimental evidence
that beryllium oxide will react with
uranium tetrafluoride at high tempera-
tures to form uranium dioxide and
beryllium fluoride, according to the
following formulas:

2Be0 + UF, = U0, + 2BeF, ,

(U0,) (BeF,)?
K = ~——r————
(Be0)? (UF,)

forméd. In fact, it can easily be

shown thermodynamically that any (the activity of a molecular species

trivalent iron or chromium formed 1is denoted by parentheses). From
TABLE 34

Free Energies and Equilibrium Constants for Reduction
| of Uranium Tetrafluoride with Metals

 

 

 

 

 

 

 

 

 

M AF (kcal) X a, a, ay
Fé+2 +20.0 10-4-4 10-1-6 10-2-°9 10-3-6
Fe'? ‘ 45.0 1093 1027 10-4 2 10" S
Cr+2 4.0 10-0-88 10-9.5 10-1-8 10-2-5
Cr*3 13.0 10-2-85 10-1 10-2-5 10-3-2
Ni 928.0 10-6- 1 10-2-2 10-3:5 10°4- 2

el f
134 Gomid it
 

available data,(3'%'7) the free energy
and equilibrium constant of this
reaction were calculated to be t+4.,2
kcal and 10°%+?, respectively, at
1000°K.

If the beryllium oxide and uranium
tetrafluoride are assumed to be of
unit acF1Y1ty, and %50, = %Ber, a,
the equilibrium constant is given by

4a° .

This results in a value of 10°°+* = 0.4
for the approximate activity of uranium
dioxide and 0.8 for beryllium fluoride
at equilibrium.

, Although the accuracy of the data
is unknown, there seems to be sufficient
evidence for the assumption that the
above reaction occurs to a reasonable
extent. '

EMF Measurements in Fused Fluorides
(L. E. Topol, L. G. Overholser,
Materials Chemistry Division). A
series of experiments are planned for
studying the possible mechanisms of
fluoride corrosion by measuring the
potential differences of various cells
containing the fluorides of nickel,

chromium,or iron dissolved in fused
alkali fluorides.

Concentration cellsof ;hé following
type will be investigated:

M/MF+(C) in fused AF/MF-(C) in fused AF/M

where M is iron, chromium, or nickel,
and A 1s sodium, potassium, or lithium.

 

(G)L. Brewer, L. A. Bromley, P. W. Gilles, and
N. L. Lofgren, The Thermodynamic Properties and
Equilibria at High Temperatures of Uraenium Helides,
Oxides, Nitrides, and Carbides, MDDC-1543, Sept.
20, 1945. :

(7)0. Kubaschewski and E. L1, Evans, MHetal-

lurgical Thermocheristry, Academic Press, New
" York, 1951.

- FOR PERIOD ENDING MARCH 190,

1952

Binary and tertiary mixtures of
the alkali fluorides will be studied,
and the effect of small additions of
uranium fluoride will be evaluated,
if possible.

Electrode Potentials in Fused
Sodium Hydroxide (Ambrose R. Nichols,
Jr., Materials Chemistry Division).
The apparatus previously described(®’
has undergone continued modification
over the past three months, It
consists of a nickel vessel within
which 1s placed a porous cup, which
divides the contents into two electrode
compartments. A nickel electrode is
suspended in each compartment., Chromel -
Alumel thermocouples in nickel walls
are located against the outer wall of
the nickel cup and in each of the
compartments. The whole assembly
si1ts at the bottom of a closed stain-
less steel container that rests in a
5-in. pot furnace. Helium purified
by passing over copper turnings at
450°C and through magnesium perchlo -
rate tubes 1s passed through the
container at a pressure slightly in
excess of atmospheric.

When a measurement 1s to be made,
the desired amounts of purified sodium
hydroxide and nickel oxide are weighed
in a dry box and placed in the two
electrode compartments of a nickel
container. The closed container is
then removed to the furnace and the
necessary connections made. Helium
circulation is started before heating
begins. After the chosen temperature
has been reached, the potential between
the two electrodes is determined by
using a type-K potentiometer, although
in some casesa recording potentiometer
has also been used.

 

(S)A. R. Nichols, ‘*“EMF Measurements in Hy-
droxides,” Aircraft Nuclear Propulsion Project
Quarterly Progress Report for Period Ending
December 10, 19571, ORNL-1170, p. 110.

e g

’ ks

Pil 2
¥,

Y

135

 
 

 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

The results have thus far not been
reproducible. The measured potentials
have changed with time, even to the
extent of reversal of polarity. It
was observed in each case that the
electrode in the more concentrated
solution developed a deposit of fine
nickel crystals, whereas that in the
more dilute solution appeared to have
undergone polishing. This mass
transfer has so far appeared to be
independent of temperature over the
narrow temperature range used.

Three principal experimental
difficulties exist and each could
account for the poor results: (1) fail-
ure to find a suitable diaphragm
material; (2) lack of knowledge of
the true nickel oxide concentration
in the solutions in whichthe electrodes
are placed; and (3) difficulty of
maintaining a uniform temperature

‘throughout the cell. Until these

problems are overcome, it 1is not
possible to ascertain the effect of
such factors as the purity or the
thermal history of the nickel, nickel

‘oxide, or sodium hydroxide.

Pélarbgraphy of Sodium Hydroxide
in Silver and Platinum (R. A. Bolomey,

Materials Chemistry Division). In
‘previous reports¢{®) it was stated

that anhydrous sodium hydroxide heated
to a temperature of 350 to 600°C in
either a silver or platinum crucible
in vacuum resulted in the production
of ionic species that gave character-
istic polarographic waves. It was
also mentioned that the voltage at
which the peaks occurred was tempera-
ture-dependent. To date, 85 polaro-
grams on sodium hydroxide in silver
crucibles and 88 polarograms on
sodium hydroxide in platinum crucibles
have been obtained. Even though all
the data were not obtained at the
same temperature, enough information
has been obtained to analyze the
reproducibility of the results. Such
an analysis revealed considerable

136

randomness in the waves and that the
curves of a given series show only
qualitative similarity.

The source of randomness in the
position of the polarographic waves
is not understood. It is possible
that it is to be found in the instru-
ment itself as a result of the method
employed to collect the data. At-
tempts to obtain the data by allowing
the readings to come to equilibrium
after each setting proved to be
impractical because of the slow rate
of equilibrium attainment. The present
me thod of collecting the data requires
that the potential on the cell be
varied at a uniform rate, since any
fluctuation in the rate of potential
changes can produce anomalous effects
on a current-voltage curve of the
type obtained with a polarograph and
stationary microelectrodes. It may
be that these anomalous effects are
more apparent when operating the cell
at the high temperatures required in
this work than when operating at room
temperatures in aqueous solutions.

Equipment for differential thermal
analysis has been gathered so that
preliminary tests may be made to study
the feasibility and applicability of
this method to the study of corrosion
mechanisms. It 1s intended first to
apply the method to the system sodium
hydroxide-nickel oxide under an inert
atmosphere.

Magnetic Susceptibility of Stain-
less Steel Exposed to Fluorides (W. C,
Tunnell, ANP Division). It has been
observed repeatedly that the surfaces
of all normally nonmagnetic steels
become magnetic after exposure to
high-temperature fluorides. Surface
layers have been removed and analyzed,
and it appears that the iron-to-
chromium ratio increased. A micro-
spectrographic analysis, under electro-
magnetization using a colloidal
dispersion of iron particles, revealed

¥

!
Y

o

!

Ty
BV, ?‘

0w

P
eidgacs

[
 

the magnetic material to be at the
grain boundaries where intergranular
penetration had occurred.

In one test using a gasket ring of
type-316 stainless steel, the magnetic
surface was submerged in concentrated
nitric acid, and a 1- to 2-mil layer
of stronglymagnetic material separated
that was chemically analyzed as >90%
iron, 2.5% chromium, and 1.5% nickel.
The material under the layer was not
magnetic. In another test the hot

section of an Inconel loop that had

been exposed to Fulinak was examined,
since 1t seemed more magnetic than the
cold section. Immersion in con-
centrated nitric acid left pits and
made the material less magnetic, and
the remaining surface layer appeared
to be pure nickel. Another specimen

FOR PERIOD ENDING MARCH 10, 1952

of type-316 stainless steel that had
been in contact with Flinak became
porous after treatment with concentrated
nitric acid and lost all its magnetic
properties.

It may be concluded that, as 1in
most intergranular corrosion of stain-
less steels, the chromium 1s prefer-
entially leached at the grain bound-
aries, possibly by the usual device of
carbide precipitation, which exposes
excess 1iron and nickel for attack and
gives the magnetic susceptibility,
Photomicrographs appear to bear out a
phenomenon of this type. The attack
on iron always appears greater than on
nickel, as has been shown, for example,
by tests where the rate of penetration
appears to be proportional to the
amount of iron in the metal,

137

 
 

T T——

TR T ON TSR WY IV TR0 1 et

 

 
i, o S e 1 £ A Ak

e g

A e

o i i, S SR R

Nk BN G e i
)

.

 

S o M R B i Bt s S, Bl B e e SRR i iom o i . 2

 

ik R g SR

i

12. METALLURGY AND CERAMICS

W. D. Manly, Metallurgy Division
T. N. McVay, Consultant

The drawing of tubing from the
previously prepared solid fuel plates
has been accomplished, and the results
of the drawing on the bond continuity
and oxide distribution of the solid
fuel elements are discussed. From
this work it was found that the best
core material to use in the uranium
oxide mixture was iron powder. Com-
patibility experiments have shown
boron carbide to react rather ex-
tensively with the 300-series stain-
less steels and Inconel but to be
inert to type-430 stainless steel;
therefore the ARE safety rods will
probably be constructed of type-430
stainless steel. The solid-phase
bonding of metals is always difficult
with materials that are in intimate
contact at elevated temperatures.
Consequently, self-welding experiments
are being conducted on the possible
¢combinations of materials to be used
in the safety rods. '

The use of the cone-arc welding
process for the production of tube-
to-header joints is being investigated
to determine the applications and
limitations of the process, since
there is a need for a reliable auto-
matic or semiautomatic welding method
for the production of the many tube-
to-header joints in ANP core and heat
exchanger designs. A description of
the equipment and operating procedure
and a discussion of the variables
being studied are presented. Because
of the complexity of the ANP type of
heat exchangers, it i1s probable that
extensive brazing will be necessary
in fabrication. Preliminary work for
heat exchanger assemblies indicates
that brazing can produce sound joints
if proper control of the brazing
variables 1s exercised. Nicrobrazing

alloys, Pd-Ni, Mn-Ni, Ag-Pd, Ni-Cr-Sr,

I h
i‘;e‘ '.Mé' §~£:’

i

and Ni-Cr-Si-Mn, are being studied

and their corrosion in hydroxides and
fluorides is being investigated.
Since Inconel has been designated as
the structural material for the ARE
extension, rupture and creep data are
being obtained at the operating tem-
perature of the reactor. Data are
presented to show (1) the time re-
quired to produce deformations of
various percentages and for rupture to
occur as a function of the applied
stress for fine- and coarse-grain
Inconel, and (2) the minimum creep
rate and per cent deformation per
hour for the two types of Inconel

plotted as a function of applied

stress.

The oxide ceramics or combinations
of ceramics and oxides appear most
promising for reactor application be-
cause they offer the best combination
of structural integrity and thermal
properties. Ceramic coatings for
Inconel and stainless steel are being
tested. Of those tested, the NBS
Ceramic-Coating A-418, with its high
temperature and oxidation resistance
characteristics, shows promise as a
coating for the ARE radiator and pos-
sibly other structural parts.

FABRICATION OF REACTOR ELEMENTS

E. S. Bomar J. S. Coobs
Metallurgy Division

Tubular, solid, fuel elements in

which the fuel compact is between
two concentric tubes are being de-
ve loped but have not yet proved en-
tirely satisfactory. The bonding and
distribution of the uranium dioxide
powder between the tubes after cold

drawing has varied from very poor to’

P
LR vy
e Ty

139
 

ANP DIVISION QUARTERLY PROGRESS REPORT

fair. It is hoped that better initial
bonding and a different cold-drawing
technique may yet yield good, small,
tubular, solid, fuel elements. '

Tests for solid-phase bonding
(self-welding) of the movable parts
of the ARE control rods at reactor
temperatures (1472°F) indicate a fair
amount of reaction between the boron
carbide type-316 stainless steel and
boron carbide Inconel systems in 100
hours. Of the few materials tested
the reaction between boron carbide and
type-430 stainless was the least
marked

‘Cold Drawxng of Tubular SOlld Fuel
Elements. Tubular, solid, fuel ele-
ments fabricated by joining two seml-
circular laminated plates either with
seam welds or by "rubberstatic"(1)
pressing methods have been subjected
to cold-drawing operations at the
Superior Tube Company. All samples,
except the one prepared by rubberstatic
pressing, were given an initial hot
reduction of 62 to 75% to bond-pressed
core, picture frame, and cladding.
This operation was carried out on the
flat stock before forming into tubes.
The cores of the samples were 30 vol %
uranium dioxide and 70 vol % metal
(type-302 stainless steel, iron, or

nickel).

A reductlon schedule based on com- _

mercial practice was set up at values
ranging from 18 to 28% after a test
reduction of 38% caused a tube failure.
"The weld seams were satlsfactorlly
- smoothed in this pass, and the tubes
were subsequently reduced 87.5% in
seven passes, which yielded tubes of
1/4-in. OD by 0.015-in. wall thickness.
Three of these tubes were then further
reduced to yleld tubes 1/8 in.

 

 

Bemar,
“Fuel-Element Fabrication,” Adircraft
Nuclear Propulsion Project Quarterly Progress
Report for Period Ending September 10, 1951,
~ ORNL-1154, p. 149.

(I)G. M. Adamson, E. S.

Coobs,

140

W 4 e T e

OD by*"‘ :

and J. H.

0.015-in, wall thickness. Metallo-
graphic examination of specimens taken
from the tubes showed that bonding and
uranium dioxide distribution varied
from very poor to fair for tubes pre-
pared from rolled plate and was also
fair for one specimen prepared by
rubberstatic pressing. Sections trans-
verse to the direction of drawing are
shown in Fig. 42 at reductions of
87.5%, that is, for 1/4-in.-0D tubes.
In general, further reductions to
1/8-in.-0D tubing only exaggerated the
defects found in the 1/4-in.-0D stock.

Although plug drawing was not tried
on any of the composite tubes, 1t was
the opinion of some of the technical
staff at Superior Tube Company that
plug drawing would not disturb the
core of a laminated structure so much
as drawing on a mandrel, which was
the technique used in all drawings to
date. Rod drawing produces slip
cracks in the core because of slightly
uneven reduction of the two cladding
layers, but it i1s believed that plug
drawing should give uniform reduction.
Variation in the reduction schedules
used for these tubes will be tried by
using the Metallurgy Division draw-
bench and emphasizing less severe
reductions. Mandrels are now on order
to supplement dies already on hand.

. Equipment for plug drawing 1is not

avallable at present

ARE Control Rod. At the operating
temperature of the ARE, there exists
the possibility of solid-phase bonding
of metallic components that are in
close contact. This phenomenon could
possibly lead to malfunctioning of the
control rods if it occurred between
the control rod cladding and the walls
of the tubes in which they are housed.

 

Several tests have been made in
which samples of two different metals
were pressed together at 1472°F. One-
eighth-inch-diameter specimens of the

oo s f, e
> i, { A

E:\ k ; fw.r f‘

oo F

ka;

 
 

 

FOR PERIOD ENDING MARCH 10, 1952

   

Y5299 1

  
  

SO T R & TYPE-3(6 STANLESS
: o o | STEEL CLADDING
. XY, . S

 

CORE
BY VOLUME
30% UO2 (-100 MESH
AS RECEIVED
70% Ni(-325 MESH)

PR | " | o | ' ;-*TYPE~316 STAINLESS
e | S / STEEL CLADDING

CORE
BY VOLUME
30% U0, (~200,
+325 MESH)
70% TYPE-302
STAINLESS STEEL
(-325 MESH)

 

 

Fig. 42. Transverse Sections Through %-in.-OD,Cold-Drawn, Tubular, Solid
Fuel Elements. (a) Tube prepared from rolled plate. 200X. (b) Tube prepared
by rubberstatic pressing, 200X, (Prints reproduced 83% of original size. )

141

 
i,

™

T

e g ke

A e

 

|

¥

systems.

dissimilar materials were subject to
250 psi for 100-hr intervals. This
pressure is probably higher than would
be encountered radially at the control
rod and control-rod-tube contact
points. Type-316 stainless steel and
Inconel, in various atmospheres, were
used in the initial tests. However,

results of compatibilify'tests made

available after a portion of the
welding tests had been run indicated
a fair amount of reaction between the
boron carbide-type-316 stainless
steél and boron carbide-Inconel
systems. A less marked interaction
was found when boron carbide was

"contained in type-430 stainless steel.

The solid-phase welding schedule was
revised to include the type-430 vs.
type-316 stainless steel and Inconel
~ However, the only bonding
evidenced that could not be removed

‘with finger pressure was between

similar metals. Tests of the type-430
vs. type-316 stainless steel system
with tank helium atmosphere are now
being conducted. Exposure to the

‘other atmospheres will be carried out

if results indicate that this 1is
desirable. |

CONE-ARC WELDING

P. Patriarcar G. M. Slaughter
Metallurgy Division

The feasibility of the use of the

cone-arc welding process for production
of tube-to-header joints has been
demonstrated by previous 1nvestiga-
tions.(%?’ The welding group of the

‘Metallurgy Division is investigating

the applications and limitations of
the process, since the need for a
reliable automatic or semlautomatic
process for production of the many
tube-to-header joints required by

 

(2)E. R. Mann, Means for Making Unifornm
Circular Heliarc Welds by Deflecting the Ion Beanm
Continuously, ANP-63, Apr. 9, 1951.

142

ANP DIVISION QUARTERLY PROGRESS REPORT

ANP core and heat exchanger designs
is evident.

Equipment. The apparatus consists
of a G-E, 200-amp, direct-current
welding generator with the addition of
a resistor circuit that permits the
use of currents as low as 3 amp, 1if
desired. A Miller Electric Company,
high-frequency, spark-gap oscillator
provides the high frequency required
for starting the inert welding arc
without having to touch the 1% thoriated
tungsten electrode to the work. The
apparatus has been provided with a
timer and contactor control circuit
to render the arc timing an automatic
operation.

Principle of Operation. The cone-
arc welding technique is particularly
suited to welding a tube to a header
plate. The nozzle of the welder 1is
a permanent Alnico magnet that sets
up lines of flux to the tube and plate
over which the nozzle 1s held. Since
the welding arc will strike from the
electrode tip to a point on the
periphery of the tube, the arc stream
is cut by the lines of flux at a small
angle. This angle sets up resultant
forces tangential to the tube periphery
in the plane of the header sheet,
and the forces rotate the arc around
the tube at very high speeds. To the
observer the arc appears as a cone;
hence, the name "cone arc." The arc
raises the temperature of the tube
edge and the periphery of complete
header hole in a uniform manner until
the melting point its reached and a
circumferential weld results.

Experimental Procedure. Some of
the variables that may affect the
operation are:

1. Arc current and time,

2. Arc distance, which affects the
arc voltage and the angle between
the arc stream and the magnetic
lines of flux,

1y £,

P i‘b ‘3 Ll \“"3
T L A
g’:}; ot E
 

~arc was struck.

3. Magnetic nozzle to work distance,

4. Choice of inert gas and rate of
flow, which affect the arc voltage
(higher for helium than argon) and
the turbulence,

5. Work geometry, that is, tubing
' size, header thickness, and heat
transfer uniformity of header hole-
‘to-hole distance, hole-to-header-
edge distance, and centering of
work and electrode.

In the initial experiments header
holes without tubing were used, since
it was assumed that the behavior of
the arc in the formation of a circum-
ferential molten pool would yield
representative information with or
without a tube. The header material

‘used was type-304 stainless steel sheet

1/8 in. thick. Holes were 3/16 in. in
diameter and spaced 3/8 in. center to
centéer. Welding conditions were as

" follows:
1. 1/16-in. diameter electrode,

2. Electrode to plane of work dis-

tance, 0.070 in.,

3.. Magnetic nozzle (tip of soft iron)
to plane of work distance, 0.41
il’l,, ’

a4."Argon flow, 30 ft /hr,
%5,  Open 01rcu1t voltage, 70

- 6}H'Arc'voltage,_11,

7. Curfént,'#ariéblé;'
8. Tlme, varlable
. 4‘3 w" '

Dur1ng weldlng,pools of molten

'metal formed around the periphery’
of the hole about 20 sec after the

The pools grew until
they encompassed the entire periphery
of the hole and effected the desired

~decreasing arc txme,}

FOR PERIOD ENDING MARCH 10, 1952

weld. © As would be expected, 'the size
of the heat-affected zone, as de-
termined by the size of the heat-tinted
zone surrounding the header hole,
diminished with increasing current and
The limiting
value of permissible 'hole-to-hole
distance will probably be a function
of the arc current and time for a
given header thickness and hole size.
Experiments will be conducted to de-
termine these limiting values, since
ANP designs may require as small a
hole-to-hole distance as possible in
order to obtain the maximum number of
heat exchanger tubes per unit area of
header sheet.

The results from the few experiments
made were inconclusive, but it 1s ex-
pected that further work during the
coming quarter may yleld sufficient
information for a comprehensive study
of the quality of cone-arc welded,
tube-to-header joints as a function of
the welding variables.

BRAZING

P. Partriarca G. M. Slaughter
Metallurgy Division

Experiments were made to determine
the feasibility of brazing since 1t
is quite probable that this method may
be used extensively in the fabrication
of the ANP type of heat exchangers.

.‘Tube'bompacts with tubes, 0.100 in. OD

with 0.010-in. wall thickness, have

“‘been successfully brazed to baffle
- plates that were 0.020 in, in thick-

ness. Preliminary work on such as-

... semblies indicates that brazing can
" "be used advantageously in the produc-
--.~tion of sound joints if proper control

is maintained of such brazing variables

~as joint fit, joint geometry, degree

of base-metal cleanliness, and qualaty
of the furnace atmosphere during the
brazing process. Corrosion tests of

Nicrobraz and a 60% palladium-40%

) | 143

 

 
Sk, A 2

RGR Ka  dinl
v 2

 

gl s i,

 

 

‘@yend of a 6-1n. —long butt joint.
;legs
+ Inconel strips,

S
oo e, L0

'de31rable

ANP DIVISION QUARTERLY PROGRESS REPORT

nickel alloy in sodium hydroxide and
in UF,-NaF-KF-LiF mixture showed the
palladlum -nickel alloy to be far more
corrosion resistant in both fluids at
1472°F, but the attack by the fluorlde

on elther braze was severe.

Flow Iests. A relat1vely long-term
program for evaluating various brazing
alloys for the high-temperature ANP
type of application is anticipated
because of the promising results ob-
tained in the experiments on the feasi-
blllty of bra21ng A series of tests
were conducted to determine the mini-
mum temperature at which the various
bra21ng alloys flow readily. The test
spe01men consisted of a moderate
amount of brazlng alloy placed at one
Both
of the joint were 0.062-in.
one of which was
milled flat on the mating surface to
facilitate the production of a tight-
fitting 301nt The specimens were
then heated untll the temperature was

‘i‘found atwhlch the braz1ng alloy flowed
“freely up the entire joint length of
.6 inches.
“fi}these experlments that any oxide scale

It became apparent during

ofi the metal specimens is highly un-
- The brazing alloy does not
ea51ly wet scaled surfaces and flow
around a 301nt is 1mpeded A very
dry hydrogen atmosphere, ‘with a dew
point in’ the range of -60°F or below,
prevented scallng in most cases;

therefore moisture was ‘apparently the

‘prominent factor in scale formation.

It was suspected that if the hydrogen
continuously reduced any oxide and
there were subsequently moisture forma-

"tion in the hydrogen atmosphere, the

exit dew point of the hydrogen would
decrease with increasing flow rate.
A series of experiments proved this
hypothesis. Tests with an inlet-gas
dew point of -78°F and high hydrogen

" flow rates showed that exit dew points
‘approximating those at the inlet were

attained. The experiments were sig-
nificant in that they made it obvious

o

with a higher magnification,

that a relatively large hydrogen flow
rate during furnace brazing was de-
sirable. ' “

Corrosion of Brazing Alloys. The
resistance to corrosion of the various
brazing alloys, alone and in combina-
tion with various base metals such as
found in joints, is important in the
selection of suitable joining media;
therefore emphasis has been placed on
the results of corrosion tests of the
brazing alloys in fluorides and hy-
droxides. Extensive tests wereé con-
ducted on Nicrobraz, a boron-containing
alloy, which is an excellent brazing
alloy for use in conventional high-
temperature applications. Samples of
the pure alloy were subjected to
corrosion experiments in fluorides and
hydroxides and the baths were later
analyzed chemically for boron.

A sample of Nicrobraz treated in
sodium hydroxide for 100 hr at 1500°F,
showed the heavy corrosive attack
characteristic of this medium. There
was generally heavy surface attack to
a depth of 60 mils and severe local
attack. The sample tested in the
NaF-KF-LiF-UF, mixture for 100 hr at
1500°F was moderately pitted, par-
ticularly near the intermetallic
components, to a depth of 4 mils.
Other surface areas were relatively
untouched. Results of the chemical
analyses of the contents of both the
fluoride and hydroxide baths 1indicate
that boron is preferentially leached
from the alloy. An attempt will be
made to determine the actual percentage
loss of the boron from the alloy.

.The corrosion resistance of Nicro-
brazed tube-to-header joints with

Inconel as the base-metal has been

studied. A section of a tube-to-
header joint exposed to sodium hy-
droxide exhibited excessive attack
both at the brazed joint and on the
base-metal, as shown in Fig. 43e¢ and,
in Fig.
43b. The relatively mild attack of

o TR
D :
% v +

144 DU A
 

1952

'FOR PERIOD ENDING MARCH 10,

 

Corroéion Tést of Nicrobrazed Inconel Tube-to-Header Speéimens

43.

Fig.
Exposed to NaOH for 100 hr at 1500°F.

(b) 250X.

(a) 25X,

145

e

[
bi 3

gy

e

 
 
 
 

e i, . R itk e

st e B

 

 

 

 

i s ol Bt et e i il i R N

o, i i B B i, ALl . B M B 300 50 b e ot B

oo IR s BC i i Bl B B it w

TARY R TR TN e S e
*

ANP DIVISION QUARTEBLY PROGRESS REPORT

the mixture NaF-KF-LiF-UF, on these
joints is shown in F1g 44, which 1s

a photomlcrograph of a joint tested in’

the fluoride mlxture. Chemical analy-
ses of the baths will be given in
later reports.

A few experlments have also been
conducted on the 60% palladium-40%
nickel and 60% manganese-40% nickel
alloy systems. Small ingots of the
palladium-nickel alloy were tested for
100 hr at 1500°F in both the sodium
hydroxide and the NaF-KF-LiF-UF, mix-
ture and gave promising results. The
corrosion resistance of this alloy
was excellent the attack was less
"than 2 mils in both cases. An Inconel
'tube to- header joint brazed with the
60% manganese 40% n1ckel alloy was

Fig. 44. Corrosion Test of Nicrobrazed Inconel Tube-to-Header Joint Ex-

 

tested for 100 hr at 1500°F in sodium
hyd}oxf&é; and there was moderate
attack ‘at the joint (Fig. 45). More
extens1ve'corr051on tests on,these
alloys will be performed, and the
baths will be analyzed for the presence
of the various elements in the brazlng
alloy being 1nVest1gated

~Similar tests will be conductéed on
other high-temperature brazing alloys
of current interest, including 75%
silver-20% palladium-5% manganese,
64% silver-33% palladium-3% manganese,
60% palladium-37% nickel-3% silicon,
16.5% chromium-73.5% nickel-10%
silicon, 16.5% chromium-71.5% nickel-
10% silicon-2.5% manganese. No data
are yet available on the resistance to
corrosion of»these ‘alloys. Methods

 

posed to NaF-KF-LiF-UF, for 100 hr ‘at 1500°F. 250X.

146
 

 

 

FOR PERIOD ENDING MARCH 10, 1952

LR R

Fig. 45. Corrosion of Inconel Tube-to-Header Joint Brazed with a 60%
Manganese~-40% Nickel Alloy after 100 hr at 1500°F in NaOH. 100X,

have been devised for coating the
inside of the standard nickel test
capsules with brazing alloy so that
corrosion resulting from dissimilar
metal combinations can be eliminated
when testing samples of the pure
brazing alloy. When brazed joints
are tested for corrosion, a capsule
of the same metal as that of the joint
base-metal is used to eliminate dis-
similar metal combinations.

MECHANICAL TESTING OF MATERIALS

R. B. Oliver  J. W. Woods
: - C. W, Weaver

Metallurgy Division

The structural integrity of Inconel
at the operating temperature of the

ol

reactor is being accurately determined.
Elongation, rupture, and creep rates
vs. stress curves have been obtained
for fine- and coarse-grained Inconel
specimens tested in argon at 1500°F,
Three tube-burst tests with stresses
up to 4000 psi have been running con-
tinuously for 1300 hr without develop-
ing detectable leaks.

Inconel Creep and Stress Data.
Elongation, rupture, and creep rate
Vvs. stress curves were obtained with
fine- and coarse-grained Inconel sheet
specimens 0.065 in. thick, 0.500 in.
wide, and 3 in. long. "The extension
was measured during the test by ob-
serving the relative displacement of
two platinum strips fastened at opposite
ends of the test section. A Gaertner

"
tcovas

1 147

o1
£

 

 
 

 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

micrometer microscope having a least
division of 50 pin. was used to evalu-
ate this relative motion.

The time to produce deformations of
0.1, 0.5, 1, 2, 5, and 10% and the
time for rupture to occur, as a func-
tion of the applied stress for fine-
grained Inconel, are shown in Fig. 46.
Also presented is the minimum creep
rate in percentage of deformation per
hour as a function of the applied
stress. The percentages appearing
above the rupture curve are the total
elongations at the several stresses.
The fine-grained material was cold
rolled and bright annealed at 1650°F
to produce a grain size of approxi-
mately 0.105 mm (90 grains/mm?).
Similar data for the coarse-grained
Inconel-sheet specimens are presented
in Fig. 47, but the curve representing
the times to produce deformations of
10% is omitted since it is nearly
coincident with the rupture curve.
This Inconel sheet was annealed for
2 hr at 2050°F in a hydrogen atmos-
phere to produce a grain size of ap-

proximately 0.250 mm (15 grains/mm?).

Two separate creep rates were ob-
tained for the fine-grained specimens
tested at 3000 and 1850 psi; these
rates were observed during three
different tests. Linear sections
were observed on the strain-time
plots at 50 to 150 hr and again at
‘about 300 and 600 hours. The earlier
of the two linear periods exhibited
creep rates of one-third to one-half
of the rate found during the later
linear period.

Tube-Burst Tests. Three tube-burst
tests with tangential stresses of
about 1000, 3000, and 4000 psi, re-
spectively, have been running for ap-
proximately 1300 hr without developing
detectable leaks in the type-316 stain-
less steel tubes. These tubes, with
480 mils OD and 10-mil wall thickness,

148

were loaded internally'with‘argon
under pressure and exposed to an

‘environment of stagnant air. Since

no method exists to gage tubular
specimens during the test, the re-
sults will be based on before and
after measurements, ‘ - T

A second phase of this work is to
obtain data on physical properties of
metals in the fluorides. Apparatus
used for a similar study of sodium has
been modified and calibrated so that
fluorides may be used. Two tube-burst
tests were run in NaF-KF-LiF-UF, (10.9
mole % NaF, 43.5 mole % KF, 44.5 mole
% LiF, 1.1 mole % UF,) by using the
apparatus described in the previous ANP
quarterly report. Type-316 stainless
steel has withstood a hoop stress of
2320 ps1i and the Inconel 2170 psi. Both
these tests have been 1n operation
over 400 hours. Stress-to-rupture
tests in the same environment and
temperature but of 1000-hr duration
were also run on Inconel and type-316
stainless steel tubes. These materials
withstood hoop stresses of 1200 psi
and 1700 psi, respectively, and the

~diameters increased 1.7 and 0.7%,

respectively.

Operation of Creep and Stress-
Rupture Equipment. The center of
gravity of the lever-arm counter weight
on each testing machine was raised
approximately 2 in. above the plane of
the knife edges prior to making a
series of creep and stress-rupture
tests. With the counter weights thus
raised, the tare weight of the lever
arm was 1ncreased by the same amount
that the spring load of the compressed
sealing bellows was decreased, so that
a constant load could be maintained
during test. The 1initial, or tare,
load of each testing machine was
measured with wire strain gages
mounted on a duplicate specimen and
the constancy of the load was also
verified. '

i

 

 

 
6¥1

STRESS (psi)

0.00014
30000

20000

10000
9000
8000
7000
§000
5000

4000

3000

2000

1000
o4

I3

Fig.

 

CREEP RATE (% per hr)

0.004 Q.01 0.1 10

46.

INCONEL SHEET, GOLD-ROLLED
AND ‘ANNEALED AT 1650°F
GRAIN SIZE, APPROXIMATELY
90grains /mm2, 0105 mm DA
TESTED IN ARGON AT 4500°F

TIME 1S SHOWN FOR 04,05, 1,2,5, AND 10%
EXTENSION AND RUPTURE VS. STRESS;
EXTENSION MEASURED OPTIGALLY

10 TIME (hr) 100 1000

Creep and Stress-Rupture Data for Fine-Grained Inconel Sheet.

 

wor

UNCLASSIFIED
DWG. 14445

10

10000

‘0T HOMVIW ONIONT @0I¥dd ¥od

¢cS6l1

 

TEERTI e ey
 

0ST

 

. o UNCLASSIFIED
_ C ' _ DWG. 14446
: : : ‘ CREEP RATE (% per hr)

0.0001 0.004 oN 01 : 1 10

© 20,000

INCONEL SHEET, ANNEALED
FOR 1hr AT 2050°F

GRAIN SIZE APPROXIMATELY
15grains/mm? 0.250mm DIA

TESTED IN ARGON AT 1500°F

(0000 _

9,000 239

8,000 ~_ 22%
7,000 —t >

    
 

~ 6,

£ 5000 I T~ =

4 4000 S o>

5 1 Se— \
~35

3000

TIME iS SHOWN FOR 04,05, 1, 2, AND 5%
2,000 EXTENSION AND RUPTURE VS. STRESS;
EXTENSION MEASURED OPTICALLY

ol { 10 100 1,000 - 10,000
TIME (hr)

Fig. 47. Creep and Stress-Rupture Data for Coarse-Grained Inconel Sheet.

J40ddH SSHHSOH_d ATHALYVNO NOISIAIQ ANV
 

 

.

CERAMICS LABORATORY

T. N. McVay, Consultant

Consideration of the possible appli-
cation of ceramics-coated materials

to reactors has indicated several po-

tentialities that require further
development. A ceramic coating has
been applied to a high-temperature
radiator and appears to offer excellent
oxidation resistance. Petrographic
examination of fluoride fuels is being
undertaken and additional equipment -~
mainly high-temperature furnaces -
have been added to the laboratory
facilities.

Ceramic Applications to Reactors.
Some general conclusions can be given
concerning the usefulness of ceramics
in reactors on the basis of an analy-
sis of their physical properties and

other information. ¢

1. Glass and glass-bonded ceramics,
which include nearlyall "conventional”
ceramics, are not useful in high-
temperature reactors as structural
elements because of a tendency to
soften at relatively low temperatures,
brittleness, poor mechanical and thermal
shock properties, and poor thermal
conductivities. However, 1f they can

"be used as liquids, grains, powder,

or coatings they may be considered.

2. Oxide ceramics generally have
most of the above-mentioned disad-
vantages, often to a lesser degree,
but they do not exhibit the tendency
to soften at low temperatures. All
the oxides with the exception of beryl-
lium oxide have thermal conductivities
considerably less than those of most

metals. If the thermal and Strfictural_

requirements are not too great and
the coolant corrosion problems can be

 

(3)J. R. Johnsen,. Ceramic Materials as Related
to the Reactor Program, ORNL CF-52-1-144, Jan,
18, 1952.

FOR PERIOD ENDING MARCH 10, 1952

met, oxides may be used in reactors,
perhaps even as structural material.
Oxides in combination with other ma-
terials offer the greatest promise.

3. Carbides, nitrides, borides,
hydrides, and sulfides have been
investigated, but most of them need
further study. Beryllium carbide was
intensively investigated and developed
in the NEPA project and will be studied
further under G.E. It has many de-
sirable properties and if successfully
coated to prevent corrosion (particu-
larly oxidation) and loss of fissicn
products, should prove useful in high-
temperature reactors. Although many
of these unusual materials are very
refractory, they are also very re-
active at high temperatures, particu-
larly in the presence of oxygen and
water vapor.

4. Ceramic-metal combinations
(Cermets) have been studied intensively
by groups sponsored by the Air Force.
Both structural materials and coatings
have been developed. A few that have
shown promise are Al,0,-Cr, Al,0,-Fe,
Al,0,-Ni, MgO-Ni, and TiC-Co. The
Cr-Al1,0, Cermets have some interesting
properties such as reasonably high
strength at temperatures up to 2372°F
and very good oxidation resistance at
these temperatures. They are somewhat
brittle, however, and are not too
resistant to thermal shock. This
property depends on the oxide content.
Fabrication techniques have been well
developed for some shapes. Consider-
able further study will be necessary
to determine their usefulness in high-
temperature reactors.

Coatings for the Radiator. A con-
siderable amount of work has been
done this quarter in applying ceramic
coatings tostainless steel and Inconel,
since these may be required in a
liquid-to-air radiator. Much promise
for high-temperature and oxidation
resistance for reactor materials has

151

 

 
 

 

 

been shown by NBS Ceramic-Coating

'A-418. Samples coated to about 2 mils

thickness are currently undergoing
testing. The computed oxide composi-
tion of the frit (glass phase of the
coating) is:

R

Si0, '37.5
B,0, | 6.5
-BaOQ 44.0
CaO , , 3.5
" ZnO 5.0
Al,0, 1.0
ZrO2 2.5

100.0

Similar coatings developed for the
Air Force have been shown to protect
stainless steel at 2000°F for over
200 hr, or 1800°F for an almost in-
definite period. Testing procedures

for these coatings are described in
. the literature. ’

(4)

 

(_4)W. N. Harrison, D. - G. Moore, and J. C.
Richmond, *“Ceramic Coatings for High-Temperature
Protection of Steel,’”” RP1773, J. Research Nat,
Bur. Standards 38, 293-307 (March 1947).

152

-ANP DIVISION QUARTERLY PROGRESS REPORT

Ceramic Laboratory Egquipment. A
high-temperature molybdenum-wound
furnace designed by Thomas Shevlin,
Consultant from Ohio State University,
will be used for Cermet work. The
shell and accessory components are
being fabricated in the Y-12 shops.
The refractories and windings are
either on hand or on order. =~

The high-temperature x-ray furnace

and the small, high-temperature vacuum
furnace have been built. The thermal-

diffusivity and thermal-expansion
equipment have been designed and are
in the shops. The high-temperature
dilatometer 1is 1n use.

Microscopic Examination 6f Fluo-

rides. Petrographic examination of

fluoride fuels is under way, and the
solubility of uranium tetrafluoride
in beryllium fluoride glass is being
studied. Five mole per cent of the
uranium tetrafluoride appears to be
soluble in the beryllium fluoride.
No glass was present in the one fused
mixture of NaF-KF-ZrF,-UF, studied.

 

 

 
 

 

L

My

&t

FOR PERIOD ENDING MARCH 10, 1952

13. HEAT TRANSFER AND‘PHYSICAL PROPERTIES RESEARCH

H. F. Poppendiek, Reactor Experimental Engineering Division

Viscosities of NaF-KF-LiF eutectic
have been determined over a wide
temperature range by utilizing three
types of viscometers that range from
about 8 centipoises at 550°C to 3
centipoises at 800°C. The addition of

~up to 30 wt % of uranium tetrafluoride

increased the viscosity of this
fluoride from 5 to 7 centipoises at
700°C but only from 4.0 to 4.3 centi-
poises at 800°C. Some preliminary
thermal conductivity information on
this coolant has also been obtained.
Heat capacity determinations have been
made for two different compositions of
NaF-BeF,-UF, mixtures over wide temper-
ature ranges. A table summarizing
physical properties of materials of
interest to ANP 1s included.

A series of corrosion failures has

made it impossible, for the present,

to obtain fundamental heat transfer
data for fused salts and hydroxides.

‘The heated-tube, lithium system has

been completed and 1s to be used to
determine heat transfer characteristics
for this coolant in the near future.
Some experimental boilingheat transfer

data for a mercury system have been

obtained.

The mathematical analysis of circu-
lating- fuel heat transfer for turbulent
flow has been evaluated for a series
of Reynold’'s and Prandtl’s moduli for
the case of no wall heat transfer.
Mathematical solutions for the natural
convection of liquid fuel elements have
been developed for turbulent flows.

VISCOSITY OF FLUORIDE MIXTURES
M. Tobias S. I. Kaplan

Reactor Experimental Engineering
Division

J. M. Cisar
ANP Division

F. A, Knox F. Kertesz
Materials Chemistry Division

The viscosity of the various fluoride
fuels proposed for the circulating-
fuel reactor is of great importance,
since a low value (less than 10 centi-
poises at 800°C) 1is required. The
mixture NaF-KF-LiF-UF, appears to meet
this requirement, whereas NaF-KF-UF,
does not.

Viscosity of NaF-KF-LiF. The
viscosity of the NaF-KF-LiF eutectic
(11.5 mole % NaF, 42.0 mole % KF,
46.5 mole % LiF) was measured by using
three different viscometers, a modified
Brookfield viscometer, an efflux unit,
and the previously described falling-
ball instrument.(!? The data ob-
tained(2+3) range from about 3 * 1
centipoises at 550°C to 3 + 1 centi-
poises at 800°C.

 

(I)F. A. Knox and F. Kertesz, *“Brookfield
Viscometer,” Aircraft Nuclear Propulsion Project
Quarterly Progress Report for Period Ending
September 10, 1957, ORNL-1154, p. 136.

(2)y, Tobias, Measurements of the Viscosity of
Flinak, Y-F30-6, Feb. 26, 1952.

(B)S. I. Kaplan, Viscosity Measurements of
Flinak by the Falling-Ball Viscometer (to be
issued),

153

Lt idiaat

TERT O )

 
e i

 

 

/

ANP DIVISION QUARTERLY PROGRESS REPORT

Viscosity'of NaF—KF-LiF-UF4. The

previously described(*) apparatus con-

sisting of a modified Brookfield
viscometer and a controlled atmosphere
furnace with certain improvements to
guard against hydrolysis and oxidi-

‘zation was used during the past quarter

to determine the viscosity of NaF-KF-
LiF-UF,. Increasing the amount of
uranium fluoride, up to 30wt %, failed
to cause a substantial change in the
viscosity, as shown in Takle 35.

TABLE 35

Viscosity of NaF-KF-LiF-UF, MIXTURES

as a Function of yYranium Tetrafluoride

 

 

 

Concentration
UF, CONCENTRATION*| VISCOSITY + 0.5 (centipoises)
T (wt %) 700°C 800°C
0 5.2 4.0
2 5.2 4.1
15 | 6.0 4.3
30 7.2 4.3

 

 

 

*Solvent (mole %): 11.5, NaF; 4.2, KF; 46.5, LiF.

Modifications ofViScosity*Apparatfis.

. The efflux and Brookfield viscometers

are being installed in a dry box filled
with inert gas to study the viscosity
of beryllium-containing fluoride
mixtures after hydrofluorination to
remove insoluble oxides formed by
oxidation and hydrolysis. A new,
nickel viscometer tube has been 1in-
stalled in the falling-ball apparatus
for further high-temperature salt work.

 

(4)F. A. Knox and F. Kertesz, ‘*Viscosity of

Fluoride Mixtures,” Aircraft Nuclear Propulsion .

Project Quarterly Progress Report for Period
Ending December 10, 1951, OBRNL-1170, p. 126.

154

The length of the measured fall path
has been increased by about one-fourth

and the amount of material needed
decreased by one-third,

THERMAL CONDUCTIVITY OF LIQUIDS

AND SOLIDS
L. Cooper M. Tobias
W. D. Powers S. J. Claiborne

Reactor Experimental Engineering
Division

The thermal conductivity apparatus
was further checked by calibration with
a molten metal. The value of the
thermal conductivity of lead obtained
with this device agreed within 13% of
the value given in the literature.
Experiments to study Flinak (LiF-NaF-
KF eutectic) were initiated; however,
at the end of one run a corrosion
failure (attack on a weld and a stain-
less steel bellows) halted the experi-
mental work. The preliminary results
of this single run indicated that the
thermal conductivity of Flinak was
severalfold greater than that of the
fluoride heavily laden with uranium
(46.5 mole % NaF, 26.0 mole % KF, and
27.5 mole % UF,) that had been studied
previously. Modification of the
system will be attempted so that
corrosion will be minimized, Upon the
completion of this modification,
further thermal conductivity measure-
ments of the fluorides will be made.

A new, longitudinal flow apparatus,
originally developed to investigate
solids, has been erected. This device
will now be used for studing ligquids
with high thermal conductivities. Tt
is believed that the minor free-
convection currents that might occur
in this system will not significantly
influence the heat transfer (conduction
being the important heat transfer mode
in this case).

 

 

 

 
 

 

#

HEAT CAPACITIES

‘W. D. Powers R. M. Burnett
G. C. Blalock

Reactor Experimental Engineering
Division

The enthalpies and heat capacities
of the following salt mixtures have
been determined by the use of Bunsen
ice calorimeters. For the 76 mole %
NaF-12 mole % BeF,-12 mole % UF,
mixture at 250 to 465 C,

Hy (s) - Hyo( (s) = 0.22T - § (cal/g)

¢p = 0.22 £ 0.02 (cal/g°°C) ,

‘and at 520 to 1000°C,

Hy (1) ~Hyop (s) = 0.32T - 35 (cal/g) ,

c, = 0.32 £ 0.03 (cal/g °C) .

For the 25 mole % NaF- 60 mole % BeF -

15 mole %UF4 mixture at 280 to 1000° C

HT - HO,OC = 0.32T - 45 (cal/g) ;

4~

¢, = 0.32 + 0.02 (cal/g-°C) .

The enthalpy temperature curve of the

second salt mixture exhibited no dis-
contlnultles and thus 1ndlcated the
character1stlcs of a glass.

At present the heat capac1t1es of
Inconel potass1um hydroxide, barium

fhydroxlde, and a lithium-sodium

potassium- fluorlde mixture are belng
determined.

'VAPOR PRESSURE_OF LIQUID FUELS
R. E. Moore

Materials Chemistry Division

Previous determinations of the
vapor pressure of uranium tetrafluoride

FOR PERIOD ENDING MARCH 10, 1952

are in disagreement; therefore new
measurements were made between 1037
(slightly above the melting point)
and 1185°C., The apparatus and pro-
cedure employed were described previ-
ously.¢%*6) The data may be repre-
sented by the equation

log,, P(mm Hg) = (-9130.5/T) + 7.774 ,
which gives a maximum variation from
the measured values of 4.6%. The heat
of vaporization, as calculated from
the equation, is 42.1 kg-cal/mole.

In earlier work Johnson(7) used
a static method for determining
the vapor pressure of solid uranium
tetrafluoride, and Ryon and Twitchell(8)
measured a series of boiling points
at reduced pressures., The highest
vapor pressure obtained with the solaid
was approximately 6.5 mm at 1000°C,
but in the present work this pressure

"was observed at about 1040°C. The

boiling-point method gave a much
higher heat of vaporization than that
calculated from the data in this
report; that is, the pressures ob-
tained at higher temperatures were
considerably higher, and those obtained
at lower temperatures considerably

- lower, than the values reported herein,

The difficulty of obtaining reliable
boiling points in the low pressure
range and the fact that the points
were scattered m1ght account for some
of the differences.

Vapor preéstres of sodium fluoride
and potassium fluoride, the other two
components of the NaF-KF-UF4'eutectic

 

(S)R. E. Moore and C. J. Barton, “Vapor
Pressure,” op. cit., ORNL-1154, p. 1317.

(S)R. E. Moore, “Vapor Pressure of Ligquid
Fuels,” op. cit., ORNL-1170, p. 126.

(7)K. 0. Johnsson, The Vapor Pressure of
Uranium Tetrafluoride, Y-42, Oct. 20, 1947.

(B)A. D. Ryon and L. P. Twichell, Vapor
Pressure and Related Physical Constants of lIranium
Tetrafluoride, H-5.385.2, July 25, 1947.

155

 
 

 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

mixture, have been determined by
 Wartenberg and Schulze.(®) The data
for sodium fluoride and potassium
fluoride, as well as the data reported
here for uranium tetrafluoride, were
extrapolated to 1267°C, the highest
temperature at which the vapor pressure
of the eutectic mixture was measured.

The ideal partial pressures at 1267°C of

each of the components of the mixture,
as calculated by Raoult’'s law, show
6.4 mm for sodium fluoride, 31 mm for
potassium fluoride, and 19 mm for
uranium tetrafluoride, The ideal
total pressure is therefore approxi-
mately 56 mm, whereas the experimental
vapor'pressure'at 1267°C is 11.6.¢5)

Assoc1at10n in the mixture probably
accounts for the large deviation from
ideal behavior. A 11ke1y poss1b111ty
is the presence of complex ions
formed by association of fluoride ions
and uranium tetrafluoride molecules.

PHYSICAL PROPERTY DATA

M. Tobias
Beactor Experlmental Engineering
D1v1s1on

' Summaries of the available data on
the physical properties of fluoride
salts and other reactor materials are
"presented in Tables 36 and 37. Most
~of the data were obtained by the
‘physical properties group at ORNL;
‘however, pertinent physical property
measurements obtained by other organi-
zations are also included. The ORNL
physical property measurements are
being made for the purpose of quickly
supplying the ANP project with data of
‘a reasonable accuracy. The general

 

(9)
Pressure of Some Salts.
568 (1921)

v, Wartenberg and H. Schulze, “Vapor
11" Z. Elektrochem. 27,

156

. 5. Density:

assigned to the data
tables are noted as

accuracies
presented in the
follows:

1. Melting point: within about +10°C.

2. Heat Capacity: within about *10%.

3. Thermal conductivity: preliminary
checks on the thermal conductivity
device for liquids, made by using
molten lead 1indicated that the data
fell within about +30% of the known
values. An error analysis, assum-
ing pessimistic Chromel-Alumel
thermocouple deviations, suggested
that the NaF-KF-UF, data were
accurate to within about +30%.
Further checks are belng made

4. Viscosity: the results so far are
preliminary. A definite accuracy
cannot yet be assigned.

within about +5%.

Only the existence of preliminary
physical property data has been noted
on Tables 36 and 37; when the results
have been confirmed, the data will be
listed.

NATURAL CONVECTION IN CONFINED SPACES
WITH INTERNAL HEAT GENERATION

D. C. Hamilton L. Palmer

F. E. Lynch R. F. Redmond

Reactor Experimental Engineering
Division

Analytical solutions have been
developed for natural convection
systems in which the heat source and
the wall flux are uniform and the
aspect ratio is very high (i.e., L/d
very large). Three memorandums have
been published regarding the laminar

 

 
LST

wy

 

TABLE 36

Physical Properties of Fluoride Salts

 

 

 

 

 

 

 

 

 

 

 

APPROXIMATE (¢}
MELTING POINT HEAT CAPACITY THERMAL CONDUCTIVITY VISCOSITY DENSITY VOLUME EXPANSION
MOLE % (°C) {cal/°C-g} {Btu/hr- ft2-°F/ft) (centipoise) (g/cc) COEFFICIENT
NaF-BeF,-UF, 76, 12, 12 480 Solid, 0.22 at 250°C < T < 465°C;
liquid, 0.32 at 520°C < T <
1000°Ct®)
NaF-PeF,-UF, 25, 60, 15 635 0.32 at 280°C < T < 1000°C ‘*!
{starts
solidifying)
NaF-KF-UF, 46.5, 26, 2.5 530 Solid, 0.15 at 240°C < T < 535°C, 0.53 at 550°C < T < 750°C'® 470 - 1.15x 1073 T 2.82 x 1074 at 535°C
(eutectic) tiquid, 0.23 at 535°C < T < at $35°C < T < 1000°C )|  3.24 x 10"% ac 1600°C
1000°Cte?
NaF-KF-UF, 48.2, 26.8, 25 558 4,54 - L1 x10°3 T at 2.84 x 104 at 600°C
600°C < T < 900°C N 3.10 x 10°% at 900°C
NaF-KF-LiF 11.5, 42, 46.5 455 Preliminary results Preliminary 2.39 - 5,9 %x 1004 T at 2.84 x 10"% at 530°C
(Flinak) results: 8 at | 530°C < T < 850°C'#8} 3.12 x 10°% at 850°C
550°C to 3 at
800°C
UF,-LiF-NaF-KF | 1.1, 44.5, 10.9, 43.5 455 2,65 - 9.0 x 104 T at 4.15 x 1074 at 530°C
(Fulinak) : 530°C < T < gso°c'™ 4.79 x 10-4 at 850°C
UF, 103541 0.073 + 6.1 x 1075 T at 280°C
< T < 950°C'*!
(G)C. J, Barton, ORNL Materials Chemistry Division, personal communication.

(6)y, D, Powers, ORNL CF-51-11-195, Nov. 30, 1951.

(C)w

. D. Povers, ORNL CF-51-9-64, Sept. 13, 1951.

(d)1, Basel and M. Tobias, ORNL CF-51-7-169, July 31, 1951.

(e)s, 1. Kaplan, ORNL CF-51-8-97, Aug. 13, 1951.

(f)3. Cisar, ORNL CF-51-11-78, Nov. 14, 1951

(8)5. Cisar, ORNL CF-51-12-91, Dec.

14, 1951,

(h)5, Cisar, ORNL CF-51-11-198, Nov. 30, 1951.

(i),

o

D. Ryon and L. P. Twitchell, H-5,385.2, July 25, 1947,

‘0T HOUYW ONIGNA aoIydd dod

cS6tl

 
 

 

8S1

ol gl

TABLE 37

*

Physical Propertiés of Miscellanéous Materials

 

 

 

 

Type-316 stainless steel

1370 to 1400 (k)

<1< 980°C

Solid, 0,1093 + 5.66 x 10°5 T at
150°C < T < 1000°C \*

32 at 300°C

9.02 at 100°C (A)
12. 4 at 500°C

APPROXIMATE -
MELTING POINT HEAT CAPACITY THERMAL CONDUCTIVITY VISCOSITY
(°c) (cal/°C*g) (Btu/hr*ft2+°F/f¢) (centipoise)
" Sodium hydroxide 323(a) Liquid, 0.49 at 340°C < T < 990°C " | 0.81 at 520°C(¢) 4.0 at 350°c(4)
: 2.2 at 450°C
1.5 at 550°C
_ | e 1.0 at 650°C
Lead-bismuth alloy 125(¢) Liquid, 0.035 at 175°C < T < 1000°C | 5.32 at 160°C ¢
(55.3 mole % Bi) 6.53 at 320°C
Nickel 1452 F) Solid, 0.12 + 3.1 x 10" T at 240°C | 36 at ¢°c(8)

Li thium 186 {f) Liquid, 1.001 + 2.76 x 10°° T at 20.6 to 26.6 at m.p.'"’ 0.60 at 205°C (<)
250°C < T < 1000°C 0.41 at 1000°C
Zirconium 1700 (£ Solid, 0.0697 + 3.62 x 10°5 T at 11.3 at room temp. /)
150°C < T < 1050°¢ (i) 11.2 at 200°C
10.7 at 500°C
Mol ybdenum 2620 ) Solid, 0.0675 at 150°C < T < 1050°C‘*?| 70.3 at 540°C (%)

 

 

 

62.9 at 870°C
56.5 at 1145°C

 

 

(a)

personal communication.

C. J. Barton, ORNL Materials Chemistry Division,

(f)Chemical Engineer’s Handbook, 2d, ed., p. 313-367.

€)rbid., p. 949.
(b) .
W. D. Powers, ORNL CF-51-11-195, Nov. 30, 1951. (h)United States Steel Corporation, Fabricetion of U.8.8.

. ST 5
(C)H. R. Deem, Battelle Memorial Institite, personal Stainless and Hgat Resisting Steels, 1950 ed.

communication. . (i)w. D. Powers, ORNL-1154, p. 134.
(PH. R. Stephen, NEPA IC-50-4-20, Apr. 10, 1950. “7)G. Bing, F. W. Fink and H. B. Thompson, BMI-65, Apr. 16,

g
‘¢)iiquid Metals Handbook, p. 31, June 1, 19%0. 1951

(E)E. Mikol, ORNL-1131, Feb. 14, 1952.

 

LH0d3d SSIYI0Ud ATHALHUVAD NOISIAIG dNV
 

 

 

*

i

&)

flow case,(10+11.123  Ope analysis in-
volves a parallel plane system and
consists of solving the heat con-
duction equation by using a simplified
velocity distribution; a similar
analysis has been made for a circular
pipe system; and a third analysis con-
sists of solving the hydrodynamic and
heat flow equations simultaneously,
The temperature distributions and
critical Reynold’s moduli for the third
analysis, which was more exact than
the first, were almost 1dentical.

The following comments apply only
to the ideal systems treated in these
memorandums. JIn small tubes (0.2 1in.
in diameter) a temperature reduction
as a result of natural convection does
not seem to occur for practical values
of the variables. The ratio of the
difference in center line wall temper-
ature and the temperature when con-
duction is the only mechanism present
1s represented by ¢ . For laminar
flow in very large tubes, @, may be as
small as 0.3. In the case of turbulent
flow, ¢, is smaller by an order of
magnitude, To attain turbulent flow
and the resulting low values of ¢0 in
small tubes, large temperature dif-
ferences are necessary. It should be
noted, however, that systems using
large-diameter tubes would yield
turbulent flow with much smaller
temperature differences.

An apparatus is being constructed
to obtain an over-all temperature
difference for a cylindrical system in

 

(10)yy . Hamilton, H. F. Poppendiek, and

L. D. Palmer, Theoretical and Experimental
Analyses of Natural Convection Within Fluids in
which Heat is Being Generated, Part I, ORNL
CF-51-12-70, Dec., 18, 1951.

(Il)From A. Simon to H. E. Sterm, Letter
Regarding Agreement on Tower Calculations, ORNL
CF-52-1-1, Jan. 2, 1952.

(12)D. C. Hamilton, R. F. Redmond, and L. D.
Palmer, Theoretical and Experimental Analyses of
Natural Convection Within Fluids in which Heat 1is
Being Generated, Part III, ORNL CF-52-1-2, Jan,
11, 1952,

FOR PERIOD ENDING MARCH 10, 1952

the turbulent flow region., The new
design will facilitate the measure-
ment of wall temperature. Mercury
will be used as the heat generation
medium, '

A flat plate apparatus 1s being
designed to permit visual study of the
velocity distribution in the laminar
region, (13> A dilute sulfuric acid
solution will be used as the heat
generation medium,

ANALYSIS OF HEAT TRANSFER IN A
CIRCULATING-FUEL SYSTEM

H. F. Poppendiek L. Palmer
Reactor Experimental Engineering
Division

Mathematical solutions for laminar
and turbulent heat transfer in the
circulating-fuel systems were described
in a previous report,(!*) The wall-
mixed mean fluid temperature difference
in the established flow region for
laminar flow 1is

Gr02 [llF ‘- 8]
t -t = ——
w mm k 48

 

 

 

where

G = volume heat source,

ro - pipe radius,

k = fluid thermal conductivity,

2 d
Gr, (d4],

dq :
— = pipe wall heat transfer rate.
dA},
{13)R. F. Redmond, Theoretical and Experimental

Analyses of Natural Conwvection Within Fluids in
wvhich Heat ts Being Generated, Part V, ORNL
CF-52-1-5, Feb, 12, 1952.

(14)H. F. Poppendiek and L. Palmer, Forced
Convection Heat Trensfer in a Pipe System with
Volume Heat Sources Within the Fluids, Y-F30-3,
Nov., 20, 1951.

159

T

 
Lo

 

 

 

 

~ ANP DIVISION QUARTERLY PROGRESS REPORT

The wall-mixed mean fluid temperature

~difference for turbulent flow is found

to be a function of Reynold’s modulus
(Re), Prandtl’s modulus (Pr), the
volume heat source, the fluid thermal

2x1073

10

10-2

10° 104

 

conductivity, the pipe radius, and the
function F., The turbulent solution
for the case of no wall heat transfer
(F = 1) has been evaluated for a range
of Re and Pr and graphed in Fig. 48.

UNCLASSIFIED
DWG. 14447

 

105 10°

Re

AFig. 481 Dimensionless, Wall-Mixed Mean Temperatureflnifferéncé assiFunction
of Reynold’s and Prandtl’s Moduli for a Heat Transfer System with Insulated

Pipe Walls,

160

 

 

 

 

 
 

L s,

 

O

‘the Metalloy Corporation.

These solutions for laminar and tur-
bulent flows can be used in designing
circulating-fuel systems,

HEAT TRANSFER COEFFICIENTS

Heat Transfer in Molten Lithium
(H. C. Claiborne and G. M. Winn,
Reactor Experimental Engineering
Division). The apparatus for measuring
heat transfer coefficients with an
electrically heated tube was completed.
Internal cleaning-and degreasing were
accomplished by thorough flushing with
ethanol and trichlorcethylene. The
system was outgassed by heating and
repeated flushing with argon scrubbed
with lithium and eutectic sodium-

"potassium alloy to remove the oxygen

The flush-
ing procedure consisted of raising the
pressure to about 35 psig with scrubbed
argon and bleeding the pressure down
to about 2 psig. After filling with
lithium, the system, when not in
operation, was kept under 35 psig with
scrubbed argon to prevent air con-
tamination,

and nitrogen contaminants,

Sixteen pounds of lithium was
added to the system through the filter
in 4-1b batches. The lithium used was
the low-sodium grade purchased from
To keep
contamination down to a minimum, the
film of nitrides and oxides adhering
to the 1-1b cylinders of solid lithium
was shaved off and the solid metal
soaked 1in trichloroethylene. This
operation was done in a dry box filled
with argon. The lithium was success-
fully pumped around the system (via
the test section by-pass) by the two
electromagnetic pumps in series, The
electromagnetic flow meter was cali-
brated by using the drain tank that
had been previously calibrated with
water,

~ After the system was prepared for
obtaining heat transfer data, it was

FOR PERIOD ENDING MARCH 10, 1952

found that an electrical current could
not be passed through the test section,
Examination revealed that the test
section had melted at one point,.
Apparently too much current was passed
through the section while testing the
transformer circuit. A new test
section is being fabricated and other
minor changes are being made. Another
attempt to get heat transfer data will
be made as soon as this work 1is
complete.

Heat Transfer to Fused Salts and
Hydroxides (H. W. Hoffman andJ. Lones,
Reactor Experimental Engineering
Division). Flinak was removed from
the heat transfer system and replaced
with sodium hydroxide during the last
quarter., Corrosion by the Flinak had
caused leaks in several welds and in
one of the mixing pots, and'it was
necessary to replace these parts of
the system., A redesigned mixing pot
consisting of a section of 1-in. nickel
pipe 2% 1in, long and capped at both
ends was installed. The fluid enters
at one end of the pot tangentially to
the inside surface of the pipe, passes
through a perforated nickel disk, and
leaves at the bottom. A thermocouple
is located in front of the exit.

Entrance Region Heat Transfer in a
Sodium System (W. B, Harrison, Reactor
Experimental FEngineering Division),
The modifications and additions to
the experimental system proposed in
the last quarterly report{(13) have
been made, and preparations are being
made for loading the system with
sodium,

Extremely high values of the heat
transfer coefficient should be achieved
with the use of sodium in an entrance
region (e.g., 200,000 Btu/hr ft2:°F);
therefore good wetting of the copper

 

(IS)W. B. Harrison, ““Entrance-Region Heat
Transfer in a Sodium System,” op. cit., OBRNL-1170,
p. 118. .

161

 
 

- ANP DIVISION QUARTERLY PROGRESS REPORT

test section by the sodium is of con-
siderable importance. There are no
sodium wetting data available for
copper, but the data for similar
metals indicate that copper may be
well wetted by sodium even at the low
operating temperatures (up to 300°F).
A test will be conducted to determine
the degree of wetting and possible
ways of improving it 1f it should be
poor. The study will consist of
measuring relative electrical re-
sistances across a few copper-sodium
interfaces as functions of temperature,
The interfaces proposed at present are
sodium in contact with copper that has
been silver plated, mercury plated,
and oxidized. The equipment has been
assembled for this experiment, and
data should be available in the near
future. Work on the heat transfer
system 1is being deferred in order to
‘capitalize on any positive results of
the wetting studies,

HEAT TRANSFER OF BOILING
*_ LIQUID METALS

W. S. Farmer

Reactor Experimental Engineering
Division

~ The previous boiling-mercury data
for the horizontal plate system(16)
have been analyzed and compared with
existing information in the literature
on nucleate boiling and free con-
vection, A plot of this data is shown
in Fig. 49. The Nusselt numbers
(hL/kR) obtained experimentally fall
below the values predicted by cor-
relations of Insinger and Bliss(17)
for nucleate boiling and Jakob¢(!¥®) for
free convection, Over most of the

 

(16)“’. S. Farmer, “Heat Transfer in Boiling-
Liquid-Metal Systems,' op. c¢it,, ORNL-1154, p. 138.

(17)T. H. Insinger, Jr. and H. Bliss, “Trans-
mission of Heat to Boiling Liquids,”” Trans. A4m.
Inst. Chem, Engrs. 38, 491 (19490).

(IB)M. Jakeb, Heat Transfer, I, 640, Wiley,
New York (1949).

162

temperature difference range the
experimental values are approximately
one-fourth the predicted values,
Existing correlations may not correctly
take into account Prandtl’s number and
wetting, which may be of importance in
liquid metal beoiling.

UNCLASSIFIED
DWG. 14448

 

o
o
S

MERCURY TEMPERATURE, 340° F

500

 

 

300
@
® A4Jé

WHERE
7¢=METAL SURFACE TEMPERATURE

7, =BULK LIQUID TEMPERATURE

 

 

 

 

 

HEAT TRANSFER COEFFICIENT, 4 (Bfu/hr'ffz-oF)

[e]e;

 

20 50 100 200
o
T,-7, {°F)

Fig. 49. Heat Transfer Coefficients
for Boiling Mercury.

Minor changes are to be made in the
horizontal flat plate system in order
to evaluate wetting characteristics
and heat transfer surface resistances.
In addition, a small, inexpensive,
mercury circulation unit is to be con-
structed. Mercury will then be used
as the condenser coolant for experi-
ments in which the heat transfer
coefficients of boiling sodium will be
measured,

The experimental apparatus for in-
vestigating heat transfer coefficients
of boiling liquid metals by using a
horizontal tube geometry was completed
this quarter. Preliminary tests using
water as a boiling fluid are now in
progress. '

 

 

 

 

 
 

 

FOR PERIOD ENDING MARCH 10, 1952

14. RADIATION DAMAGE

D. S. Billington, Solid State Divisien
A, J. Miller, ANP Division

Studies have continued on the
the effect of radiation on the sta-
bility of airplane reactor constitu-
ents. Considerable data have been
obtained from experiments in the X-10
graphite pile, the higher flux LITR,
and the Y-12 éyclotron. In all phases
of the work, preparations are being
"made to carry out additional studies

in the LITR and to utilize the MTR
facility when it becomes available.
This work has been concerned largely
with the stability of fused fluoride
"salt mixtures of the type proposed as
fuel for circulating fuel reactors.
In several of the inpile experiments
with the salts in Inconel capsules,
the rate of attack on the container
material has been considerably higher
than that observed in the out-of-pile
controls, However, the evidence is
not conclusive that the increased

corrosion rate is caused by the radia-

tion field.

Other radiation damage studies have
included inpile liquid metal loop
experiments and creep of metal. So-
dium was circulated in the 1inpile
loop for 165 hr, during which time
radioactive decay curves were taken.
The thermal condictivity of metals
under irradiation has also been
studied. The earlier creepand thermal
conductivity data have been partially
substantiated by recent tests. One
such creep test with nickel indicates
that the inpile creep rate becomes
higher than its bench counterpart after
about 115 hours. Additional details
on radiation damage studies are con-
tained in a quarterly report of the
Physics of Solids Institute. (!’

 

(I)Physics of Solia’s'Institute Quarterly
Progress Report for Period Enrnding January 31,
1952, ORNL-1261 {in press}.

IRRADIATION OF FUSED MATERIALS

G. W. Keilholtz
Materials Chemistry Division

The effects of radiation on the
stability of fused fluorides in con-
tact with Inconel have been investi-
gated by use of the X-10 graphite
pile, the LITR, and the Y-12 cyclotron.
Reasonably detailed inspection of
materials from experiments carried out
in the X-10 graphite pile and cyclo-
tron at 1500°F showed no positive
evidence of fuel decomposition or
increased corrosion of the Inconel
when compared to out-of-pile control
tests. However, two fused fluoride
fuels irradiated in the LITR at power

densities of 800 and 84 watts/cc in

approximately 1/8-in.-ID tubes have
exhibited increases in the corrosion
rate as indicated by analyses of the
fuel for corrosion products. One of
these has the composition NaF-KF-UF,,
46.5, 26, and 27.5 mole %, respec-
tively, in which the power dissipated
when irradiated in the LITR is about
800 watts/cc of material. The second
fuel has a composition NaF-BeF,-UF,,
47, 51, and 2 mole %, respectively,
and a power dissipation of about 84
watts/cc. This is in the power range
of that occurring in the large-diameter
tubes of the proposed ARE. 1In both
cases there 1s some question as to
whether the inpile tests are being
carried out in a manner sufficiently
isothermal to prevent increased cor-
rosion rates by thermal gradients in
the capsules. In the case of the
low-uranium-concentration fluoride
mixture, irradiated at 84 watts/ce,
analytical results are available from
only one experiment. These results
are 1n sharp contrast to those from

163

B ialh s ok roriat Lo WP

T

 
ANP DIVISION QUARTERLY PROGRESS REPORT

an experiment on a similar beryllium
type of fuel containing a higher per-
centage of uranium, which had a power
dissipation of 554 watts/cc. The
latter experiment showed no evidence
" of increased corrosion, on the basis
" of an impurity analysis of the fuel,
as compared with the control.

Preparations are being made to
carry out studies on the escape of
xenon from the melt in the X-10 graph-
ite pile and on fuel stability in the
MTR.

Pile Irradiation of Fuel (J. G.
Morgan, P. R. Klein, C. C. Webster,
- B. W. Kinyon, M. J. Feldman, H. E.
Robertson, Solid State Division).
Experiments during the past quarter
have been conducted entirely on

fused fluoride fuels in the LITR.

‘Results of these experiments as com-

pared with the control runs are shown
in Table 38. The corrosion of the
Inconel in some LITR runs is not only
deeper and more general than that in
the controls, but there are also two
kinds present - the intergranular-type
found in the control runs and a type
of corrosion that is confined within
the grain and shows evidence of being
preferential in the grains attacked.

The capsules are considerably cor-
roded above the point where the liquid
level is assumed to be. There 1is a
small amount of such corrosion in the
capsules irradiated in the X-10 graph-
ite pile but little evidence of cor-
rosion of Inconel in contact with the
vapor phase in control runs. Chemical
analyses in Table 38 indicate that
less chromium is dissolved in the
irradiated fuel than in the control

TABLE 38

" LITR Tests on Fused Fluoride Fuels in Inconel at“1500°F

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ c : , INCONEL. COMPONENTS IN
'FUEL COMPOSITION | TIME | IRRADTATION FUEL AFTER TEST
~ {mole %) (hr) (watts/cc) Ni Cr Fe CAPSULE CONDITION
46.5 NaF- 115 800 26,878 | 1878 | 16,329 | Generally corroded to
26 KF-27.5 UF, ' depth of 2 to 3 mils
o 115 Control 1,034 | 754 | 1,270 | Occasional 1 to 2 mil
, intergranular attack
161 800 45,169 950 6,414 Generally corroded to
depth of 2 to 3 mils
161 Control 1,100 645 2,655 | Occasional 1 to 2 mil
intergranular attack
136 800 1,380 160 1,220 No corrosion*
47 NaF-~ 143 84 80,000 | 2450 3,100 Inspection incomplete
51 BeF,-2 UF,
T 136 Contrel 1,123 | 3300 811 Inspection incomplete
95 NaF- 139 554 1,490 | 1060 | 1,200 | Inspection incomplete
60 BeF,-15 UF,
131 Control 1,540 | 7380 | 2,300 Inspection incomplete

 

*At 824°F (solid)

164

 

 

 
 

 

.y

samples but that there 1s a large
increase in the nickel and iron con-
tent of the irradiated fuel compared
with that in the control. '

Cyclotron Irradiation of Fuel

(W. J. Sturm and M. J. Feldman, Solid

State Division, R. J. Jones and
R. L. Knight, Electromagnetic Research
Division). Fuel irradiations in the
cyclotron with 20-Mev protons described
in the previous quarterly report¢?’
were carried.out at 30 to 415 watts/cc
of fuel for a short period of time,
usually an hour. During the past
quarter smaller Inconel capsules have
been developed toavoid terminations of
runs because of uneven temperatures.
A helium cooling system has been de-
signed and is being constructed to
allow higher power dissipations.
With the Inconel microcapsules i1t was
possible to bombard a lithium-bearing
fuel - KF-NaF-LiF-UF,, 43.5, 10.9,
44.5, and 1.1 mole %, respectively -
up to 8 hr with 100 to 400 watts/cc.
Fuel analyses and metallographic ex-
aminations are being made.

Additional fuel irradiations are
being carried out by North American
Aviation with the use of the 60-in.
Berkeley cyclotron.

INPILE CIRCULATING LOOPS

0. Sisman , C. Ellis
W. W. Parkinson W. E. Brundage
A. S. Olson R. M. Carroll

C. D. Baumann

Solid State Division

Sodium was circulated at a velocity
of 1 ft/sec in the X-10 graphite pile
through a loop of Inconel for 50 hr

at 1000°F and 115 hr at 1500°F. An

 

(2}W. J. Sturm, M. J. Feldman, R. J. Jones,
J. 8. Luce, and C. L. Viar, “Cyclotron JIrradiation
of Fuel and XOH Capsules,” Aircraft Nuclear
Propulsion Project Quarterly Progress Report for
Period Ending December 10, 1951, ORBNL-1170, p. 143.

FOR PERIOD ENDING MARCH 10, 1952

electromagnetic pump was used, and the
temperature of the material in the
pump cell was 1000°F. The flow rate
of sodium through the loop gradually
decreased, and eventually the loop
could not be operated. When the ac-
tivity of the loop has decayed suf-
ficiently, an examination will be made
to determine the condition of the loop
and the cause of the flow stoppage.
At various times while the flow was
stopped, radiocactive decay curves were
obtained for the part of the loop
outside the pile. The curves are
being examined for long-lived corrosion
products from the Inconel tube walls.

A second sodium loop for the X-10
graphite pile and a sodium loop for
operation in the LITR are being con-
structed. A fused fluoride fuel loop
for operation in the MIR is being de-
signed.

CREEP UNDER IRRADIATION

J. C. Wilson J. C. Zukas
W. W. Davis

Solid State Division

It was previously reported¢?®’ that
a cantilever creep test at 1500°F and
1500 psi showed that X-10 graphite
pile irradiation at a flux of 4 X 10'°
fast neutrons/cm? caused an increase
in total creep strain of about 20% in
type-347 stainless steel after about
250 hr of exposure, which was the
duration of the tests. Extrapolation
of the bench and inpile curves to
longer times indicated that the dif-
ference between them increased with
time. In order to obtain further
information on this point, a 500-hr
test was run in the pile with no

 

(3)J. C. Wilson, J. C. Zukas, and W. W. Davis,
“Creep Under Irradiation,” Aircraft Nuclear
Propulsion Project Quarterly Progress Report for
Period Ending September 10, 1951, ORNL-1154,
p. 170.

165

 
|
|

strain measuring microformer (since
in 500 hr the creep strain would ex-
ceed the maximum microformer travel).
The beam deflection was measured with
cathetometer after withdrawal from
the reactor. The strain in the bench
test exceeded that of the inpile test
by 5%. Since this is contrary to what
would be expected from extrapolation
of the earlier data, the experiment
will be repeated.

‘One 1inpile test was run on an
electrolytic nickel sample at 1300°F
at a maximum fiber stress of 2000 ps1.
The material was annealed at 1500°F
for 5 hr following a cold reduction
of 40%. The temperature of 1300°F
was chosen as the maximum at which the
grain structure was known to remain
stable during the test. Figure 50
shows a plot of creep rate vs. time
for both inpile and bench tests. In
common with the earlier work on type-
347 stainless steel, the curves show
that the inpile strain-time curve

SESRE
DWG. 14122R1

 

 

N\

\ . BENCH TEST ~
\ o—— IN/PILE TEST

 

 

  

 

>

 

-
w

 

 

CREEP RATE (pin./in./ hr)
© )

 

 

 

 

ELECTROLYTIC-NICKEL,CANTILEVER CREEP SPECIMENS
2000-psi MAXIMUM FIBER STRESS

1300°F (ANNEALED 5 hr AT 1500°F AFTER 40% COLD
REDUCTION .

W

 

 

 

 

 

100 150
TIME (hr)

Fig. 50. Comparison of Bench and
Inpile Creep Rates of Nickel.

166

becomes more linear and shows a lesser
strain but a higher rate than its
bench counterpart after about 115
hours. At about 120 hr the inpile
rig had reached an approximately con-
stant strailn rate, whereas the rate in
the bench test was still decreasing.

RADIATION EFFECTS ON THERMAL
CONDUCTIVITY

A. F. Cohen L. C. Templeton
Solid State Division

In a preliminary relative thermal
conductivity experiment on Inconel,
previously reported,(4) a large de-
crease in thermal conductivity was
observed after three days of irradia-
tion at approximately 1517°F in the
X-10 graphite pile. To check this
result, a carefully annealed Inconel
specimen was irradiated at 482 and
1067°F and showed apparently no effects
from their radiation. When raised to
1517°F there was again some apparent
lowering of thermal conductivity, the
cause of which has not yet been de-
termined. In relative thermal con-
ductivity tests of high-purity cobalt-
free nickel in the X-10 graphite
pile, no lowering has been observed
during long periods at low tempera-
tures and several weeks of testing 1in
the region of 1500°F. An absolute
thermal conductivity test is being
run in the LITR at 1500°F on Inconel
that has been heat-treated to pre-
cipitate all the carbides that can
be precipitated by thermal treatment.

In an absolute thermal conductivity
test on type-316 stainless steel in
the fast flux of the LITR at 212 to
392°F the thermal conductivity was
shown to be unaffected by the irradia-
tion.

 

(4)A. F. Cohen, ‘“Radiation Effects on Thermal
Conductivity,” op., ctit., ORNL-1154, p. 171.

 

AP - 1 A - e Wi e

- -

T A 0 g T m———

e e

Ty v e e R ow

"y Ty Sl ¢

 
 

5

T
4.

e

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

5
X

i
Tl
e

 
 

 

 

 

@’

SUMMARY AND

The survey of the supercritical-
water reactor by the Oak Ridge National
Laboratory’s subcontractor, Nuclear
Development Associates, Inc.,, is com-
plete. NDA has concluded (sec. 15)
that a reactor of this type may be
developed for supersonic propulsion
but that the reacteor i1s an 1intricate
machine and will present a difficult
design job. Although OBNL will not
pursue this reactor cycle because of
the prevailing belief in the greater
potentialities of low-pressure liquid-
coolant reactors, NDA will continue
its work, together with Pratt and
Whitney, under direct AEC contract,

The large, analytical chemistry
program required in support of the
materials research program included
the routine analysis of 574 samples
and the development of new analytical

INTRODUCTION

procedures where necessary. This
development (sec. 16) 1is largely con-
cerned with the analysis of fluoride
mixtures for corrosion products and
changes in composition,

The "List of Reports Issued" (sec.
17) includes formal reports and
informal documents on all phases of

the ANP Project.

A directory of the research projects
of the Aircraft Nuclear Propulsion
Project of the Oak Ridge National
l.aboratory 1is given in sec., 18. The
research projects of the Laboratory’s
subcontractors on the ANP Project are
listed, as well as the research now 1in
progress at the Laboratory. The
research projects being performed by
ORNL for the ANP programs of other
organizations are included and marked
as such,

169

 

 

 

 
 
 

 

I e e o B i ot it i et b i o b b i R B M ik il bt el B b o e ook R M et i Tt TR AT T e i oo B . bk AR

 
 

 

15. THE SUPERCRITICAL-WATER REACTOR

Nuclear Development Associates, Inc.

The supercritical-water cycle,
first proposed in Report Wash-24,(1)
has been under continuous 1investigation
by the Nuclear Development Associates,
Inc., for over a year.
completed their analysis of the super-
critical-water reactor for ORNL and a
final report has been written. (%)
This report concludes the subcontract
work by NDA for ORNL on the super-
critical-water reactor. NDA will,

‘however, continue its work in this

field in conjunction with Pratt and
Whitney Aircraft Division in a study
of the engine and hardware aspects of
this system. The conclusions and
recommendations in the final NDA report
are quoted in detail below. This
study was concerned only with the
reactor and shield; the power plant
and engine aspects of the proposal
are outside the scope of the present
summary. :

‘DESCRIPTION OF REACTOR
The reactor consists of a structure
of stainless steel plates immersed
in a vessel of water that is above
the critical pressure, so that it can
be heated to high temperatures without

- phase change. The plates contain the

fuel and provide surfaces for trans-
ferring the heat to the water. The
water serves several functions: it

is the coolant that carries away the

heat from the plates; it is the moder-
ator that slows down the neutrons in
the chain reaction; through variations
in its density 1t controls the re-
activity of the machine; it is the
neutron reflector that surrounds the

"reactor core; and it is the innermost

portion of the reactor shield.

 

(I)Application of a Water Cooled and Moderated
Reactor to Aircraft Propulsion, AEC Reactor
Development Division, Wash-24, Auwg. 18, 1950.

f2)The Supercritical Water Reactor, Nuclear
Development Associates, Inc., ORNL-1177, Feb. 1,
1952.

NDA has now

CONCLUSIONS OF THE NDA STUDY

It is the conclusion of NDA that
it is possible to develop a reactor
of this type but that it is an intri-
cate machine and will present a dif-
ficult design job. NDA does not be-
lieve that this reactor represents an
easy short cut on the difficult road
to supersonic flight but that it may
offer one not-impossible path to that
goal. It is further stated that this
study represents only an early step
toward evolving such a reactecr and
definitely does not provide a pre-
liminary design ready for detailing.
An attempt is made to display the
potentialities of the machine and to
set forth at least some of the prob-
lems - the study has not solved the
problems. The impressions that have
been gained 1in regard to particular
items are:

1. The heat output (400,000 kw) re-
ported¢!’ appears attainable with
a reactor having a modest size
core (2.5-ft-square cylinder) and
a reasonable fuel inventory (20

kg).

2. The present estimate (85,000 I1b)
for the weight of a divided shield
with a crew compartment of 1in-
termediate size compares favorably
with the allowance (90,000 1b)
made in the report, or with the
somewhat higher weights allowed
for in the preliminary Boeing
studies.

3. This shield, like other divided
shields, does not permit normal
approach to the airplane after it
has landed. Additional awkward
shielding provisions are needed
for ground handling.

4. In view of industrial experience

with high pressures, the pressure

171

T W T P

Tt

T T

 
 

ANP DIVISION QUARTERLY PROGRESS REPORT

(5000 psi) called for in this re-.

actor does not 1in 1itself appear
to constitute a major difficulty.
Departure from standard pressure

vessel practice will be neces-

sitated by the low-weight require-
ment in the present application;

“although this will pose important

design problems, it 1is thought
that they can be worked out satis-
factorily.

There is considerable experience
with water and stainless steel,
ahd, at least in the absence of
radiation effects, the use of these
materials appears to be quite
promising.

A type of fuel element filate under
development at ORNL appears promis-
ing for use in this reactor.

Incorporation of the basic fuel
plates into satisfactorily cooled
assemblies is a complex problem
that will require major effort.
The reactor calls for a very fine-
scale fuel-bearing structure
worked at tremendous heat load.

A particular difficulty is that

of maintaining equal temperatures
in parallel cooling streams be-
cause of their sensitivity (aris-
ing from the large expansion of
the fluid) to differences in heat
load or other quantities. This
tends to increase the amount by
which the maximum wall temperature

" exceeds the exit mixed temperature

172

of the coolant. Thus, the im-
portant temperatures are sensi-
tive to complex details of the
reactor power pattern and to 1im-
perfections in the design and
fabrication of the machine.

Variations in water density pro-
vide a substantial amount of
self-regulation, as well as a
convenient mechanism for slow

10.

11.

external control of the reactor.
The machine appears quite amenable
to control under steady condi-
tions.

A detailed study of startup pro-
cedure, which would also involve
the power plant, has not been
made. However, it appears pos-
sible, with only density controls,
to start up the reactor when it
is attached to a simple external
system, - '

A useful amount of shim control
can, in principle, be obtained by
variation of water density 1in
the reactor. However, the effects
of density change upon the power
pattern (cf. 8 above) have not
vet been investigated, and 1t may
prove desirable to incorporate
mechanical or other slowly acting
controls to assist 1in startup or
in shim control.

RECOMMENDATIONS OF THE NDA STUDY

If the supercritical-water reactor

is deemed sufficiently promising to
be of further interest for aircraft
or other applications, the following
recommendations for further work are
in order: :

1.

Refinement of reactor and shield
design generally, ‘

Development of fuel assembly de-
sign and fabrication,

Out—of~pilé work on heat transfer,
corrosion, and fluid’f}ow, '

Inpile tests of corrosion, water
decomposition, structural materi-
als, and fuel element life in the
relevant range of neutron fluxes,
temperature, and pressure,

Experimental and theoretical work
on the géneral "flat flux" concept
of reactor design, '

 

 
 

[ 3]

Critical-assembly experiments for
the supercritical-water reactor
in particular, including such
items as fuel content, lumping
effects, power pattern, and the
effects of water-density variation
and nonuniformities,

Semi mockup measurements of over-
all shield effectiveness and of
the radiation heat load in the
neighborhood of the pressure shell,

Re finement of stability, control
and startup considerations, with

FOR PERIOD ENDING MARCH 10,1952

attention to the coupling between
reactor and power plant,

Consideration of a reduced-power
reactor experiment for studying
the dynamics and control of startup,
and also consideration of any
other conceivable experiments that
might shed light on these problems,
e.g., the coupling (via a water-
density sensing device) of an
electrically heated supercritical-
water system to a neutron simulator.

173

T

 
 

3

 

 

   
 

 

16 . ANALYTICAL CHEMISTRY

C. D. Susano
Analytical Chemistry Division

The major portion of the effort in
the field of Analytical Chemistry 1is
concerned with the analysis of ternary
and quaternary eutectics composed of
alkali-metal, beryllium, and uranium
fluorides that have been subjected to
corrosion tests 1n stalinless steel
and nickel alloy containers. Studies
made during the past quarter led to
revisions in the methods for the de-
termination of uranium, iron, nickel,
chromium, manganese, and molybdenum.
The revised methods are yielding satis-
factory results. The determinations
of beryllium, total alkali metal,
fluoride, and silicon present more
serious problems, but progress 1is
being made in the modification of
existing methods or the development
of new methods for determining these
constituents. The feasibility of
separating beryllium from uranium by
the use of anion exchange resins is
being studied, with very promising
results. A method has been worked out
for the determination of fluoride,
which, it is believed, will show a
marked improvement in accuracy and
expenditure of time over the presently
used pyrohydrolysis method. Efforts
are being made to adapt the colori-

metric silico-molybdate method for
‘the determination of silicon in fluo-

ride eutectlcs

Several further attempts to remove

‘boron from diatomaceous earth by

washing with hydrochloric acid have
shown that the boron content can be
reduced to 50 ppm or less.

: STUDIES OF DIATGMACEOUS EARTH

J. C. Whlte W, J Boss
Analytical Chemistry Division

Several attempts to remove boron
from one type of diatomaceous earth

(with the trade name "Sil-0-Cel") by
washing with hydrochloric acid in
concentrations from 33 to 67% resulted
in lowering the boron content from
300 ppm to 50 ppm or less. This in-
vestigation has been hampered by the
lack of a more sensitive method for
the determination of boron in quanti-
ties less than 50 ppm; therefore the
ultimately possible reduction of the
boron content is indeterminate at this
time.

A method for the determination of
boron that involves the color complex
between boron and 1,1'-dianthrimide in
sulfuric acid(!’ is more sensitive
than the spectrographic method being
used, but its application to diato-
maceous earth samples has not been
successful,

ANALYTICAL STUDIES OF
FLUORIDE EUTECTICS

J. C. White C. K. Talbott
W. J. Ross C. M. Boyd

Analytical Chemistry Division

The major portion of the effort for
the ANP program in the field of Ana-
lytical Chemistry 1s concerned with
the analysis of ternary and quaternary
eutectics composed of alkali-metal,
beryllium, and uranium fluorides that
have been subjected to corrosion tests
in stainless steel and nickel alloy
containers. Determinations of uranium,
beryllium, total alkali metals, and
fluorides are made for the purpose of
following possible changes in compo-
sition of the eutectics, and detérmina-
tions of impurities such as iron,

 

(I)J. C. White and W. J. Ross, “Studies of
Diatomaceous Earth,” Aircraft Nuclear Propulsion
Project Quarterly Progress Report for Period
Ending December 10, 1951, OBRNL-1170, p. 153.

175

T

T

T

 
 

 

 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

nickel, chromium, manganese, and molyb-
denum are required in order to de-

termine the extent of corrosion of
the metal container. Since the pres-

ence of silicon in these mixtures
has been detected by examination with
. the mass spectrometer, methods of
determining silicon are also being
deve loped. The adaptation of existing
methods of analysis to the fluoride
eutectic type of sample has been the
principal problem of the group.

Uranium. The determination of
uranium in fluoride eutectics is
being satisfactorily accomplished by
the potentiometric method of titrating
uranium(IV) with ferric sulfate. (2’

Béryllium."The separationtxfberyl-
lium from uranium is generally based
on the selective precipitation of one
component, such as the precipitation
of beryllium hydroxide in carbonate
solution. The determination of beryl-
liam is ordinarily completed gravi-
metrically by precipitation of the
hydroxide, but a volumetric procedure
‘would be preferable. The feasibility
of separating beryllium from uranium
by the use of anion exchange resins is
being studied. Both column and batch
tests, with the use of Dowex-1 1in the
chloride and sulfate forms, have in-
dicated that uranium is adsorbed by
such resins and that beryllium is not
adsorbed. Such a separation will
permit the direct titration of beryl-
lium.

Another method for the separation
of uranium and beryllium, recommended
by Fischer (reported by Rodden(?’),
depends upon the extraction of the
uranium salt of l-nitroso-2-naphthol
with amyl alcohol. This method will
be 1nvestigated.

 

'(2)C. J. Rodden (Editor-in-chief}, Analytical
Chemistry of the Manhattan Project, p. 70-71,
. McGraw-Hill, New York (1950).

176

‘Total Alkali Metals. In the pro-
cedure now in use, all the cations
except the alkali metal ions are pre-
cipitated by NH,OH from a sulfate
solution of the eutectic and evaporated,
and the alkali metals are weighed as
sulfates.

Total Fluoride. The determination
of total fluoride in the eutectics is
accomplished by hydrolysis of the
powdered eutectic in the presence of
U;04 and steam at 1000°C, followed by
the titration of the hydrogen fluoride
contained in the distillate.

The préblem of removing interfering

~ions 1s well on the way to being

solved by a method that involves fusion
of the material with sodium carbonate.
Leaching the carbonate fusion mixture
with water dissolves the fluorides,
which can then be separated from the
insoluble residue by filtration.
Beryllium, uranium dioxide, and all
the metal impurities except molybdenum,
generally a very minor component, are
retained in the residue. Fluoride 1is
precipitated in the filtrate as lead
chlorofluoride, which may be dried and
weighed. Alternatively, a volumetric
determination of chloride, after dis-

solution of the precipitate in nitric’

acid, also yields the fluoride content.
Recovery of fluoride was from 99.3 to
100.6% complete when the chloride
concentration ranged from two to five
times the theoretical amount. The
volumetric method is less satisfactory,
and further work is being done to
correct its deficiencies. =

The carbonate fusion method offers
two main advantages over the hydrolysis
procedure: (1) it is more adaptable
to use on a large scale, 1in that 1t
requires no special equipment and 1is
more rapid, and (2) it permits the
determination of a number of constitu-
ents on the same sample; in fact, when
necessary, all the desired constituents

~

e

 

 
 

 

R i i i

-

can be determined on a single sample
of the fused mixture.

Nickel. The dimethylglyoxime method
being used for the colorimetric de-
termination of nickel employs special
treatment for application to fluoride
eutectics.¢®

Chromium. The use of diphenyl-
carbazide to form a violet complex
with chromium(IV) for the colorimetric
determination of chromium 1is well
established; choice of a suitable
oxidant, unfortunately, is not. Per-
manganate has been used satisfactorily
in this capacity, but sodium bismuthate
has also been extensively studied.

Manganese. The colorimetric method
for the determination of manganese (*’
prescribes the use of sodium bismuthate
as the oxidant to produce permanganate
ions. This method, however, has
proved unsatisfactory for the corro-
sion-test samples currently being
received, because they have been found
to contain high chromium-to-manganese
ratios. As an alternative, the
periodate method has been studied and
found to be more applicable, owing
chiefly to the possibility of re-
tarding the oxidation of chromium
without affecting the oxidation rate
of manganese through the control of
the acidity of the system.

Silicon. In the colorimetric de-
termination of silicon, the complex
molybdosilicic acid is extracted with
either butyl or amyl alcohol, and the
molybdenum blue complex is measured
spectrophotometrically. Although this

‘method yields results that are some-

times difficult to duplicate, it is

 

(3)From J. C. White to L. J. Brady, Effect
of pH and Concentration of Ammonium Persulfate on
Colorimetric Determination of Nickel by the
Dimethylglyoxime Method, Y-B31-324, Jan. 17,
1952. _

'(4)1. C. White, *Study and Analysis of
Fluoride Eutectics,” Analytical Chemistry Division

Quarterly Progress Report for Period Ending
December 26, 1951, ORNL-1233, p. 62.

FOR PERIOD ENDING MARCH 10,1952

widely used for the determination of
soluble silicon in low concentration
ranges.

Dissolution of the fluoride eutectic
to yield all the silicon in the form
of soluble silica is a critical step.
Hoffman and Lundell(®’ made a thorough
study of the determination of silicon
in the presence of fluoride and recom-
mended sodium carbonate fusion of the
solid. Small concentrations of sili-
con can be leached out of the fusion
mixture with water. The use of this
method in preliminary tests 1indicated
the presence of substantial quantities
of silicon 1n certain fluoride eutec-
tics. Aneffort 1s being made to adapt
the method for application to the
fluoride eutectics and to the indi-
vidual components of the eutectics.

SOLUBILITY OF BORIC ACID IN WATER

A. P. Fraas
ANP Division

G. F. Petersen

The solubility in water of the
boric acid, H;BO;, is so low that the
increase in the water density 1is
slight. However, by adding 11.85 g
of lithium hydroxide to 100 ml of
water, 1t was possible to effectively
dissolve 66 g of boric acid per 100
ml of water, with a resulting density

of 1.2 g/cc. This method has been
considered as a device for reducing
shield weight.

CLARITY OF BORATED WATER

H. P. House
Analytical Chemistry Division

Little work was carried out on this
project during the quarter, and no
further work is planned because the
use of borated water is no longer
contemplated in the ARE.

 

(5) ;. 1. Hoffmen and G. E. F. Lundell, “Determi-
nation of Fluorine and of Silica in Glasses and
Enamels Containing Fluorine,”” RP110, J. Research
Nat. Bur. Standards 3, 583 {(June 1929),.

177

T

AT EC T TR R T

T TR BT IR T e

 
 

 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

STUDIES OF ALKALI AND ALKALINE
EARTH HYDROXIDE COOLANTS

J. C. White

Analytical Chemistry Division

No work has been done on analytical
methods applicable to alkali and alka-
line earth hydroxide coolants this
quarter because of the increased em-
phasis on the study of fluoride
eutectics. '

ANALYTICAL SERVICES

H, P. House L. J. Brady
: J. W. Robinson

AnélYfidal Chemistry Division

Analytical service work is now
centralized in the ANP analytical
Chemistry Laboratory in Building

178

9201-2. Over 70% of the samples.

analyzed this quarter were from reactor
chemistry studies of reactor fuels and
coolants, Emphasis shifted from
alkali and alkaline earth hydroxides
to fluorides, uranium, alkali metals,
and corrosion products in various
fluoride salt mixtures. Other determi-
nations, such as the purity of salt
mixtures, oxygen in helium and argon,
and 1ron in beryllium oxide, were also

‘made. A summary of service analyses

performed is shown in Table 39.
TABLE 39

Summary of Service Analyses

Samples on hand 11/2/51 149
Number of samples recéived' 638
Total number of samples - 787
Number of samples reported 574
Backlog as of 2/2/52 213

 

 

 

AT A et e ey
 

 

REPORT NO.

Y-F15-10

ORNL-1177

OBNL- 1255

Y-F26-29

ANP-66
Y-F10-75
Y-F10-76
dRNL~1176

Y-F10-77

Y-F10-81
Y-F10-85
Y-F10-63

Y-F10-78
Y-F10-82

Y-F10-89

FOR PERIOD ENDING MARCH 10, 1952

17. LIST OF REPORTS ISSUED

TITLE OF REPORT

Design

Design Study of Closely-Coupled Reactor-
Heat-Exchanger-Shield Combinations for
Use with a Fused Fluoride Circulating
Fuel

The Supercritical Water Reactor
Basic Performance Characteristics of the
Steam Turbine-Compressor-Jet Aircraft

Propulsion Cycle

Regarding Homogeneous Aircraft Reactors

Reactor Physics

Some Results of Criticality Calculations
on BeQO and Be Moderated Reactors

Some Results of Kinetic Studies on the ARE
Design of 10 June 1951

The Circulating Fuel ARE Core Series of
December 13, 1951

Criticality Calculations for Hydrogen
Moderated Reactors from Microscopic Data

Physics of the Aircraft Reactor Experiment

Statics of the ANP Reactor - A Preliminary
Report

Effect of Potassium ih fihe Fue]-Coolafit
Solution in Two ARE Reactors

A Flux Transient Due To a Positive

Reactivity Coefficient
Water Moderated Reactors
The ARE with Circulating Fuel-Coolant

Effect of Structure on Criticality of the
ARE of January 22, 1952

AUTHOR(S) DATE ISSUED

{(to be issued)

A. P. Fraas

G. F. Wislicenus

2-1-52

Nuclear Development
Associates, Inc.

{to be issued)

©

H. Cohen
A. P. Fraas

W. B. Cottrell 1-29-52
C. B. Mills

J. W. Webster 10-15-51
0. A. Schulze -
J. W. Webster 10-19-51
M. J. Nielsen

C. B. Mills 12-27-51
M. Shapior 11-29-51
S. Preiser

C. B. Mills 1-5-52
C. B. Mills 1-8-52
C. B. Mills 1-21-52
C. B. Mills 1-14-52
C. B. Mills 1-7-52
C. B. Mills 1-11-52
C. B. Mills 1-29-52

179

 
 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

REPORT NO.

Y-F10-79

CF-52-1-2

CF-52-1-3

CF-52-1-5

CF-52-1-4

CF-52-1-76

CF-51-11-195

CF-51-11-198
CF-51-12-91

Y-F30-5

Y-F30-6

180

TITLE OF REPORT

Effect of the Delayed Neutrons on the
Kinetic Response of the S;atidnary Liquid
Fuel ARE l

Theoretical and Experimental Analyses of
Natural Convection Within Fluids in which
Heat is being Generated. Part III. Heat

Transfer from a Fluid in Laminar Flow to

the Walls to a Cylindrical Tube: A
Simplified Velocity Distribution was
Postulated

Theoretical and Experimental Analyses of
Natural Convection Within Fluids in which
Heat is being Generated. Part IV. Heat
Transfer from a Fluid in Turbulent Flow
to Two Parallel Plane Bounding Walls

Theoretical and Experimental Analyses of
Natural Convection Within Fluids in
which Heat is being Generated. Part V.
Natural Convection Velocity Measurement
in a Parallel Plane System

" Theoretical and Experimental Ana]&ses of

Natural Convection Within Fluids in which
Heat is being Generated. Part VI. Heat
Transfer from a Fluid in Turbulent Flow
to the Walls of a Cylindrical Tube

Heat Transfer in Nuclear Reactors

Physical Properties

Heat Capacities

Density of Fulinak

Density of Flinak

Physical Property Charts of Some Reactor
Coolants, Fuels, and Miscellaneous
Materials

Measurements of the Viscosity of Flinak

Viscosity Measurements of Flinak by the
Falling-Ball Viscometer

AUTHOR( S}

C. B. Mills

. C. Hamilton
R. F. Redmond
L. D. Palmer

o
o

H. F. Poppendiek
Hamilton’

L. D. Palmer

o
0

R. F. BRedmond

D. C. Hamilton
. Poppendiek
L. D. Palmer

i
3

=
=

. Lyon
¥W. B. Harrison

W. D. Powers
G. Blalock
J. M. Cisar
J. M. Cisar
M. Tobiasr

H. F. Poppendiek

,M' Tobias

S. I. Kaplan

DATE ISSUED

1-7-52

1-11-52

{to be issued)

2-12-52

(to be 1issued)

No &ate

11-30-51

11-30-51
12-14-51

1-28-52

2-26-52

(to be issued)

o

{1

b

 

 

 

 

 

 

 

 

 
 

 

REPORT NO.

" ORNL-1131
Y-F10-86

 Y-F10-87

Y-F10-91

Y-F5-73
ORNL- 1046

CF-51-8-290

CF-51-12-185
Y-F5-75

ORNL-1142

CF-52-3-1

ORNL-1217

CF-51-12-70

" FOR PERIOD ENDING MARCH

TITLE OF REPORT

The Thermal Conductivity of Molybdenum
over théKTemperapure Range 1000-2000°F

Critical Masses of Some Alkali-Hydroxide
Moderated Reactors

The Power Géneration in a B4C Curtain ARE
" on One End of the ARE No. 1 Reactor

ARE with Ffie]-Cdolant in the Reflector

Shielding Research

Notes on the Ideal Unft Shie}d

A Nuclear Plate Camera for Fast Neutron
Spectroscopy at the Bulk Shielding '
Facility

An Estimete of the Photoneutron Flux in

the Water Surrounding the Bulk Shielding
Facility Reactor - '

Estimate of Background at Tower Shielding

 Facility

Rough Notes From Selected Shielding
Literature .

Multiple-Crystal Gamma -Ray Spectrometer

Gamma-Ray Spectra from the Divided Shield
Part I

Neutron Attenuation Through Air-Filled
Ducts in Water

Heat Transfer Research

Theoretical and Experimental Analyses of
Natural Convection Within Fluids in which
Part I. Heat
Transfer from a Fluid in Laminar Flow to
Two Parallel Plane Bounding Walls: A
Simplified Velocity Distribution was
Postulated

‘Heat is being Generated.

AUTHOR(S)
E. P. Mikol
C. B. Mills
C. B. Mills
C. B. Mills
C. B. Ellis
L. A. Wills
J. L. Meem
E. B. Johnson
J. M. LaRue
G. P. Letz
T. J. Morley
-J. N. Renaker
A. Simon
C. B. Ellis

F. C. Maienschein

F. C. Maienschein

C. E. Clifford, et al.

D. C. Hamilton
. Poppendiek
L. D. Palmer

=
g

10, 1952

DATE ISSUED

1-21-52

1-21-52

2-27-52
7-10-51
1-11-52
8-24-51

12-17-51

12-5-51

(to be issued)

3-3-52

(to be issued)

12-18-51

181

 
{ ANP DIVISION QUARTERLY PROGRESS REPORT

 

 

REPORT NO. TITLE OF REPORT AUTHOR(S) DATE ISSUED
Addendum to Classified Addendum to Unclassified.Memb ~ D. C. Hamilton - ! 12-19-51
CF-51-12-70 CF-51-12-70 on Natural Convection in H. F. Poppendiek »
Confined Spaces | L. D. Palmer '
CF-52-1-1 Theoretical and Experimental Analyses of R. F. Redmond 1-22-51
Natural Convection Within Fluids in which D. C. Hamilton '
~ Heat is being Generated. Part II. Heat L. D. Palmer

Transfer from a Fluid in Laminar Flow to
Two Parallel Bounding Walls

 

 

 

 

 

 

 

Chemistry
Y-F35-2 The Vapor Pressure of Some Salts II. Z. fur H. v. Wartenberg 1-21-52
Electrochemie 2%, 568-573 (1921) (translated H. Schulze
by Mary E. Lee)
‘ Y-B31-324 Effect of pH and Concentration of Ammonium J. C. White 1-17-52
f ‘ Persul fate on Colorimetric Determination .
of Nickel by Dimethylglyoxime Method-
Y-F35-1- Concerning the Thermal Behavior of Sodium E. G. Bunzel 1-3-52 sy
: ~ Compounds, Specifically of Sodium Oxide F. J. Kohlmeyer :
and Sodium Sulfide and Their Reactions s
with Metals, Z. fur Anorg. Chem. 254, '
1-30 (1947) (translated by Mary E. Lee)
;: Y-F35-3 The Phase Diagrams for the Systems KF-MgF2 H. Remy . 1-23-52V
1 and RbF2. Rec. Trav. Chim. 59, 516-525 W, Seeman
] {1940) (translated by Mary E. Lee)
% | .
' Metallurgy and Ceramics
; CF-51-11-186 Operation of a Ni-NaOH Thermal Convection Metallurgy Division 711;29-51 r
; Loop
Y-B32-88 Changes in Melting Point of Fluorides G. J. Nessle | - 2-12-52
During Corrosion Tests C. J. Barton
CF-52-1-144 Ceramic Materials as Related to the Reactor J. R. Johnson - 1-18-52
Program
CF-52-1-192 Safety Rods for the ARE : W. D. Manly 1-29-52
' E. S. Bomar ¢
Miscellaneous “ b
Y-F26-31 ANP Information Meeting of February 20, 1952 W. B. Cottrell | 2-28-52
Y-¥26-23 ANP Ihformation Meeting of November 14, 1951 W, B. Cottrell 11-21-51

182

  
 

FOR PERIOD ENDING MARCH 10, 1952

‘ 1
18. DIRECTORY OF ACTIVE ANP RESEARCH PROJECTS AT ORNL()

March 1, 1952

I. REACTOR AND COMPONENT DESIGN

A. Aircraft Reactor Design

1.
2.

B. ARE
1.
2.

Survey of Cirdu]ating-Fue] Cycle
Survey of Air-Water Cycle (G.E.)

Reactor Design
Core and Pressure Shell

Fluid Circuit Design

Pressure and Flow Instrumentation
Structural] Analysis

Thermodynamic and Hydrodynamic Analysis

Remote Handling Equipment
Shielding Studies for the ARE

Flectrical Power Circuits

Control Studies
High-Temperature Fission Chamber
Control System Design

Control Rod Design ~

Buildifié

Cafiséruétion.

Internal Design

E. Reaétbr Stéticsr N

1. :Sia@iés of ghg Circulatihg-FhéfVRéactors

.‘2:

'Péféfietfié&8£fidiéé SfHH;O:Mbdéféted.Circfi]atifig-r |
- fFue] Bgacpor | ,

. Statics of the-CfItiéalcEiperiméhfs'

. Statics of the NaOH Moderated Reactor

Préparation'of‘Reactor Calculations and Cylindrical

Coordinates for the IBM

9704-1
9204-1

9201-3
9201-3

6201-3
9201-3

9201-3
9204-1

9201-3
9201-3
1000

2005
2005
9201-3

7503
1000

9704-1
o704-1

9704-1
9704-1

9704-1

Fraas

Lane, Noderer

Hemphill, Wesson

Cristy, Lawrence,
Jackson, Eckerd,
Scott

Hluchan

Maxwell, Walker

Lubarsky, Greenstreet -

Longyear, Palmer
Hutto, Alexander
Enlund
Walker

Hanauer
Epler, Kitchen, Ruble

Estabrook

Nicholson Company

Browning

" Mills, Edmonson,
- - Uffelman, Johnson

Mi]ié, Edmonson,

Uffelman, Johnson

Mills, Holmes

Mills, Edmonson,

Uffelman, Johnson

Edmonson

-(1)This directory was previously issued as Directory of Active ANP Research Projects ot ORNL, by W. B.

Cottrell,

Y-F26-30, March 1, 1952,

183

 
ANP DIVISION QUARTERLY PROGRESS REPORT |

6. Problem of Minimum Critical Mass ' - 9704-1 = Coveyou

 

 

; 7. Efiiggioggstrxbut1on of Therma] and Eplthermal 9704i1 ‘ Coveyou ¢
8. IBM Calculatlons for the ORNL ARE Proposals '  9704-1 Mills, Edmonson . :
9. 1IBM Calcu]atlons for(}EZBeactors (G E.) o 9704-1 Leeth (G.E.), Johnson 7
10. Age Calculatlons of Hydrogen Moderated Beactor (G E.) 9704;1 7 Macauley, Leeth (G.E.)
11. Temperature Coeff1c1ents of Reactivity 9704-1 ~ Osborne

12. Effect of Voids o 9704-1 Tamor

F. Reactor Dynamics

1. Kinetics of the Circulating-Fue]-Béactof N 9704-1 Ergen, Prohammer,
' : ' ' , , ' Thompson, Coveyou

G. Critical Experiments _
1. ARE Critical Assembly o 9213 | Callihan, Zimmerman,

Keen, Williams, Haake,
Scott, Kennedy (P&W)

 

bk il s

2. Air-Water Reactor Critical Assémb]y (G.E.). T 9213 ‘Same as ébove_

H. Pump Development‘

 

 

 

 

1. Gas-Sea]‘Cehtriffigal Pump | - | 9201-3 McDonald, Cogb
2. Frozen-Seal Pump for Sodium and Fluorides 9201-3  Cobb, Huntley, Smith, i
: : ‘ ' ; - ‘ . Taylor
3. ARE Pump Design and Development | o 19201-3  Cobb, Grindell, Taylor,
. ‘ ) : : S Southern _
4. E]ecnromagnetic Pump ' | 9201-3 McDonald, Southern, Wyld
5. Canned-Rotor Pump - 9204-1 Richardson
6. Rocking-Channel Sealless Pump ' S _ BMI Dayton
7. Seals for NaOH Systems | | BMI Simons, Allen
- 8. .Seals for High-Tbmperafiure Systems | ' 9201-3 CoBb,'Johnson,'Sfiith,
: . . : oo : A ‘ 7 : Tay]or, Grindell
1 I. Valve Development |
j | 1. Self-Welding Tests ' ' 1 9201-3 Adamson, Petersen, Reber
2; Bellows Tests at High Temperatures | - | 9201-3 Johnson '
3. Valves for High-Temperature Systems o 9201-3  Cobb, Ward, Johnson

J. Heat Exchanger and Radlator Development

 

1. Nak- to-NaK Heat Exchanger o | . 9201-3 Fraas, Wyld, LaVerne,
: - Petersen
2. Sodium-to-Air Radiator : A 9201-3 Fraas, LaVerne,
| : . , Whitman, Petersen
3. Boeing Turbojet with Na Radiator ) " 9201-3 Fraas, LaVerne
4. Fuel-to-Liquid Heat Exchanger , ' . 9201-3 Fraas, Whitman,
' ' ‘ . ' i ) McDonald, Salmon,
Bailey

184

  
 

ks

FOR PERIOD ENDING MARCH 10, 1952

5. Fuel-to-Gas Radiator

6. Radiator and Heat Exchanger Design Studies

Instrumentation

1. Heater Development and Application

2. Development of High-Temperature Measurement
Techniques \

3. Devejopment of Flow and Pressure Measuring
Devices for Fluoride Systems

4. Leak Detection for Fluoride Systems

5. Monitoring Equipment for Sodium Leaks

6. Development of High-Temperature Strain Gage

9201-3

9201-3

9201-3
9201-3

9201-3

5201-3
K-1005

Cornell
Aero.

1X. SHIELDING RESEARCH

Cross-Section Measurements

1.

B W N

5
6.
7
8

Neutron Velécity Selector
Analysis for He in Irradiated Be
Total Cross Sections of N'* and o' (G.E.)

Elastic Scattering Differential Cross Section
of N** and 0% (G.E.)

. . 6 .
Total Cross Sections Of.Ll , L17, Beg, BiO, B12, cl2
Fission Cross Sections
Cross Sections of Be and C

Inelastic Scattering Energy Levels

shielding Measurements

1.
2.

~NLON U e

. Na Bremsstrahlung Measurement

Divided-Shield Mockup Tests (G.E.)
-Bulk.Shiéfding Reactor Power Calibration
Bulk Shiejdiné Reactor Operation

‘Heat Release per Fission

VMetaf_Ducfi.Tésfis

Air Duct Tests (G.E.)

‘Shielding Theory and Calculations
1. Survey |

Survey Report on Shielding
Shielding Section for Reactor Technol ogy

Calculations of Removal Cross Sections

Theory of Neutron Traasmission in Water

2005
3026
9201-2
9201-2

9201-2
9201-2
3001

9201-2

3010
3010
3010

3010

3001

3025

3001

3022,
9204-1

3022
3022
3022

McDonald, Bailey,
Salmon, Whitman,
Tunnell

Fraas, Bailey, Salmon,
Whitman

McDonald, Affel
McDonald, Affel

Cobb, Southern, Taylor

Cobb, Southern, Taylor
Cameron, McKown

Puffer, Grey, Donovan

Pawlicki, Smith
Parker
Willard, Bair, Johnson

Willard, Johnson, Bair

Johnson, Willard, Bair
Lamphere, Willard

Clifford, Flynn, Blosser

Willard, Bair, Kington

Meem and group
Johnson, McCammon

Holland, Leslie,
Roseberry

Meem

Hullings, Hull

Sisman

Clifford, Flynn, Blosser

Blizard, Welton

Blizard
Blizard
Blizard, Enlund

185

 
i e 'i.‘i’i;.;i S

Sk ik

i i st

 

 

ANP DIVISION QUARTERLY PROGRESS REPORT

3. Interfiretation of Pb-HZO Lid Tank Data
6. Divided-Shield Theory and Design

7. Air Duct Theory (G.E.)
8. Shielding Section for Reactor Handbook

9. Cbnsu]tation on Radiation Hazards (G.E.)

D. Shielding Instruments
1. Gamma Scintillation Spectrometer
2. Neutron Dosimetgr'Deve]opment

3. Proton Recoil Spectrometer for Neutrons

4. He3 Counter for Neutrons
5. LiI Crystals for Neutrons

6. Neufirqn Spectroscopy with Photographic Plates

" E. Shielding Materials

1. Preparation of High-Hydrogen Rubber

2. Development of Hydridés for Shields

2005
NDA

3001
3022

2001

3010
3010
3010

3010
3010
3006

Goodrich
Company

MHI

II1. MATERIALS RESEARCH

A. Liquid Fuel Chemistry
1. Phase Equilibrium Studies of Fluorides

2. Preparation of Standard Fuel Samples
3. Special Methods of Fue]-Purification

4. Evaluétion of Fuel Purity

5. Thermodyhamic Stability and Electrochemical
Properties of Fuel Mixtures -

6. Hydro]ysis and Oxidation of Fuel Mixtures

7. Simulated Fuel for Critical Experiment

8. Stability of Slurries of U0, in NaOH

9. Phase Equilibria Among Silicates, Borates, etc.
10. Fuel Mixtures Containing Hydrides

11. Chemicél Literature Searches

12, So]utjon of Metals in Their Halides

B. Liquid Moderator Chemistry

1. Preparation and Evaluation of Pure Hydroxides

186

9733-3
9733-3
9733-3

9733-3
9733-3

9733-3
9733-3
BMI
BMI
MHT
9704-1
3550

9733-3

' Maiefiéchein, Schenck

Johnson, Haydon

Overholser, Cuneo

Simon

w

Goldstein, Feshback,
Stern

Clifford, Simon
Blizard, Hungerford,
Simon, Ritchie, Meem, i
Lansing, Cochran, B

Maienschein
Morgan

Maienscheih
Hurst, Glass, Cochran

Cochran, Henry,
Hungerford

Cochran

 

 

Davidson . !

Banus

 

Barton, Bratcher,
Traber

Nessle, Truitt, Morgan,
Love

Grimés, Blankenship,
Nessle

Overholser, Sturm

Overholser, Tofiol

 

B]ankenship; Metcalf

Patterson

Crooks
Banus

 

Lee

Bredig, Johnson,
Bronstein

Overholser, Ketchen

 
 

*,

2.

3.

FOR PERIOD ENDING MARCH 10, 1952

Electrochemical Behavior of Metal Oxides in
Mol ten Hydroxides

Moderator Systems Containing Hydrides

Corrosion by Liquid Metals

1.

w

4
5
6.
7
8

Static Corrosion Tests in Liquid Metals and Their
Alloys

Dynamic Corrosion Research in Harps

Effect of Crystal Orientation on Corrosion

Effect of Carbides on Liquid Metal Corrosion
Mass Transfer in Molten Metals

Diffusion of Molten Media into Solid Metals
Structure of Liquid Pb and Bi

Alloys, Mixtures, and Combustion of Liquid Sodium

Corrosion by Fluorides

1.

o -~ O\ W b
. . . - .

10.

11.

Static Corrosion of Metals and Alloys in Fluoride
Salts

Isothermal Static Corrosion Tests in Fluoride Salts
Fluoride Corrosion in Small-Scale Dynamic Systems

Dynamic Corrosion Tests of Fluoride Salts
Corrosion of Meta]s.and Their Oxides in HF
Magnetic Susceptibility Due To Fluoride Corrosion
Fluoride Corrosion in Rotating Cylinder Apparatus

Reaction of Metals with Fluorides and Contaminants

Loop Tests.with Fluorides

_ Dynamic'Cbrrosion Testing of Fluorides

'Equlllbrla Between Electropositive and Transition

Meta]s in Halide Melts

Corr051on by Hydrox1des

1.

2,

Statlc CorrOS1on of Meta]s and A]ons in Hydrox1des

Mass Transfer 1n Molten Hydr0x1des

'Ph s1céj'CHéhisbry of the Hydroxide Corrosion
P enomenon

Statlc Corr081on by Hydrox1des

Static Corr031on by Hydroxides

" Physical Properties of Materials
1.
2.

Density of Liquids

Viscosity of Liquids.

9733-2

 MHI

2000

9201-3
2000
2000
2000
2000

2000
2000

2000
9766
9766

9201-3
9766

9201-3
9201-3
9733-3

2000
2000

3550

2000
2000

2000

9766

9204-1
9204-1

Bolomey, Nichols

Banus

Vreeland, Day,
Hoffman

Adamson, Reber

Smith, Cathcart,
Bridges

Brasunas, Richardson
Brasunas, Richardson
Richardson

Smith

Bridges, Smith

Vreeland, Day, Hoffman

‘Kertesz, Buttram, Smith,

Meadows

Kertesz, Buttram, Smith,
Meadows

Adamsén, Reber

Kertesz, Buttram, Croft
Tunnell

Tunnel ]

Overholser, Redman,
Powers

Smith, Cathcart, Bridges
Brasunas, Richardson

Bredig, Johnson,
Bronstein

Vreeland; Day; Hoffman

Brasunas, Richardson,
Smith, Cathcalt

Cathcart, Sm1th

Kertesz, Croft

Jaffee, Craighead

Cisar, Kap!lan

Tobias, Cisar, Kaplan,
Jones

187

 
ANP DIVISION QUARTERLY PROGRESS REPORT

 

3. Thermal Conductivity of Solids 9204-1 Powers, Burnett
i 4. Thermal Conduct1V1ty of L1qu1ds 9204-1 c1a{boffié, Cooper e -
i 5. Specific Heat of Solids and Liquids 92041 Powers, Blalock E
| 6. Thermal Diffusivity 9204-1 - Tobias S | t,?

7. Electrical Resistance of Fluoride Salts 9201-3 Affel - | k E
| 8. Viscosity of Fluoride Fuel Mixtures 9766 Kertesz, Knox | %
é 9. Vapor Pressure of Fluoride Fuels | 9733-2 : Barton,‘Moore |

10. Vapor Pressure of BeF, BMI Patterson, C]egg

G..'HéatlTranofer .

 

1. Connection in Liquid Fné] Elements | 9204-1 | Hamil ton, Redmond,
- _ o ‘ . ' Lynch

2. Heat Transfer in Circulating-Fuel Reactor 9204-1 Poppendiek, Pa]mén
3. Heat Transfer Coefficients of Fluorides and 9204-1 Hof fman, Lones

; Hydroxides

i 4. Heat Transfef Coefficient of Lithium 9204-1 'C]aiborné, Winn

; 5. BoilingrLiQuid Metal Heat Tranoferr o ' ' 9204;1‘ Farmer. o

! 6. Sodium Heat Transfer Coefficients in Short Tubes 9204-1 Harrison

’ 7  Heat Transfer in Special Reactor Geometrles 9204-1 Claiborne

_ H. Flnoride Haqdling

 

 

1. Fluoride Production | © 9201-3 'Whito Courtney, Mann
é.‘ Deaién of Spacial Fluoride Handling Equipment 9733-3 Grimes, Blankenshlp,
B D _ - y I Nessle
3. Prétroatment.of Fluoride-Containing Systems 9201-3 Mann, Wischhusen
4. Inspection of Components of Fluoride Systems 9201-3 Reber, Mann
5. Filling Techniques for Fluoride Systems 9201-3 Mann, Wischhusen
6. _Protroaping“of Fluoride Systems 9201-3 McDonald, Affel B
,'7. Decontamination of Fluoride Systems ) 9901-3 .Mann, Wlschhusen | i
:“8. >F]uor1de Salvage and Disposal o ' 9201-3 ' Mann, Courtney, wnitex' f
1. L1qu1d Metal Handllng ’ '_ i
: 1. Equlpment Cleaning Techniques 9201-3 | Mann E
E 2. Continuoua.and Batch Sodium.éurification | | 9201-3 Mann h ;
% 3. Samp]inéjTeohniques | 9201;3 | Mann, Blake]y E
4. Sodium Vapor Trapping 9201-3 Mann |
5. xLiquidlMooa] Salvage and Disposal ' 9201-3 _ Deven1sh Mann
6. L1qu1d Meta] Safety Equlpment 9201-3 ereven1sh Mann
7. B]anket Gas Purlflcatlon 9201;3 | Mann

J. Dynamic Liquid Loops
1. Operationrof Convection Loops 9201-3 Adamson; Reber

2. Opefation.of-Figure-Eight Loops 9201-3 Cough]en

 

188

 
 

"y

FOR PERIOD ENDING MARCH 10, 1952

3. UOa;NaOH Slurry Loop

4. Operation of Thermal Convection Loops

K. Materials Analysis and Inspection Methods

1. Analysis of Fluorides for Metallic Corrosion

Products
2. Analysis of Alkali Fluoridé Eutectics
3. Trace Impurities in Alkali Fluorides

Determination and Removal of Boron in Diatomaceous

4.
Earth
5. Metallic and Oxide Impufities in Alkali Hydroxides
6. Chemical Methods of Fluid Handling
7. Meta]]égraphié Ekafiinations
8. Identification of Compounds in Solidified Fuels
9. Preparat1on of Tested Spec1mens for Examination

- 10. Identification of Corrosion Products from Dynamic
Loops

11. Assembly and Interpretation of Corrosion Data from
Dynamic Loop Tests

. L. Radiation Damage

1. Liquid Compound Irradiations in LITR

2. Liquid Compound Irradiations in Cyclotron

3. Liquid Compound Irradiations in MTR
4. Fluoride Fuel Irradiation in Berkeley Cyclotron

5. LiquidrMeta] Corrosion in X-10 Graphite Pile Loops
6. Stress Corrosion and Creep in LITR Loops

7; Creep of Metals in X-10 Graphite'Pi]e and LITR

8. Thermal Conduct1v1ty of Metals in X-10 Graphite
Pile and LITR

9, Diffusion of Fission Products from Fuels

10. Neutron Séeégfufi'of LITR
11. Irradiation of Water (G.E.)
M. Strength of Materials

1, Creep Tests in Fluoride Fuels

2., Creep and Stress-Rupture Tests of Metals in
Vacuum and in Fluid Media

BMI
2000

9201-2

9201-2
9201-2
9201-2

9201-2
9201-3
2000

9733-2
9733-2
9733-3

9733-3
9201-3

3005

9201-3

3025
NAA
3001

3005
3001

3025

3001,
3005

3001

3005
3550

9201-3
2000

Simons

Cathcart, Bridges,
Smith '

White, Talbott

.White, Boyd, Ross

White

White, Ross

Whi#e

Mann, B]akeiy

Gray, Krofise, Roeche
Barton, Anderson
Nessle, Truitt, Didlake
Hoffman, Blankenship

Blankenship, Blakely
Adamson

Keilholtz, Morgan,
Webster, Robertson,

~Klein, Kinyon

Keilholtz, Feldman,
Sturm, Jones

Keilholtz, Klein, Kinyon

Pearlman

Sisman, Bauman, Carroll,
Brundage, Parkinson,
Ellis, Olsen

Sisman, Bauman, Carroll,
Brundage, Parkinson,
Ellis, 0]sen

Wilson, Zukas, Davis
Cohen, Templeton

Keilholtz, Morgan,
Webster, Robertson,
Klein, Klnyon

Sisman, Trice, Lewis

Taylor

Adamson, Reber

Oliver, Woods, Weaver

189

e i b

 
. ANP DIVISION QUARTERLY PROGRESS REPORT

 

 

 

 

 

3. High-Temperature Cyclic Tensile Tests ' 2000 ' Olivef, Woods, Weaver
4. Tube Burst Tests 9201-3 ' Adamson, Reber 7 ¥
5. Tube Burst Tests 2000 Oliver, Woods, Weaver
6. Relaxation Tests of Reactor Materials 2000 "QOliver, Woods, Weaver -
7. Creep Tests of Reactor Materials (G.E.) 2000 'O]iver, Woods,:Weaver
N. Metals Fabrication Methods
1. Welding Techniques for ARE Parts 2000 . Patriarca, Slaughter
2. Brazing Techniques for ARE Parts 2000 Patriarca, Slaughter
3. Molybdenum Welding Research | BMI Parke
4. Molybdenum Welding Research ' MIT Wul ff
5. Resistance Welding for Mo and Clad Metals RPI Nippes,.Safage
6 Welds in the Presence of Various Corrosion Media 2000 Vréélénd, Patriarca,
_ _ Slaughter
7. Nondestructive Testing of Tube-to-Header Welds 2000 Patriarca, Slaughter
8. Basic Evaluation of Weld-Metal Deposits in Thick 2000 Patriarca, Slaughter
Plates
, 9. Evaluation of the Cone-Arc Welding Téchnifiue 2000 Patriarca, Slaughter E
; 10. Development of High-Temperature Brazing Alloys Wall-  Peaslee *
. , Colmonoy »
11. Evaluation of the High-Temperature Brazing Alloys 2000 Patriarca, Slaughter _ &
0. New Metals Development
1. Mo and Cb Alloy Studies | 2000 Bomar, Coobs
2. Heat Treatment of Metals 2000 Bomar, Coobs
P. Solid Fuel Element Fabrication
1. Solid Fuel Element Fabrication ' 2000‘ Bomar, Coobs
2. Diffusiofi—Corfosion in Solid Fuél Elements 2000 Bomar, Coobs
3. Determination of the Engineering Properties of 2000 Bomar, Coobs
: Solid Fuel Elements
; 4. Electroforming Fuel-Tube to Header Configurations Gerity Graaf
; , Michigan
5. Electroplating Mo and Cb Gerity Graaf
Michigan ‘
6. Carbonyl Plating of Mo and Cb 2000 Bomar 5
7. Rdlling”of Ffie] Plate Laminates (G.E.) 3012, Bomar, Cunningham, -
: 2000 Leonard :
L o ! ' ?
4 Q. Ceramics and Metals Ceramics |
; 1. BeQ Fabrication Research Gerity Graaf %
: : : Michigan
- 2. Metal Cladding for Be0 ' ' Gerity Graaf
; , Michigan _
? 3. 7B4C Control Rod Development | 2000 Bomar, Coobs
: 190
}

 

 
 

 

~

FOR PERIOD ENDING MARCH 10, 1952

4. Hot Pressing of Tungsten Carbide Bearings 2000 Bomar, Coobs
5. High-Temperature Firing of Uranium Oxide to 2000 Bomar, Coobs
Produce Selective Power Sizes - : ‘
6. Development of Cr-UO2 Cermets for Fuel Elements 9766 Johnson, Shevlin
7. Ceramic Coatings for Stainless Steel 9766 White
8. Ceramic Valve Parts for Liquid:Metals and Fluorides 0.S.U. Shev]in
9. Application of Ceramic Materials to Reactors 9766 Johnson
10. Crucible Development for High Temperatures Norris Wilson, Doney
¢ Electric
Lab.
IV. TECHNICAL ADMINISTRATION OF AIRCRAFT NUCLEAR PROPULSION PROJECT
AT OAK RIDGE NATIONAL LABORATORY
PROJECT DIRECTOR R. C. Briant*
ASSISTANT DIRECTOR FOR COORDiNATION A, J. Miller*
ASSOCIATE DIRECTOB FOR ARE J. H. Buck
Assistant for Coordination B. T. Macauley
Administrative Assistant L. M. Cook
Project Editor W. B. Cottrell

PROJECT DIRECTORY
SECTION NUMBER

STAFF ASSISTANT FOR PHYSICS

W. K. Ergen*

Shielding Research E. P. Blizard IT B,C,D,E

Reactor Physics W. K. Ergen* I E,F

Critical Experiments A. D. Callihan I G

Nuclear Measurements A. H. Snell 11 A
STAFF ASSISTANT FOR RADIATION DAMAGE A, J. Miller*

Radiation Damage D. S. Billington III L
STAFF ASSISTANT FOR GENERAL DESIGN A. P. Fraas*

General Design A. P. Fraas* I A
STAFF ASSISTANT FOR ARE E. S. Bettis*

ARE Design R. W. Schroeder I B,D

Reactor Control E. S. Bettis* 1 C

Special ARE Projects ' unassigned I K

STAFF ASSISTANT FOR ENGINEERING RESEARCH R. C. Briant*

Experimental Engineering H., W. Savage I H,1,J,K,
11T D,H,I,K
Heat Transfer Research H. F. Poppendiek III F,G

191

 
. ol s, e it diibkie i iy

 

 

Ceramics
STAFF ASSISTANT FOR METALLURGY
Metallurgy

STAFF ASSISTANT FOR CHEMISTRY
Chemistry
Corrosion Research

Chemical Analyses

o
Dual capacity,

192

= =

T. N. McVay

Manly*
Manly*

oo

Grimes*

o =

. Grimes*

Kertesz

w.
w.
F.
C.

D. Susano

IT1

111

IIT.

IT1
ITI

C!D}E)J’M’N’
0,P,Q

A,B
D,E -