ORNL-1439.txt

MARTTH MARTETTA ENERGY SYEYEMS LigRARIES
M; {

_H"{} :l

3 yys5g 0350535 5

 
 

ORNL-1439

This document censists of 224 pages.

Copy (¢ %fi of 246 copies. Series A,

Comtract No. WeT7405-eng- 26

AIRCRAFT NUCLEAR PROPULSION PROJECT

QUARTERLY PROGRESS REPORT

for Period Ending December 10, 1952

R. C. Briant, Director
J. H. Buck, Associate Director

A. J. Miller, Assistant Director

Edited by:
W. B, Cottrell

DATE ISSUED

 

OAK RIDGE NATIONAL LABORATORY
Operated by
CARBIDE AND CARBON CHEMICALS COMPANY
A Division of Union Carbide and Carbon Corporation
Post 0Office Box P

0ak Ridge, Tennessee

 

I

3 4yy5k 03L0OL3S 5

 

 

 
O o~ NI e
- o - . 2 s @ - -

P

A
A

L

ngnmgnm'_:flé:»?:ééné'bm_ér'sc;pmunp?u-fl?fi_mgmmcmno:flm

=
'

 

INTERNAL DISTRIBUTION

., Adamson

:Eéffel

Baldock

Baftpn

Bettis

Billington
Blankenship

Blizard

Bredig

Braiant

Briggs

Bruce

Buck

Callihan

Cardwell

. Center

Cisar

Clewett

Clifford

Cottrell

. Cowen

Eister

Emlet (Y-12)

Ergen

Felbeck (C&CCC)-
Fraas

Gall

Graham y
Grigorieff {consultant)
Grimes ’

ollaender
Householder

Humes (K-25)

Jones /

Keilhq'tz

. Kellty

Kertes$
M. King
E. Latson

?:::;:msé oW O TmnOTO

h&”Ufl_CdCUEDE:U_:SCU';U:'U:-JR‘_UJPT.UCU

ORNL-1439

Progress
41. R. S. Livingston
42. R. N. Eyon
43. W. D..Manly
44. W. B. McDonald
45. J. L. Meem
46. AJ J. Miller
47. K. Z. Morgan
48. E. J. Murphy
49..  H. F. Poppendiek
50. P. M. Reyling
5. H. W. Savage
52. E. D. Shipley
<53, 0, Sisman
© 54. L. P. Smith (consultant)
55. A. H. Snell
56. F. L. Steahly
57. R. W. Stoughton
58. C. D. Susano
59. J. A. Swartout
60. E. H. Tayler
61. F. C. Uffelman
62. F. C, VonderLage
63. J. M. Warde
64. A. M. Weinberg
65. J. C. White
66. E. P. Wigner (consultant)
67. H. B. Willard
68. G. C, Williams
69. J. C. Wilson
70. C. E. Winters
71-80. ANP Library
81. Biology Library
82-87. Central Files
88. Health Physics Library
89. Metallurgy Library
90. Reactor Experimental
Engineering Library
91-95. Technical Information
Department (Y-12)
Central Research Library

96-97.

  

 
198-108.
L 109,
110-117.
. 118.
119-123.,
124,
125,
126-127.
128,
129.
130-134.
135-159,
160-163.,
164,
165.
166-173.
174.
175-182.
183-184.
185-187,
188.
189.
190-192.
193-196.

197.
198-199.
200-201.

202.

203.
204-205.

206.

207.
208-209.

210.
211-214.
215-231.

 

EXTERNAL LISTRIBUTTON

Argonne National Laboratory (1 copy to Kermit Anderson)
Armed Forces Special Weapons Project fiSandla)

Atomic Energy Commission, Washington &

Battelle Memorial Institute

Brookhaven National Laboratory

Bureau of Aeronautics (Grant)

Bureau of Ships

California Research and Development Company

Chicago Patent Group

Chief of Naval Research

duPont Company !

General Electric Company, ANPP(3 coples to AF Engineering Office)
General Electric Company, Richland

Hanford Operations Office

USAF -Headquarters, Office of Assistant for Atomic Energy
Idabo Operations Office (1 copy to Phillips Petroleum Co.)
Iowa State College

Knolls Atomic Power Laboratory

-Lockland Area Office

Los Alamos

Massachusetts Institute of Technology (Benedict)

Massachusetts Institute of Technology (Kaufmann)

Mound lLaboratory

National Advisory Committee for Aeronautics, Cleveland
(3 copies to A. Silverstein)

National Advisory Committee for Aeronautics, Washington

New York Operations Qffice

North American Aviation, Inc.

Nuclear Development Associates,

Patent Branch, Washington

Rand Corporation (1 copy to V. G. Henning)

Savannah River Operations Office (Augusta)

Savannah River Operations Office (Wilmington)

Inc.

University of California Radiation Laboratory

Vitro Corporation of America

Westinghouse Electric Corporation

Wright Air Development Center
2 copies to B, Beaman
1 copy to Col. P. L.
1 copy te Lt. Col.

Hill

M. J. Nieléena

 

1i1
 

coples to Consolidated Vultee Aircraft Corporation
copy to Pratt and Whitney Aircraft Division '
copy to Boeing Airplane Company

copy to K. Campbell, Wright Aeronautical Corporation
chnical Information Service, Oak Ridge

@O = b e DD

232-246. | T

iv

 
 

Reports previously issued

ORNL-528
ORNL-629
ORNL-768
ORNL-858
ORNL-919
ANP-60
ANP-65
ORNL-1154
ORNL-1170
ORNL-1227
ORNL-~1294

ORNL-1375

Period
Period
Period
Period
Period
Period
Period
Period
Period
Period
Period.

Period:

in this series are as follows:

Ending
Ending
Ending
Ending
Ending
Ending
Ending
Ending
Ending
Ending

Ending

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

June 10, 1952

Ending September 10, 1952

 
 

CONTENTS

PAGE

FOREWORD . « & & v ¢ 4 v 4 s v 6 6 o v o o o o o o o o s o s « s o o o 1
PART T. RFACTOR THEORY AND DESIGN

SUMMARY AND INTRODUCTION . . ¢ & o v ¢ 4 o v o 5 5 5 o o s ¢ 5 o o o & = 5

1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT . . + . . + ¢ v « o . 7

Fluid CIiTOUIT v 4 v 6 4 6 o v o v v v o s o v s o o o o o o o o 7

Stress Analysis of Piping « v « + 4 4 o 56 o o o o o s o « o 2 « = 8

RBeactor o o & o ¢ 4 4 v 4 v v u b e e e s e e e e e e e e e e e 11

Instrumentation o « o & « v & o 5 o 5 5 6 0 v s e e s 4 s 2 s s o 11

Off-gas System o 4 ¢ v & & o 4 & 4 s o w4 6 8 s e s s e 8 e w . 11

Description of the system . « v o v ¢ v o & o o o 2 & 6 o o & @ 11

Maximum activity of stack gases « + ¢ 4 v « « « o o« 2 » o o « o 12

Normal discharge from the surge tank . . ¢« ¢ « ¢ o o o +» & « & 15

Gas vented from the fuel dump tank during dumping « « o + . . . 15

Reactor Control System . . & v &« & v v ¢« % o o o » v o o o o o 15

2. EXPERIMENTAL BEACTOR ENGINEERING . & « & & o + 4 o v o o « « o o « 17

Pumps for High-Temperature Fluids . . . . ¢ « ¢ & ¢ o v o o o & 17

Pump with combination packed and frozen seal . . . . ¢« . .« . . 17

Pump with frozen-sodium seal . « ¢ & o ¢ « o o ¢ o o « = o » 19

Allis-Chalmers pump o o v « v 4 o % & o o « « s o« o o s o s s o 19

Laboratory-sized pump with gas seal . . ¢ .+ o 4 « v o s o « o & 19

ABE pump ¢ o v & o v 4 s e s e 4 e v 4 s e s e s e s e e 21

Rotating Shaft and Valve Stem Seal Development . « & « « + o « o 21

Screening tests for packing materials and lubricants , , . , . 21

Packing compression tests o« ¢ « v 4 « » 4 s « s s 5 s o o« o o . 23

Packing penetration tests . v o « o o & « s o + o o s’ ¢ o o o o 23

Face seal tests o & & o & v v & s v o 4 ¢« o o v 4 v e 8 e a e 25

Combination packed and frozen seal . & 4 o & ¢ o & o « o « o & 25

Frozen-lead seal . . . . & ¢ v ¢« & o 4 4 o« ¢« o o s o o v <« o & 26

Frozen-sodium seal . . . . . & ¢« 0 v o s v 6 v 4 e 0 e e e e 26

Bellows-type of valve stem seal « & o & o« « o« o o o « o o o o & 26

Heat Exchangers . . o . « & ¢ v « v v v v o o o o o v o o o o« o & 27

Sodium-to-air radiator .« ¢ s ¢« o 4 o & o 2 & o « o & 3 0 o @ 27

Bifluid heat transfer system . & o o v o o« « o o s o o » o & o 28

InStrumentation + o« o & o s « o 4 o s o 4w o 6 6 8 4 b a0 s « &+ s 29

Rotameter type of flowmeter + . ¢ ¢« & & ¢« v « o o o & o o « o 29

Rotating-vane flowmeter . « v & ¢ & & v &« o & o o o o o « o o« o 30

Diaphragm pressure-measurement device o« o « o s o o o o s o v =« 30

 

vii
 

Moore Nullmatic pressure transmitter .« . o« + « o =
Handling of Fluorides and Liquid Metals . . « « . .« .

NaK distillation test .,
Gas-line plugging . .« &

Q a 9 @ a e 4 9 ? ° a - 9

e * e 9 * e < ° 9 ° ° e *

ZrF, vapor condensation test . < + ¢ ¢ ¢ & & ¢ o o

Project Facilities . . .
Helium distribution . .
Fabrication shop . . .
Instrument shop facility

Gas-fired furnace facilaty

3. GENERAL DESIGN STUDIES . .
Air Radiator Program . .
Radiator Perfermance . .,
Engine Performance . . .

Heat Exchanger Design Charts

4. REACTOR PHYSICS . . . . . .

- - e . . . . . ® 2 . ® 9

2 a e . - @ e e ° e L . e

? e . - » . e ° * 9 * @ a

e * 9 * - e ® . 9 * @ A

* . 9 a a o . e 9 . - . .

Oscillations in the Circulating Fuel Reactor . . . .

LLimits on the oscillations

Periodic oscillations .
Specific examples . . .
Reactor Calculations . .
Temperature Dependence of
a Resonance . « . . .

5. CRITICAL EXPERIMENTS . . .

*
]
°

a

-« e e o ° e ° . @ L * @
< * * * 9 - - - . " ® - .
° e 9 * L a . Qe » . . - °
* - * e * ° ° e @ . Qe - -

Cross Section Exhibiting

9 . * T . . e . - . L] @ 4

@ e e ° e 4 * ° e 9 - . e

Reflector-Moderated Circulating-Fuel Reactor Assembly

ARE Critical Assembly . .
PART I1.
SUMMARY AND INTRODUCTION . . . .

SHIELDING BESEARCH

6, DUCT TESTS IN LID TANK FACILITY . . « ¢ ¢ &« « ¢ ¢ o o &

G-E Outlet Air Duct . . .

- * ° - * . ¢ - @ - e a 9

Neutron dose measurements . e . * . . 9 e . . ° . .
Calculated neutron dose .

G-E Inlet Air Duct . .« .

. e - - Al - 2 < 9 - 9 4 °

. . < ° * . - . L - - e e

Induced Activity Around a Duct . o . « ¢ « ¢ « o + &
Radiation Around an Array of Cylindrical Ducts . . .

7. BULK SHIELDING FACTLITY .

. * - »

Measurements with the Divided-Shield Mockup . . . . .

viiil

 

PAGE

30
32
32
32
32
33
33
33
33
33

35
35
35
35
40

41
41
42
42
44
44

46

48
48
52

61

62
62
62
62
71
74
76

85
85
PAGE

Air-Scattering Experiments . . . . . .+ ¢ 4 v 4 e 4 ee e e e 85
Irradiation of Animals . . .+ . + + « & « & o o 4 e s 4 e e 4w 85
Other Experiments . . . . . « . . + .+ + & C e e e e e e e e e e 92
8. TOWER SHIELDING FACILITY . . . . & v 4 v ¢ o o v o o o o s o o = » - 95
Facility Design . . . . & & v v o o v s« o s o o s s 0 s s e s s 95
Structure Scattering . . .+ ¢ « « « « o + « 4 4 4 4 e 4 2 s e e 95

9. NUCLEAR MEASUREMENTS . . . . . e h e e s e 96
Angular Distribution of Neutrons Scattered from Nltrogen P e e 96

Results of nitrogen-scattering experiment . . . . « . + « .+ « . 96
Analysis of nitrogen-scattering data . . . . <« « » s . & o s U4
A Fast-Neutron Dose Measurement . . . . . . . « « « « « « = « « & 97

PART II1. MATERIALS RESEARCH |
SUMMARY AND INTRODUCTION . .+ » o o e e e e e e e e e e e e e e w103

10. CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS . . . . . « . . + « + &+ « + & 105
Fuel Mixtures Containing UF, . . . . . . ¢« .« « + . « + v « . 105
LiF-BeF,-UF, . . . . ¢ o o o v v v v 0 o v o e e e e e e 105
NaF-BeF,-UF, . . . . . .« o o v o v o v v v oo oo o v o e 105
LiF-ZreF -UF, 0 0 0 L 0 0 v s v e v e e e e e e e e e e e 107
NaF - L1F ZrF L L e e s e 107
NaZrF -NaUFg . . . . o o o 0 v v v v e v o o e v n e e e 107
RbF - AIF —UF e e e e e e e e e e e e e e e e e e 107
Fuel Mlxtures Conta1n1ng UCI e e e e e e e e e e e e e e e e 108
Fuel Mixtures Containing UF, . . . . . . . . . . . . e s e e . . 109
NaF-UF; .« v v v v v v h h e o e e e e e e e e e e e e e 109
KF-UF, . . . C e e e e e e e e e e e e e e e e .. LI0
Alkali Fluoborate Systems e e e e e e e e e e e e e e e e e e 110
Differential Thermal Analysis . . . C e e e e e e e e . 1T
Simulated Fuel for Cold Critical Experlment S O
Coolant Development . . . . . « o« v v & « o« = s o o w0 e e - 112
NaF-ZrF, . . . 0 o . v v v v b v e e e e e e e e e e e e 112
LiF-ZrF, . . . . v v v e s e e h e e e e e e e e s e e 113
NaF-KF-ALF; . . . .« o 0 o @ v v v v o s o o o v s 6 o v v o e 113
NaF-BeF,-ZrF, . . . . . « o v o v o 0 v v s v v v v e e 114
NaF-LiF-ZrF, . . . . . < ¢ v o« o o o v o v o o v 00 a e 114
NaF-BaF, . . e e e e e e e e e e e e e e e e e e e e e e I11S
NaF-KF - L1F er e e e e e e e e e e e e e e e e e e e e e e 115
NaF-RbF-A1F, . . . . e 115
Studies of Complex Fluor1de Phases e e e s e s e s v e s e e s 115

 

1%
 

PAGE

K,CrFg-Na ,CrF, -Li,CrF, . . . . . . . ¢ o o v o 0 o v o v v o 115
Solid phases in NaF-Bek, -UF, and NaF-ZrF, systems . . . . . . . 115
UF;-ZxF, . . . « o o v o o s e e e e e e e e e e e e e e 115

NaF-ZrF, . . . « o o o o i e e e e e e e e e e e e e e e e 118
Other fluorlde complexes . . P 118
Reaction of Fluoride Mixtures w1th Redu01jg Agents e e e e e 118
Reducing power of various additives . . . + v « « « o « &« o « 118
Identification of reduction products . . . . « « + ¢ 4 o v . o 119
Reaction of fluoride mixtures with NaK ., . . . . . . . . . . . 120
Reaction of NaF-ZrF, -UF, with ZrH, . . . e e e e e e 120
Solubility of potaqs1um in NaF-KF L1F eutectlc e e e e e 4 e e 121
Production and Purification of Fluoride Mixtures . . . . . . . . 122
Laboratory-scale fuel preparation . . . . . . . . « + + + « + & 122
Pilot-scale fuel purification . . . . . . +« « « + + v v « « « . 122
Fuel production facility . . . . e e e e e e e e s 123
Preparation of hydrofluorinated fuel samples et e e e e e e 124
Hydrofluorination of Zr0,-NaF mixtures . . . . . . . .« . . . . 124
Purification of Hydroxides e et e e e e e e e e e e e e e e e 125

11. CORROSTION RESEARCH . . . . . . . . . .. e e e e e e e e e e 127

Fluoride Corrosion in Static and Seesaw Tegts e e e e e e e e 127
Oxide additives . . . . . v e e e e e e e e e e e 127
Comparison of liquid- and vapor - phase COrroS1on . . . « « o+ & 127
Crevice corrosion ., . Ve e e e e e e e e e e e e e 128
Carboloy and Stellite alloys e e e e e e e e e e e e e e 130
Reducing agents . . . . .« « .« 0 v 0 4 v e e e e e e e e 130
Fluoride treatment . . . . o ¢ + ¢ v & 4 & & & o o 4 e 0 e . 131
Temperature dependence . . . + & v 4 v 4 v 4 e e e e e e e e 132
Ceramic materials . . . . e e e e e e e e e e e 132

Fluoride Corrosion in Thermal Convectlon loops . . « .« « o o .. 133
Mixtures containing ZrF, . . e e e e e e e e e e e e e e 133
Hydrofluorinated NaF KF LiF Uf e e e e e e e e e e e e 136
Corrosion inhibitors . . . . . . + . v v v 0« 0 v e e e e 137
Inserted corrosion samples . . . . . + ¢ . ¢ ¢ 0 0 0 e e e e 138
Crevice corrosion . . . . . . e e e e e e e e e e e e e e 138
Variation in loop wall comp051t10n C e e e e e e e e e e e 139
Self-insulating properties of fluorides . . . . . . . . . . . . 139
Other 1oop tests . . v ¢ v v 4 v v o o o o o o o s s e a4 e e 139

Hydroxide Corrosion . . . + « & & v v o« o o o o o o s o o o o & 141
Corrosion inhibitors . . . & + © v v v v v s 4 e e e e 141
Temperature dependence . . . . . . . . . 4 0 4 e e e e e e e 141

Nickel alloys in NaOH . . . . . . . . . « o o ¢ o v o v o o v . 141

 
 

PAGE

Compatibility of BeO in KOH . . . . e e e a e e e e e e s 141
Nickel 1n NaOH under hydrogen atmosphere C e e e e e e e e e 142
Liquid Metal Corrosion . . . . ¢ & & v « v 4« v o v s « & o o+ « 144

High-velocity corrosion by sodium . . . « « « « + & v v « + & 144
Ceramic materials im sodium . . « & « & &« « & « 4 4 v o 4 4 4 . 146
Lead in metal convection loops . . . . + ¢« v & « v & 4 4 e e 147
Lead in quartz convection loops . . .« & ¢ « « 4 s ¢ s a4 s s e e 148
Compatibility of BeO in NaK . . . . . ¢« « « « v & « v v v s o 150
Sodium in forced-convection loops . . . . . « . . « .. . . . 150
Fundamental Corrosion Research . . . . . « . . « « ¢« 4 « « « « . 151
Interaction of fluorides with structural metals . . . . . . . . 152
Air oxidation of fuel mixtures . . . . . + « « & - ¢ ¢ . . . 152
EMF measurements in fused fluorides . . . . . . . . . v .+ +« . . 153
Preparation of trivalent nickel compounds in the hydroxides . . 154
12. METALLURGY AND CERAMICS . . . . v e s e e a s 4 s aa e s e s 155
Fabrication of Solid Fuel in bpheres &« e b e e e e ee e e s 155
Suspension in refractory powder . . . . . « & . 4 s 4 e 4 e 4 s 156
Momentary melting in a high-temperature arc . . . . . . « . . . 156

Spraying from a metallizing gun . . . . . . . « « + ¢ o« . . . 156
Solid Phase Bonding of Metals . . . . . . . « . + + & « &+ & « « & 156

Columbium Research . . ., . . . . . . . . . . . o . . e e e e 157
Gaseous Treactions . + + o 4 o o s o & o o o o o = 4 e e 0. 157
Oxidation in air . . . . . c e s e e 5 e e e s ae e e e 158

Creep Rupture Tests of Structural Metals . . . . . .+« « ¢« . + . . 159
Inconel 1n air . . & 4 & ¢ o i e 4 e e v e e e e e e e e e e e 159
Inconel in hydrogen . . . e e e e e e e e e e e e e e e 159
Inconel in molten fluorldes e 159

Type 316 stainless steel in argon . . . . + « v « &« o v & o & 159
Type 316 stainless steel in molten fluorldes e e e s e e e s 159
Brazing and Welding Research . . . . . . . . + + « . « v o o 4 & 161
ARE welding . . . . &« v v 0 v h e e e e e e e e e e e 161

Cone-arc welding . . . « . . « « & 4 « i e e w e e e e e e e 161
Reésistance welding . . . T 162
Automatic heliarc machine weldlng T 164
Fabrication of heat exchanger units . . . . . . . « . « .+ + .+ . 164
Brazing of copper to inconel . . . . . . . . . o . 0 0 o . . 165

Evaluation Tests of Brazing Alloys . . . . . + « « « « « « &+ « . 165
Oxidation of brazing alloys . . . . « . . « « ¢ ¢ o s « o+ + . 165

Corrosion of brazing alloys by sodium hydroxide . . ... . ., . . 168
Corrosion of brazing alloys by sodium and lithium . . . . . . . 168
Tensile strength of brazed joints . . . « . . . . . + « « . .+ & 169

X1

 
 

PAGE

Melting point of 60% Pd-40% Ni alloy . . . . . . . « « « « . . 171

Ceramics Research . . . . . « « ¢ v v v v o v v 0 e e e e e . e 171
Development of cermets for reactor compomnents . . . . . + + .+ & 171
Ceramic coatings for an aircraft-type radiater . . . . . . . . 171
Ceramic coatings for shielding metals . . . . . . . . « . « « . 171

Uniformity of beryllium oxide blocks . . . . . . . « . « .« « . 171

13. HEAT TBANSFER AND PHYSICAL PROPERTIES BESEARCH . . . . « « ¢« & « & 174
Thermal Conductivity of Liquids o « ¢ ¢ o & o o o o s o o o s s = 174
Heat CapaCitY Of Liquids T @ 8 % e & & 4 @ @ 8 94 % @ @ 4@ @ @& e 9 ].76

Viscosities of Fluoride Mixtures . . ¢« « &« o« ¢ s o v « o o o o & 176
Measurements with the Brookfield viscometer . « o« ¢ ¢ « o + « & 176
Capillary viscometer . & o« o & o o« o s s o o ¢ o o s & o o o 176

Density of Fluoride Mixtures . « & ¢« ¢ ¢ ¢ ¢ o ¢ v v o o o« o o = 178

Vapor Pressure of Molten Fluorides . .« « + ¢« ¢ v ¢ o ¢ o o o o o 178

Convective Heat Transfer in Fluoride Mixture Nalb-KF-LiF . . . . . 179

Analysis of Specific Reactor Heat Transfer Problems . . . . . . & 180
Heat Generation in the reactor reflector .« ¢ ¢ & o o ¢ o« & o o 180
Analysis of fluid-to-air radiators . « « o « o « o« o o o « s 181

Natural Convection in Confined Spaces and Thermal Loop Systems . 182

Turbulent Convection in Annuli . . ¢« ¢« ¢ &« ¢ &« o o o o s o s o 182

Circulating-Fuel Heat Transfer . « ¢ ¢« ¢« o« ¢ o o ¢ o & o o o & = 185

14, RADIATION DAMAGE . . ¢ ¢ ¢ ¢ o o o ¢ o o o s s o s 5 s s s o o o 186

Irradiation of Fused Materials . . &« ¢ &« ¢ ¢ ¢ ¢ o o o o« s o s o 186

In-Reactor Circulating Loops « ¢ « ¢ ¢ « ¢ « o o o o « o o o o s 187

Creep Under Irradiation « o« o« o o o o o o o o« o o o o o s « o« o @ 188

15, ANALYTICAL STUDIES OF BREACTOR MATERIALS . . + & ¢ ¢ o ¢ o « o« o o & 190

Chemical Analysis of Reactor Fuels . . ¢« ¢« & ¢ ¢ ¢« o v ¢ o o o o 190
Zirconium « « « o s s o o s s © s+ s s s e s+ s a o 2 s e e s o s 191
Chromium .« + ¢« ¢ ¢ ¢ o ¢ o o « o s ¢ o s s & « o s o o a o « s 192
Aluminum o v & & & ¢ 4+ ¢ o s s s 4 6 s s s 8 v s s s e e e = s 192
OXYEEN & « & & s o o o o 5 s v s s 5 o o s o v o o o o s 3 o 192
Water + ¢ ¢« o ¢ o o o 4 s o 3 © 5 s e 6 ° v s s s e & e 3 a4« 192

Determination of Beryllium in NaK . . . ¢« & « ¢« ¢ o« o ¢ ¢ o o & 192

Spectrographic Analysis . « ¢ ¢« « o o ¢ ¢ o ¢ & 4 o s s s 8 s s e 193
Uranium oxides « & o « o o o o o o & s o s s » s o s s o o o o 193
Nickel fluoride « « « &« « o o o o o o s « o o s o o o « 5 o s o 193

Petrographic Examination of Fluorides . « « « ¢ ¢ ¢« ¢ ¢« « & « & & 193

ZI‘Oz—HF . . . e . . . @ v o % % - ° @ L) e L e s « 9 . ] ) ° . - 193
NaF“ZrOFz . s e . e . ° @ ° ° ° e . e 4 - ° . ® e ° e s 9 ° ® . 193

 

x11
 

-

Optical Properties of Some Fluoride Compounds . . . . . . .
NaQUFG < L] @ .’ . » a - # 2 9 ® 2 a 2 @ - ] - " ° - ® * . -
KaUF7 - * - a o » a4 ? a - ¥ a . a 9 w -« - a . * . - & * o
NaaZrFT » - o @ L] L] < > - . < e - o o a . 2 * e L] < .I - *
2: ZI'F4'UF3 . e - ° - ) ) ° & .

Chemical Analyses for UO, in UF, , . . . . . . . ..

Service Chemical Analyses « ¢ v ¢ o 4 « o 4 « o 5 o & o o s

| PART IV, APPENDIXES
SUMMARY AND INTBODUCTION + & » v v v v v v e e v e e e e e e a o
16, LIST OF REPORTS ISSUED . . v v v v 4 ¢ v s v v o o o o o o 4
17. DIRECTORY OF ACTIVE ANP RESEARCH PROJECTS AT ORNL . . . . .

I. Reactor and Component Design . . . + ¢« v +v 4 o« o + &

IT. Shielding Research .« . . ¢« ¢ ¢ & ¢ ¢ & v o a « o o &

ITI. Materials BResearch . . . . ¢« o o' v ¢ v v v 4 o + &
IV. Technical Administration of Aircraft Nuclear

Propulsion Project at Oak Ridge National Laboratory .

 

PAGE

194
194
194
194
194
194
194

199
199
202
202
203
204

209

X111
ANP PROJECT QUARTERLY PROGRESS REPORT

FOREWORD

This quarterly progress report of: the Aircraft Nuclear Propulsion Project at
OBNL records the technical progress of the research on the circulating-fuel
reactor and all other ANP research at the laboratory under its Contract W-7405-
eng-26. The report is divided into four parts: I. Reactor Theory and Design;
1T, Shielding Research; III. Materials Research; and IV, Appendixes. Each part
has a separate Summary and Introduction.

The Aircraft Nuclear Propulslon Project is comprised of some 300 technlcal
and scientific personnel engaged in many phases of research directed toward the
nuclear propulsion of aircraft. A considerable portion of this research 1is
performed in support of other organizations participating in the national ANP
effort. However, the bulk of the ANP research at ORNL is directed toward the
development of a circulating-fuel type of reactor.

The nucleus of the effort on circulating-fuel reactors is now centered upon
the Aircraft Reactor Experiment - a 3-megawatt high-temperature prototype of a
circulating-fuel reactor for the propulsion of aircraft. The current status of
the ARE is summarized in section 1; however, much supporting research and
developmental information on materials and problems peculiar to the ARE will be
found in other sections of Part I and Part III of this report, in addition to
the general design andmaterials research contained therein. Shielding Research,
Part 11, is devoted almost entirely to the problems of aircraft shielding.
 
SUMMARY AND

The Aircraft Reactor Experiment
(sec. 1) is now well into the transi-
tion period between design andreality.
The design is essentially complete,
almost all the components are onorder,
and a substantial number of these
have been received and installed in
the ARE Building. The significant
modifications during the past quarter
include completion of the off-gas
system design (incorporating holdup
tanks rather than charcoal adsorbers)
and the inclusion of a reactor by-pass
(so that the fluid circuits may be
checked out independently of the
reactor). Coincident with the com-
pletion of the reactor design, the
Aircraft Reactor Experiment Hazards
Summary Report, ORNL-1407, was sub-
mitted to the AEC for approval. It
is anticipated that even though the
experiment may be completely assembled
by the summer of 1953, a significant
and indeterminant period will be re-
quired for shake-down operation before
the reactor becomes critical. |

Valves, pumps, instrumentation,
and other components of both the
fluoride fuel (NaF-ZrF,-UF,, 50-46-4
mole %) and reflector coolant (NaK)
circuits are being developed for the
Aircraft Reactor Experiment (sec. 2).
In most instances, these components
have been tested on smaller than full-
scale prototypes of the actual ARE
components. Tests are now under way;
however, on the full-scale pumps,
valves, and some instrumentation
designed for the reactor experiment.
At this time, centrifugal pumps with
both gas seals and frozen seals have
operated satisfactorily for extended
periods at temperatures between 1200
and 1500°F. A combination packed and
frozen seal has been specified for
the ARE pump. Although a bellows type
of seal has been specified for the
ARE, a considerable program has been
undertaken on high-temperature, self-
lubricating seals.

INTRODUCTION

The rotameter type of flowmeter
and the modified Moore Nullmatic
pressure transmitter have both operated
satisfactorily at high temperatures
(~1400°F). Neither of these instru-
ments is affected by the ZrF, vapor
above the fuel. Vapor traps of the
type that will be required in the gas
system above the fuel surge tanks have
been satisfactorily developed.

The heat transfer coefficient of
an aircraft type of sodium-to-air
radiator, in which the radiator fins
were sectioned every 2 in. 1in the
direction of air flow, was increased
20% over that of the same radiator
with plain flat fins. :

The general design studies (sec. 3)
were confined to performance analysis
of a Sapphire turbojet engine in which
the engine radiator performance was
extrapolated from a sodium-to-air
radiator section tested at ORNL.
Performance data for both the radiator
and engine are presented. Arrange-
ments have been completed with the
Wright Air Development Center for
their participation in the development
of high-temperature liquid-to-air
airvecraft radiators.

The several reactor physics studles
(sec. 4) include those of oscillations
in a circulating fuel reactor, a
technique for reactor calculations,
and the temperature-dependence of a
cross section exhibiting a resonance.
The damping influence of fuel circula-
tion has been demonstrated even when
the earlier assumptions are replaced
by more realistic conditions. With
regard to reactor calculations, 1t 1s
shown that the slowing down of neutrons
in parallel slabs of materials with
different properties can be described
in some instances by a set of 1mages
of the original neutron source,

Critical assemblies of both the
Aircraft Reactor Experiment and a
reflector-moderated reactor have been

tested (sec. 5). Critical mass,
ANP PROJECT QUARTERLY PROGRESS REPORT

control rod effectiveness, flux and
fission distribution, and self-
shielding factors have been determined
for this 1nitial reflector-moderated
assembly. As was expected, this as-
sembly gave rather peaked flux dis-
tribution curves, and the lumped fuel

resulted in rather large self-shielding
effects. Measurements on the ARE
critical assembly continued. Cali-
bration curves for the ARE control
rods were obtained, and the effects
of various core components on reac-
tivity were compared.
PERIOD ENDING DECEMBER 10, 1952

1. CIRCULATING FUEL AIRCRAFT REACTOR EXPERIMENT
| J. H. Buck
Research Director’s Division

E. S. Bettis
ANP Division

A few relatively minor changes in
the design of some phases of the
Aircraft Reactor KExperiment have been
made 1n the past gquarter., These
changes involved the off-gas system
and the fuel disposal system, pri-
marily, and are discussed below.
There have been no changes in the
general design, and detailing of con-
struction and installation designs
has been largely completed. Installa-
tion of equipment in the building has
proceeded at an accelerated rate
throughout the quarter but has not
yet reached the expected peak rate.
Components from outside vendors, as
well as those fabricated 1n ORNL

shops, are being received in increas-
ing quantity; but some of the major
items, notably, the heat exchangers,

have not yet been received. It 1s
expected that all components will be
on hand by the end of the year.
Component testing is proceeding in
the experimental engineering labora-
tories (sec. 2}, but the final component
testing 1s to be done in the ARE
Building. It is new planned to by-
pass the reactor so that a thorough
system test can be made before the
nondrainable reactor is tied into the
~ circuit. This procedure will make
 possible a complete shake-down run
without the possibility of incorporat-
ing a nondrainable component 1in the
system. The design change necessitated
by this alteration in operational
procedure is minor and will be effected
in the field. Considerable effort
during the quarter went 1nto the
preparation of the “Haz ards Report, (1)
which, in addition to serving the

(1)J. H. Buck and W. B. Cottrell, Aircraft

Reactoer Experiment Hazards Sumeary Report,
OBRNL-1407 (Nov. 24, 1952).

purpose implied in the name, provided
the first draft of an operations manual
for the experiment.

It is practically impossible to
predict a date for satisfactory opera-
tion of the ARE. There are no means
available for predetermination of the
time required to correct difficulties.
It is not so difficult to attempt to
predict a date for the completion of
the installation of equipment, and
therefore a date for beginning the
system test in the building. At
present it is expected that the system
will be ready for initial testing
near the first of June 1953. No
uranium will be added to the system
until all reasonable checks have been
made to ensure the integrity of the
entire assembly.

FLUID CIRCUIT

G. A. Cfisty, Engineering Department

The previous report(z) discussed
design changes necessitated by the
higher fuel viscosities. In brief,
the higher viscosities reduced the
Reynold’ s number of the heat exchanger
tube side and changed the heat transfer
coefficient to such an extent that the
calculated minimum film temperatures
were 1n the vicinity of the freezing
point. Two corrective measures were
discussed; redesigning the heat ex-
changers and increasing the flow rate
through the entire circuit. The first
change has been accomplished. The
need for the increased flow rate has
been obviated by an accurate measure-
ment of the fuel thermal conductivity;
the new value of £kis 1.5 Btu/hr*ft"°F,
whereas 0.5 had been used in earlier

 

{ZJG. A. Cristy, ANP Quar. Prog. Rep. Sept. 10,
1952, ORNL- 1375, p. 6.
ANP PROJECT QUARTERLY PROGRESS REPORT

analyses. With this higher conduc-
tivity value and the redesigned heat
exchangers, the calculated minimum
film temperatures are comfortably
above the freezing point, with the
original flow rate, Restoring the
flow rate to its original value will
reduce the reactor inlet temperature
to 1150°F and reduce the maximum
system pressures and pressure shell
stresses.

It has been decided to employ NakK
as the reflector and pressure-shell
coolant. This has necessitated com-
plete redesign of the reflector-
coolant heat exchangers, primarily
because NaK will extract more heat
from the fuel tube elbows and other
components washed by the coolant than
would the lower conductivity salt
around which the earlier calculations
were based. The heat exchangers have
been re-engineered and currently are
being constructed. The changes re-
ferred to above are reflected 1n the
revised flow sheet, Fig. 1.1.

Detailed engineering designs for
most of the outstanding i1tems have
been released during this quarter.
The items included were such components
as the fuel system surge tanks, the
reflector-cooling system surge tanks,
the reflector-cooling system heat
disposal loop and ducting, the thermal
barriers, the water system piping, the
glycol surge tank, the off-gas dis-
posal system, and some of the fuel
system piping. The redesigned ARE
pump is described in section 2, “Ex-
perimental Reactor Engineering,” and,
as stated therein, initial tests are
now in progress. Delivery of the
bellows type of valves that will be
used in the fuel and reflector systems
is expected within a month. The parts
for these valves are being supplied by
Fulton Sylphon Co., but the valves
are to be welded at ORNL.

STRESS ANALYSIS OF PIPING

R. L. Maxwell J. W. Walker
Consultants, ANP Division

The maximum stresses that will be
encountered in the piping for the ARE
have been calculated to be approxi-
mately 24,000 1b/in.%. Since these
stresses will be somewhat relieved by
creep at the operating temperature,
this stress value 1s only significant
in defining the stress range of the
pipe in going from the cold to the hot
condition. If the stresses were
completely relieved by creep, this
would correspond to a strain of about
1%.

In order to reduce the loads and
stresses in the pipe in the hot con-
dition, the piping will be pre-
stressed, or presprung, by an amount
equal to 75% of the total thermal
de formation. This will have the
effect of putting the maximum loads
on the pipe in the cold condition and
reducing to a large extent the load
in the hot condition. Because of the
uncertainty of realizing the designed
cold spring in the actual instal-
lation, only partial credit should be
taken for prespring. However, the
loads and stresses in the cold con-
dition were determined by considering
the full amount of the prestressing,
and the loads and stresses in the hot
condition were determined by con-
sidering only 50% prespring. This 1is
in agreement with accepted practice.

The maximum stresses in the cold
condition will be higher for the same
deformation than in the hot condition
because of the difference 1in the
modulus of elasticity. The maximum
values of the stresses are:

Cold § .. = 25,000 lb/in.?
12,000 1b/in.?

HOt Smax =
The maximum stress in the cold con-
dition is about one half the yield-
point stress, and the maximum stress
L 2-in. GATE VALVE

  
 
  

 

 

    
   

 
   

 

 

 

 

TO DRAIN

   

TO STACK

    

 

 

 

 
 

 

 

 

 

  

 

  

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

   
    
   

 

 

 

 

 

 

 

 

 

 

   
  
  

 

 
 
  

 
 
 

   

 

 

   
  
   

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

GLYCOL ]
o EXISTING EQUIPMENT SHOWN DOTTED - SURGE TANK g i L
————— -~ (‘j —e—
r ~ _ \ B i N
f \ 4-in. PIPE = - i {
r+li:ESERVOIR e e A ——= . : = ‘ - ! o | 55-gal DRUM
& \\ 3 ) ] Y | - 12 "_‘_] . i T % I i
P e e - ) Lot ' r= w o w O ! V2
T y L | w i F E|Z 15:2 i i
- i ] k- : V= : )
= Y, -in. PIPE SL,r ; | g% i 18l i L :
&t —L~—- Sive g4, P30 . : : / G i
2 - ; * wik® iz 519 @ . z ? //
! PRESSURE I vaLves | a ¥ g [ g - 2 >
| SWITCH | VALVES TO POND 5 Sl g : a '
: el B i - : ! 1
gl L 2 5 £ ;lj r e Yo
£l X i Y ! * % {
o i 1 f
| | ! ;
! |
P ) l 5 55-gol DRUM ouTsIDE OF '
—{He} — s | BUILDING
. FROM He CONTROLLED SUPPLY - ' $
P
: i
FROM Hz CCNTROLLED SUPPLY i : i
| I
E i
i [ !
|
i
|
i
}
B
@
@
!
S :
oz ECONOMIZER _—
S5 o
. 2
a x
wi i
el g MAINTENANCE]
w w i
wl oy DRAIN
7@ i
51 3 ! 3000 | 7,300
uy
¥ ! ]
®
=z
<
=
S:
?
TO DRAIN
TO DRAIN
- FILL
= = ! — LINE
1P Si @ Iz
- o 4 - 1
- S I - 4
He = i 4 2 a
I Lo ; ENED = x
. I ! i
& ”P ik |
Lo
—>a 1 ' TANK NO. 3 TANK NO.
i : TANK NO. 4 TANK NO.5 TANK NO, & L Mok FLEL CARRIER
; KoK (o) NaK NoK i i e fe
- ! ; T
’ @ He i @ DUMP TANK !
: i 3 i
i : : 7\
Y% SUPPLY HEADER Z fi- ‘ % 1 |
: s ~ ® . L& t l
o % MAXE-UP HEADER ol f | | E ) i
e o ef I He
oy
R FILL AND pqfFILL AND
" DUNMP LINE DUMP LiKE NOTES
{ :

 
  

 

 

   

i FROM He SUPPLY TRAILER
THe SCRUBBER RESERVE MANIFOLD

12 He BOTTLES

FLOW IN gpm

SO0 S

FLOW IN cim

¢

Fig.

PRESSURE IN ps.g

TEMPERATURE i °F

i. 1.

PRESSURE IN INCHES OF WATER, GAGE

Fluid-Circuit

FUEL

NoK

HELIUM
WATER
GLYCOL
INERT SALTS
VAPOR TRAR

NORMAL OPERATING POSITION
OF VALVES

~—=3— OPEN

——»— CLOSED

————=— THROTTLING

Flow Sheet.

FLOW CONDITIONS ARE BASED ON THE ASSUMPTION
THAT THE COOLANT HAS THE FOLLOWING PROPERTIES:

o =187 brett
1= 7 TO 3 cp AT OPERATING CONDITIONS
Cp=0.26 Btu/b°F

VOLUME OF MAIN SYSTEM (APPROX.):

INTERNAL: 1.5 13
EXTERNAL: 6.0 f+3
TCTAL 7.5¢13
in the hot condition corresponds to a
maximum creep, or deformation, of 0.5%.

REACTOR

The reactor design, with the ex-
ception of very minor changes, remains
as previously described. All detailed
drawings have been released to the
shop. The beryllium oxide blocks have
been sized and are ready for assembly
into the core. The pressure shell has
been received from Lukenweld, ¢*’ the
reinforcing segments for the head have
been welded in place, and the head
holes have been bored. The various
core pressure-shell components are
currently being fabricated and as-
sembled.

INSTRUMENTATION
R. G. Affel, ANP Division

Installation of all instrument
panels has been completed. Approxi-
mately 80% of the process instruments
have been mounted and supplied with
electric power and/or compressed ailr.
The remalning instruments should be
installed by January 15, 1953. Minor
instrumentation changes have been made
to accommodate the design changes in
the pump seal. All major process
instrumentation components are de-
tailed and either on hand or on order.
It is planned to complete the instal-
lation of instruments on the panels
so that as work in the pits proceeds,
the sensing elements, when installed,
may be checked directly to the panels.

To date, 21 of a series of instru-
mentation prints have been issued and
shop fabrication started. The series
includes such items as the fuel
circuit flowmeter, the reflector-
cooling system (NaK) electromagnetic
flowmeter, the NaK purification system
electromagnetic flowmeter, the ta-
chometer mountings for the reflector
coolant and fuel helium fans, and the
liguid-level indicators for the re-

flector coolant and fuel surge tanks.

(3’The Lukenweld Division of The Lukens Steel Co.

Detailed prints of the vapor traps
required by the ZrF, condensate (cf.,
sec. 2, “Experimental Reactor Engi-
neering’”) for the surge tanks and the
hot fuel dump tank have been issued
and sent to the shops for fabrication.
Location of thermocouples on the
system has proceeded at a satisfactory
pace. Two prints showing thermocouple
construction have been issued and sent
to the shops. Since the system will
require approximately 750 thermo-
couples, every effort is being made to
run the thermocouple extension wires
to their approximate pit locations
before the pits have major components
placed within them. It is believed
that this procedure will expedite
final installation and test of the
temperature instrumentation.

OFF-GAS SYSTEM

S. A, Hluchan, Instrument Department
H. L. F. Enlund, Physics Divasion

The previous report(4) described an
activated-carbon off-gas scrubber
cooled by liquid nitrogen. Further
analyses have shown that the liquid
nitrogen consumption would be 1in
excess of 1000 1b/day because of the
decay heat of xenon and krypton 1f the
gases were held for less than 100 min
before reaching the scrubber. The
nitrogen consumption did not become
reasonable until the gases were held
up for many hours before admittance to
the scrubber. In fact, with little
additional difficulty, the gases could
be held up until they decayed suf-
ficiently to be discharged from the
stack and the need for the scrubber
would be obviated. Accordingly, it
has been decided to abandon the carbon
scrubber in favor of mechanical re-

tention.
Description of the System, A

helium atmosphere is maintained over
the surge tanks in the fuel system
and the fuel dump tanks. This helium

(4)T. Boseberry, ANP (Quar, Prog. Rep. Sept. 10,
1952, ORNL-1375, p. 12. ) '

 

11
ANP PROJECT QUARTERLY PROGRESS REPORT

gas, which will contain the volatile
fission products (Be, I, Xe, and
Kr), ¢5) passes through a NaK vapor
trap (where the Be and I are removed),
then into two holdup tanks (to permit
the decay of Xe and Kr), and is then
released in the stack. However, the
release of gases to the stack 1is
dependent upon two conditions: the
wind velocity is greater than 5 mph
and the activity, sensed by a monitor,
is less than an established maximum
value. The monitor is located between
the two storage tanks, which are
connected 1in series. Provision is
also made to exhaust the reactor pits
through the holdup tanks in the off-
gas system at a rate of 27 cfm (the
limit of the exhaust system). The
off-gas system 1s shown schematically
in Fig, 1.2.

With a static helium atmosphere in
the surge and dump tanks, the volu-
metric flow rate of fission gases 1is
a maximum of 0.0014 cfm. In order to
have a measurable flow of gas, the
fission gases are mixed with a fixed
air bleed of 10 cfm between the first
holdup tank and the monitors. The
minimum flow rate past the monitors
then 1s 10 cfm, and the maximum is
37 ¢fm (that 1s, 10 c¢fm from the fixed
air bleed plus 27 c¢fm 1f the room 1is
being exhausted at the same time).
The monitor setting can then be based
on the premise that the flow 1s 37 cfm,
and the setting will be conservative
by a factor of 3.7:1 wher the minimum
flow rate prevails.

The second holdup tank 1s placed
between the monitors and the stack
valve to increase the gas transit time
between the monitor and the valve to
ensure that the valve will have time
to close after the monitor signals the
presence of excess activity.

Maximum Activity of Stack Gases,
The maximum activity of the stack
gases has been calculated by assuming

 

(SJM. M. Mills, A Study of Reactor Hazards,
NAA-3R-31 (Dec. 7, 1949),

12

that the average energy of disinte-
gration is equal to 1 Mev.*? The
design of the off-gas system includes
two vacuum pumps that deliver a maximum
of 27 c¢fm and a blower that discharges
10 ¢cfm to give a total of 37 cim
maximum discharge through the monitor,
The maximum permissible ground concen-
tration for the 1-Mev fission-product
activity 1s

MPC = 1.6 x 10°% ue/cm® .
At 37 cfm, or 1.05 x 105 cm?®/min, and

with a permissible activity discharge
rate of 0.832 curie/min, which produces
the above MPC, the activity per cubic
centimeter should not exceed

m_g;gégm__: 0.8 Mc/cms ’

1.05 x 10°
as determined by the radiation moni-
tor. The monitors that control the
positions of the stack valve will be
set to prevent discharge as soon as
the activity of the gas rises above a
predetermined level.

The discharge rate of xenon and
krypton fission products to the atmos-
phere must be such that the activity
1is within the above limit. The
effective system holdup, before the
monitors, is 11.4 ft3, or 3.53 x 10°%
cm?® (the volume of one tank plus
piping). If it can be assumed that
the decay time of the fission products
is equal to the system volume divided
by the flow rate, the total energy at
a particular time can be calculated.
This energy must be converted to
disintegrations per unit time by
considering effective energy changes
in relation to the variation of 1so-
topic concentrations with time. The
values of permissible and calculated
stack activity have been plotted in
Fig. 1.3, together with the permissible
discharge rate, as determined by the
National Committee on Radiation Pro-
These data, together with

 

K. Z. Morgan (Chairmen), Maximum Permissible
Armounts of Radioisotopes in the Human Body and
Hax imum Permissible Concentrations in Air and
Kater, NBS Handbook No. 52 (to be published).
€T

214

el

‘@3 sAS fesodsyd se9-3JJO

cka

 

FROM PiTS

 

HF
DY Y

 

 

 

 

UNCLASSIFIED
DWG. 16957

AR FROM
BASEMENT —b—g
10 ctm BLOWER

SHIELDING WALL ;% 7

 

 

  
  
     
   
 

mt}
FILL TANKS B
LINE VOLUME APPROX 2 ft3 (GAS)
G.
Y4-in. BOURDON RESTRICTOR -
{ VALVES £
He SUPPLY o
o
&
FUEL o COOLING E )
SURGE TANK o N
054 3 GAS OR £ REFLECTOR GOOLANT VAGUUM PUMP | R ADIATION
LIQUID VOLUMES o SURGE. TANK MONITORS

Ys~in. BOURDON RESTRICTOR
VALVES

He SUPPLY

 

VAPOR TRAP

   
    
 

FUEL
SURGE TARK

  

054 13 GAS OR
LIQUID VOLUMES

REFLECTOR COOLANT
SURGE TANK

  

 
    

NaK VAPOR TRAP
AND STORAGE TANK

=== NORMAL VENT PATH
o
e~ HOT GAS BY-PASS

 

 

   

 

3

VACUUM PUMP

 

VACUUM DISCHARGE PATH

\_ e

   
 

VALVE CONTROLLED ¥
BY ANEMOMETER &

¢S6T ‘01 ¥IAWADIA ONIGNA (0T HAd
Vi

A4

‘€T

‘axoydsouw}y 94l 03 AJrAIIOY wWoRrdAI) pue uoudx jJo IFIBYISIQ

DISCHARGE RATE {curies/min}

1000

500

200

t00

50

20

| !
REF., BU. STDS HANDBOOK NO 52
{PREPUBLICATION) K.Z. MORGAN,
COMMITTEE CHAIRMAN

REACTOR POWER, 3mv
OPERATING TIME, 200 hr
I {APPROX SATURATION)

DWG. 15958

|
DISCHARGE ACTIVITY
{curies /min)

TOTAL DISCHARGE RATE (41.6 ¢¢/min)
PER RESTRICTOR (208 cc/min)

DISCHARGE RATE TO YIELD MAXIMUM PERMISSIBLE .

GROUND CONCENTRATION ACCORDING TC REF.

|

20 50 100 200
HOLDUP TIME {min)

5.9 days = 8500 min

 

 

       

500 1000 2000 5000 10,000

      
   
 

H90Ud ATHALYVNO L3AL0Hd JdNV

L

LHOdHd SS
the disintegration rates and energies
involved, are tabulated in Table 1.1.
It will be noted that after 8500 min
of holdup, corresponding to a discharge
rate of 20.8 cm®/min through each
restrictor, the ground tolerance will
not be exceeded. : }

Normal Discharge from the Surge
Tank. The maximum pressure within the
surge tank is limited by a 5-psig
pressure regulator in the gas supply
system. The maximum discharge rate
from the surge tank is limited by a
Bourdon restrictor to 20 cm®/min with
a 3-psi pressure differential. At
this flow rate through each of two
restrictors, the holdup volume of the
system 1s such that it will allow
sufficient decay time, as required by
the above calculation, to permit the
discharge of the off-gas directly up
the stack. The change of effective
energy owlng tc the persistence of the
longer lived isotopes has been ac-
counted for when ground concentration
of this gas is considered.

Gas Vented from the Fuel Dump Tank
During Dumping. In considering waste
gases from the surge tanks, 1t has
been assumed that all gaseous products
escape 1nto the surge tanks during
operation. Actually, 1t is expected
that some portion of the fission-
product gases will remain in the fuel
and that some fraction of this portion

PERIOD ENDING DECEMBER 10, 1952

will be released when the fuel is
admitted to the radicactive-fuel dump
tank. When fuel i1s admitted to the
fuel dump tank, 1t is expected that
helium will be displaced at a rate of
the order of 5 ¢fm. This helium,
after passing through the NaK scrubber
and the first holdup tank, will be
admitted to the monitors. Should the
monlitors sense excess activity, the
stack valve will prevent discharge,
and the vacuum pumps will permit
recirculation through the holdup tanks
and monitors. The recirculation
should homogenize the gases and ensure
that the monitor i1s sensing a repre-
sentative sample. When the gases have
decayed sufficiently to permit release,
the monitor-controlled interlock will
open the stack valve.

REACTOR CONTROL SYSTEM

E. P. Epler
Research Director’s Division
The high-temperature
chamber, without U?3% plating, has
completed 1000 hours at 800°F., The
temperature will now be raised at the
rate of 50°F per week to determine
the high-temperature characteristics
of the MgO,*Si10, insulator. The
fission chambers to be used for the
ARE startup are scheduled for delivery
February 1, 1953, and will be tested
at 800°F.

fission

 

 

 

 

 

TABLE 1. 1. ACTIVITY OF XENOXN AND KRYPTON AS A FUNCTION OF HOLDUP TIME

STACK DISCHARGE RATE | EFFECTIVE | DISINTEGRATION : :
AOLDUP TINE (curies/min) ENERGY RATE DECAY POWER pc IN ;IB
Days ¥Minutes | Calcul ated MPC {Mev) (dps) Watts Mev/sec (uefem™) :
Y% 720 344 2 0. 41 9.15 x 1015 500 3.75 x 10%% [ 3.9 x 10-8
1 1, 440 110 2. 45 0.34 5.88 x 101° 320 2.00 x 10% | 4.71 x 10-6
2 2, 880 39.6 3.51 0.237 4.92 x 1018 160 1.00 x 10*% | 6.75 x 10~
3 3, 320 28.5 3.94 0.211 3.79 x 1013 128 8.00 x 109 | 7.58 x 10-®
15 15 6

5 7, 200 6.91 4. 54 0.183 1.84 x 10 54 3.38 x 10 B.74 x 10
7 10,080 3.02 4. 54 0.183 1,13 x 1015 33 2.06 x 10%% | 8.74 x 10-8

 

 

 

 

 

 

 

 

 

15
ANP PROJECT QUARTERLY PROGRESS REPORT

The first production model of the
servo amplifier has been tested with
the OBNL Power Plant Simulator, and
it proved to be satisfactory. This
test was made with the actual com-
ponents, or duplicates, to be used 1in
the ARE. A memorandum describing the
on-0ff servo for the ARE 1is being
prepared.(7)

16

The wiring diagram for the control
system has been completed and working
drawings for the relay panels, con-
sole, and chamber installation have

-been issued to the field.

(T)E. R. Menn, J. J. Stone, and S, H, Hanover,

An On-Off Servo for the ARE, ORNL CF-52.11-228 (to
be published).
PERIOD ENDING DECEMBER 10, 1952

2. EXPERIMENTAL REACTOR ENGINEERING
H. W. Savage, ANP Division

Developmental work on components
for high-temperature dynamic fluid
systems has continued, with the effort
- being primarily on pumps, seals, heat
“exchangers, and instrumen tation.
Centrifugal pumps with both gas seals
and frozen seals have operated suc-
cessfully over the temperature range
of 1200 to above 1500°F, and operational
dif ficulties with these pumps have
- become less frequent. Techniques have
been developed for remotely stopping
~and restarting pumps 1lncorporating a
frozen-fluoride seal.

Emphasis has also been placed on
seal research 1n an effort to find
better materials for packing rotating-
shaft seals and valve-stem seals.
In the search for materials that will
~have negligible change in physical
properties at high temperatures and
will furnish some lubrication to the
- shaft at these temperatures, several
promising materials were found. A
. program is under way to develop a
face seal that will seal high-temper-
ature fluids against a gaseous atmos-
" phere, with the minimum leakage.

Various rotating-shaft seals for
containing high-temperature liquids
have been tested. To date, the best
results have been obtained with a
- packed seal in which the pumped liquid
~may be frozen. A program is under way
for developing a special type of seal
~in which a material different from
" the liquid being pumped and having a
higher melting temperature 1s injected
1nto the seal and frozen. Tests
conducted to date indicate that this
type of seal shows promise.

Heat transfer tests have been
conducted with the sodium-to-air
~radiator, and the sodium inlet temper-
" ature has been increased to 1700°F.
The radiator tested had 15 fins per
inch, which were sectioned every
2 inches. This gave a net improvement
~of 20% in the over-all heat transfer

coefficient, as compared with that for
an unsectioned radiator, A bi-fluid
heat transfer system has been con-
structed, and shake-down tests have
been conducted satisfactorily.

The rotameter type of flowmeter
that uses a tapered core riding inside
an induction coil as the float-position
sensing element has given very satis-
factory performance.’ This type of
flowmeter will be used in the ARE.
The Moore Nullmatic pressure trans-
mitters, modified for high-temperature
operation, have given trouble-free
performance during several hundred
hours of operation with pressure
transmitter temperatures as high as
1400°F. Pressures up to 70 psi were
measured.

Techniques for the vacuum distil-
lation of small guantities of Nak
fromliquid systems arebeing developed,
since all the NaK used for cleaning
the ARE fuel system cannot otherwise
be removed. Tests have continued for
determining the temperature-dependence
of gas line plugs above systems
employing a ZrF, -bearing fluoride
mixture. Vapor condensation traps for
this ZrF, vapor, in which the vapor
is bubbled through the eutectic
NaF-KF-LiF, have been satisfactorily
tested. _

Improvements made in the plant
facilities are the installation of a
helium distribution system, the
expansion of the fabrication shop
facilities, the completion of the
instrument shop facilities, and the
installationof six gas-fired furnaces.

PUMPS FOR HIGH-TEMPERATURE FLUIDS

W. B. McDonald G. D. Whitman

W. G. Cobb W. R. Huntley

A. G. Grindell J. M. Trummel
ANP Division

Pump with Combination Packed and
Frozen Seal. In the first test,
the pump with the combination packed

17
ANP PROJECT QUARTERLY PROGRESS REPORT

and frozen seal ran for 550 hr befare
1t had to be stopped because of a valve
failure in the loop. Average operating
conditions during this run were:

Pump suction pressure 5 to 15 psig
35 to 50 psig
1500 rpm

15 gpm

750 to 1050°F

Pump discharge pressure
Shaft speed

Flow

Packing temperature

Temperature at extreme
end of frozen seal

600 to 800°F

The packing of the pump was made up
of Inconel braid with mixtures of
graphite powder and nickel powder
between the layers of braid. The pump
shaft was coated with Stellite No. 6.

Operation of the pump was charac-
terized by frequent power surges of
the driving motor because of shaft
seizing in the seal area. Tricresyl
phosphate and lead glass were periodi-
cally injected into the seal as
lubricants, with indeterminant results,
Postrun examination of the shaft
showed very severe scoring in the
frozen seal region (0.036 in. on the
diameter of the 1 1/2-in, shaft).

The second run of this pump was of
600 -hr duration, Loop shutdown was
caused by leakage of the pump flange.
Changes made 1in the pump and in the
operating procedure before this run
wer e:

1. pump shaft resurfaced to remove
scoring of first run,

2. packing changed to nickel-foil
wrapped braid with graphite powder
be tween packing layers,

3. no lubricants injected.

Operation was much smoother than
that of the first run. This seemed
to be primarily the result of closer
temperature regulation of the frozen-
seal area. 1t was found that 750 to
BOO°F was the most satisfactory range
for relatively smooth operation.
lLeakage of solid fluoride from the
seal at these temperatures varied
approximately from 1 to 3 g/day. The
packing temperature was maintained at

1050°F.

18

A stop-start technique for the
frozen seal was developed during this
second run. Upon turning off the
motor power, the pump coasts freely to
a stop. However, the shaft becomes
immovable in approximately 30 seconds.
Therefore calrod heating is applied to
the frozen seal area and the tempera-
ature brought to an indicated 850°F,
The seal then loosens, and the operation
may be resumed with no leakage. Timing
1s important, since the application
of too much heat may result in heavy

seal leakage. A quick means of
cooling the seal, such as an air
blast, must be used i1f the frozen

seal is completely removed,

A second Durco pump has been
modified to i1nclude a combination
packed and frozen seal. This pump has
accumulated 84 hr of operation at
1200°F and flow rates up to 60 gpm.

Operation of this pump has been
very difficult because of unexplained
sei1zing 1in the seal area. Temperature
rises along the seal area indicate
that the taking place 1in
the packing area of the seal. This
cannot be definitely concluded,
since the shaft has not been
examined in that area.

Some stop-start data were obtained
during one of the runs. It appeared
that a very definite temperature range
(750 to 820°F in this case) was needed
for remote loosening of the shaft and
subsequent startup without leakage,.
The start-stop procedure was repeated
five times, and successful operation
followed each trial, However, this
performance could not be duplicated
in later trials. Part of this incon-
sistency 1is thought to be caused by
the chip retainer, which 1s being
removed for the next run.

seizing 1s

however,

One other change 1s being made in
the seal geometry in an attempt to
obtain more consistent operation.
The radial clearance between shaft
and wall in the frozen seal region is
being increased from 0.010 to 0,040
in. to more closely duplicate the
design of the first combination seal
"pump, discussed above, which ran
satisfactorily. ‘ :

Pump with Frozen-Sodium Seal.
Approximately 4000 hr of operation
had been accumulated when the test
of the pump with the frozen-sodium
seal had to be terminated because of
leaking flanges and failure of a
thermocouple well. The final run
was of 1000-hr duration at a pump
temperature of 1100°F, '

Postrun examination of the pump
showed no measurable wear of the
shaft in the frozen-seal area. Scoring
of the rear face of the impeller,
- which indicated inter ference with the

housing, was very severe and was
probably the source of pump noises
during the last run. ' :

Allis-Chalmers Pump. During this
quarter the Allis-Chalmers pump that
has a hydrostatic bearing was tested
with NaK as the system fluid. When
this pump was first tested, the shaft
gas seal provided by Allis-Chalmers
was used. The seal consisted of three
Graphitar rings, each of which had
three 120 ~deg segments. The room-
temperaturediametral clearance between
shaft seal rings was designed to be
between 0.0065 and 0.008 inch. At
operating conditions (approximately
170 gpm, 45-ft head, and elevated
temperature), the diametral clearance
and the gas leakage to atmosphere past
these rings were to approach zero.

In the first test, the gas leakage
past the seal was excessive, and a
large amount of gas was entrained in
the system fluid. Therefore two
modifications were made to the pump:
a rotary face seal was added to
decrease gas leakage to atmosphere,
and additional baffling, suggested by
Allis-Chalmers, was placed in the
region above the pump impeller to
reduce gas entrainmentin system fluid.

The second test revealed that the
additional baffling did materially
reduced gas entrainment in the system
fluid and also that the rotary face

PERIOD ENDING DECEMBER 10, 1952

seal reduced gas leakage toa tolerable
level. The unlubricated rotary face
seal, however, squealed and chattered
aftera very short period of operation.
Further testing indicated that the
seal operating temperature was sensitive
to changes in pump suction pressure,
a condition that may be overcome by
using a balanced bellows seal. These
tests indicate that this pump will
give satisfactory performance when a
good gas seal is incorporated in the
design. A frozen-sodium gas seal 1is
presently being fabricated for test
in this pump.

Laboratory-Size Pump with Gas Seal.
The laboratory-size centrifugal pump
with a gas seal, described previ-
ously,(1) has been operated for a
total of 1275 hr with the fluoride
fuel NaF-ZrF -UF, (46-50-4 mole %) at
temperatures up to 1500°F. The average
fluid temperature during this period
was approximately 1350°F. The satis-
factory performance data obtained are
presented in Fig. 2.1. The pump
discharge pressure and pressure drop
across the venturi were measured by
means of the null-balance level-
indicator systen.

The initial difficulty of liquid-
level detection in the pump bowl at
high shaft speeds has been overcome.
The turbulence in the liquid surface
has been reduced by additional baffling,
and the action of the probes used
for level detection has been quite
satisfactory up to shaft speeds of
4400 rpm.

The rotary gas seal ~ Graphitar
No. 30 lubricated with spindle o0il
running against hardened tool steel -
has functioned with ‘extremely small
gas leakage and without maintenance
or appreciable wear during the total
running time. Oil leakage past the
seal assembly was collected, and 1t
was found to amount to less than
1 ¢m?® per 24-hr period. The only

(1y. 6. Cobb, P. W. Taylor, and G. D. Whitman,
ANP Quer. Prog. Rep. Sept. 10, 1952, ORNL-1375,
p. 1lé.

19
ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 2.1,
mole %) at 1300°F.

mechanical failures in this pump have
been broken guide vanes in the dis-
charge case. These vanes were origi-
nally held in place by small tack
welds, which have now been reinforced,
and additional welds have been added.

20

M
20 ] DWG. 17439
DENSITY ~ 195 Ib/ft3
80 - — -
4300 rpm
-‘MN
M""""—-_
Mr_"“-«.
I,
"? O ....... A — . I
4000 rom
e
iy
6 O I R e —
50
2 ~—
o
b=,
T
40
\g
30 - IR R 30
|
EFFICIENCY AT 3600 rpm
2200 rpm /‘ ------ i
20 e b 20
L~ ™ o
T T &
S
. .
{0 / 10 W
////f//? 1250 rpm
0 z’// 0
O 2 4 & 8 10 12 14 16
FLOW (gpm)

Performance Curves for Centrifugal Pump with NaF-ZrF, -UF, (46-50-4
Density of fluoride mixture is about 195 1b/ fe 3.

A second pump has been fabricated
and installed in the fluoride loop
of a bi-fluid heat exchanger system.
No radical changes have been planned
in the present design. However, some
effort has been put forth to simplify
the construction of the shaft and
radiation heat shields.

ARE Pump. It is now planned that
the pumps for the ARE fuel circuit are
to be similar to the ANP experimental
engineering model FP. Figure 2.2
shows an assembly-section view of the
pump. A standard commercial pump base
with ball bearings that support a
horizontal, overhung shaft provides
the foundation., The pump casing is
supported through the sealing head and
casing extension. The sealing head
1s sealed to the pump casing with a
commercial oval-ring gasket in a
bolted flange joint, By removing this
Joint from the hot region near the
pumped liguid and away from regions
requiring preheating, satisfactory
sealing 1s readily accomplished. '

Pump suction and discharge con-
nections to the circulating system
are made by welding. The pump impeller
is an Inconel casting from standard
Worthington Pump Co. The
discharge case 1s machined from heavy
Inconel forgings and has a concentric
volute and a welded tangential dis-
charge. The impeller is double-keyed
to the overhung shaft and i1s held
in place by a thrust nut and a locking
screw. Normal radial clearances are
employed between the impeller hubs and
the sealing labyrinths, and an axial
clearance of 5/32 in. 1in either
direction prevents jamming resulting
from differential thermal expansion,
The shaft seal is of the frozen-packed
design; the gland for the stuffing
box provides the region for the frozen
seal. To prevenl cracking or checking
of the hard surface material on the
shaft, the material is applied as a
loose-fitting, prefabricated sleeve,
The concentricity between seal and
shaft and the differential expansion
are controlled by angle-beveling the
inside of both ends to the geometrical
angle of the sleeve.

A cartridge type of electric heater
~1s provided in the center of the shaft
for control of temperatures. A turbo-

patterns.

PERIOD. ENDING DECEMBER 10, 1952

slinger arrangement is provided just
outboard of the seal gland to maintain
a suitable temperature on the oil
seals and bearing. Two independent
heaters are provided on the outer
surface of the seal shell for further
control of the temperatures in the
seal region. Provision is made for
the circulation of a coolant through
selected portions of the shaft. Circu-
lation of lubricating o1l through
the bearing housing 1s provided to
minimize effects of radiation damage
to the lubricant.

The pump (Fig. 2.2) provides a
seal for NaK for system cleaning by
maintaining molten sodium in the gland
groove and freezing on both sides of
the annulus. During operation with
fluorides, this annulus will serve as
a shield to prevent the loss of
enriched material by guiding the
expected slight leakage to a con-
tainer. The coolant for freezing the
sodium is to be kerosene at approxi-
mately S0°F, A beveled surface on the
gland makes a metal-to-metal s'eal
against the inner edge of the seal
shell to form a NaK-tight seal around
the outer surface of the gland. This
frozen-sodium seal also seals the
system during vacuum removal of the
NaK. During operation with NaK, the
cartridge heater is removed from the
center of the shaft, and a squirt-
tube arrangement is inserted to circu-
late cooled kerosene.

The only alteration necessary to
make the pump suitable for circulating
NaK as a moderator and reflector
coolant is the elimination of the
shaft sleeve and the seal heaters.

ROTATING SHAFT AND VALVE STEM
SEAL DEVELOPMENT

McDonald -R. N. Mason -
Tunnell P. G. Smrth
- W. R. Huntley

ANP Division

W. B.
w. C.

Screening Tests forPacking Materials

and Lubricants. Work has been done

21
GG

UNCLASSIFIED
DWG. 17440

SHAFT SLEEVE

GLAND GROOVE SEALING HEAD

SEALING GLAND
PAGKING

   
  
  

TURBO-SLINGER
.~ TUBULAR HEATERS

COMMERCIAL PUMP BASE
CASING

 

IMPELLER

 

 

 
 
 

'}« \m\xw\wm\\m\“w“\“w.\ .\\M\M i)
A
= b~

s
i

 

 

 

 

 

 

 

 

T DiSCHARGE

   
 

 

 

CARTRIDGE HEATER

 

 

Fig. 2.2. Assembly of ARE Pump. Sectional view.

~
L

904d ATYALYVN0 LOIdl0¥d dNV

LY0dAd SS
on packed seals for pumps and valves,
and screening tests have been made on
a number of packing materials for
possible use as lubricants. The test
consisted of heating the material
under compression to 1500°F for a
period of time and then comparing the
heated material with the original
material. The materials tested
included C (graphite), PbO, Ni,0,,
Ni,0, + PbO, PbO + Mo, Ni,O, + Mo,
CrF,, Ni,0, +C, PbD + C, Pb,(OH),CrO,,
MoS,, Ni, 0, +MeS,, MeS, + Mo, BN+ MoS,,
BN, Ni,0, + BN, BN + C, ZnS, CeQ,, and
CaF,. The tests indicated that the
following materials may possibly be
useful as high-temperature lubricants:
C (graphite), N1203,~Ni203 + Mo,
Ni203 + C, MoSz, MoS2 + Mo, MoS2 + BN,
and BN + C. CeO, may be a fine
abrasive. It 1s planned to further
test these materials in contact with
fluorides,

One test apparatus for screening
materials has been built and operated
that consists of a fluoride pot with
stuffing boxes on the top and bottom.
A single shaft is inserted through
both the stuffing boxes, and the
material to be tested 1s conhtained in
the spring-loaded glands. The shaft
1s reciprocated by an air cylinder,
and the power requirements can be
measured. Graphite and MoS, have
been vested at 1000°F in this device
with no fluoride leakage out of the
stuffing boxes. However, the fluoride
wets the surface of the shaft and 1is
carried i1into the packing where it
causes the shaft to bind with the
packing, and the power reguirement
becomes excessive. Mixtures of
graphite and MoS, with powdered
fluoride have also been tried in
these glands, but there was no fluoride
in the pot. The results were similar
1n that the powdered fluoride melted
and caused the gland walls and the
- shaft surface to bind. _

One other test device that is being
assembled has 1dentical packings
located at each end of a lantern gland

PERIOD ENDING DECEMBER 10, 1952

in a heated container with a rotating
shaft through the assembly. The
equipment is designed so that liquids,
such as fluorides or lubricants, can
be introduced under pressure into the
lantern gland. This assembly 1is
mounted so that spring tension prevents
rotation, and the torque required for

operation can be measured.
Packing Compression Tests, As

reported previously,(?? experience
with high-temperature packing materials
indicates thatin almost every instance
stuffing boxes have required re-
tightening after they reached operating
temperature. Therefore a series of
tests have been run onvarious materials
to determine their compression charac-
teristics. A dial indicator for
measuring the expansion or contraction
of the material being tested was
mounted on the test apparatus. In
the tests, a l-in.-thick layer of the
material being tested was subjected
to a compressive force of 75 psi and
the compression was measured. The
material was then heated to 1500°F
and held at that temperature until
the dial indicator showed no further
change. After reachingthis equilibrium
condition, the material was held at
1500°F for a further period of 1 hr,
and the compression was again measured.
The change in compression, that is,
the difference between the compression
before heating and the compression
after heating, 1s given in Table 2.1.
This test procedure was repeated
several times for each material. In
all tests thus far, with the exception
of the test of nickelic oxide, there
was little, 1f any, further dimensional
change after the initial heating.
Packing Penetration Tests. A
series of tests has been run to
determine the wetting susceptibility
of various materials. The materaal
to be tested is compressed by a heavy
washer and screw in a contailner.

(2)9. R. Ward, H. R. Johnson, aend R. N. Mason,
ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL-1375,
p. 19.

23
ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 2.1. COMPRESSION TESTS OF VARIQUS PACKING MATERIALS
AT 1300°F AND 75 psi

 

 

 

REMARKS

 

COMPRESSION DUE TO
- HEATING
MATERI AL THERMAL EFFECT
CYCLE
(in.)
Graphite 1 0.038
0.009
3 0.005
Boron nitride 1 0.071
2 0.003
3 0.003
4 0.002
Nickelic oxide 1 0.012
2 0.004
3 0.003
4 0.002
Molybdenum disulfide 1 0.058
2 0.008
3 0.007
4 0.001
5 0.002
6 0.000
10% Graphite plus 1 0.165
90% nickel powder
(by weight)
25% Aluminum powder 1 0. 165
plus 75% iron
powder (by weight)

 

 

 

Material was easily removed from the apparatus
and was apparently unchanged.

Material was easily removed from the apparatus
and was apparently unchanged.

Material was easily removed from the apparatus
although a large force (4 tons) was required
to pusE the stem through the packing material
in order to disassemble the apparatus. The
material contracted several thousandths of an
inch when cooled after each heating period.
Upon the fourth heating, thematerial expanded
sTightly.

Material was easily removed from the apparatus
and was apparently unchanged.

Material had to be chiseled from the apparatus.

Material had to be chiseled from the aspparatus.

 

Fluoride 1s introduced above the
material being tested, and gas pressure
1s applied to drive the fluoride
through the material.

Graphite, when compressed, was not
gas tight, and when the fluoride was
loaded there appeared to be a definite
interface between the fluoride and
graphite, which indicated that the
graphite was not wetted by the fluoride.
The test ran 240 hr with a pressure of
30 psi above the fluoride.

Boron nitride, when compressed,
also was not gas tight; however, after
the fluoride was loaded the test ran
26 hr with 5-psi pressure above the

24

fluoride without leakage of the
fluoride i1nto the boron nitrade.
The pressure was then increased to
10 psi and leakage began. There was
some greenish color a1n the boron
nitride. The main leakage path of the
fluoride appeared to be along the
walls of the container and the screw
stem toward the hole 1n the bottom of
the container for the screw stem.
Stainless steel braid was compressed
and heated to the annealing temperature
twice and further compressed after
each heating. The material permitted
free flow of gas before 1t was filled
with fluoride. ITmmediately after
filling, fluoride leakage occurred at
only 1-psi pressure. The leakage soon
stopped, and the gas flow was somewhat
retarded. When the pressure was
increased £o3 psi, most of the fluoride
was forced from the container., Exami-
nation showed that only a small amount
of fluoride had adhered to the braid.

Attempts were made to run tests of
molybdenum disulfide and nickelic
oxide, but these materials were so
“fluid” that the container would not
hold them. Parts with smaller clear-
ances are being fabricated. ,

Face Seal Tests. Face seals are
being considered for high-temperature
use with both gases and liquids. All
tests have been run in air at room
temperature with speeds of about
1600 rpm and pressures of up to 14 psi.
The following combinations of materials

were tested 1n air, without added
lubrication:
Carboloy (gréde 779} vs. Carboley {(grade 779)

Carboloy {grade 779) vs. graphite (grade C-18)

Graphite (grade C-18) vs. graphite {grade C-1§)
In the test of Carboloyvs. Carboloy,
a loud, grinding noise developed
within 5 min, and inspection showed
that annular grooves had been worn
into each surface. Operation with the
other two combinations of materials
was satisfactory until chattering
developed when a temperature of about
550°F was reached at the seal because
of the heat of friction. If graphite
depends upon a layer of tightly held
moisture for some of its lubrication

PERIOD ENDING DECEMBER 10, 1952

qualities,(3*4) then this chattering

may be due to the driving-off of
surface-held moisture at a more rapid
rate than 1t can be supplied by
diffusion from the interior of the
graphite. The time required for the
test seals to reach the chattering
stage was usually between 1/2 and
4 hr; however, in one test the Carboloy
vs., graphite seal operated satis-
factorily for 42 hours. The area of
seal contact was 1 1/2 in.? in each
test, and the seal contact force was
varied from 5 to 20 pounds. Similar
results were obtained when Carboloy
vs. graphite seals were tested with
MoS, and RbOH used as lubricants.
Although some seals showed relatively
little chattering, therewas, generally,
a marked increase in the torque
required to overcome the seal frlctlcn
as the seals heated up.

Combination Packed and Frozen Seal.
A seal consisting of a stuffing box
packed with powdered graphite in which
the slight leakage of fluoride was
frozen in the packing compression
member was operated for a period of
215 hr at temperatures from 970 to
1500°F with system pressuresthat
varied from 16 to 37 ps Leakage
rates were measured at_various temper ~
atures and pressures in an attempt to
determine the most favorable operating
conditions.

(E)Pure Carbon Cempany, Inc., Properties of

Pure+Bond, Bulletin No. 52, p. §
Y)R. H. Savage, J. Appl. Phys. 189, 1-10
(1948). |

 

 

 

TABLE 2. 2. LEAKAGE RATES FfiR’COMBINATION PACKED ANP FROZEN SEAL
SYSTEM PREEEE&E TEMFERATURE AVERAGE LEAKAGE LEAST AMOUNT OF LEAKAGE
(psi) ( °F) (g/hr) (g/hr) |
16 1150 1 | 0.15
30 1200 1 0,4
37 1200 3 0.4

 

 

 

 

25
ANP PROJECT QUARTERLY PROGRESS REPORT

During this test there was no
indication of shaft seizing by the
seal. Spectrographic analysis of the
leakage showed the constituents to be
Co, B, Cd, Cr, Fe, K, Ni, and Si, as
well as Na, U, and Zr. Chemical
analysis of the leakage indicated that
the Na, U, and Zr constituents were
present in very nearly the same
proportions as in the fluoride in the
system. A second test, with a similar
packing, in which the compression
member 1s spring loaded to keep tension
on the graphite at all times has been
in operation for approximately 350 hr
with only small leakage of solid
fluoride and with no operational

difficulties.
¥rozen-Lead Seal. A test, similar

to previous tests with sodium, was
conducted to determine the operating
characteristics of a frozen seal when
lead is used as the system fluid.
The equipment consisted of a chrome-
plated stainless steel shaft rotating
in a finned sleeve in which the frozen
seal was formed. A pot to contain
the hot lead to supply the seal was
attached to the finned section. The
finned sleeve was provided with a
heater for melting out the frozen
seal 1in order to resume operation
after a shut-down. A portion of the
finned section was shrouded so that a
blast of air could be directed across
the seal when required. This test
operated for a period of 526 hr with
leakage occurring when the finned
section temperature was permitted to
exceed 600°F. The lead temperature
in the pot was between 1000 and 1200°F,
and the system pressure was 8 ps1.
Very smooth operation was experienced,
with no tendency toward shaft seizure.
The test was terminated when the seal
became too hot and caused an excessive
loss of lead. This test 1indicates
that a lead pump with a frozen-lead
seal can be expected to give as satis-
factory performance as a sodium pump
with a frozen-sodium seal.

Frozen-Sodium Seal. Preliminary

26

tests have been conducted to determine
the feasibility of using a frozen-
sodium seal for sealing a high-temper-
ature NaK pump. This design consists
of a heated annulus around the shaft
to which molten sodium 1s supplied
from an attached container. Sodium 1in
the annulus i1s kept molten at all
times. A coolant passage 1s located
on either side of the molten sodium
for freezing the sodium around the
shaft and preventing sodium leakage
from the annulus. The temperature of
the NaKis reduced to below the melting
point of the sodium at the frozen
sodium-to-NaK interface. In the
initial tests, some difficulty has
been encountered with heat conduction
by the solid shaft to the region of
the frozen-sodium seal that 1s suf-
ficient to melt out the sodium forming
the seal when the NaK bath 1is at
temperatures greater than 500°F;
however, 1t 1s expected that internal
cooling of the shaft will eliminate
this difficulty.

These tests indicate that such an
arrangement will satisfactorily seal
the rotating shaft of a NaK pump 1f
the temperature of the NaK at the
NaK-to-sodium interface 1s below the
melting point of the sodium and
sufficient internal cooling of the
shaft 1s provided to prevent the
melting of the frozen-sodium seal.
This design 1s being incorporated
into the pumps for the moderator-
cooling circuit of the ARE and also
in the pumps for the 1nitial circu-
lation of NaK through the ARE {fuel
circult.

Bellows-Type of Valve Stem Seal
(G.M. Adamson, R.S. Crouse, Metallurgy
Division). Two, 3-ply, Inconel
bellows 2 in. in diameter and 3 1in,
long were examined for the Experimental
Engineering group. The bellows had
been fabricated at the Fulton Sylphon
Co. by using a special rolling technique.
The first one had been i1mmersed in the
fuel mixture NaF-ZrF,-UF, (46-50-4
mole %) for 975 hr and the second one
for 1975 hr at 1500°F.¢%> The first
bellows examined had fluorides between
the outer and center ply and the
second one, which was full of fluorides,
had failed completely.

The most important observation made
during this investigation was that the
wall thickness of the bellows was
much less than that expected. Instead
of the thickness being 0.027 in.,
the average thickness was only 0.022
inch. What was even more serious was
that the bends were even thinner. In
the first bellows, the bends had been
thinned an additional 34% and in the
other 21%. In the first bellows,
several spots in the outside ply were
less than 0.002 in. thick. |

The first bellows probably failed
through one of the thinned down areas
in the outer ply. No definite cause
for the complete failure of the second
bellows can be given. When examined,
this bellows contained many circumfer-
ential cracks. Although it i1s certain

that some of these cracks were made

after the test, 1t is probable that
some of them were fabrication cracks
and were the cause of the failure.
In both these bellows, considerable
self-welding between the plies was
found at the bends. _

The manufacturer was contacted, and
two similar bellows were obtained for
examination in the "“as received”
condition. Two other 3-ply bellows
1 3/8 1in. in diameter and 6 in. long
were also sent for examination. The
two 2-1n. bellows are even thinner
than those discussed above. Both
these bellows had an average thickness
of 0.018 inch. The plies were more

nearly uniform and the thinning in the
- bends was 16.7 and 12.5%. The two
smaller bellows also had an average
thickness of 0.018 inch. The reductions
- found 1in the bends of these two were

16.7 and 11.1%. No self-welding or
fabrication cracks were found in any
of these bellows.

Sp oy Taylor, ANP Quar. Prog. Rep. Sept. 10,
1952, ORNL-1375, p. 19.

PERIOD ENDING DECFMBER 10, 1952

The manufacturer is now attempting
to fabricate 4~-ply bellows to obtain
the desired wall thickness. The
individual plies will be as thick as
those of the bellows described above,
or thicker, if posszble

HEAT EXCHANGERS

G. D. Whitman D. F. Salmon
ANP Division
Sodium-to-Air Radiatoer. The third

sodium-to-air radiator performance
and endurance test was started during
the quarter and ran for 1212 hr before
failure., The running time at the
varioussodium inlet temperaturee is
given in the following:

TEMPERATURE RUNNING TIME

(°F) (hr)
700 - 24
900 48
1100 . 72
1300 34
1500 _ 655
1600 336
1700 43

Total 1212

Failure occurred at a sodium inlet
temperature of 1700°F, and the resulting
fire did very little damage because
the system was dumped as soon as the
leak was indicated by a heavy smoke
discharge from the air duct. No fire
fighting was necessary, and the sodium
outside the system was allowed to burn
1tself out. _

Again, as 1in both of the previous
tests, a center tube between the
header and top fin on the sodium inlet
side of the radiator failed. The
tubes 1in this radiator were fabricated
from type 316 stainless steel and had
0.025-2n. walls in contrast to the
type 304 stainless steel tubes with
0.015-1n. walls used in the previous
designs.

Since all three failures occurred
in the same location, 1t is possible

27
ANP PROJECT QUARTERLY PROGRESS REPORT

that a localized stress was set up in
the shorter tubes between the header
and the fin matrix during the brazing
operation. Such stress may be relieved
by increasing the spacing between the
headers and top fin.

Considerable difficulty was en-
countered after about 400 hr of
operation of this test in keeping the
radiator tubes free fromoxide buildup,
which would have eventually caused
plugging. The plugging that occurred
in the cooler sections of the radiator
could be temporarily relieved by
shuttingoff the air flow and operating
under isothermal conditions, at 1200°F,
for approximately 1 hour. After
normal operation resumed, the
plugging would start again, and after
several hours a drop 1a heat transfer
performance would be observed.

A by-pass filter circuit was
designed into the loop to prevent such
oxide plugging, but 1t was rendered
almost inoperable becausecf inadequate
flow regulation. A small orifice was
used to reduced the flow through the
filter caircurt, but the orifice
plugged easily at the low flow rates
necessary for lowering the fluid
temperatures in the filter.

A throttling valve was substituted
for the orifice, and the fluid entering
the filter was held at 600 to 800°F
below the main stream temperature,
Approximately one-eighth of the total
system flow was continuously passed
through the filter, and the radiator
was cleaned sufficiently to rumn about
500 hr without serious tube plugging.
When the temperature of the sodium
entering the radiator was raised
to 1700°F, some tube plugging started
again and had not been satisfactorily
cleared before the failure.

The radiator was of the same
general design as that of previous

was

models, and the fin spacing was 15
fins per inch. The fins were cut
through, interrupted, and slightly

in the direction
The over-all coefficient

of fset every 2 in.
of air flow.

28

of heat transfer 1s shown plotted
against mass air flow rate per square
foot of face area (Fig. 2.3). An
improvement of about 20% in the over-
all heat transfer coefficient was
achieved over the continuous fin
exchanger, as shown by the dashed
line. The accompanying increase 1in
air-pressuredrop was between 10 and 15%.

Bifluid Heat Transfer System.
Assembly of the bifluid loop reported
previously(®) has been completed. The
NaK system has been filled and the
electromagnetic pump has been checked.
Various minor changes were necessary,
such as correcting reversed thermo-
couples and clearing gas lines.
Considerable effort was required to
get the electromagnetic pump started.
The trouble, however, stemmed entirely
from nonwetting or uncleanliness of
the pump cell. When the NaK was
heated 1n the sump tank so that the
pump cell temperature reached 450°F,
the pump started i1immediately. A flow
rate of 11 gpm at 250 v input to the
pump primary coil was achieved. This
pump is a G-E model G-3, with a modified
cell that has nickel lugs welded into
the sides of a type 316 stainless
steel throat.

The electromagnetic flowmeters will
be calibrated against the NaK venturi
to complete the shake-down of the NaK
system.

The fluoride system has been filled
with the fluoride fuel, NaF-ZrF4-UF4
(50-46-4 mole %), and check runs are
being made before combined operation
of the two-fluid systems 1s undertaken.
The gas seal of the fluoride pump was
operated dry up to 5000 rpm to check
for gas or o1l leakage, but none was
apparent. Vibration of the pump and
support structure at the high speed
was slaght,

Dynamic corrosion data of interest
to the ARE will be obtained in the
heat exchanger of this loop (a small-
diameter nickel tube). A temperature

(G)D. F. Salmen, ANP Quar. Prog. Rep. Sept. 10,
1952 OBNL-1375. p. 21.

 
PERIOD ENDING DECEMBER 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ORI
DWG. 17441
20
;h w////
: olw
o 0 © -~
= 10 ~ _—
C " | © z
: i
™~ -~
= P
5 Q’y// ,-K
o >
- _ -~ CONTINUOUS FINS (TEST NO.2)
= -
w -
(_2 .
- 5 ’,/'
wl /,"
Q -~
O il
o Na INLET TEMPERATURE (°F)
g o 700
1 ® 900
- A 1100
I...-
= ¥ 1300
L O 1500
-l e 8 1600
< m {700
i
!
=
O
1
1000 2000 5000 10,000 Z0,0QO
MASS VELOCITY (Ib/ hr -ft2)
Fig. 2.3. Heat Coefficients of Sodium-to-Air Radiator with Interrupted Fins.

drop of the fluoride passing through
the nickel tube in excess of 100°F
and Reynold’s numbers above the laminar
'range should be achieved. The 1impeller
"of the fluoride pump was fabricated
from Inconel to compare its corrosion

‘resistance with that of a type 316

stainless steel 1mpeller that was

used in another loop with an identical

Spump.

INSTRUMENTATION

W. B. McDonald P. G. Smith
P. W. Tavylor A. L. Southern
J. M. Trummel
ANP Division
Rotameter Type of Flowmeter. It
was reported previously¢(’?? that a

 

(U)p, G. Smith and A. L. Southern, ANP Quar.
Prog. Rep. Sept. to, 1952, ORNL-1375, p. 23.

29
ANP PROJECT QUARTERLY PROGRESS REPORT

rotameter type of flowmeter had been
installed in a high-temperature
fluoride loap and that 1t had not
operated successfully. The difficulty
encountered with the flowmeter was
binding of the tapered-core position
indicator in its housing as a result
of thermal distortion. The details of
construction of this flowmeter and 1ts
principle of operation were given
previously.(7)

Other such flowmeters have been
constructed with increased clearance
around the tapered core to prevent
binding. One has operated for a
period of 600 hr and measured flows up
to 17 gpm at 1300°F. The temperature
of the static fluoride in which the
tapered core operates is 1100°F. Flow
measurements have been accurate to
within 10%. It was found that such
accuracy can be maintained for long
periods 1f close control is maintained
over the temperature of the fluorides
in which the tapered core 1s operated.

A similar flowmeter, with a capacity
to 60 gpm, has been installed and is
operating successfully 1n a larger
test loop. This 1instrument can be
adapted for ARE use.

Rotating-Vane Flowmeter. The
rotating-vane flowmeter manufactured
by the Potter Instrument Co. and
modified by the Experimental Engineering
group, as reported previously,(®) has
been installed in a high-temperature
loop. Attempts to measure flow with
this instrument have been unsuccessful.
Under loop operating conditions, no
signal that can be calibrated in
terms of flow can be obtained from
the flowmeter., One difficulty ex-
perienced i1s that the loop heater
circuits cause 1nduced voltages in the
pickup coil that are greater than the
signal expected from the flowmeter;
however, when the heater circuits are
turned off for short periods, no
useful signal 1s detectable with the
present i1nstrumentation. It 1s

(8)y. B. McDonald and J. M. Trummel, ANP Quar.
Prog. Rep. Sept. 10, 1952, ORNL-1375, p. 24.

30

thought that thermal distortion at
this temperature (1300°F) caused the
bearings to seize and render the
instrument inoperable, This instrument
will not be available for examination
until tests with the loop in which 1t
has been placed have been terminated
and the loop has been dismantled.

Diaphragm Pressure-Measuring
Device. The diaphragm pressure-
measuringdevice reported previously{®>
which utilizes a linear differential
transformer to measure the deflection
of a diaphragm, has been i1in operation
for more than 100 hr at temperatures
up to 1040°F., The diaphragm in this
instrument is made of 1/16-in.,, flat,
Inconel stock 3 in. in diameter.
Preliminary tests indicate that the
operation is reliable up to 1000 °F
over the range from 0 to 40 psi. To
date, the instrument has been tested
only with inert gas against the
diaphragm; however, it 1s not expected
that the introduction of molten
fluorides will radically change the
operating characteristics of the
instrument, The only observable
effects of 1ncreased temperature are a
shift of the zero point and an increase
in the slope of the performance curve
(Fig. 2.4). A second instrument 1is
being fabricated that will operate
with fluorides at 1000 to 1100°F
against the Inconel diaphragm.

Moore Nullmatic Pressure Trans-
mitter. The Moore Nullmatic pressure
transmitter reported previously,(19)
which was modified to prevent the
fluorides from coming 1in contact with
the bellows pressure-sensing element,
has been installed in a high-tempera-
ture fluoride loop and has operated
successfully for approximately 200
hours. The pressure-transmitter
temperatures have ranged as high as
1400°F, and pressures have been
measured up to 70 psi. A slight zero
shift has been encountered with

9y, W. Taylor, ANP Quar. Prog. Rep. Sept. 10,
1952, OBRNL-1375, p. 25.

S TYP)
PERIOD ENDING DECEMBER 10, 1952

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 1744

580 N DWG .2
. 240
%
LLk
tl
o
o
i >
D -
z 200
o
<
o
T
o
3 160 0
« 5
e ok
= QO
Q
5 120 :
g
-
L
L)
o
=
S 80 .
v
T
o
<
a

4C¥

0 : L ,
0 10 . 20 : 30 40
GAGE PRESSURE (psi)
Fig., 2.4. Performance Curves for Diaphragm Pressure-Measuring Device,

increased temperatures, aswas expected.
Accurate temperature control of the
instrument should eliminate this
source of inaccuracy. , _

Static corrosion tests of 3-ply
Inconel bellows 1n the fluoride
mixture NaF~ZrF4-UF4 (46-50-4 mole %)
at 1500°F indicate that such bellows
may be incorporated in the Moore
Nullmatic pressure transmitters and
operated with the high-temperature
fluorides 1in contact with the bellows.
Six such instruments, rated at Q0 to

100 psi, have been ordered from the
manufacturer.

It is also possible that an instru-
ment constructed with single-ply
type 316 stainless steel, type 347
stainless steel, or nickel bellows may
be used with the ZrF,-bearing fluoride
fuels, sinceit is possible to maintain
the fluid in static condition and
to keep the temperaturein the pressure
transmitter between 1000 and 1100°F
and thus minimize the effects of
corrosion. (Results of preliminary

31
ANP PROJECT QUARTERLY PROGRESS REPORT

static corrosion tests of stainless
steel, Inconel, and nickel 1in these
fluorides indicate little corrosion.)

The possibility of locating the
instrument beneath the loop and
completely submerging the bellows 1in
a bath of molten lead with a tempera-
ture gradient in the lead from 1100°F
at the top to Just above the melting
point at the bellows is being in-
vestigated. This would provide a
molten lead - molten fluoride interface
that would make possible pressure
transmittal from a 1500°F fluoride
stream without exposing the bellows
to damaging temperatures. Static
tests have indicated that the fluoride
mixture will remain on top of the
lead and will not mix with the lead.

HANDLING OF FLUORIDES AND LIQUID METALS

L. A, Mann D. R. Ward
J. M. Cisar P. W. Taylor
W. B. McDonald
ANP Division

NaK Distillation Test. The use of
cutectic NaK alloy (78 wt % K, 22
wt % Na) for cleaning the ARE {fuel
circuit 1s planned. It is anticipated
that about 95% of the NaK in the entire
fuel circuit can be drained and the
remainder will be removed by vacuum
distillation under heat, An experi-
mental unit has been constructed, and
the first test of vacuum distillation
of NaK has been conducted. 1In this
test, approximately 6 1b of eutectic
NaK at 1200°F was distilled from a
boiler by wusing a small vacuum pump
that pulled continuously from the
receiver and maintained a pressure of
less than 2 mm Hg. A steady flow of
approximately 3 scfmof helium was
maintained through the system. This
operation continued intermittently for
approximately three days; however,
there was some down time on the system
because of gas leaks and other equipment
failures. This test indicated that
under attainable operating conditions,
this quantityof NaK could be distilled

32

from a boiler in somewhat less than
16 hours.

Gas-Line Plugging., It was reported
previously(1!) that continued tests
were being conducted to determine the
extent of gas-line plugging from the
vapors of a bath of the fuel mixture
NaF-ZrF4-UF4 (46 -50-4 mole %) at
various elevated temperatures. A
summary of the results of the tests,
to date, follows:

1. In operation at 1050°F, there
wasno gas-line plugging during 1488 hr
of testing.

2. 1In operation at 1300°F, all
lines partly plugged after 666 hr of
operation. Upon examination, the
lines appeared to be totally plugged,
but the plugs were sufficiently porous
to allow the passageof gas at approxi-
mately onc-tenth that of the initial
flow rate.

3. In operation at 1500°F, all
lines plugged solid ain 100 hr of
operation.

These tests indicate that gas
lines may be expected to operate
satisfactorily in a system contailning
this fuel, providing the free surface
of the fuel adjacent to the entry gas
lines 1s at a temperature in the range
between 1050 and 1100°F. Above these
temperatures, some difficulty may
be expected.

ZrF, Vapor Condensation Test,(12)
In the absence of reliable vapor-
pressure data for ZrF,, tests were
conducted to determine the minimum
container-wall temperature needed to
minimize or prevent condensation of
ZrF, on the walls. Each of four
1-in.-IPS tubes 18 1n. long was
charged with approximately 180 g of
the fuel mixture NaF-ZrF4-UF4 (46 -50 -4
mole %). The fuel was maintained at
1500°F in each tube, and the portions
above the fuel tubes were maintained
at different temperatures, ranging

______ for each tube.

(Il)w. B. McDonald and P. W. Taylor, ANP Quar.
Prog. Rep. Sept. 10, 1952, OBRNL-1375, p. 32.

(12)\\(. B. McDonald and J. W. Trummel, ANP Quar.
Prog, Rep. Sept. 10, 19592, ORNL-1375, p. 34.
' These conditions were maintained for
300 hours. The tubes were then
sectioned and examined for crystal
~deposits of ZrF,. Correlation of
“crystal deposits on the inside tube
walls with the various tube-wall
 temperatures during the test indicates
that the ZrF, vapor given off by the
fuel at 1500°F will condense on any
" surface that is at a temperature of
1250°F or lower. When the tube-wall
temperature 1s maintalned at a temper-
ature greater than 1250°F, no deposits
of ZrF, crystals will be formed.
These data check reasonably well wath
available vapor-pressure data for
ZrF, and for the fluoride mixture
NaF-ZrF,-UF, (46-50-4 mole %).

PROJECT FACILITIES(!®)

Helium Distribution. High-purity
heliumis now delivered to distribution
systems in the experimental engineering
laboratories, Building 9201-3, and
the Fuels Production Pilot Plant,
Building 9928, by pipe line from tank
cars through a pressure-reducing
station located at a considerable
distance from the building. The
installation cost for the several
thousand feet of pipe line has been
offset completely by the saving already
accrued from complete elimination
of the labor formerly required to
carry out the time-consuming process
of filling empty ecylinders with helium
at the tank car, transporting the
cylinders to intermediate storage,
performing individual purity tests
spectrographically on each cylinder,
distributing the cylinders to work
locations, and finally returning the
empty cylinders to the tank car for
refilling. :

Fabrication Shop. The shop facili-
ties of the Experimental Engineering

 

(13)The fuel production ard purification
facilities are described in section 11, “Chemistry
of High-Temperature Liguids."” '

PERIOD ENDING DECEMBER 10, 1952

Department, Building 9201-3, have been
relocated and consolidated to permit
expansion of the machine shop and
relocation of every fabrication work
area to new locations adjacent to the
mechanical stores crib. With the new
arrangement, the production of component
parts 1s expedited through the various
steps of fabrication, handling is
reduced, better housekeeping 1is
achieved, and a more workable ar-
rangement for the shops is obtained.

The use of oxygen and acetylene
cylinders has been eliminated by
installation of piped supplies of
these gases to the work benches in
the fabrication area from outdoor
cylinder-manifold systems. The
ventilation system has been expanded
to relieve the new shops of smoke,
fumes, and the excessive heat generated
during fabrication procedures. ‘

Instrument Shop Facility. A new
facility in Building 9201-3 has
recently been completed to provide a
suitable environment for conducting
under constant and most rigid external
conditions the design, development,
adjustment, and calibration of sensitive
electronic instruments and equipment,
The nonmagnetic laboratory benches
are supplied with electric outlets,
grounding busses, and piped supplies
of helium, natural gas, oxygen, and
acetylene. The illumination installed
in the area 1s adequate. Installation
of an air-conditioningsystem designed
to maintain this laboratory at constant
temperature and constant humidity will
start in the near future.

Gas-Fired Furnace Facility. A
battery of six gas-fired furnaces,
each rated at 68,000 Btu/hr, is being
installed on the main track floor of
the experimental engineering labo-
ratories, Building 9201-3. The
furnaces will serve as heat sources
for the various experimental programs
under way at the present time.

33
ANP PROJECT QUARTERLY PROGRESS REPORT

Each furnace 1s provided with
positive safety protection and control
instrumentation, including purging
blowers, high-voltage spark-ignition
systems, differential pressure regu-
lators with fail-safe and manual

34

reset features, dilution exhaust
stacks, and fail-safe interlocking
controls with provision for tempera-
ture control and automatic shut-down
in the event that the associrated
experimental equipment fails,
PERIOD ENDING DECEMBER 10, 1952

53. GENERAL DESIGN STUDIES
A, P. Fraas, ANP Division

A program for the development of
liguid-to-air radiators for turbojet
~engines has been outlined. Because of
the magnitude of the work involved,
the Air Force (Wright Air Development
Center) is planning a program for the

development of radiator fabrication
" technigues., The performance of a
radiator already tested at ORNL(!? has
 been extrapolated to that of a full-
scale radiator, and the data have been
~used to analyze the performance of a
Sapphire turbojet engine incorporating

such a radiator. Performance data for
~ both the radiator and engine appear to
be very encouraging.

Further design and analysis of
reflector-moderated reactors(2) have
been held in abeyance pending the
completion of critical experiments
to determine feasibility., Preliminary
results of the first critical assembly
are described in section 5, “Critical
Experiments.” ‘ ;

ATIR RADIATOR PROGRAM
A. P. Fraas, ANP Division

Further analytical and design
studies have been carried out on tube-
and- fin radiators. The results of this
work, coupled with the experimental
test results, indicate that 1t 1s not
possible to predict the heat transfer
pertformance of various radiator core
geometries closely enough to ascertain
which of the several designs 1is the
best one without testing them. A set
of rough scrting calculations indi-

cates, however, that tests on the
cores listed in Table 3.1 should be
made, Since this constitutes a con-

siderable program of test work that is
more closely related to the engine

 

1 . . .
¢ )See Experimental HReactor Engineering
Sections of this and previous guarterly reports,

(2)4. p. Fraas, ANP Quar. Prog. Rep. June 10,
1952, OBNL-1294, p. 6. _ '

than to the reactor, the Wright Air
Development Center has agreed to try
to carry out, through contracts with
vendors, part of the program; heat
transfer information and endurance
life data should thus be obtained, and
one or more sources of radiator supply

should be established. :

RADIATOR PERFORMANCE

G. D, Whitman H. J. Stumpf
ANP Division

Performance calculations show that
a radiator core with 30 fins per inch
should be very satisfactory even
though it would operate at Reynold’s
numbers far down in the laminar flow
range., Therefore a unit was built and
prepared for test, which 1s described
as type L in Table 3.1.

The performance characteristics of
a full-scale radiator were calculated
on the basis of test data obtained for
the core element described as type C
in Table 3.1 (cf., sec. 2, “Experi-
mental Reactor Engineering’). Figure
3.1 shows the performance data 1in the
form of a chart that gives ‘“heating
effectiveness’  as a function of
radiator depth in the direction of air
flow for a series of air-mass flow rates.
Figure 3.2 gives pressure-drop data
for this type of radiator for a 12-in.
depth in the direction of air flow.
Nickel fins were assumed 1instead of
stainless steel fins, The ““heating
effectiveness’ was taken as the ratio
of the air temperature rise to the
temperature difference between the air
entering the radiator and the sodium
entering the radiator,

ENGINE PERFOBMANCE
H. J. Stumpf, ANP Division

Estimated per formance characteristics
for a turbojet engine fitted with

35
9¢

 

 

TABLE 3.1. 1TYPES OF TUBE AND FIN RADIATOR CORE ELEMENTS PROPOSED FOR DEVELOPMENT TESTING
Tube outside diameter, 3/16 in.
Tube material, type 316 stainless steel (1f available)
Tube spacing transverse to air flow, 2/3 in,
Tube spacing in direction of air flow, 2/3 in,
Air-passage length, 12 in,
Air-inlet face, 2 by 3 1/3 in.
RADIATOR NO. OF ' TUBE
TYDE FIN DESCRIPTION FINS PER FIN MATERIAL ABRRANGEMENT STATUS OF TEST
INCH
A Plain plate 10.5 Stainless steel In-line Completed May 1952
B Plain plate 15 Stainless steel In-line Completed June 1952
C Plate interrupted every 2 in. 15 Stainless steel In-line Completed October 28, 1952
D Plate iaterrupted every 2/3 in. 15 Stainless steel In-line Radiator being prepared for testing
E Plate interrupted every 2 1in.,
with electroformed turbulators i5 Nickel In-line Fins being fabricated for ORNL
F Disk fins, stamped and ceramic 15 Nickel In-line Fabrication technique being developed
coated at ORNL
G Disk fins, stamped 15 Stainless steel In-1ine Suggested Wright Field project
H Disk fins, stamped 15 Nickel In-line Suggested Wright Field project
1 Disk fins, stamped 15 Nickel Staggered | Suggested Wright Field project
J Disk fins, coined 15 Nickel Staggered | Suggested Wright Field project
K Disk fins, stamped and chrome
plated {possibly 0.0003 in.) 15 Nickel In-line Suggested Wright Field project
L Plain plate 30 Nickel In-line In progress

 

 

 

 

 

 

LUOdIY SSTYI0Md ATHALHVNO LDA[0Hd dNV
90

85

80O

-~
tn

~J
<

O
: Q

miaie

HEATING EFFECTIVENESS, 7 (%)
0
v

n
o

40

30

20

PERIGD :ENBING DECEMBER 10, 1952

o
5
o
o

 

) We
0.010-in. NICKEL FINS INTERRUPTED EVERY Zin. A
t5 FINS per in. :
¥\g-in~DIA TUBES ON %-in-SQUARE CENTERS. |

WEIGHT FILLED WITH Na =801b /3
AIR INLET TEMPERATURE : 400°F
Na INLET TEMPERATURE=1600°F |
L No TEMPERATURE DROP = 400°F L

/
/!

 

_\\

NN N

3 "
&
%
.{:?'

nY4
I S/
a,

o :

L2 :

q? .

<L )
o

 

 

 

 

\\

 

 

 

AANAN NN

 

 

 

 

 

A\ O\ N

 

 

N\
N
NN\
MR RNAN

 

 

 

Al

 

S

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6 8 10 12 14 16 18
AIR PASSAGE LENGTH (in) |
Fig. 3.1. Performance of Air Radiator with Interrupted Fins.

37
ANP PROJECT QUARTERLY PROGRESS REPORT

DWG. 17444

 

 

 

 

 

 

 

/
3 /

 

o AP [(in. OF WATER)

 

FINS INTERRUPTED EVERY 2 in.
RADIATOR DEPTH 12 in. IN DIRECTION OF FLOW

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

05 1.0 1.5

20 3.0 5.0

AIR FLOW (Ib per sec per ft2 INLET FACE AREA)

Fig. 3.2,

liquid metal radiators are shown 1in
Fig. 3.3. Although i1t would by no
means be an optimum configuration, for
convenience it was assumed that a
Sapphire engine could be modified to
give a compression ratio of 4:1 and
could be fitted with full-scale

38

Pressure prop Throigh Alr Radiator with Interrupted Finms.

radiators that would be similar to
radiator core element C (Table 3.1),
except that they would have nickel
fins rather than stainless steel fins.
The pumps and lines for the secondary
fluid (assumed to be NaK) were designed
to allow for the equivalent of 20 ft
PERIOD ENDING DECEMBER 10, 1952
3
CWEG. 17445
CORRECTED COMPRESSION RATIO = 4.0
160 - ENGINE WEIGHT = 2785 Ib
BT RADIATOR WEIGHT =960 Ib
140 e .
= RADIATOR DEPTH =12 in. ,
3 120 RADIATOR INLET FACE AREA =12 ft°
o SEA-LEVEL AR FLOW =47 Ib/sec
= 100 \6000i\ ENGINE DIAMETER =37.5 in.
T 80
e
g
40 45000 H
55,000 {f
20 :
L ENEL 32 000
SERY
28,000 '_g
o ft 24,000 ' Z
1500 E
20,000 . =
o]
o)
16,000 Z
5,000 12,000 E‘;
I
55,000 f1 8000
4000
6000
5 B000 SEA-LEVEL
— 4000 :
2 ¢
2 3000 15,000
}_.
g 2000 35000t
O
i....

1000

0 : 0.3 . 06 09

 

1.2 1.5 1.8 20

MACH NUMBER

Fig. 3.3. Performance pata for M(}dified Sagphire Engine with Interrupted-

Fin Radiators.

of pipe from the outside of thereactor
shield to the turbojet engine. The
weight of the engine, radiators, pumps,
lines, and contained NaK was found to
be 4256 pounds.

The engine weight, without the
radiators, was assumed to be the same

as that for anengine with the standard
compression ratio of 6.85, It was
further assumed that the saving 1n
weight effected by eliminating both
the combustion chambers and several
stages from the compressor would be
equal to the weight of the radiator

39
ANP PROJECT QUARTERLY PROGRESS REPORT

support structure and the special
ducts, baffles, etc. required for the
ligquid metal radiator. NEPA design
studies indicated that it should be
possible to do better than this; hence
the weight given above, 4256 1lb, 1is
probably a moderately conservative
estimate, Further radiator development
work may make 1t possible to cut as
much as several hundred pounds from
the 960-1b radiator weight in this
installation,

HEAT EXCHANGER DESIGN CHARTS
A. P, Fraas M. E. LaVerne

The OBRNL-ANP liquid-to-liquid heat
exchanger design and development
efforts have been based on a matrix of
closely spaced small-diameter tubes
that permit practically pure counter-
flow operation, This type of matrix

40

has exceptionally good heat transfer
characteristics. In the course of
full-scale aircraft power plant design
work, a number of charts for this type
of heat exchanger has been prepared.
These charts were intended, in part,
to show the effects of the various
parameters in a readily understandable
form and, in part, to simplafy and
reduce markedly the chore of making
detailed design calculations. The
charts, togetherwith brief explanations
and sample calculations to illustrate
their use, have recently been issued.(®
The subjects covered are (1) heat
exchanger matrix geometry, (2) fluid
flow and pressure loss, (3) heat
transfer coefficient, and (4) per-
formance of a series of counterflow
heat exchangers.

(3)A. P. Fraas, Heat Exchanger Design Charts,
ORNL-1330 (to be issued),
PERIOD ENDING DECEMBER 10, 1952

4. REACTOR PHYSICS

W. K. Ergen, ANP Division

In the theory of power oscillations
in a circulating-fuel reactor, 1t 1is
now possible to demonstrate the damping
influence of the fuel circulation, even
if some of the previous assumptions
are replaced by considerably more
general conditions. Specifically,
it 15 no longer necessary to assume
that all particles of the fuel spend
the same time in the reactor and that
the fuel and power distributions are
constant over the reactor. The more
general condition (Eq. 1), which
replaces the old assumptions, covers,
among other things, the case in which
the power distribution is constant
only in the direction of flow, but not
necessarily in the direction perpen-
dicular to 1it, and in which there are
a number of alternate fuel paths, each
with a different fuel transit time.

In the field of reactor calcu-
lations, it was found that the slowing
down of neutrons in parallel slabs of
materials with different properties
can be described, under some not too
unrealistic conditions, by a set of
images of the original neutron source.
The images are somewhat similar to
those which would be obtained 1f the
neutron source were replaced by a
light source and the interfaces between
the slabs were replaced by mirrors or
refracting surfaces.

For some time, attempts have be en
made to obtain a better understanding
of the nuclear physics involved in the
temperature coefficient of reactivity.
These attempts have now resulted in an
explicit expression for the tempera-
ture dependence of a macroscopic
absorption cross section., The tempera-
ture dependence is caused by the
thermal motion of the atoms.

Calculations specifically carried
out for the ARE and Fireball Critical
Experiments are reported in connection
with these experiments (Sec. 5).

OSCILLATIONS IN THE CIRCULATING
FUEL REACTOR :

W. K. Ergen, ANP Division

It was shown previously(l) that,
under certain conditions, the circu-
lation of the fuel introduces a damping
of power oscillations of a reactor.
These previous comnsiderations have
since been gemeralized, and the
damping can still be demonstrated,
even if one drops the assumptions that
all particles of the fuel spend the
same time 1in the reactor and that the
flux and power distributions
constant over the reactor.

The delayed neutrons are still
neglected, and it 1s still assumed
that a change AT of the average fuel
temperature T causes an instantaneous
change ~aAT in the reactivity and that
the inlet temperature is constant,
The old assumptions of constant fuel-
transit time and constant flux and
power distribution are replaced by the
much more general condition that

are

% = ¢ | P(¢t)

-~ Jr do K(o) P(t - o) |, (1)

0

where € 1s a positive constant associ-
ated with the reciprocal heat capacity
and P 1s the reactor power. The
kernel K{o) has the property dK/do g 0,
and, for o greater than some fixed
value A, K(o) = 0. The term €P
corresponds to the instantaneous
increase 1n reactor temperature due to
the power generation. The integral

 

(I)W. K. Ergen, 4ANP Quar. Prog. Rep. Marefi 10,

1952, ORNL-1227, p. 41; 8. Tamor, ANP Quar. Prog.
Rep. June 10, 1952, ORNL-1294, p. 31; 5. Temor,
Note on the Non-Linear Kinetics of Circulating-
Fuel Reactors, Y-F10-109 (Aug. 15, 1952),

41
ANP PROJECT QUARTERLY PROGRESS REPORT

represents the fact that the memory
of the earlier power history 1is
continually “washed out’’ by the
circulation of the fuel, for instance,
because fuel which has “seen’” the
earlier power history leaves the
reactor.

It is convenient to take the inlet
temperature as zero on the temperature

scale. Then Eq. 1 is equivalent to
T = € f G(og) P(t -~ o) do (2)
0
with
G(0) =1 , (3)
G(®) =0 , (4)
dG/do = -K(o) (5)

The second equation describing the
kinetic behavior of the reactor is,
as 1n the previous calculations,

*

a
P=-2P(T-T,) , (6)
T

where 7 1s the prompt-neutron gener-
ation time, and T, is the average
temperature the reactor reaches in the
equilibrium state. This state 1is
is characterized by a constant power,

P,.
Limits on the Oscillations. As in
the previous considerations, it is to
be shown first that the power remains
bounded., Since, for ¢ > A, dG/do =
~-K(c) = 0 and G(®) = 0, it follows
that G(o) = 0 for o 2 A. This means,
physically, that the power prevailing
at t ~ A, has no i1nfluence
on the temperature T at time t, for
instance, because all the fuel present
at time t - A, or earlier, has left
the reactor before time t.
-~ A is the time
at which the reactor power passes
through P, and is rising. Eventunally
the reactor heats beyond T,, the
temperature at which P = 0, and then
the power has to start dropping.

or earlier,

Assume now that ¢

The attainment of the temperature
T, and the connected drop in power
have to occur at time t, at the latest,
because, unless the power drops before

42

t, it will have exceeded P, during
the whele time interval t — A to t,
and even P, would have been sufficient
to raise the temperature to T,. This
shows that the power cannot start at
P, and rise continuously for a time
interval in excess of A,

Also, the rate of increase in log
P is limited, as can be seen from
Eq. 3, where the temperature cannot
be less than 0, the inlet temperature,
Since both the rate of increase and
the time of increase are bounded, the
maximum is bounded, too. Furthermore,
if the reactor power stays above P,
for the time A, it has to drop below
P, before it can rise again, because
the temperature could not otherwise
go below T,, When the power has gone
below P, and starts increasing again,
the above argument again holds,

A similar argument shows that the
power has a finite lower limit,
Actually, the limits found by the above
consideration are far wider than the
limit obtained in practice, but they
show that there are no antidamped
oscillations that would increase to
infinity.

Periodic Oscillations. It is now
of interest to determine under what
conditions periodic oscillations of
constant amplitude can occur. For
this purpose, one multiplies Eqs. 1

and 6:
a d d

- = = (T ~ T )? = ¢ — log P(t
o dt o) 1c Log PLO)

P(t) -—f do K(o) P(t - o) | . (1)

0

If Eq. 7 is integrated over a period
p of the oscillation, the left side
becomes zero because of the periodi-
city. The same applies to the integral
of

d
P(t)'g; log P(t) .

The remaining integral is transformed
by partial integration:
x

0

= log P(t) f d;:T K(o) P(t - o)
.

p .
f dt 2L 1o P(1) f do K(o) P(t - o)
dt o f

"PERIOD ENDING DECEMBER 10, 1952

t=p

 

t=0

p ® d
f dt log P(t) J- do K(o)=— P(t - o) .
o 0 dt

Again, because of the periodicity,
the first term on the right has the
same value at ¢ = p and t = 0, and
hence only the second term remains on
the right side. Noting that

d plt - o) d Pt )
—_ - = e —— P -
dt | dor v

and interchanging the order of inte-
gration, one transforms ‘this term into

o

_ ., , : :
do K(U)——f dt log P(t) P(t ~ o)
0 .

0

 

 

Since
K@) = 0
and fi
w
dK (o)
K(0)=-f dr Tl
o do

the term can also be written

f do dK (o)
do

0

 

p
f dt log P(t) P(t)
o :

p |
f dt log P(t) P(t - ¢o) | . (8)

0

At this point, it 1is convenient,
though not essential, to choose the

P
ko) [ dt tog P() PG - o)

 

units of P in such a way that log P
is always positive throughout the
oscillation. This is possible, since
P is bounded below. If log P is always
positive, a theorem(z) regardlng
inequalities becomes applicable. [t
has only to be noted that log P 1s a
monoton increasing function of P. The
theorem states that the expression
in the bracket 1s never negative. [t

 

2ol

o=0

fmdodk " dt log P(t) P(t - o)
e ___U< o

0 dor J; o8 _

is zero only 1f P(t — o) = P(t) for
all t, that is, i1f o is a multiple of
the period p. dK/do was assumed to
be nonpositive; hence, the whole
integral -8 1s £ 0. :

It has been shown before that, for
any periodic oscillation, the integral
over a period of the left side of
Eq. 7 is equal to zero. Hence, the
integral over the right side, which 1is
equal to the integral 8, has to vanish,
too. The condition for this is,
according to the above, that dK/do
vanish, except possibly at the points

 

 

(2)G. H, Hardy, J. E. Littlewood, and G. Pdlya,
Inequalities, 2d ed., p. 278, Theorem 378, Cambridge
Univ. Press, 1952.

43
ANP PROJECT QUARTERLY PROGRESS REPORT

where the square bracket in integral 8
vanishes, that 1is, at the points where
o 1s a multiple of the period. At
these points dK/do may even go to ~®
without vielating the condition for
periodic oscillations.,

Speci fic Examples. The previously
considered case of constant transit
time and constant power distribution
is obtained from the above general
case by giving K(o) the specific
values K(og) =1/8 for o <& and K(o) = 0
for o > 6 (compare Eq. 1 with Eq, 2
of ORNL-1227, p. 43,¢!) or Eq. 2 of
Y-F10-109, p, 2(!?), dK/do = 0 at all
points except o = & (where dK/do = —x),
and periodic oscillations persist only
if € is a multiple of the period,

If the power distribution is only
constant in the direction of fuel
flow, but not necessarily 1in the
direction perpendicular to 1it, and if
there are a number of alternate fuel
paths with transit times 6, , 8,, ...,
¢ , respectively, then the above

n
general theory applies with K(o)} = a,

for Qi_l < o < Qi’ where 7 = 1, 2,

e, n, 8,70, and the a, are positive
1

constants, a. @, 1. Periodic

oscillations are only possible i1f all
f, are multiples of the period, a
condition very unlikely to occur 1in
practice,

Tt should be mentioned that T. A,
Welton of the Physics Division has
made significant advances in reactor
kinetics by a somewhat different
technique.¢3) The above results can
also be regarded as special casesof
Welton’'s theorems.

REACTOR CALCULATIONS(*)
R. K. Osborn, Physics Division
It was noticed that the method of

images can be used to solve some
specialized but not unrealistic problems

 

of the slowing down of neutrons. The
(3)
T. A. Welton, Phys. Quar. Prog. Rep. Sept.
20, 1952, ORNL-1421 (in press),
(4)R. K. Osborn, Some Solutions of the Age

Y-F10-111 (Nov. 17, 1952).

Equation,

44

case considered involves two homo-
geneous, semi-infinite spaces of
different properties, joined along a
plane x =0. A plane, 1sotropic source
of monoenergetic neutrons is placed 1in
one of the half spaces, say the right
half space. Under the conditions set
forth below, the left half space, the
“reflector,” can then be replaced, as
far as the slowing down density in the
right half space 1s concerned, by an
“image” source located symmetrically
to the given source with respect to
x=0. The slowing-down density in the
reflector can be described by a
“refracted’’ source located to the
right of x = 0. The image source and
the refracted source have the same
energy as the given source.

Properties of the right half space
are denoted by the index 1 and those
of the left half space by the index 2,
Zs, &, 1, and g are, respectively, the
macroscoplic scattering cross section,
the average lethargy 1ncrease per
collision, the average cosine of the
scattering angle, and the slowing-down
density; 7 1is the “age’’ in the medium
that contains the source.

§22§2(1 - fiz)

 

m (1)
£,22,(1 - 7))
and
£ 222
n = — (2)
S1241
are constant with lethargy., Another

assumption 1s that the source plane is
parallel to x = 0, so its equation can
The Fermi age
with no absorption, 1is
assumed to describe the slowing-down
process:

be written as x = X,
equation,

o%q,

it

Ox 2

 

9q
L S(x - X) S(r)

r ’ (3)

and
82q2 oq,

ou3 - m > =.0 . (4)

 

 

The assumed boundary conditions are
qz(O, T) = n ql(O, ' I (5)
and '
3q2(0, T) Bql(O, T)
e et TR Q)] e et

Ox ox

The solution in the half space
x > 0 is then

. (6)

 

 

- | - X)?2
qi(x, T) = (47} % exp |~ ifi——-““
: 4
+ X)? |
b A exp |- EEOTLL o
47 1
where
1 - =
T
A = T (8)
1+—n—-
o

The image source is thus always weaker
than the given source. It 1s positive
or negative depending on whether

b2
n 1 - i,
m 1
1 - u,

is smaller or greater than 1.

The solution in the reflector is
given by

1
B 4mrr / (x - Y)? (9)
= —_— ex = m——— ],
7 m P 47
m
that is, the neutrons seem to come

from a “refracted” source of strength

2n
I f
— (10)

s

PERIOD ENDING DECEMBER 10, 1952
located at
v - 2 (11)
I |
with the whole space filled waith

material that has the same properties
as the reflector. The refracted
source 1is always positive.

That the solutions satisfy, for
their respective regions, the dif-
ferential equations 3 and 4 and that
they also satisfy the boundary con-
ditions 5 and 6 can be shown by direct
substitution. _ f

The constancy of m and n 1s not a
bad assumption in many cases, because
the scattering cross section and,
hence, £ and & are usually constant
over a wide lethargy range. The
constancy of the scattering cross section
is not a necessary condition for the
constancy of m and n. For instance,
if the £’s and u's are constant and
the scattering cross sections in both
media vary in such a way that their
ratio remains constant, m and n would
still remain unchanged. However, there
are few practical examples in which m
and n are constant and the scattering
cross sections vary. The most im-
portant, but somewhat trival, example
1s the case i1n which the two media are
of the same material and differ only
in density. Then the “image source”
will have strength zero. The refracted
source has the strength 1, and there
is the same amount of material between
the refracted source and the boundary
x = 0 as there is between the actual
source and the boundary.

If the macroscopic absorption cross
section 2 is not zero, then the image
method ¢an be applied only in the
somewhat trivial and not very realistic
case in which

2 2

. 2 (12)
z 124, |

278 2

 

 

&
=

If there are several sources, or a
continuum of source, the solutions for

45
ANP PROJECT QUARTERLY PROGRESS REPORT

the individual sources are merely
superimposed,

An interesting generalization of
the method can be made 1f there are
a number of parallel plane layers of
different properties, with one of the
layers containing the plane 1sotropic,
monoenergetic source of strength 1.
This generalization can be shown on
the example of Fig. 4.1, Here a
source 1s located at the center of a
slab of thickness 2a. The slab 1is

embedded in reflector material, which

 

 

 

extends to infinity on both sides;
x = 0 1s the source plane.

UNCLASSIFIED

OWG. {7446

42 A2 4 1 4 42 4

o L

I ' | SOU?CL | l I

| | I / | | |

| | I / | | j

L

l | | (g4 | | |

i

L LAl
S - ¥ | e e —— a4 e~

L—r———4—~44—60 Lag—a—44-4‘A60~—w ~vm~n+J

Fig. 4.1. Neutron Seceurce and Its
Images pescribing Slowing-pown Density
Inside Slab,

If the left reflector were replaced
by the material of the slab, the so-
lution in the slab would simply be
generated by the two sources - the
original source, and the image source
of strength A at 2a. However, in
order to satisfy the conditions at the

 

2 -1
NzB . ~hy(v-vy) /(1+] hz}
S (v,T) = +

v , 1
2vel]l + —
Fh2>

left boundary with the two sources
present, one has to 1introduce two
additional sources -~ one of strength A

N A

 

 

46

jlay,y) ==

at -2a and one of strength A% at —4a,
Now, in order tosatisfy the conditions
at the right boundary for the
additional sources,

two
one has to put a
source A? at +4a and a source 4% at
t6a. 1f this procedure 1s continued,
1t becomes evident that the solution
applicable to the inside of the slab
1s obtained by placing sources of
strength A' at planes x =*2ia, where
0, 1, 2, .... The solution 1in,

the right side reflector, is
described by sources of strength BA?
located at distances (21 + 1)a/J% Lo

1 =
say,

the left of the interface x = a, where
I‘- = O’ 1 ¥ 2 ¥ 4 5 3
TEMPERATURE DEPENDENCE OF A CROSS

SECTION EXHIBITING A RESONANCE®)
R. K. Osbern, Physics Division

The explicit energy and temperature
dependence of a macroscopic absorption
cross section exhibiting a resonance
was calculated on the assumptions that
the microscopic cross section 1s of
the form

~ogol o NYT . _B

ofi(fvl2|) = Ae o |
12

3

and that the atoms are in a Maxwell-
Boltzmann velocity distribution. In the
definition of o above, A, B, and [
are adjustable parameters and v,, is
the relative velocity between neutrons
The guantity v,, 1s the
relative velocity that
in the

and atoms.
particular
corresponds to 3a resonance
cross section,

The expression for the cross section
is:

~hg(utug)?/ (147 hy)
&

2 ._l_
2011 +

{S)B. K. Osborn, The Temperature Dependence of
a Cross-Section Exhibiting a Resonance, Y-F10-112
(Nov. 17, 1952).

N ,A

 

jla,,¥)

 
PERIOD ENDING DECEMBER 10, 1952

 

where . vy t [h,v
N, = number of atoms per cubic 1 1 + th
centimeter,
_ _ : : ve = Thyv
v = velocity of the neutrons, a, # i
' 1+ l_'h2
m |
h = 2, + [h
2 2T Y o= ~l_.._...__2_’
r
m, = atomic weight of the atoms, ' 2
' ) 2 1 ae 7°
ey jla,y) = {a +-—-[1+er-f Ny a)] +
Vo = |v10| ) 2y, A\

47
ANP PROJECT QUARTERLY PROGHRESS REPORT

5. CRITICAL EXPERIMENTS

A, D, Callihan, Physics Division

During the past guarter some
inaugural experiments were done on a
critical mockup of the proposed
reflector-moderated aircraft propulsion
reactor, ! and the preliminary results
are reported. The initial experiment
on the reflector-moderated arrangement
was assembled from on-hand materials
and is a correspondingly poor simulation
of the actual reactor., Nevertheless,
the assembly became critical with a
critical mass of 13.5 kg, The flux
distributions obtained show the ex-
pected high thermal flux in the
moderator island and reflector. A
more realistic mockup of the reflector-
moderated reactor
sembled.

The program of measurements on a
critical assembly simulating the ARE
reactor, described earlier, was com-
pleted, and some of the results are
reported.
rods have

is now being as-

The regulating and safety
been caltibrated, and a
comparison has been uade of the con-
tribution to reactivity of various
core components. Considerable data
from the ARE program have yet to be
evaluated and will be presented in
subsequent progress reports,

REFLECTOR-MODERATED CIBCULATING-FUEL
REACTOR ASSEMBLY

D. V. P, Wailliams R. C. Keen
J. J. Lynn
Physiecs Division
D, Scott C, B, Mills

ANP Division

A preliminary study of some of the
nuclear characteristics of the re-
flector-moderated circulating-fuel
reactor(?) has been made with an

(l)A. P. Fraas, ANP Quar. Prag., Rep. June 10,
1952, ORNL-1294, p. 6.

(2)C. B. Mills, The Fireball, A Reflector
Moderuted Circulating-Fuel Reactor, Y-F10-104
{June 20, 1952).

48

assembly comprising component materials
that were readily available. The
arrangement of the assembly was first
(3)

1ts

critical on November 1 with a mass
of about 13.5 kg of U?3°%, In
present form, to be described below,
the components were symmetrically
located and the loading was 15.0 kg of
UZSS.

The critical assembly consisted of
a central 12-in. cube of metallic
beryllium that was almost completely
enclosed by a 3-in.-thick fuel layer.
Surrounding the fuel was a composite
reflector consisting of a 12-in.-thick
layer of beryllium and an outer
of graphite., The graphite was

inner
layer
8 1in.
6 in.

thick on two opposite sides and
thick on the other ones. These
materials were arranged in the matrix
of 3-in, sguare aluminum tubing, which
has been described in previous re-
ports.¢*) The loading at the mid-plane
of the reactor model 1s shown 1in
Fig, 5.1.

The fuel consisted of
alternate layers of uranium and sodium;
the uranium was in disks about 3 1in.
in diameter and 0.01 in. thick, and
the sodium metal was canned 1n stain-
less steel boxes 2 7/8 by 2 7/8 by 1
inch., The wall thickness of the
boxes 8 mils. Two disks were
placed between adjacent cans of sodium,

section

was

The long dimensions of these 1tems
were parallel to the reactormid-plane,
The fuel region covered four sides of
the beryllium core completely and
extended partly over each end, with
all but the center 6- by 6-in. section
end being enclosed. This 1is

of each

(3)It is to be emphasized that the critical
mass of this assembly is strongly dependent upon
the arrangement of the components. For example,
the mass was changed by as much as 15% by very
minor alterations in the structure,

(4]A. D. Callihan, ANP Quar. Prog. Rep. Dec,

10, 1951, ORNL-1170, p. 35; and ANP Quar. Prog.
Rep. Mar. 10, 1952, ORNL-1227, p. 59.
PERTIOD ENDING DECEMBER

o

16, 1952

DWG. 17447

 

 

 

 

 

 

 

 

 

v%a

 

//

 

7

M/'

§
\

  

 

vt gt o, e u W a s e e et ea e o B T e T e T T e e e Pttt et e Wt e
+28%Je%e™0u a - ale et ettt el et sl e et ettt st atatn
- » - - B_% 3 8 Ll k& 4 P s B Sln & e W e - - s - - - - - - - - -
nta%el0%et a®s a a?sta atata ettt it ntaa s Bt T et e T
- - - a . - -' - - “....— ................. - - & ’.fl - ..' I. - I..
woe, P s ettt e ettt al ettt e et ettt '-'s:-'. a*e s 'c.-'-'
- - - - - I". ‘..-*fl. "".. ".‘ -‘
«%a 2%a®u *a Attt et et eta 0 s e
...’..- .'.‘ ‘I.‘. - el Ma Pt WA DL ket e s 'U-O‘. .. ..... TelaaNaa 4.
e ata®adn*a” e . ot o 0 ------- ar ar et patelelat el et ATt t0 . .
atae -t s ot s ettt e et et - .
- - - - - . - ' & . ....... . - - - o " s 4 s A me e o &
. a’a’e sletn . a®atelatet,* st ettt et e TR %0,
*«a to st st te e st . s et at u T a Tt Ty a"aNe
- A - - - * n ° o - » L - 9 - - " & =T s ]l® & _ @ N w ¥ _a _ i» - »
coalalesaletats . . h‘ N IS tereratelateTatd el e ety el ultel .
* 2 el 4 v atn’e o ga ", s et atulata A"t a® s leltws .
- // v 1 mlatvnelnl, ¥ . et .
- ~ sta"a "R a0, td Tl
-9 * - - e - - - - » - > -
- - - > - - - -
- -l -t 2 s "2Ta"s"a -"n »|
o WAt et et et
eoqtasl /// /A /// MRS P : ot
- - - . p_* - - - » - -
. cadlel ata / Setarhiizis - <ot
. e 4" a »le w < 1~ ,/ L e et e T T LT T L Y . -
- » - » - o e - - - - ” - - -
» - n - - - b - & &y & b e e & e - »
. - ate dh et e ta Tttt - -
'.D - » » - - - l"“b ............ *
Tt et el atu"da"e P a7t :
eenrieiei=latel Y e
a2 a"a"ula’s A 7 T
nar /7 //,
-
- .
'.-
..U
2as /
Pty
»
R
.:q
-
o'n:a
- * /

 

/’ 7
/z%%%g
7

7

 

_

 

 

\

» 'I‘ .I!. ‘.

.l =
Cessrethagales
- . v '.“'
P e® n Nty
e s s sl w a
> L -
Laleozomoslel
. * » =

et

N\

 

FUEL

 

 

11

7

 

12

)

 

 

7

 

7

ERYLUUM

 

 

W e bR e bRMEDN NI OTFBG
LA R A EERER SRR ERRN SRS N

 

 

 

 

 

 

 

 

   

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, .
o :
e :
oy j?//?%;f %/ ////// / :
-
.
S i
”Eéf%@@%%fi%%@%@@%i?
Qe e
i e e
o, S
2! mfl"fQA'éééznifiifé’{.;fiisziiqfiizifig;uq.. ........ i gi' |

 

 

  

 

 

 

 

 

 

 

 

 

 

-~

Fig. 5.1.
Moderated Reactor Assembly.

5.2,

shown in Fig,

above.

to preserve symmetry,

which 1s a vertical
section through the reactor perpen -
dicular to the mid-plane referred to
It is noted that the graphite
end-reflectors were not complete and,
the sections of
the fuel at the mid-plane were separated

by two 1/2-in.

fuel shell,
reactivity change,
that from the effective fraction of

delayed neutrons.

Vertical Section Normal to Axis'at Mid-plane of Reflector-

thicknesses of aluminum.

The control and safety rods were
reflector elements located ocutside: the
Each corresponded,

in

to about 30% of

49
ANP PROJECT QUARTERLY PROGRESS REPORT

SIT

DWG. 17448

 

et
Lelels

 

v o = o

ofete’.

efete’e
-

s atats
--------
------------
etes -

e "a"a's
--------
.

 

 

 

 

 

 

N\
NN
N
N
&\\X\\T.:.Z.:.'. .'.'.'.‘.:..
NN

 

.
0 00000

BERYLLIUM

 

 

 

 

§§§§§§ DN
NN
\
N\
N
N\
N\
\
\
\
\

 

 

 

NN

 

 

 

o

b ag o snfageinidrel
proiesetelels

 

 

N
\

NN
X
\

 

 

 

s
]
L

 

 

 

\
NN
NN
&\\
NN
N
N
N

 

N

 

7
%?U EL
/7
9000

 

 

 

.............. .. .... .
saasasasnseprervedontorns sisanssabhrasnnnalvvwrerinrrese
B A SO A OO0 DXONNS SO 00N 0000

\FC\Q\:C\DQFC\ DN N

:j/’

 

77

A

 

X
NN
NN
N
\
N\

 

i
-
0

 

 

?fi;g%%%%%%%% —om= 3in. =

 

 

 

j;ng_ X §§§§
25 AN

.

 

D

 

 

 

 

 

 

 

 

 

 

111111
------

PR

s e,

-
ooooo

o,
.

o e e aale «Teqse -
. - e v _e e sfa®a®s PR L M L
aMatae e ettt et at et 4 2 e " e e e e e s ol 2"
o rl et e et a el 0o s e e teleT et T d e -
et el et et ettt et et T L e "atale’s w«Ma®as

e el e e a0 . e et a e al el et u . . o
et h ettt e e s s 2%atal et aMe’u s et
sfe et q al el o e s et el et e s a0 el o ol . o
----------- LI «MeT 2"

--------------- aaTe e s el e e » - o

 

 

 

 

 

 

 

 

 

 

 

 

 

fFig. 9.2,

Some measurements have been made
that indicate the neutron flux and
power distributions 1in various parts
of the array. The flux was measured
in the usual manner with bare- and
cadmium-covered-indium foils and the
power or fission-rate data were derived
from fission fragments collected on
aluminum foils placed in contact with
the uranium.

50

Vertical Section Along Axis of Reflector-Moderated Reactor Assembly.

In Fig, 5.3 are shown neutron-flux
traverses measured horizontally along
row 12 in the mid-plane of the reactor,
and in Fig. 5.4 are the data obtained
along a central traverse of cell 0-12,
perpendicular to the mid-plane. In
both cases, the activations of bare-
and cadmium-covered-indium foils and

These

results are typical of those obtained

their difference are shown.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PERIOD ENDING DECEMBER 10, 1952
45 DWG. 17449
| l } |
TRAVERSE FROM CELL 042 TO X—42
40 .
~BERYLLIUM FUEL BERYLLIUM REFLECTOR GRAPHITE -
ISLAND ~
. .
"
o
>
];I:‘J . - —
5 BARE INDIUM ACTIVATION
= o
i
: N
¢ //
o
=
2
[
2 e e
TN - DIFFERENCE BETWEEN BARE AND
N CADMIUM - COVERED ACTIVATIONS
/ \YK
\\
™ i:\\
M
CADMIUM - COVERED bj\
INDIUM ACTIVATION -1 e -
G 5 10

Fi_ g. 5.3.

along other traverses, and they confirm
the theoretical prediction of high-
neutron flux in the moderator 1sland
and reflector., This effect is, as
theoretically expected, particularly
pronounced for thermal neutrons (the
dashed curve, Fig. 5.3).

The cadmium fraction, derived from
the indium-foil activation data and
de fined as the ratio of the activation
by neutrons of energy below the cadmium
cut-of f to the activation by all
neutrons, is plotted in Fig. 5.5 for
the horizontal traverse of Fig. 5.3.
The values of this cadmium fraction,
as calculated from the IBM multigroup
computations, are also plotted in Fig,

15 20 25 30

DISTANCE FROM REACTOR AXIS (in)

Indium Traverse Radially in Mid-plane of Reactor.

5.5. The theoretical and experimental
values practically coincide in the
center of the fuel and in the bulk of
the reflector. The discrepancy at the
fuel surface is not surprising, since
all practical calculation methods,
including the “age’’ multigroup method,
are not exactly valid near boundaries.

The relative fission rates at the
surface of the uranium metal in cell
Q-13 are shown in Fig. 5.6, where the
two points at 9 1/2 in. were measured
on opposite sides of the uranium disk
adjacent to the beryllium reflector.

In another experiment, the fuel
disks, with aluminum foils adjacent,
were wrapped in cadmium sheet, The

51
ANP PROJECT QUARTERLY PROGRESS REPORT

DWG. 17450

 

45 1

 

35-0-__

40 --m—BERYt_uUM——-—+-—FUE_L—-}-«-—BERYLUUM------~~———~—-—~ -l- ————— ~— GRAPHITE -J!

|
|

 

|

 

 

W
O
A

l
TRAVERSE ALONG CELL O-12

 

 

‘

—

 

 

 

 

“\
M—:fi\ iUM-COVERED
. . CADM cO

  

 

 

w
v
#
/

INDIUM ACTIVATION

COUNTS PER MINUTE x4073
™
o

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

~ .
~ -’
L -7
N\ -
~ -
" z:;//’ ST C
BARE INDIUM ACTIVATION
MINUS CADMIUM-COVERED
5} — . ACTIVATION
0 i !
0 5 10 15

 

DISTANCE FROM REACTCOR MID-PLANE (in)

Fig. 5. 4. Indium-Foil Activation Along Axis of Reactor.

resul tant activities collected on the
foils are a measure of the fission
rate caused by the higher energy
neutrons. From these and the data
from the preceding experiment,it 1s
observed that about 70% of the fissions
are produced by thermal neutrons.,

Two, adjacent, 10-mil-thick uranium
pieces were replaced by ten pieces
that were each 2 mils thick and
separated by aluminum foils. From the
observed distribution of activity
across this composite fuel element,
the self-shielding of the 20-mil-thick
uranium 1is found to be 34%, that 1is,
only 66% of the uranium 1s effective,
This figure of 66% is in satisfactory

52

agreement with a figure of 63% computed
from the multigroup data., Of course,
the orientation of the disks, parallel
to the neutron current from the re-
flector, and the nonuniform power
distribution make the theoretical
calculation difficult and somewhat
uncertain,

ARE CRITICAL ASSEMBLY

D. Scott C. B, Mills
ANP Division
J. F, Ellis D, V., P, Williams

Physics Division

The preliminary assembly of the ARE,
which has been described in previous
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PERTOD ENDING DECEMBER 10, 1952
DWG. 17454
——BERYLLIUM FUE%f - BERYLLIUM -GRAPHHE
1.0 —
0.8 = e
=
o
—
(%
<I
o
-
= O EXPERIMENTAL
g ' A CALCULATED
0.4 .‘b}\[\\o l :
| Inaare) ~ IMtcapmium covereo)
CADMIUM FRACTION = I
: M BARE)
{
Q.2 1 : Z/ [l i
0 | | | |
0 4 8 12 16 20 29 28 32
REACTOR RADIUS (in.)
Fig. 5.5.  Cadmium Fraction (from Indium-Foil Activity) vs. Radial Fractioht
reports, (?) consisted of BeO as re- mild steel, 48 in. long and 1/4 in.

flector and moderator with a packed
powder mixture of UF,, Zr0,, C, and
NaF as fuel. It was first made critical
with a loading of 5.8 kg of U?3% con-
tained i1n 61 of the 70 fuel tubes
available. Subsequent measurements
showed that adding fuel to the remain-
ing nine tubes would result in $2.70
excess reactivity., ($1.00 is the
reactivity change corresponding to the
effective delayed-neutron fraction,)
It was found, 1n an attempt to
evaluate the reflecting effect of the
pressure shell, that a cylinder of
(5Yp, Scott, ANP Quar. Prog. Rep. June 10, 1053,

ORNL-1294, p. 38; D. Scott and C. B. Mills, ANP
Quar. Prog. Rep. Sept. 10, 1952, ORNL-1375, p. 43.

thick, surrounding the BeO blocks,
increased the reactivity 34.5 cents.
A second 1/4-in.-thick layer raised
the reactivity an additional 21 cents,
to a total of 55.5 cents. The shell
did not extend over the ends of the
reactor. '

In another experiment, an evaluation
was made of the reactivity value of a
tube of fuel at various radial positions.
In Fig., 5.7 is shown the increase in
reactivity as a stainless steel tube,
filled with fuel, 1s placed in a void
at different points along a reactor
radius. Also shown are the corres-
ponding values resulting from the

53
ANP PROJECT QUARTERLY PROGRESS REPORT

D\SE 17452

 

6 1

‘

POWER DISTRIBUTICN IN CELL Q-13

o

 

 

 

iy
T
1

 

 

 

Y
_ =
|
l
I
I
|

 

 

 

RELATIVE FISSION RATE (counts/sec x1G°)

 

 

 

 

 

 

0 L

 

 

 

 

 

 

 

 

 

 

0 1 2 3 4

5 6 7 8

0
<

DISTANCE FROM REACTOR MID-PLANE {in)

Fig. 5.6.

insertion of an aluminum tube, 1denti-
cally loaded, in the same positions,

One of the projected measurements
with this assembly was the evaluation
of the regulating and safety rods of
the ARE reactor. It was necessary to
install the rod at the center of the
core, thereby replacing the central
fuel tube and the central column of
BeO. As shown in Fig. 5.7, the removal
of the fuel tube decreased the re-
activity $1.07, Since the BeO column
removal lowered the reactivity an
additional $2.63, 1t became obvious
that the loading was insufficient to
permit a measurement of the rod.

The available excess reactivity was
increased by adding UF, to the fuel
mixture in the stainless steel, thereby
increasing the U?3% density from 0.163
g/cm® to 0,214 g/cm3®. When the fuel

mixture was reloaded in the Be(, the

54

Power Distribution in Fuel,

system was agalin made critical with
slightly less than 42 tubes containing
approximately 5.2 kg of U235, The
installation of one of the ARE regu-
lating rods and associated equipment
required the addition of 13 fuel tubes,
or approximately 1.6 kg of UZ35 o
malintain the
condition,

in a critical
calibrated by
measuring the increment of rod required
to override a known positive reactor

array
The rod was

period. The calibration curve is given
in Fig. 5.8. In the “zero’ position,
the bhottom of the rod is at the level
of the top of the BeO, and an initial
increase 1n reactivity upon insertion
of the (poison) rod is to be noted.
Similarly, the insertion of the final
few inches of the rod also increases
the reactivity because of effects
probably resulting from the competition
between the poisoning by the rod and
PERIOD ENDING DECEMBER 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

aiiis
DWG. 1745;3
300 i
250
© STAINLESS STEEL FUEL TUBE vs. VOID
A ALUMINUM FUEL TUBE vs VOID
— 200 ;
A
> A
e
>
F o150
<{
it A
z
g A
& 100 — O ]
A
O
A
50 o
0
0
& 2 . 4 8 8 10 12 14 18 18
RADIAL POSITION (in.)
Fig. 5.7. Ak of Stainless Steel Fuel Tube at Various Points Along Reactor
Radius. ‘

neutron reflection by 1ts stainless
steel container in an otherwise un-
reflected or partly reflected region,
The rod referred to here was a “weaker”
rod than the others designed for the
ARE. Tt was 2 in. in diameter, 30 in,
long, and it contained 0,145 g of B4C
per linear inch. Tts complete in-
sertion decreased the reactivity b
about $1.25. | '
Because of insufficient available
excess reactivity, 1t was necessary to
obtain an estimate of the poisoning by
a second ABRE regulating rod and by the
ARE safety rod by the “rod drop”
method, which utilizes the prompt
transient of a sudden reactivity
decrease and gives the result in terms
of the effective delayed-neutron
fraction. The rod assembly was mounted
7 1/2 in., from the center of the core,

and both the regulating rods and the
safety rod were measured there. The
following results were obtained: “weak”
regulating rod (0.145 g of B,C per
linear inch), $0.80; “strong” regulating
rod (0.68 g of B,C per linear inch),
$1.60; safety rod, $5.50. _
In the Aircraft Reactor Experiment
Hazards Summary Report(®) it was
assumed that the regulating rod in the
center of the core would be worth 0.4%
in Ak/k, and that the safety rod at a
position 7 1/2 in. from the center of
the core would be worth 5% in Ak/k.
Hence, the regulating rod seems to be
somewhat more effective thannecessary.
The safety rod is somewhat less
effective than was previously assumed,

(6)J. H. Buck and W. B, Cottrell, Aircraft:
Reactor Experiment Hazards Summary Report,

ORNL- 1407, p. 24 and 27 (Nov. 24, 1952).

55
ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

120

 

DWG. 17454

—

 

=~

 

110

 

100

 

 

 

20

 

80

 

 

 

 

70

 

60 |-

 

ROD VALUE {cents)

50

 

 

 

 

i
? /
40 }

 

30

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

| P

 

4

16

18 20 22 24 26 28 30 32 34

ROD POSITION (in))

Fig. 3.8,

but it 1s still regarded as adeqguate.,

A comparison was made of the two
fuels used in these experiments by
substituting tubes containing fuel of
the first loading (0.163 gof U23%/cm?)
for tubes containing fuel of the second
loading (0.214 g of U235/¢cm3) at
various points 1in the reactor core
when it was otherwise loaded with the
second fuel mixture.

The losses 1in

56

Calibration Curve of ARE Regulating Rod (weak).

reactivity incurred by this sub-
stitution are shown in Table 5,1. For
a similar experiment, an Inconel tube
was loaded with the second mixture and
substituted for an identically loaded
stainless steel tube. The losses
caused by the Inconel are also shown
in Table 5.1. This comparison 1is
important since the ARE fuel tubes are

fabricated of Inconel.
PERIOD ENDING DECEMBER 10, 1952

TABLE 5.1. LOSS IN REACTIVITY AT VARIOUS RADIAL POSITIONS CAUSED BY
CHANGES IN CORE CDMPONENTS

 

RADIAL POSITION
(in.)

REACTIVITY OF LOW DENSITY FUEL TUBE
V3. NORMAL FUEL TUBE

 REACTIVITY OF INCONEL FUEL TUBE

VS. STAINLESS STEEL FUEL TUBE

 

(cents) {(cents)
3.7 29.0 27.9
7.5 21.1 24.8
11.3 14.5 17.2
15 9.0 10.1

 

 

 

57
 
SUMMARY AND INTRODUCTION

E, P. Blizard

J. L, Meem, Associate

Physics Division

The Lid Tank Facility has been
applied directly in experiments needed
to solve problems of immediate interest
in the design of the air ducts through
the reactor shield of the GE-Convair
airplane (sec. 6). These experiments
are essential, since the configuration
is so complicated thatit ispractically
impossible to calculate., Neutron and
gamma 1sodoses around mockups of the
GE-ANP inlet and outlet ducts have been
obtained. The data obtained provide
the information necessary for specifi-
cation of the duct shields. In
addition, some measurements have been
made to determine the induced activities
around the duct. The leakage around
an alternate ducting arrangement is
also being measured. |

The Bulk Shielding Facility (sec.7)
has been applied, in part, to further
alr-scattering experiments. Although
these experiments are not yet con-
clusive, indications are that the
earlier experiments gave doses that
were too high; but the designs of the
1950 Shielding Board are not yet com-
pletely vindicated. The BSF has also
been used to extend the spectral and
dose measurements on the NEPA divided-
shield mockup. These data will now be
used to improve the calculation of the
dose at the crew position, Additional
experiments now under way at the BSF
include the irradiation of animals and
the measurement of the energy release
per U**° fission, |

The Tower Shielding Facility (sec.,
8) has been approved and the Laboratory
is committed to a schedule that calls

for completion of comstruction in
November 1953, Detailed engineering
design has been contracted to the
architectural firm of Knappen, Tippetts,
Abbett, McCarthy, who have revised the
tower design to eliminate cantilever
trusses., lhe new design had a signifi-
cantly higher neutron-scattering back-
ground than earlier designs; conse-
quently, alternate designs are now
being studied. Initial tests with the
completed facility, whichwill probably
make use of the GE-Convair shield
design, are scheduled to commence 1in
1954, ]

The nuclear measurement (sec. 10)
studies include an experiment to
determine the angular distribution of
neutrons scattered from nitrogen and a
new calibration of the fast-neutron
dosimeter. The angular distribution
of neutrons was measured by utilizing
the 5-Mev Van de Graaff accelerator.
Analyses of the pulse-height distri-
butions of the nitrogen recoils indi-
cate a definite change in the dif-
ferential cross section of nitrogen
that is associated with the resonance
value and an increase in the forward
scattering. The fast-neutron dosimeter
was ‘“calibrated,” in a bioclogical
sense, at the Zero Power Reactor at
Argonne National Laboratory by taking
advantage of a situation in which a
high fast-neutron dose wasaccidentally
received by an individual. Tt was
found that the ZPR leakage fluxes and
fast-neutron dose readings 1in the
water reflector were quite similar to
those expected from a comparison with

the BSF data.

61
ANP PROJECT QUARTERLY PROGRESS REPORT

6. DUCT TESTS IN LID TANK FACILITY

J. D. Flynn
G, T. Chapman

J. M. Miller
F. N. Watson

Physics Division

C. L. Storrs, GE-ANP

The Lid Tank Facility has been
applied primarily to a study of air
ducting for the GE-ANP reactor.
Partial mockups of i1nlet and outlet
“annular” ducts have been tested for
neutron and gamma attenuation, The
isodoses wmeasured arocund the ducts
indicate the proper shield shape in
the presence of the ducts.,

In addition, some experiments have
been carried out to determine the
activation of engines, structure, etc.
to be expected from neutrons that leak
out via the ducts., Possible methods
of improving the situation were ex-
plored when 1t was seen that the
activation could be excessive.

Multiple ““wavy’ cylindrical pipes
were also tested as an alternate duct
design. Although these are no doubt
superior to the “annular’ ducts as far
as shielding is concerned, they present
more difficult engineering problems.

G-E OUTLET AIR DUCT

Neutron Dose Measurements. With
the determination of the effect of the
transition section on the fast-neutron
dose and the thermal flux, the measure-
ments on the G-E outlet air duct have
been completed.(!) Figures 6.1 and
6.2 show thermal isodoses made with a
BF, counter and the automatic plotter.
Figure 6.3 gives the results of center-
line measurements with this counter.
Fast-neutron measurements are exhibited

in Figs. 6.4, 6.5, 6.6, and 6.7,
There is a slight discrepancy
between the measurements of thermal
and fast neutrons. The center-line
data should coincide when the relaxation
length exceeds 7.5 cm. The discrepancy
has been attributed to
(1)
Rep. Sept. 10, 1952, OBNL-1375, p. 48,

changing

 

62

Farlier results ave given in ANP Quar, Prog.

sensitivity of the BF; counter during
the month-long series of readings.
Unlike the fast neutron dosimeter, this
counter was not recalibrated during
the thermal-neutron experiment.
Despite the suspected slow drift in
sensitivity, the shape of the 1sodoses
should be reliable and should furnish
adequate information on the effect of
the transition section.

Two additional sources of error
should be mentioned. The relays that
operated the 1isodose plotter introduced
a few spurious counts, which caused
the plotter to operate 1 to 2 cm
further from the source than it should
Besides this, the effective
center of counting moved about inside
the detectors, depending upon the
direction of the flux gradient, and
introduced an error in position esti-

have.

mated at not more than 2 cm over the
range from the center line to the
side of the duct.

In repositioning the duct after
changing the transition section, 1t
was inevitable that a slight variation
in the distance from the source axis
should occur. Table 6.1, which gives
the actual coordinates of four points
on the duct identified in Fig. 6.1,
may be used in making precise calcu-
lations - the drawings give only
average position.

Calculated Neutron Dose. A calcu-
lation of the effect of the duct with
no transition section on the center-
line dose has been made with some
success. As can be inferred from the
isodose plots, the center-line dose 1s
not affected much by the side arm of
the duct. The calculation therefore
deals with an 8 1/2-in. air space
(essentially a void) in front of the
source plate.
£9

160
150
40 |
130
£20
110
100
SO
BO
70

60

40

30

DISTANCE FROM FRONT OF DUCT (cm)

20

-0

-20

~30

-40

-30
160

150

 

140

DWG. 16285A

NO TRANSITION SECTION
o e $8-1, TRANSITION SECTION

FUMBERS CURVES ARE THERMAL NEUTRON FLUX
NORMALIZED TO NEUTRON DOSIMETER

SOURCE WiTH NO
TION

SQURCE WITH 18-in. TRANSITION SECTION

130 120 WO 100G 20 B8O 7O 60 5C 40 30 20 (0 Q 10 20 30 40 50 &80 70 80 90 WO WO 120 150 140 180
DISTANCE FROM CENTER LINE {cm)

Fig. 6.1. Thermal-Neutron Isodoses Around G-E Outlet Air Duct (Shutter Open).

‘0T HJHWEOAO HNIOGNA doX¥dd

26T
¥9

DISTANCE FROM FRONT OF DUCT {cm)

160

150

140

130

120

VG

300

90

80

70

60

50

40

30

20

 

150

140

130

Fi

120

g.

Sy

OWG 16220A

—_— B-in TRANSITION SECTION
— e 1 2 - . TRANSITION SECTION

NUMBERS ON CURVES ARE THERMAL NEUTRON
FLUX NORMALIZED TO NEUTRON DOSIMETER

SOURCE WITH 6-1n. TRANSITION SECTION

SECTION

"o 100 90 80 7O 80 850 40 30 20 10 o 10 20 30 40 5C¢ 60 TO B8O S0 100 110 120 430 140 150

6.2.

DISTANCE FROM CENTER LINE {cm)

Thermal-Neutron Isodoses Around G-E (utlet Air Puct (Shutter Open).

L0444 SSHU90Ud XTHALHVNO LDAT0¥d dNV
THERMAL NEUTRON FLUX, NORMALIZED TO NEUTRON DOSIMETER (mrep/hr)

Fig.
Duct

2 X10°

PERTOD ENDING DECEMBER 10,

3X10

 

 

 

18-in. TRANSITION SECTION
—~12-in. TRANSITION SECTION

 

1952

DW!. 1629148 ¢

 

 

 

s—m.TRANsnvom SECTION
NO TRANSITION SECTION

3

72 B4 96 108 120 132

Z, DISTANCE FROM SOURCE (cm)

144

e.3.
{Shutter

Open).

 

 

156

 

68 180

Thermal-Neutron Centef-Line Measurements Behind G-E Outlet Air

- 65
ANP PROJECT QUARTERLY PROGRESS

&-in. TRANSITION SECTION
z2=125 cm

FAST NEUTRON DQSE (mrep/hr)

S

18-1n.
Z=1442 cm

12-in, TRANSITION SECTIO
Z =140 cm

 

24 36

48

REPORT

DWG. 17370

NO TRANSITION SECTION
Z2=99cm

6-in. TRANSITION SECTION
z=143cm

TRANSITION SECTION
z=109 cm

12-in. TRANSITION SECTION
Z2=130 ¢m

RN

12-in. TRANSITION SECTION
Z=135cm

18+4n. TRANSITION SECTION
Z =1474 cm

 

60 Te B4 96 408

y, HORIZONTAL DISTANCE FROM SOURCE AXIS (cm)

Fig. 6. 4.

66

Fast-Neutron Dose at End of G-E Outlet Air Duct.
FAST NEUTRON DOSE (mrep/hr)

PERIOD ENDING DECEMBER 10, 1952

DWG. 162244

NO TRANSITION SECTION
y29892 cm '

B-in. TRANSITION SECTION
¥=92.9 cm

12-1n. TRANSITION SECTION
¥=947cm

1840 TRANSITION SECTHON
y=933cm

 

0 12 24 36 48 60 72 84 26 108 120
2z, DISTANCE FROM SOURCE (cm) '

Fig. 6.5. Fast-Neutron Dose at Side of G-E Outlet Air Duct,
| 92 < Y < 93.3.

67
ANP PROJECT QUARTERLY PROGRESS REPORT

FAST NEUTRON DOSE (mrep/hr)

68

NO TRANSITION SEGTION
=102 cm

12-in. TRANSITION SECTION
Y=1047 c¢m

18-in. TRANSITION SECTION
¥=103.3 ¢cm

        

12 24 36 48 60 72 84 96 108 120
Z, DISTANCE FROM SOURCE (em}

Fig. 6.6. Fast-Neutron Dose at Side of G-E Outlet Air Duct,
102 < Y = 104.7.
PERIOD ENDING DECEMBER 10, 1952

2x10°

FAST NEUTRON DOSE (mrep/hr}

EXPERIMENTAL DATA:

24 36 48 60 T2 B4 96 108
z, DISTANCE FROM SOURCE {cm)

Fig. 6.7. Fast-Neutron Center-Line Measurements
and Various Transition Sections.

 

120 132 144

with G-E Outlet Air Duct

69
ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

 

 

TABLE 6. 1. COORDINATES 0OF FIDUCIAL POGINTS FOR FIG. 6.1
LENGTH OF POSITION I 1Y 11X 1V
TRANSITION
SECTION (in.) Y{em) z(cm) Y{cm) Y(em) Y{ cm) z( cm)
0 43.1 97.0 61.0 90.9 45.9 23.3
6 44.8 111. 3 62.4 91.9 45.1 40,2
12 46.0 128.7 63.7 93.1 43.4 55.5
18 45,4 144.2 62.3 92,2 44,5 71.5

 

 

 

 

 

 

The neutron dose for this con-
figuration has been obtained from the
data with water alone by using asimple
assumption, The dose at a distance r
in water froman isotropic point-source
of strength dS{(%) js

G(r)

2

 

d D(r) = dS

A2
where G(r) is an undetermined function
of r. Integration of this expression
over a source disk of radius a gives,
for a point on the axis at a distance
Z from the source,

A
d(Z) -EfG(r)dr ,
2 . r

where o is the specific source strength
in neutrons/cm?*sec, and R is the
distance from the edge of the source
to the detector.

Now, consider the case in which
there 1s a plane-bounded void of
thickness (r-r’') between source and
detector, which are sti1ll separated by
a distance r. For a point source,
assume the dose to be

 

G(r')

dD (r,r') = dS T

 

4771
Integrating this as before gives

R
D (z.7') - o G(r')
vt S 9 r
zZ

Since the ratio of r to r’' is a con-

 

(Z)This caleculation is mainly based on the
work of E. P. Blizard, Introduction to Shield
Design, Part I, ORNL CF-51-10-70 (Jan. 30, 1952),

70

 

stant throughout this integration,

Hl
o o G(r')
Dr(Z,Z ) :“E —

 

!
- dr’
r

Zl
where R’ 1s the distance,
between the edge of the source and the
detector. By referring to a sketch of
the geometry, Fig. 6.8, and to the
previous integral, it 1s immediately
apparent that this expression 1s just
the dose at a distance Z' 1in water
alone from a source of reduced radius

in water,

a', where a' 1s given simply by the
ratio
! t
a z
R
a z

To find the dose that would be
measured with a source of different
radius, an approximation given by
Blizard to the Hurwitz transformation
from a disk to an i1nfinite plane
source 1s used; that 1s, the trans-

formation is made frow the data for

 

water alone and the actual source
radius, to an 1nfinite plane source,
and from this back to a source of
smaller radius. The approximate ratio
of dose with i1infirite source to that
with source of radius a 1is
D(Z' , ) L1
e > — t a ,
D(Z', a) 2
2>\.2 fr X4 + 1
a? A ’

where A 1s the relaxation length of
the water.
 

 

 

PERTOD ENDING DECEMBER 10,

1952
DWG. 17371
VIRTUAL SOURCE

DETECTOR

 

 

 

 

 

AN

 

 

 

"o

e

 

Fig. 6.8.

With the obvious substitutions, the
dose at a distance Z with a VOld of

 

- thickness Z - Z' becomes .
| D (Z) = D(z') 22
v - 1+ 22’
, 2)\2 \ al Z!
o = — + 1 ’ TTEF T
a'? A / a Z

where D (Z) is the dose with a void and
" D(Z) is the dose with water alone. The
~curve in Fig, 6.7, labeled “no transitioen
section]’ was calculated by using this
formula, The experimental points are
also indicated, and it can be seen
~that the agreement between experimental
~data and the calculations isexcellent.

Some very rough calculations have
" been made to explain the effect of the
transition sections on the center-line
~data. If the water sections are
assumed to be opague to neutrons, the
calculated dose 1s lower than the
. observed dose. More accurate calcu-
~lations are yet to be made. ‘

N=

 

 

e PURE WATER —— o]

Geometry for Calculating the Effect of a Void on Neutron Dose.

G-E INLET AIR DUCT

A mockup of the G-E inlet air duct
has been tested in the Lid Tank Facilitw
Some gamma shielding in the form of
slabs of 7/8 in. of iron and 1 1/2 in.
of lead was used to simulate the
design shield along the center line.
In spite of care in choosing from
available slabs of iron and lead, it
was 1mpossible to make the mockup
simulate the actual situation very
closely, As a consequence, caution
should be used in applying the gamma
data to the design.

Thermal- and fast-nentron and gamma
data were taken with the standard Lid
Tank instruments, and the results have
been converted into isodose plots,
Figs., 6.9 and 6.10. It is unfortunate
that the source strength was completely
inadequate to enable measurement of
the fast-neutron dose at the end of
the duct, *

The final step in this experiment
was to study the attenuation of the

71
Gl

Zz, DISTANCE FROM SOURCE {cm)

200

180

1e0

$40

120

100

80

60

40

20

160

40

1098

/7
1005 10t

120

Fig.

100

6.9.

80

60

4Q

20

. SDURCE

=0

..
.

3\<j

-20

R
g
<

N

-40

¥, HORIZONTAL DISTANCE FROM SOURCE AXIS (cm)

DWE. 17159

——— THERMAL MNEUTRON
— FAST NEUTRON

FIGURES ON CURVES ARE DOSE iN mrep/hr

THERMAL FLUX IS NORMALIZED 7O
FAST NEUTRON DOSE

1072

1g7"3

107!

10—0.5
N

LEAD
|RON

AIR-FILLED DUCT

TRANSITION SECTION

 

~60 -80 -100 —120 =140

Neutron-Isodose Measuremenis Around G-E Inlet Air Duct Mockup.

=180

LMOdAY SSHY90Ud ATHAIYVN0 104l 0dd dNV
Z, DISTANCE FROM SOURCE (cm)

OWG. 17160

 

260 ; ; ; T

 

 

: : |
| i | !
| ——— SIDE SHIELD REMOVED
; 1095 pshe i

 

 

 

 

 

160 ;

 

140 :
10" r/he

 

 

 

 

 

 

 

 

  
     

 

 

 

 

 

 

 

 

 

      

 

 

 

 

       
  

      
  
  
  

  
     
  
  

          

 

    

 

         
 

  

 

   

 

 

 

 

 

 

  
  

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

€l

 

 

 

 

 

 

 

 

 

 

: ;
120 7 ' \\ i
1018 r/nr é ?
- ' : R L ! . E
100 : : ? |
g E 1 i
: : ' i
! ' : i
102 r/he ; ] 5' ;
¢ ' i i ! : i ! i
. ; i { ! ; i
: :‘ i ; z ; i
! i ] e i i ;
: E L A S S S SR | !
60— : AR AR SIS : FRON §
| ; § s
i ; : !
, _AIR-FILLED DUCT i
4 ’ ! !
4 4 ;
0 7 : :
/’ : ; :
3 ‘ : :
10% r/hr : . : i
/ : \ S : !
/ f N b E s
20 o A AR EN T [ :
/ / 10% r/fr oS -} TRANSITION SECTION 5 5
- { . TRON-T o b o S S ; ;
f R R X a % ;
i R :.'. _'-' I ; i i i !
o | A 4Bk 1L A S U
i RN BOURCE TR f 5 ? E
; NN, >~ N “ | ; '} :
5 \\ \\\E"\\.\ \b\ o i | ! v
0 -140

20 0

-20

¥y, HORIZONTAL DISTANCE FROM SOURCE AXIS (cm}

-40

~80 -80

-

Gamma-Isodose Measurements Around G-E Inlet Air Duct Mockup.

-120

‘0T MAGWIDIU INIANI aolydd

cbo6l
ANP PROJECT QUARTERLY PROGRESS REPORT

gamma shield, Center-line traverses
werec made as the shielding slabs were
successively removed. The results

appear 1in Fig. 6.11.

INDUCED ACTIVITY AROUND A DUCT

An experiment has been performed 1n
the Lid Tank Facility with the primary
object of determining the order of
magnitude of the induced radioactivity
to be expected outside the air ducts
in the GE-ANP initial engine shield.
In addition, 1t was desired to estimate
the value of thin boron shields in
lowering this activity, to investigate
the utility of BF,-counter measurements
in predicting activation, and to
determine the effect of borating the
shield water around the ducts,

The mockup of the G-E outlet azir
duct that had been previously tested(®)
was modified by the addition of a tank
so that borated water could be placed
around the duct. Figure 6,12 shows
the configuration. The outside of
this tank simulated the outside of the
design shield. Measurements were made
at the end and the side of the duct,
as indicated in Fig. 6.12.

Two types of measurements were
undertaken: samples of cobalt and
manganese were exposed and their
activities measured, and traverses
were made with a BF3 counter. In all
cases, various thermal-neutron shields
were used, and the results were com-
pared with unshielded measurements,
Data were obtained with both pure and
borated water around the duct.

The induced activities under various
conditions are tabulated in Table 6.2,
and the results of the counter traverses
are presented graphically in Figs.
6.13, 6.14, 6.15, and 6.16. Figure
6.17 depicts the results of center-
line traverses with no duct in the
tank but with various thermal shields
around the counter.

 

(B)ANP Quar. Prog. Rep. June 10, ORNL-

1294, p. 72,

1852,

74

Since the activation measurements
were of an exploratory nature, a com-
plete investigation ofall combinations
of parameters was not attempted. The
samples were prepared hurriedly, and
therefore errors of 30% are to be
expected from differences in the
geometry of exposure and counting. In
the low flux available in this experi-
ment, the cobalt did not become radio-
active enough to afford an accurate
measurement of its activity,

The various ratios of counting
rates are given in Tables 6.3, 6.4,
and 6.5, which show, respectively, the

effect of thermal shields, acomparison
between the counting rates at the side
and end of the duct, and the effect
of borating the shield water.

It is apparent that a thin boron
layer is far more effective 1in shield-
ing the boron counter than inshielding
either manganese or cobalt. This 1is
to be expected, since an 1important
part of the activation of the latter
is due to resonance absorption at
epithermal energies where the boron
cross secticn is relatively low,

The variously shielded detectors
are senslitive to different neutron
energies, and therefore an indication
of the comparative neutron spectra at
the side and the end of the duct can
be obtained by comparing the counting
rates under these conditions. When the
expected experimental errors are con-
sidered, there does not seem to be
evidence of significant differences 1in
the spectra at these two positions.
With pure water in the tank, there are
more epithermal neutrons than thermal
neutrons; or looking at the situation
from the opposite viewpoint, the ducts
increased the thermal flux. Under
these conditions, a conservative
estimate can be made of the expected
activation around a ducted shield from
the readings of a BF, counter, provided
an experimental comparison has been
made at some point without ducts.

It was at first a surprise to find
that adding boron to the shield water
GAMMA DOSE {mr/shr)

PERIOD ENDING DECEMBER 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10% 1 I
........ — e e e . h amrem e s aaaaafanmmmm e M —— . mmmmmm e i 4 o ememeeeemiaeef
5 S
2
103
Lol
5
2
10?
g
2 ........
.1
10 s e N
A e g e e e \‘
. l?\
Bl — AS SHOWN ABOVE
5 , : : \\‘m
A— 4in. x 4in. IRON AND FIRST LEAD PLATE REMOVED
] ® — 4in x 4in. IRON AND BOTH LEAD PLATES REMQOVED
@— 4in.x 4 in. IRON, BOTH LEAD PLATES AND {RON PLATE REMOVED
2
{ i i
60 70 80 90 100 1o 120 130 140 150 160 170
DISTANCE FROM SOURCE (cm)
Fig. 6.11.

Gamma Center-Line Measurements with G-E Inlet Air Duct.

15
ANP PROJECT QUARTERLY PROGRESS REPORT

increased the flux of epithermal
neutrons by about 25%. However, calcu-
lations of the amount of water dis-
placed in adding the boron showed that
the result is gquite reasonable. The
borated water contained 1.005 wt %
boron and 1.7 wt % potassium and had a
specific gravity of 1.0403. Thus,
there was a displacement of 1.3% of
the hydrogen, which resulted in higher
epithermal flux. The thermal neutrons
were naturally greatly reduced by the
boron. However, placing a thin layer
of boron at the surface of the shield
would eliminate thermal neutrons with-
out displacing so much water,

RADIATION AROUND AN ABRAY OF
CYLINDRICAL DUCTS

Previous measurements of the radi-
ation around mockups of the annular
air ducts for the G-E reactor showed
that these ducts will permit the
escape of a large flux of fast neutrons
through the sides of the shield.
Furthermore, this annular duct con-
figuration is undesirable because of

In view of these
it was desired to
study the shielding properties of an
alternative ducting method, namely, an
array of bent cylindrical ducts.

its large weight.
considerations,

DWG. 17153

POSITICN OF MEASUREMENTS

 

BORATED WATER ———f_

BORATED WATER

 

 

POSITION OF MEASUREMENTS

 

AIR-FILLED DUCT

 

 

TRANSITION SECTION ————=

SOURCE -——“-1/ //

Configuration for Induced-Activity Measurements on G-~E Qutlet

Fig. 6.12.
Air Duct.

76

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.

 

 
Figure 6.18 is a photograph of the

bent cylindrical ducts in position in
the Lid Tank Facility. There are 19,
round, aluminum pipes in a hexagonal
array. FEach pipe has an outside
diameter of 3 15/16 in. and 1/16-1in.
walls. The pipes have three straight
sections 22 1in. long connected with
45-deg bends. 1In the plane of the
source plate, the centers of the
pipes define equilateral triangles
with sides 7.33 1in. long. Since the
ducts leave the source plate at an
angle of 22.5 deg to the normal, the
array 1s not symmetrical about the
source axis. With this arrangement,
the fractions of the total volume that
are alr, water, and aluminum are 26.6,
71.8, and 1.6%, respectively, 4
The radiation measurements around
this array have been obtained in con-
siderable detail. An analysis of the
results shows that the gamma dose can

PERIOD ENDING DECEMBER 10, 1952

be accurately computed by considering
the ducts and water to be a homogeneous
shield of appropriately reduced density.
The neutrons, however, stream through
the ducts in sufficient numbers to
increase the flux at the end to one or
two orders of magnitude greater than
that predicted on a reduced density
basis. This is in accordance with the
duct theory proposed by Simon and
Clifford, that is, geometrical attenu-
ation along the straight sections and
inverse sine reflection at the bends.
The design of this particular array is
not optimum, since the attenuation
through the ducts is not high enough
to exceed that through the equivalent
amount of water, An optimum design
might have another bend in the ducts
to 1ncrease their attenuation, or
closer spacing between ducts to further

reduce the density of the equivalent

homogeneous shield.

TABLE 6.2. SUMMARY OF ACTIVITIES INDUCED IN THE LID TANK:FACILITY

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

f ACTIVITY INDUCED
SAMPLE NEUTRON SHIELD POSITION. WATER AROUND DUCT (me/ g)
: : me/ g
Outlet Duce
Cobalt None End Borat%d 58.1 t 9.6 x "1t
Boron cover End Borated 5.5 9.1 x 10710
None Side Borated 138.3 + 10.1 x 1071
Boron cover Side Borated 18.3 + 9.3 x 1p~10
Manganese None End Plain - 178.5 + 1.5 x 10”8
Boron cover End Plain 4.64 + 0.17 x 1078
Boron backing End Borated 68.3 t 1.0 x 10°®
Boron cover End Borated 3.16 + 0.16 X 10°8
Boron backing Side Borated 124.3 + 0.8 x 108
Boron cover Side Borated 11.7 £+ 0.3 x 10°8
Inlét Duct
Manganese None End Plain 1.04 + 0.15 x 107
None Side Plain 2.57 + 0.02 x 10°*
Notes: (1) Manganese activated to saturation at 6-watt source strength. (2) Cobalt activated 100 hr

at 6 watts. (3)
clude systematic errors.

The erfor factors given are standard deviation of counts taken and do not in-

17
ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TABLE 6. 3. EFFECT OF THEGMAL SHIELDS ON EXTERIOR ACTIVATION
RATIO OF COUNTS
DETECTOR POSITION RATIO MEASURED
In Plain Water In Borated Water
Boron Side Bare to Cd covered 57 26
Bare to B covered 250 130
Cd covered to B covered 4.4 5.0
End Bare to Cd covered 41 26
Bare to B covered 230 150
Cd covered to B covered 5.6 5.7
No ducts Bare to Cd covered 40
Bare to B covered 180
Cd covered to B covered 4.4
Cobalt Side Bare to B covered 5 to 15
End Bare to B covered 4 to 100
Manganese End Bare to B covered 39
B backed to B covered 22
Side B backed to B covered 11
TABLE 6. 4. RATIO OF EXTERIOR ACTIVATION AT SIDE TO EXTERYOR ACTIVATION AT END
RATIO OF COUNTS
DETECTCR SHIELD RATIO MEASURED
In Plain Water In Borated Water
Boron None Side to end 3.4 2.9
Cd cover Side to end 2.5 2.9
B cover Side to end 3.1 3.3
Cobalt None Side to end 2.4 +1
B cover Side to end 3
Manganese B backed Side to end 1.8
B cover Side to end T
TABLE 6. 5. EFFECT OF WATER BORATION ON ACTIVATION
RATIO OF COUNTS
DETECTCR SHIELD RATIO MEASUBRED
1' At Side of Duct At End of Duct
Boron None Plain to borated 1.4 1.2
Cd cover Plain to borated 0.64 0.75
B cover Plain to borated 0.73 0.77
Manganese B cover Plain to borated 0.73 1.5

 

 

 

 

 

78
counts /min

 

PERIOD ENDING DECEMBER 10, 1952

SO

DWG. 17154

COVER
Zgp#13830m}

UM COVER
Zgo7136.3 cm

COVER
£go7136.3 em

~(—— PURE WATER AROUND DUCT
—A—— BORATED WATER AROUND DUCT Z,136.3 cm |

a0 45 50 55 60 65 70
¥, HORIZONTAL DISTANCE FROM SOURCE AXIS (cm)

Fig. 6.13. Boron-Counter Measurcments at End of G-E Outlet Air Duct.

79
ANP PROJECT QUARTERLY PROGRESS REPORT

3x108

remsmae e PURE WATER AROUND DUCT
(s — BORATED WATER AROUND DUCT

108

- y=1145 em___ | T

R NO COVER

-

counts /min

R\ CADMIUM COVER q

BORON COVER

 

30 40 50 60 70 80 20 100 110 {20 130 14Q
2, DISTANCE FROM SOURCE {cm)

Fig. 6.14. Boron-Counter Measurements at Side of G-E Outlet Air Duct.

80
PERIOD ENDING DECEMBER 10, 1952

DWG. 17156

 

PURE WATER ARCUND TANK

—— e —e BORATED WATER AROUND TANK

5
=
&
e
£
1) - 2 2136.3¢m
w 2 - A
D2 et A
o
2z
O
o
= -1
o 10
=z
—
[7p]
<l
L T
g 5 ‘ 221463 cm
&
Ll
~1
O
<
g
O 2
=z
>
2
i
[T
- 2
é 10
[ee
i_‘
=
Ll
=

Z2=156.3 cm

< 5 S
= N
o
¥ N
'.._

z

107

5 %104

-0 0 1o 1 20 30 40 20 60 70 30 20 100
y. DISTANCE FROM SOQURCE-CENTER LINE (cm)

?

Fig. 6.15. Thermal -Neutron Measurements at End of G-E Outlet Air Duct.

81
ANP PROJECT

THERMAL NEUTRON FLUX NORMALIZED TO FAST NEUTRON DOSE {mrep/hr)

Fig. 6.16.

82

 

QUARTERLY PROGRESS REPORT

DWG. 17157

——PURE WATER AROUND DUCT
—~=BORATED WATER AROUND DUCT

40 60 80 100 120 140 160
z, DISTANCE FROM SOURCE (cm)

Thermal ~-Neutron Measurements at Side of G-E Outlet Air Duct.
counts / min

PERIOD ENDING DECEMBER 10, 1952

DWG. 17158

NO COVER

CADMIUM COVER

 

2
BORON COVER
10°
5
2
10% :
12 24 36, a8 60 72 B4 96 108 120

2, DISTANCE FROM SOURCE (em)

#i9.6. Boron Counter Centerline Meosurements in Pure Water.

Fig. 6.17. BoronFCounter Center-Line Measurements in Pure Water.

83
¥8

Fig.

6' 18-

 

Multiple Cylindrical Ducts in the Lid Tank Facility.

t phoTO 10289 1

" ;;e,«,y:‘v‘m?;
]

LHOdAY SSHY90Hd ATHALYVNO LOAf0Yd dNV
7
J. L. Meem
R, G. Cochran
M. P. Haydon
K. M. Henry
H. E. Hungerford
Physics
Gamma-ray spectral measurements on
the divided-shield mockup have been
extended, and the neutron and gamma
dose measurements on the shield mockup
are reported. The air-scattering
experiments with the reactor at 100 kw
are now under way, and the program of
irradiating monkeys with the reactor
has been started.

MEASUREMENTS WITH THE DIVIDED-SHIELD
MOCKUP

Two further sets of measurements of
gamma-Tay energy spectra and angular
distributions from the divided-shield
mockup (with lead disks) were made
after a preliminary attempt by the NDA
group to calculate the dose to be
expected at the crew position. The
measurements were made at iy = 60 deg,
6 = variable, (!’ and ¥ = 60 deg, ¢ =
variable., The energy spectra obtained
are shown in Figs, 7.1 and 7.2, and
the angular distribution for the angle
g is in Fig. 7.3. For all
measurements, the spectrometer col-
limator remained horizontal,

Table 7.1 shows all gamma-ray
spectral measurements completed since
the last tabular summary.(?) It 1s ex-
pected that the information is now
sufficiently complete to enable an
accurate calculation of gamma dose at
the crew position.

shown

Since the collimator 1s always kept
horizontal in the measurements, Y, @,
and 0 are sufficient to specify the

(I)For a definition of angles Y, ¢, and F see
Table 7.1, For a discussion of the experimental
setup, see F. C, Maienschein, Gamma-Ray Spectral
Meesurements with the Divided-Shield Mockup,
Paert I, ORNL-CF-52-3-1, Part II, ORNL-CF-52-7-1,
Part II'I, ORNL-CF-52-8-38,

(2)ANP Quar. Prog. Rep. June 10, 1952, ORNL-
1294, p. 45.

PERIOD ENDING DECEMBER 10, 1952

. BULK SHIELDING FACILITY

E. B. Johnson

J. K. Leslie

T. A. Love

F. C. Maienschein
G. M. McCammon

Division

position and orientation of the spec-
trometer. The angle & for the measure-
ments with ¢ = 60 deg and ¥ = 0 was
determined so that the spectrometer
collimator would always look at the
edge of the lead disks in the divided-
shield mockup.

Center-line measurements of gamma-
ray and fast-neutron dosages and also
thermal-neutron flux have been com-
pleted and are shown in Figs. 7.4, 7.5,
and 7.6. The attenuation curves are
shown with and without the shield in
place. It should be remembered that
for these measurements the reactor had
a beryllium oxide reflector. (3’ The
dashed curves on Figs. 7.4, 7.5, and
7.6 represent the center-line data(*’
with the water-reflected reactor.

AIR-SCATTERING EXPERIMENTS

The air-scattering experiments with
the reactor at a power of 100 kw are
well under way, Indications are that
the previous measurements, reported
last quarter, were somewhat pessimistic,
but confirmation of the ANP-53 shield
design{®) is not yet clear. It is to
be expected that these experiments
will soon be concluded, since the BSF
1s poorly adapted to air-scattering
measurements.

IRRADIATION OF ANIMALS(®?

The program of irradiating monkeys
by using the reactor, as outlined in

(B)ANP Quar.
ORNL-1227, p. 76.

{4)ANP Quar. Prog. Rep. June 10, 1951, ANP.65,
p. 110,

(S)Repart of the Shielding Board for the Air-
craft Nuclear Propulsion Program, ANP-53 (Oct. 16,
1950).

(G)This is a joint program in which the USAF
School of Aviation Medicine, Consolidated Vultee
Aircraft Corporation, and Wright Air Development
Center, as well as ORBRNL, are participating. -

Prog. Rep. March 10, 1852,

85
ANP PROJECT QUARTERLY PROGRESS REPORT

Y =60°
8 = VARIABLE

GAMMA-RAY FLUX (gammas/cm@/sec/Mev/watt/steradion)

0 2 4 6 8 10
GAMMA-RAY ENERGY (Mev)

Fig. 17.1. Gamma-Ray Flux at Edge of Lead

86

 

Disk.

DWG. 16777
PERIOD ENDING DECEMBER 10, 1952

T
DWG. 16776

GAMMA-RAY FLUX (gammas/eme/sec/Mev/watt/steradian)

 

"0 2 4 6 8 10 2 14
GAMMA-RAY ENERGY (Mev)

Fig. 7.2. Gamma-Ray Flux at Edge of Lead Disk as a Function of the Azimuth
Angle.

87
ANP PROJECT QUARTERLY PROGRESS REPORT

e
SO DWG. 16778
>« HYPOTHETICAL SHIELD
\ BOUNDARY

N\
\

DIVIDED SHIELD MOCKUP

    
  
  

LEAD SHADOW DISKS

d'\o'(\\
owsxef
DIVIDED SHIELD | o o
AND REACTOR ¢ 80°
70°
60°
50°

a0°

rFig. 7.3. Gamma-Ray Flux as a Function of Angle for Energies as Shown
(BSML, ¥ = 60°).

88
PERIOD ENDING DECEMBER 10, 1952

103

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g
_ ,,V_A,V,J._.___”_..,J,,,,_ b
|~ Be0 REFLECTOR FOR ! *
(. DIVIDED-SHIELD REAGTOR ® - EXP.6. MEASUREMENTS MADE ON BeO-REFLEGTED DIVIDED-
e SHIELD REACTOR IN OPEN WATER -
T — Jeee
A P . MEASUREMENTS MADE WITH REAGTOR IN POSITION IN -

DIVIDED--SHIELD MOCKUP, 0.4 %, -BORATED WATER,
NO SHADOW SHIELD

. MEASUREMENTS MADE WITH REAGTOR IN POSITION 1IN |
DIVIDED~SHIELD MOCKUP, 0.4 % - BORATED WATER,
WITH LEAD SHADOW SHIELD

S e

 

 

 

 

   
   
 

 

 

 

 

 

IO IR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  
   
     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10
1
1o
5
3
=
S ;
= REAR PERIPHERY OF
W 1072 | DIVIDED-SHIELD ---- -
g [~ MOCKUP { EXP. 7) 2
a i I
3 — } |
: !
© | J_--_..“._L _______
Tol S r -------- Y-
e /
o S | ' Ny oL bt ]
i 1 L 5%
N L i
e I EaD SHADOW-SHIELD - S — -

 

 

 

 

CISKS iN POSITION “"— B —

 

 

10"5 e R

 

 

 

 

 

 

 

 

 

 

 

 

 

1078 |

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1077

 

0 40 80 120 is0 200 240 2P0
CENTER-LINE DISTANCE FROM ACTIVE CORE {cm)

Fig. 7.4. Gamma-Radiation Measurements Along Center Line of the Divided-
Shield Mockup in the Bulk shielding Facility.

89
ANP PROJECT QUARTERLY PROGRESS REPORT

      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  
      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

       
 
 
 
 
 
 
  
 
 
 
 
 
 
 
 
  
   
  
     
 
   
    

 

    

 

 

106 o I ot
e
----- . — BeQ REFLECTOR FOR —
5 / DIVIDED-SHIELD REAG
. T @ — EXP. 6 MEASUREMENTS MA
REAGTOR IN OPEN WATER

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o T @ —ExP 6 MEASUREMENTS MADE WITH DIVIDED-SHIELD |
0% Zi . ; REACTOR IN OPEN WATER, 4% in OF LEAD —
e INTERPOSED {0 cm FROM REACTOR TO DECREASE F——
PR —— — GAaMMA FLUX ]
AN T 1 A——EXP B MEASUREMENTS MADE WITH DIVIDED-SHIELD Lo
//,- REACTOR IN OPEN WATER, 8in OF BISMUTH —
Aéé INTERPOSED 20 cm FROM REAGTCR TO DECREASE
e ‘ GAMMA FLUX _
' MEASUREMENTS MADE WiTH REAGTOR IN fiééi
POSITION IN 0.4 %-BCRATED-WATER DIVIDED~ |-~
SHIELD MOGKUP, NO SHADOW SHIELD [—
MEASUREMENTS MADE WITH REAGTOR IN o]
FOSITION IN 0.4 % - BORATED-WATER DIVIDED- t
SHIEL MOCKUP, WITH LEAD SHADCW SHIELD
N POSITION
- N - DATA FOR WATER-REFLECTED REACTOR, CENTER-
- e LINE MEASUREMENTS IN WATER S
.6 ——— b L . L : o FE
=
T 10 . AT;_
& - -
S
o
@
E
wJ
8
0
=
Q
o
—
o
st
=
.\I— 10_.1 s
b |15 =
=
102 = 1 Ny o=t
e A e
1 . ..,J,_*‘._._..v—,,,,d,.*j fFooo _ |
- ot ——REAR PERIPHERY OF V%
3 J DIVIDED-SHIELD MOCKUP %
10 —— if"“f’::'..T'..___.__,ff\._:Jf.__':;_L ~T 3 o ;’_
W b b b e e N e L e
L o \‘.:i\‘ ) ]
_____ ] | : L
- — TR
10_5 103e- IiIprTI— & — A'—_~J coiiz ~
o i l
0 20 40 60 80 100 120 140 180 180

CENTER-LINE DISTANGE FROM ACTIVE CCRE (cm)

Fig. 7.5. Fast-Neutron Dose Measurements Along Center Line of the Divided-
shield Mockup in the Bulk Shielding Facility.

90
PERIOD ENDING DECEMBER 10, 1952

REFLECTCR FOR
DIVIBED-S

® — EX
REACTOR IN OPEN WATER
108 o sl ©o— EXP 6 MEASUREMENTS MAGE WITH DIVIDED-SHIELD
e REACTOR IN OPEN WATER, 4 ' in. OF |EAD
INTERPOSED 10 em FROM REACTOR TQ DECREASE
GAMMA FLUX
MEASUREMENTS MADE WiTH REACTOR IN POSITION
IN DIVIDED-SHIELD MOCKUP, NO SHADOW SHIELD
MEASUREMENTS MADE BEHIND DIVIDED-SHIELD
MOGKUP, WiTH L EAD SHADOW DISKS IN POSITION

AND 7. IN-FOIL MEASUREMENTS

DATA FOR WATER-REFLECTED REACTOR

107

102

THERMAL-NEUTRON FLUX (n%/wm‘f)

12

g3

1974

 

]
10
0 20 40 60 80 100 120 140 160 180 200
CENTER-LINE DISTANCE FROM ACTIVE GORE (cm)

Fig. 17.6. Thermal - Neutron Measurements Along Center Line of the Divided-
Shield Mockup in the Bulk Shielding Facility.

91
the last quarterly report, (7? has been
started., A plan view of the instal-
lation is shown 1in Fig. 7.7. The
animals are contained in watertight
cages that are lowered under water into
positioning frames on either side of
the reactor as shown. The cages are
located 50 5/8 in. from the center of
the reactor with 7 1/2 in. of lead
between the reactor and the cages.

The animal irradiation is done only
on the week end, and the reactor 1is
free to be moved forward or backward
for shielding experiments during the
week,

In cooperation with the Health

Physics Division, an extensive set of

(T)ANP Quar. Prog. Rep. Sept. 10, 1852, ORNL-
1375, p. 66,

TABLE 7.1.

dosimetry measurements in the cages
has been completed. The results will
be reported later,

OTHER EXPERTIMENTS

A report on the results obtained
from the irradiation of electronic
equipment has been issued.(®? The
experiment for determining the energy
released per U233 fission has been
completed, but the calculations have
not yet been finished. Because of low
prioerity on reactor time, no pProgress
has been made on measuring neutron
spectra or capture gamma-ray spectra.,

 

(B)A. N. Good and W, T. Price, Effects of
Nuclear Bediation on Electronic Components and
Systems, WADC-TR-187 (Aug. 18, 1952).

SPECTROMETER POSITIONS FOR GAMMA-RAY SPECTRAL MEASUREMENTS

WITH THE DPIVIDED-SHIELD MOCKUP (WITH LEAD DISKS)

 

 

 

yr Pt G »
(deg) (deg) (deg)

0 0 0, 10, 20, 30, 40, 50

50 0 ~20, -10, 0, +10, +20, +30, +40,
+50, +60, +70

60 0 ~30, -20, -10, O, t10, +20, +30,
+40, +50, +60, +70, 180

60 5, 10, 15, 20, 25, 30 See footnote

 

 

 

"

nose of the spectrometer collimator.

¢’is the angle between the aircraft axis and a line connecting the pseudo reactor cemter with the

*s> is the angle between the horizontal plene anda plane that includes the aircraft axis and thenose

of the aspectrometer collimator.

ss¢@ js the angle between the spectrometer collimator and the line joining the pseudo reactorcenter

and the nose of the spectrometer collimator.

92
UNCLASSIFIED
OwWG. £-13262

 

 

 

 

 

 

1L

.
]
||

7

 

 

 

  
  

 

 

 

 
    
 
 
 

    

ET= Seg!h S, £
EPr ~ aF e
I g_ Cw TLET e[ 66 in. WIDE x 59% in. HIGH x 1Y% in. YHICK LEAD PLATES
v S5 gp oW 95T \ SET IN POSITIONING FRAME AS SHOWN
o ] <t Ny i
& CENTER LINE OF POOL AND REACTOR =g § °%  =meacTom l \y; l
T 3
= w |
K i’ g2 &5 z 25 LL"—I\ : CAGE POSITIONING FRAMES
CONTROL PANEL L % a4y W o4y &I\
: PLoeE 3. =N
i € Eaw o £ .
Jie =97 ks -

 

 

 

 

 

 

L

 

 

 

 

CENTER LINE OF RERACTOR

 

O

 

 

 

{
!
——— 1 Yin

 

 

 

 

 

 

L AL

 

 

e
L_‘X .

AIR COMPRESSOR

FLOOR PLAN OF SWIMMING POOL

Fig. 7.7. Plan View of the Installation for the Irradiation of Animals in the Bulk Shielding Facility,

 
 
8. TOWER SEBIELDING FACILITY

C E. Clifford

T. V. BRlosser

Physics Division

JO Y.

Estabrook

ANP Division

Design of the tower configuration
and the necessary hoisting equipment
has been contracted to the firm of
Knappen, Tippetts, Abbett, McCarthy,
architectural engineers of New York,
and is to be completed by Februwary 1,
1953, The basic tower configuration
and location are to be established
definitely by December 1, at which
time ORNI. will complete plans for the
services, the building, and site
improvement. The neutron-scattering
background for the tower designed
by the architectural engineers 1is

significantly higher than for earlier
designs; consequently, aiternate
designs are being studied.

FACILITY DESIGN

After making a study of the geologi-
cal conditions affecting the foun-
dations, the KTAM engineers relocated
the tower site tothe top of the knoll,
a few hundred feet from the previous
location. The tower configuration is
now conceived to be a four-legged,
guyed structure. Such a structure was
chosen by KTAM to eliminate the large
cantilever trusses used inthe previocus
designs. The guying should also
appreciably reduce the weight of the
structure, and hence the cost. Pre-
liminary estimates obtained by KTAM of
the cost of the required hoisting
equipment are gratifyingly low.

All the electrical components of
the reactor control and instrumentation

have been chosen and are being ordered
in accordance with estimated delivery
dates. Engineering of the mechanical
portions of the reactor controls, tank,
and accessories 1s approximately 75%
complete. Engineering of the reactor
control system is being executed by
the Reactor Controls group. The
engineering of the experimental equip-
ment, which is assigned to the Instru-
ment Department, is approximately 20%
complete and 1s expected to be com-
pleted by March 1953.

The project for development of a
more sensitive fast-neutron dosimeter
is progressing satisfactorily., A flow
type of counter that uses ahot-calcium
purifier for the ethylene gas shows
sensitivities better than twice those
of the types in use at present.

STRUCTURE SCATTERING

A. Simon, Physics Division

The background from neutrons
scattered by the tower structure is
being calculated for each configuration
according to methods(!) developed at
the laboratory during the original
design period. In addition, the tower
is being made sufficiently flexible so
that this background will be measurable
directly.

(1)A Simon and R. H. Ritchie, Background
Calculations for the Proposed Tower Sh:eld;ng
Facility, ORNL-1273, p. 41.

95
ANP PROJECT QUARTERLY PROGRESS REPORT

9. NUCLEAR MEASUREMENTS

ANGULAR DISTRIBUTION OF NEUTRGNS
SCATTERED FROM NITROGEN

J. L. Fowler C. H.

J. R. Risser

Physics Division

Johnson

The method of studying elastic
scattering of neutrons by performing
pulse-height analysis of nitrogen
nuclei recoils in a counter gas(!) has
the 5-Mev Van de

The active volume

been utilized on
Graaff accelerator.
of the proportional counter used i1s a
cylinder 12,7 c¢m long and 7.62 ecm in
diameter. Field tubes®?) included in
the counter design eliminate end
effects owing to nonuniform electric
fields. The counter wire is separated
from the field-tube voltage by grounded
shield tubes. The details of the
counter design and fabrication were
handled by Zedler of the Instrument
Department.

Results of Nitrogen-Scattering
Experiment. The test with the 630-kev
proton peak from the N!'%(n,p)CH4
reaction produced by thermal neutrons
indicates an energy resolution of the
counting system of the order of 6%
(full width at half maximum of the
pulse-height distribution). Figure
9.1 shows the pulse-height distributions
from mono-
energetic neutrons produced by the
T(p,n)He® reaction.
cell was

of the nitrogen recoils

Tritium 1n a gas
bombarded with analyzed
protons from the 5-Mev Van de Graaff{
accelerator. The neutron energy
spread owing to thickness of target
and to straggling of protons 1in the
gas-cell window was of the order of
120 kev., The energies chosen are

between the resonances i1in the total

 

1
¢ )B. D. Rossi and H. Staub, Ionization Chambers

and Counters, Experimental Technigues, p. 135,
McGraw Hill, New York, 1949,
(Z)A. C. Cockroft and S. C. Curran, Rev. Sci.

Instr. 22, 37 (1951},

96

cross section reported in the liter-
(3.4.5) These distributions
an experi-

ature,
have been corrected for
mentally determined nonlinear de-
pendence of maximum recoil pulse height
on neutron energy. Near-zero pulse-
height background has been subtracted
from the distribution. This back-
ground was determined by substituting
helium for tritium in the gas cell and
was found to be completely negligible
three-fourths of the
The distribution at
0.8 Mev has been checked a number of
times under a variety of experimental
conditions: (1) the nitrogen pressure
in the counter was changed by a factor
of 3, (2) the counter voltage, and
consequently the gas multiplication,
was varied over a considerable range,
and (3) the tritium neutron source was
altered; that is, nitrogen was sub-
stituted for aluminum as the tritium-
cell window, and the experiment was
performed with a zirconium tritide
With the exception of fluctu-
near zero pulse height, the
distributions obtained were essentially
the same.

Analysis of Nitrogen-Scattering
Data. By using the known values of
total cross section,(3'4'5) the number
of nitrogen recoils per incident
neutron can be predicted from the
geometry and the nitrogen pressure.
The number of neutrons was determined
with a long counter calibrated against
a polonium-beryllium sounrce. The
comparison between prediction and the
number of counts in actual distri-
butions, extrapolated to zero energy,

for the upper
distribution.

target.
ations

indicated agreement within the accuracy
of the neutron measurements.

 

(S)C. H. Johnson, B, Petree, and R, H. Adair,

Phys. Hev. 84, 775 (1951},
(4)J. J. Hinchley, P. H. Stelson, and W, M.
Preston, Phys. Rev. 86, 483 (1952).

C. H. Johason, H. B. Willard, J. K. Bair,
and J. D. Kington, Phys. Quar. Prog. Rep. June 20,
1959, ORNL-1365, p. L.
UNCLAGSIFIED
DWG. 16615A
COSINE OF CENTER-OF~MASS ANGLE
cos O cos60” cos 30° cos{20® cosiBO°

 

 

O
9
8
4 ’ E=C.80Mev
5
g
é 3
x 2
t
o .
0 . 50 99 142 199
8
7
6
5
4
3
2
1
0
G T9 159 248 38
6
5
4
3
2
i
0 :
0 106 212 318 423

 

S =

O - N W p M ~N DO

 

NUMBER OF PULSES PER UNIT ENERGY INTERVAL OR DIFFERENTIAL CROSS SECTION CENTER-OF-MASS SYSTEM

 

6
5
4
3
2
I
0
0 151 301 452 . 603
NITRCGEN RECOIL ENERGY ( kev)
Fig. 9.1. Pulse-Height Distribution

Nitrogen Recoils from Monoenergetic Neu-
trons.

PERIOD ENDING DECEMBER 10, 1952

The cosine of the center of mass
angle of scattering of the neutron 1is
plotted in Fig. 9.1 on the same
horizontal scale as the recoil energy.
The ordinate is thus proportional to
the differential scattering cross
section in the center of mass system.(l)
The rise in the cross section at low
angles is believed to be real and due
to shadow scattering. This is being
investigated further.

Several distributions taken ingoing
over the 1,595-Mev resonance’*®) show
marked change in the differential cross
section associated with the peak of
the resonance.(®) Forward scattering
of the neutrons becomes more pronounced
at the resonance.

The dose received at the crew
position from air-scattered neutrons
will vary directly with the cross
section if this scattering is isotropic.
I1f it is not, then the dose will vary
in a quite complicated way with the
“distribution of cross section,” and
therefore it is essential to know the
angular distribution of the scattered
neutrons. For this purpoese, the
differential scattering cross section
of nitrogen, as opposed to 1ts total

sectlion, must be
forward-scattered

scattering cross
measured. The
neutrons contribute essentially nothing
to the flux i1incident on the crew
shield, but the neutrons scattered
through angles greater than about 30
deg are very important., Fortunately,
the experiment is most reliable in
this region, as will be seen from in-
spection of Fig. 9.1,

A FAST-NEUTRON DOSE MEASUREMENT

T, V. Blosser, Physics Division

An unusual opportunity for cali-
brating the fast-neutron dosimeter was
afforded by the recent unfortunate
accident with the ZPR at Argonne
Natienal Laboratory,

in which one

(G)E. Baldinger, P. Huber, R, Ricamo, and V.
Zunti, Helv. Phys. Acta. 23, 503 (1950).

97
ANP PROJECT QUARTERLY PROGRESS REPORT

SECRET
DWG. 17375

 

 

 

 

 

 

 

 

 

 

PERSONNEL. PLATFORM

4-in. DOSIMETER COUNTER —

| s— PLASTIC CONTAINER

 

g, b, ¢, d, AND & ARE CONTROL RODS

I‘“” S\
A1 :
[

cORE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 9, 2. Schematic Diagram of Zerc Power Reactor.

98
person received about four times the
military dose for one mission in fast
nevtrons (25 rem), in addition to an
even larger gamma dose. Since the
total power released in the “runaway”
critical assembly was quite well
measured, it was possible, by setting
up the experiment again, te run at
controlled power levels and measure
the dose in the position occupied by
the individual most seriously affected.
From these measurements it is possible
to determine the dose received by the
individual in terms of a reading on
the Hurst fast-neutron dosimeter.
Since this instrument is basic to all
measurements at this Laboratory,
direct “calibration” at a near-serious
dose level 1s very important. Conse-
quently, one of the ORNL dosimeters
was taken to ANL for measuring the
dose in the reconstructed situation.
It was operated by a member of the
OBNL shielding group to ensure that
the measuring technique was identical
to that used 1n the ORNL program.

The ZPR is a hydrogen-moderated,
cylindrical reactor with an 8-in,
water reflector. Since this is much
like the Bulk Shielding Reactor, it
would be expected to give quite similar
leakage fluxes, as was observed.
Figure 9.2 is a schematic sketch of
the reactor and shows the vertical
line (labeled *z”), about 7 in. from
the core axis, on which the data were
taken. The dosimeter could not be
placed closer to the core axis, since
the only location large enough was
pre-empted by a control rod. Measure-
ments taken in the 8-in. waterreflector
and 1n the air above the reflector are
shown in Fig. 9.3; all data have been
normalized to a power of 1 watt in
accordance with ANL power measurements
supplied by J. R. Dietrich. The

PERIOD ENDING DECEMBER 190, 1952

instrument was calibrated at the site
with a polonium-beryllium source, and
gamma suppression was adjusted to 10
r/hr, which is safely greater than that
to be expected in the region above the
reactor.

The fast-neutron dose readings in
the water reflector corresponded
almost exactly to those to be expected
from scale-up of BSF data. However,
the measurements 1in air may have been
somewhat high owing to scattering from
the personnel platform and control
mechanisms. ‘

SECRET
OWeG, 1737%¢

 

  

 
 

onolee [

I
| ‘ |

     

FAST-NEUTRON DOSE (mrep /hr/watt)
S
n

 

 

 

 

| _
2 -]‘ e i ----- e | {t
| - ]
0 10 20 230 40 B0 &0 70 80
2, DISTANCE FROM SCURCE (cm)

 

Fig. 9.3. Fast-Neutron Dose Measure-

ments from Reactor Core Face.

99
 
SUMMARY AND INTRODUCTION

The research on fused fluoride
systems {(sec. 10) led to the selection
of the fuel and a charging technique
for the Aircraft Reactor Experiment.
As now planned, the ARE fuel will be a
NaF-ZrF,-UF, mixture with a composition
of 50-46-4 mole %. Development of
laboratory-scale, pilot-plant, and
production-size facilities for making

the carrier for this fuel, or some
modification of it, continues. The
final mixture would be made 1n the

reactor by the addition of NaF-ZrF, -UF,
(50-25-25 mole %) charge material to
the NaF-ZrF, (50-50 mole %) carrier
material with which the reactor would
be initially loaded. For the longer
range fuel development program, fluoride
systems containing LiF, ZrF_ , AlF,, or
BeF,, singly or in combination, are
being examined.

The corrosion research (sec., 11)
has been directed primarily toward the
determination of the corrosion charac-
teristics of the fluorides (in par-
ticular, the ZrF,-bearing mixtures)
and secondarily to a study of hydroxide
and liquid metal corrosion. All
studies are made at temperatures
around 1500°F, The increased attack
experienced in recent tests with the
Zr¥,-bearing fuel mixtures is of
particular concern; the attack is to a
depth of about 9 to 10 mils after
500 hours. Although this depth of
~attack 1s tolerable, the trend is not;
" the increased attack is believed to be
due to the use of the newly installed
pilot-plant eguipment in which the
fuel was prepared. The beneficial
~effect of small additions of ZrH, as a
"corrosion inhibitor has been again
demonstrated in the thermal convection
loop, whereas additions of NaK have
not had so pronounced an effect.
‘Dynamic corrosion tests with lead in
thermal convection loops resulted in
"severely corrosive attack on Inconel
above 500°C, even though deoxidized

lead was employed. The stability of
beryllium oxide in NaK has not yet
been thoroughly investigated. Under
the conditions of some of the initial
tests, significant beryllium oxide
weight losses occurred. The effects
of various additions to hydroxides
were studied, but no decrease in cor-
rosion was obtained. The maintenance
of a positive hyvdrogen atmosphere is
still the only certain technique for
minimizing hydroxide corrosion.

The metallurgy and ceramics research
(sec. 12) 1ncludes the development of
spherical, solid, fuel elements; creep-
rupture tests of structural metal;
welding and brazing techniques; and
cermets and ceramic coatings. Alloys
for cone-arc welding, resistance weld-
ing, heliarc welding, and high-
temperature brazing have been suc-
cessfully developed for high-tempera-
ture-corrosion resistance and structural
integrity of brazed joints. In creep-
rupture tests at 815°C, Inconel and
stainless steel indicated decreasing
lifetimes in argon and hydrogen
environments as compared with the
lifetimes in air. The manufacture of
spherical, solid, fuel elements,
requested by Pratt and Whitney, is
being attempted by several techniques.
Cermets of Cr-A1,0, and ZrC-Fe have
been developed and are being tested
for corrosion resistance. A dipping
technique has been developed for the
application of oxidation-resistant
ceramic coating to nickel radiator
fins.

The heat transfer and physical
property measurements have beendirected
almost entirely toward the determination
of the various properties of several
fluoride mixtures (sec. 13). The
viscositlies and densities of the fuel
carrier, NaF-ZrF_ , and the enriched
fuel, NaF-ZrF,-UF,, have been measured.
Although these fluoride mixtures have
comparable melting points (about 510°C)

103
ANP PROJECT QUARTERLY PROGRESS REPORT

and densities (between 3 and 4 at
800°C), the viscosity of the fuel 1is
significantly higher than that of the
carrier (about 8 vs. 13 cp at 800°C),
In general, it appears that the thermal
conductivities of fluoride mixtures
decrease and the viscosities increase
with an i1ncrease in the uranium
fluoride content, The heat transfer
coefficient for the eutectic mixture,
NaF-KF-LiF, in turbulent flow may be
described by the usual expressions for
ordinary fluids. In other studies,
the flow rates in convection loops are
being determined, and thermal analyses
are being made of areflector-moderated
reactor system and the heat transfer
characteristics of a circulating-fuel
system.

Radiation damage studies were con-
cerned primarily with the exposure of
samples of fluoride fuel in the various
irradiation facilities (sec. 14),
Other studies included in-pile sodium
loop and in-pile creep measurements,
A capsule of fluoride fuel has been
irradiated in the MTR at a flux level
in the range of that of an aircraft

104

reactor, but the capsule and its con-
tents have not yet been examined,.

Short-time, high-power-dissipation
(up to 4000 watts/cm?) irradiations in
the cyclotron suggest that some radi-
ation-induced corrosion damage may
exist. HRecent in-pile cantilever-
creep measurements of Inconel in an
air atmosphere indicate that the LITR
radiation has little effect on creep

strength.
Analytical

materials (sec.
spectrographic,

studies of reactor
15) included chemical,
and petrographic
corrosion
and

identification of impurities,
products, reduction products,
constituents of reactor fuels. An
empirical volumetric method for the
determination of zirconium has been
developed. Improved tests have also
been developed for the determination
of aluminum and chrowmium in fluorides.
The mass spectrograph in combination
with a high-temperature (up to 2850°C)
oven will detect with high precision
the presence of uranium oxides,
Petrographic studies have led to the
classification of many complex com-
pounds not found in the literature.
10,

PERIOD ENDING DECEMBER 10, 1952

CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS

W. R. Grimes, Materials Chemistry Division

The main effort of the ANP Chemistry
group has been devoted to studies of
fluoride mixtures for use as fuels and
coolants for possible use in an
aircraft reactor. The development of
a fuel suitable for use in the ARE
appears to have been achieved with the
NaF-ZrF, -UF, (50-46-4 mole %) mixture.
Accordingly, more attention has been
devoted during recent weeks to the
longer range objectives of development
of fuels with improved properties.

Since there is reason to believe
that low-melting-point systems of
superior quality will be obtained,
despite the isotope separation re-
qulrements, consliderable attention
has been focused on phase equilibriums
among systems containing lithium
fluoride, In addition, systems
containing‘ZrF4, AlFS,:and BeF, are
being studied, and some work has been
initiated on systems containing UC1,.

Studies of systems containing UF,,
which seems to be a by-product 1in
some of the corrosion reactions of
the fuel, have been continued. Synthe-
sis of some materials that contain
trivalent uranium and that seem to be
identical to compounds produced in
corrosion studies has been accomplished.

Research on liquid fuels and their
production in highly pure form on a
pilot-plant scale have continued to be
a major function of the ANP Chemistry
group. The production of the large
quantities of these materials needed
for engineering testing 1s a joint
responsibility of this group and the
Experimental Engineering group. The
base material (NaZrF,) for the ARE
fuel 1s to be prepared in equipment
that is scheduled for completion in
mid-December.

Tubes of simulated fuel mixture
were supplied for the ARE cold critical
experiment by the Y-12 Production
Division, along with one tube filled

with slugs of the actual fuel compo-
sition, Efforts to i1dentify the
chemical species in the cooled melts
before and after corrosion testing
were continued. Studies were expanded
on reducing agent additions to various
fluoride mixtures and on the resultant
reactions. Preliminary results have
indicated that small additions of
NaK, Zr, or ZrH2 may have beneficial
effects on the corrosion of structural
metals. - These additions have been
shown to cause harmful reduction of
the UF, in the fuel when made in
larger quantities. Phase equilibrium
studies of systems centaining UF, have
been continued as a part of this
program. The reactions obtained by
addition of reducing agents to possible
fuel or coolant materials are being
studied in considerable detail by
means of a combination of x-ravy,
petrographic, spectrographic, and
chemical examinations.

FUEL MIXTURES CONTAINING UF,

.. M. Bratcher C. J. Barton
Materials Chemistry Division

LiF-BeF,-UF,. It appears that
compositions that melt below 450°C
cover the range from 0 to 30 mole %
UF, in this system, as shown in
Fig. 10.1. The minimum melting point
observed. was 335°C, which was for a
mixture containing 2.5 mole % UF,,
48 .75 mole % LiF, and 48.75 mole %
BeF,; this minimum value has not been
checked by heating curves, and it
may be found to be considerably lower
than the true melting point. Other
compositions at the 2.5 and 5.0 mole %
UF, levels showed higher melting
points. _

NaF-BeF «UF,. HRecent indications
that the binary eutectic of NaF and
BeF2 (43 mole % Ber):has a viscosity
of about 15c¢p at 600°C have stimulated

105
ANP PROJECT QUARTERLY PROGRESS REPGRT

    

LiF

..... f----
I_lz Be Fq

475°G

Fig. 10,1,
itnterest in this material as a fuel
solvent. The addition of 5 mole % of
UF, to this mixture yielded a material
that could be stirred readily above
350°C and showed no thermal effects
on heating or cooling curves above
375°C. However, in the preparation
of 3-kg batches of this matetrial,
two layers developed; one layer was
dark and the other light green.
There was a thermal halt on cooling
at about 470°C,

The separation into twe layers,
along with the lack of thermal effects
in small portions of the separated

106

UFy,

 

EEG. 17402

/' e

\./4527/“ -
T INDETERMINATE
T R

AN

e

BeFy

365°C

The System LiF-BeF2~UF4.

layers, has led to speculation comn-
cerning the possibility of there being
two immiscible liquids below 470°C
in this system., Filtration studies
with this material seem to discount
this hypothesis, since there 1is
evidence of a steady decrease 1in
uranium concentration in the filtrate
as the temperature is reduced below
500 °C.

It seems likely that in this system
in which thermal effects are apparently
quite small, the methods used are not
sufficiently sensitive to detect the
first separation of a solid phase on
cooling or the ‘““end of melting® on
heating of the mixture.
Studies of this system,
as attempts to improve
techniques,

as well
the research
are continuing at present.

LiF- ZrF ~UF, It appears that a
melting polnt of about 420°C may be
expected in this system. From the few
experiments performed to date, it is
not possible to specify the composition
of the eutectic or eutectics.

Material con-

20 mole_%

NaF-LiF-ZrF, -UF,.
taining 40 mole % NaF,
LiF, and 40 mole % ZrF, appears to
melt at 430°C. Additions of various
quantities of UF, to this system
yielded the thermal data shown in
Table 10.1. :

It appears that the lowest melting
point in this region (415°C) is
probably that of the quaternary
euatectic. Much more data will be
needed to establish the compositions
in the lowsmelting-point region; it
is likely, however, that the uranium
content of the eutectic 1s reasonably
low.

NaZrF, -NaUF Since present plans
call for addition of a mixture con-
taining 50 mole % NaF, 25 mole % UF;,
and 25 mole % ZrF, to the fuel solvent,
NaZrF, to brlng the reactor to
criticality, study of the NaZrF_-NaUF,
is of considerable i1nterest.

TABLE 10.1. EFFECT OF UF, ON MELTING

 

 

POINT OF NaF-LiF-Zr¥, MIXTURE*
UF, CONTENT THERMAL EFFECTS
(mole %) (°C)
0.0 430
2.0 460, 435, 415
4.0 480, 440, 415
6.0 494, 445
8.0 500, 435
10.0 502, 430

 

 

*40 mole % NaF, 20 mole % LiF, 40 mole % ZrF4'.

PERIOD ENDING DECEMBER 16, 1952

The mixture containing 50 mole % of
each complex compound melts at about
610°C and has a density at 1500°F of
3.98 g/cm? A tentative equilibrium
diagram for these compounds 1s
in Fig. 10.2. X-ray difiraction has
revealed that solid solutions are
formed in this system and that the
solution of NalUF, in NaZrF|
at about 15 mole % of NaUF, Al though
the liquidus line for this system
may be considered as well established,
much further study will be required
to define the solidus lines with
accuracy. The results of careful
examination of the 8 mole % NaUF,
region by x-ray and petrographic
techniques and the virtual absence
of segregation on slow cooling of
very large preparations of this
material, indicate that the areca
between llquldus and solidus lines 1n
the dilute solid solution region (and
differences in composition of the two
solid solutionsin equilibrium) must be
very small. In uranium-rich regions
of this diagram, however, the situation
is considerably different, and segre-
gation of phases of considerably
different composition would not be
surprising. Additional study of this
system is being emphasized at present.

given

saturates

RbF-ALF, -UF, A mixture containing
58 mole % HbF and 42 mole % AIF, 1s
oneof the lowest-melting-polnt (535°C)
alkali fluoride~aluminum fluoride
mixtures found. This mixture was used
to test the effect of low concen-
trations of UF, on the melting point
of mixtures in this system. In the
range from 2 to 20 mole % UF,, there
appeared to be a rapid rise in melting
point with increasing UF, concen-
tration, and even the 2 mole % UF,
mixture showed a marked increase 1in
melting point over that of the base
mixture. At temperatures below 550°C,
solubility of UF, in this mixture
appears to be very low. Solid phases
resulting from these studies have not
vet been examined.

107
ANP PROJECT QUARTERLY PROGRESS REPORT

A

DWG. 17403

 

 

700

 

 

TEMPERATURE (°C)

600

 

 

 

 

 

 

500

 

 

 

 

 

 

 

 

 

 

 

NquF5

 

60 80

NaUF, CONCENTRATION {mole %)

Fig. 10.2.

FUEL MIXTURES CONTAINING UCI4
R. J. Sheil C. J. Barton

Materials Chemistry Division

Thermal data for a number of binary
chloride systems containing UCl, or
UCl, mixed with various alkali and
alkaline earth chlorides have been
reported previously.{(!) In the
previous work, only the binary systems
of the uranium compounds with NaCl and
KC1l were studied in sufficient detail
to permit construction of equilibrium
diagrams. The eutectics shown for the
NaCl-UCl, and the KC1-UCl, systems
melt at much lower temperatures

(370 and 350°C, respectively) than do

1

( )C- A. Kraus, Phase Diagram of Some Complex
Salts of Uranium with Halides of the Alkali and
Alkaline Earth Metal, M=251 (July 1, 1943).

108

The System NaZrFS-NaUFS.

the corresponding fluoride eutectics.
Although the eutectics reported for
UCl, with these materials have higher
melting points than those of the
corresponding UCl, systems, they still
compare favorably with the fluorides.
The high vapor pressure of UCl, will
present obvious difficulties in the
use of these mixtures at very high
temperatures; however, the vapor
pressure of the mixtures containing
UC1l, should be
favorable.
Preliminary studies have shown
that some refinements 1n technique
will be required before the very
hygroscopic and reactive UCl, can be
handled in the apparatus without being
contaminated. Graphite crucibles were
found to be quite permeable to UCI,

considerably more
vapor; however, nickel vessels have
been satisfactory. For satisfactory
operation, the nickel vessels must be
loaded in helium dry boxes, and a dry
helium atmosphere must be maintained
over the samples at all times.

The available UCl, has served for
preliminary experiments for defining
the techniques; however, the Cl-to-U
ratio of 3.9 and the low melting point
of 5365°C (literature value for pure
UCl, is 590 + 1°C) make this material
unsuitable as a standard for study.
This material will be sublimed under
vacuum and subsequently handled under
dry, inert atmospheres to provide
pure UCl, for future studies. _

Preliminary data have been obtained,
meanwhile, on NaCl-UCl, mixtures. The
eutectic and eutectoid halts reported
by Kraus¢!) have, in general, been
confirmed, Actual data will be
reported when pure materials become
available. '

FUEL MIXTURES CGNTAINING UF,

W, C. Whitley V. S. Coleman
C. J. Barton

Materials Chemistry Division

Preliminary results of studies of
alkali fluoride-uranium trifluoride
systems were reported previously.(?)
The effort in this field was concen-
trated on the NaF-UF;, and KF-UF,
binary systems during the past quarter.

Data were obtained with two types
of apparatus: the large vacuum-~tight
apparatus previously described¢?} and
small welded stainless steel reactors
fitted with a thin-walled thermocouple
well 1in the center. In order to
assure an inert atmosphere, the
fluoride mixture was introduced into
the reactor through a 1/4-in. tube,
and the reactor was evacuated at room
temperature, filled with inert gas,
and then sealedby pinching and welding
the tube. Mixing of the contents of
the reactor was accomplished by

 

(2)V. S. Coleman and W. C. Whitley, ANP Quar.
Prog. Rep. Sept, 10, 1952, OBNL-1375, p. T9.

PERIOD ENDING DECEMBER 16, 1952

inverting after it had been heated
enough to bring all the material into
solution. The mixture was not stirred
during the cooling cycle,.

A comparison was made of cooling
curves for the same composition 1in
the two types of apparatus. The
thermal effects occurred at about the
same temperature with both types of
apparatus, but the breaks were some-
what better defined with the small
reactor. This is probably due to
better contact between the thermocouple
and melt in this apparatus and to the
smaller heat capacity of the thin-
walled thermocouple well. The thermal
data were supplemented by examination
of the solid phases by means of the
petrographic microscope and, for a
limited number of samples, by x-ray
diffraction techniques.

NaF-UF,. Cooling curves have been
run on mixtures covering the range
from 5 to 85 mole % UF;. It was

observed that the cooling curve breaks,
except for the lowest break at about
590°C, seemed to decreaseupon prolonged
heating at elevated temperatures.
This change in thermal effects ap-
parently results from the development
of UF, and/or U0, in the melt. Both
products are found 1n the fused
material, in which the UF, is combined
with NaF., An intensive effort was
made to find the source of the oxygen
and to eliminate it from the system;
but, to date, this effort has not
been entirely successful, and some
doubt exists as to the accuracy of
the thermal data that have been ob-
tained. However, several conclusions
about the system seem warranted:

1. A compound is formed between
NaF and UF,. 1t is lavender in color,
uniaxial positive, and hasa refractive
index for the ordinary ray of 1.552
and for the extraordinary ray of
1.564,¢%)

2. The most probable composition
for the compound, based on both thermal

 

(3)Petrographic exagmination by T. N, McVay.

109
ANP PROJECT QUARTERLY PROGRESS REPORT

data and examination of the solid
phases, is 3NaF-2UF;. This composition
could be in error because of the
presence of U'* in the melts, as
mentioned above.

3. A eutectic exists between NakF
and the compound, which 1s probably
close to 27 mole % UF,, and it melts
at approximately 685°C.

The data obtained with mixtures
containing more than 50 mole % UF,
do not fit any recognizable pattern.
The significance of the break at
approximately 590°C, whichwas observed
with all compositions in this system,
is still not clear. Further work
toward solving these problems 1s 1in
progress.

KF-UF;. Tt is clear that a complex
compound between KF and UF; is formed
and that a eutectic between KF and the
compound melting at approximately
690°C occurs at about 18 mole % UF,.

The solid phases have not been thoroughly
studied, and the optical properties
of the compound have not been de-
termined. Study of this system will
be resumed when the oxidation phe-
nomenon mentioned above has been
brought under control.

ALKALY FLUOBORATE SYSTEMS
J. G. Surak

Materials Chemistry Division

Thermal analysis studies of alkali
fluoborate systems, which have been
described in previous reports,(*)
were continued during the early part
of the quarter. Additional data on
the NaBF, -KBF, system have permitted
construction of the tentative equi-
librium diagram shown in Fig. 10.3.
The liguidus curve appears to go

Moore, C.
10, 1952,

(4).]. G. Surak,
ANP Quar. Prog. Rep.
p. 82.

R, E.
Sept.

J. Barton,
OBRNL- 1375,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FOO oo DWG. 17404
600
cb.-.“.*_____7
500 - >‘-h‘h\‘g MELTING POINT
5 .
w )\\
é 400 @\\“*%z ------
o .
=
=
300
G
>-§-.-.?s.h,-_~§~( PHASE TRANSITION
— 7 \\ o clf’Q”{)
200t T 4] & /
5
100
KBE, 10 20 30 40 50 60 70 80 90 NaBE,
NaBF, {rnole %)
Fig. 10.3. The System NaBF -KBF,.

110
through a minimum at about 90% NaBF,.
Halts at the minimum melting tempera-
ture of 360°C were not observed for
other compositions; apparently these
materials are completely miscible in
the solid state. :

Attempts have been made to obtain
phase equilibrium data in the NaBF,-
NaUF, and NaBF, -NaZrF, systems, with
little success. Mixtures of these
materials were heated, in each experi-
ment, to temperatures up to 800°C,
where BF, pressures were about 100 psi,
before the cooling curves were recorded.
Although the data are quite incomplete,
1t appears that more than 5 mole % of
either compound raises the melting
point of NaBF, considerably. There
is some evidence for eutectics at
about 360°C in each system at quite
low uranium and zirconium concen-
trations. Actual compositions of the
eutectics have not been established.

DIFFERENTIAL THERMAL ANALYSIS
R. E. Traber, Jr.

Materials Chemistry Division

The differential thermal analysis
apparatus for l- to 2-g samples was
used regularly during the quarter for
the careful examination of important
compositions, for testing samples that

TABLE 10. 2. THERMAL EFFECTS OBSERVED

PERIOD ENDING DECEMBER 10, 1952

were availableonly in small quantities,
and for searching for thermal effects
that are difficult to find with the
coocling-curve technigue. The apparatus
has been found to have limited appli-
cationin the study of reduced systems,
such as those containing UF,. Although
a flowof purified heliumwas maintained
in the apparatus, the erratic behavior
of the differential trace obtained
when the reduced samples were heated
was appavently due to oxidation,
Examination of the solid phases re-
maining at the end of the tests showed
that oxidation had occurred. Some
difficulties that were experienced
with the electrical system may be
attributed, at least in part, to the
high sensitivity of the system.

Table 10.2 summarizes the thermal
effects observed with differential
apparatus for a number of compositions
of current interest. The temperatures

at which the differential trace begins
to leave the base line, reaches a

peak, and returns to the base line are
recorded. Where more than one “hump”
was observed in the differential
trace, only that at the higher temper-
ature is included in the table. 1t
should be noted that the temperature
at which the differential trace returns
to the base line on the heating curve

WITH DIFFERENTIAL THERMAL APPARATUS

 

 

 

 

l 1 THERMAL EFFECTS (°C)
COMPOSITION (mole %) '
Start Peak (m.p.) End
50 NaF-50 UF, 600 685 705
54 NaF-41 BeF,-5 UF, 310 340 360
55 NaF-40 BeF ,-5 UF,-

(light-green layer) 320 360 400
55 NaF-40 BeF -5 UF,

(dark-green layer) 310 355 375
57 NaF-43 BeF, 310 340 350
50 NaF-25 UF,-25 ZrF, 500 575 625
50 NaF-50 ZrF, 490 520 535

 

 

 

 

111
ANP PROJECT QUARTERLY PROGRESS REPORT

is usually higher than the melting
point of the composition determined
from the cooling curve. This may be
due to the time required to melt all
the sample, and the difference between
the two values 1s probably a function
of the heating rate.

An effort is under way to increase
the sensitivity of the differential
thermocouple arrangement so that a
very slight thermal effect may be
found, such as that which apparently
was overlooked in the NakF-BeF, -UF,
fuel mixture (discussed previously
in this chapter).

SIMULATED FUEL FOR COLD CRITICAL
EXPERIMENT

D. R. Cuneo J. D. Redman
.. G. Overholser

Materials Chemistry Division

Filling the fuel tubes with homo-
geneous powder containing Z4r0,, NaF,
UF,, and graphite was accomplished
in the manner described previously.(3)
One tube was required, however, 1in
which the fuel density was uniform
and at least equal to the ARE fuel and
in which the fuel had the proper
Na:Zr:U ratio.

Attempts to cast slugs of this
material by heating premelted powder
of the proper composition under inert
atmospheres 1in graphite molds were
not successful. Only the bottom 1l to
2 in. of an B8-in. section so formed
was free from “pipes.” Re-use of the
material from the porous sections
led to considerable fractionation
because of segregation during the slow
cooling.

The slugs were prepared 1in Building
9212 by melting the premelted fuel by
induction heating under an 1nert
atmosphere, pouring the liquid into
10 -in. molds, and rapidly cooling the
casting, Slugs prepared in this
fashion were free from pipes over
about 8 in, of length and had uniform

(S)D. R. Cuneo and I.. G, Overholser, ANP Quar,.
Prog. Rep. Sept. 10, 1952, ORNL-1375, p. 87,

112

uranium concentrations over the entire
length,

The slugs were cut with an electri-
cally heated knife and wrapped 1in
0.25-m1i]1] aluminum foil. A total of
41.5 in. of this material, which
contained (by weight) 74.0% ZrF,,
17.2% NaF, and 8.82% UF,, was supplied
to the Critical Experiment group. The
material had a density of 3.75 g/cm?
at room temperature.

In addition, tubes were filled with
$10,, NaF, KF, and Cr for danger

coefficient measurements.

COOLANT DEVELOPMENT

L. M. Bratcher C. J. Barton
Materials Chemistry Division

No new coolant components were
considered during the quarter except
Ban,which proved unattractive because
of the high melting point of the
eutectic in the one system studied.
Studies of systems containing AlF,,
ZrF,, and BeF, continuned.

NaF-Zrf,. Study of the complicated
and important binary system NaF-ZrF,
has been in progress for some time.
Early studies were complicated by
the formation of oxide that caused
error in the calculation of compositions
and resulted in the appearance of
thermal effects above the melting
points of the pure fluoride mixtures,
Some data on this system have been
presented 1in previous reports, (8
Although further study i1s necessary to
more clearly define the composition of
the solid phases in certain portions
of the system, the location of the
liguidus line 1s believed to be
sufficiently well established to
justify the publication ofthe tentative
equilibrium diagram for the system,
as shown in Fig. 10.4. Some of the
gquestions remaining to be explained

 

about this system are: (1) the

(6)L. M. Bratcher, R. E. Traber, Jr., and C,
J. Barton, ANP Quar. Prog. Rep. June 10, 1852,
ORNL-1294, p. 90; and ANP Quar. Prog. Rep. Sept.

10, 1952, ORNL-1375, p. 89.
PERIOD ENDING DECEMBER 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oo
OWG. 17405
10O :
1000
o
200 \C\ ~ e
PR
-~
s
7
o f\\ <
$ 800 2 £
Wi O £
& oo"wfl o /fi)y
L) o o o
< 700
A "
= 0oy
00
o
Qoo
500
400 -
NoF 10 20 30 40 50 &0 70 80 20 2rFy,
NagZrf, No,ZrFg NaZrf, NaZraFg(?)
Zrf, (mole %)
Fig. 10.4.

" composition of the compound that is
- believed to exist between NaZrF_, and
ZrF,; (2) the possible solid solution
formation between NaZrF_ and Na,ZrF,;
(3) the significance of the thermal
effect at 525°C in the 33 to 38 mole %
ZrF, region, which may indicate
polymorphism of Na,ZrF, or, possibly,
a compound such as NajZ2r,F ,; (4) the
peculiar behavior {marked expansion
on freezing) of mixtures in the region
of the eutectic at 43 mole % ZrF,. _

LiF-ZrF_ . Study of the LiF-ZrF,
system has been attended by some of
thedifficulties mentioned in connection
with the NaF-ZrF, system. Although

The System NaF-ZrF,.

thermal data have been obtained for a
number of compositions in the 10 to
70 mole % ZrF, range, the melting
points of some compositions are still
uncertain. It is clear, however, that
this systemis quite different from the
other alkali fluoride-zirconium
fluoride systems that have been
studied. The minimum-melting-point
composition appears to be close to
50 mole % ZrF,, with a melting point
of about 510°C. The solid phases in
this system have received little
attention, as vet,

NaF-KF-Al1F,. The lowest melting-
point compositions in the binary

113
ANP PROJECT QUARTERLY PROGRESS REPGRT

system NaF-AlF; and KF-AlF; reported
in the literature’? are 685 and
570°C, respectively. The tentative
equilibrium diagram shown in Fig. 10.5
indicates that in the ternary system
no melting points lower than 570°C are
available.

NaF-BeF,-ZrF,. Thermal data were
obtained from cooling curves for about
25 compositions in the NaF-BeF,-ZrF,
system. Many of the compositions
studied were within the NaF-NaZrF, -
NaBeF, part of the diagram. The
lowest melting point observed was

(T)F. T. Hall and H. Insley, J. Amn. Ceraa. Soc.
Nov. 1947, Supplement.

AlFs
p 2

   

 

/

 

i
<,Aé274
! T
4’i¢/%fiéfié/{
LLNOT

375°C for a mixture containing 5 mole
% ZrF,, 40 mole % BeF,, and 55 mole %
NaF. It appears that a Bel, concen-
tration of approximately 20 mole % 1is
needed to obtain a melting point
substantially lower than that of
NaZrF, (510°C).

NaF-LiF-ZrF,. The NaF-LiF-ZrF,
system has been studied fairly in-
tensively, especially in the 30 to 50
mole % ZrF, range in which the lowest
melting points occur. The lowest
melting point obtained was 426°C for
the mixture containing 20 mole % LaiF,
40 mole % NaF, and 40 mole % ZrF, .

Isotherms cannot be prepared for this

DWG. 7406

7

 

 

 

 

 

 

 

685°C
570°C <7
b="" - ' e gpoteATTTTTTTT A
\ LN\
\\ 900°¢
K3AlFg \\ \ . NazAlFg
1025°C \ \/ 500 0%\ 1020°C
/ O e N N o
1000 / AN
950°C #
_ °C
840°C o | 80070~
O° .—?5000 /
KF > N N / NoF
710°G
Fig. 10.5. The System NaF-KF-AlF,.

114
system until the LiF.ZrF, system is
clarified. _

NaF-BaF,. The alkaline earth
fluorides have received comparatively
little consideration as possible
components of fuels and coolants,
mainly because of their high melting
points Since data on the NaF-BaF,
system apparently have not been
published, a few compositions were
tested to determine the minimum
melting point. The data indicate that
this is a simple eutectic system; the
eutectic at approximately 35 mole %
BaF, melts at about 820°C. This is
much too high to be of interest.

The KF-BaF, eutectic has been reported
to melt at 750°C. Therefore the
NaF-KF-BaF, ternary seems unlikely

to yield usefully low melting points.

NaF -KF-~LiF-ZrF, Increasing amounts
of ZrF, were added to the NaF-KF-LiF
eutectic (11.5, 42.0, and 46.5 mole %,
respectively). After a slight initial
decrease 1in melting point of about
7°C at 2 mole % ZrF,, the melting
point increased quite rapidly to
680°C at 10 mole % and reached a
maximum of 715°C at 20 mole % ZrF,
This indicates that if molten NaF-KF-~
LiF eutectic 1s used to trap ZrF,
vapors from the ARE, as has been
suggested, either the concentration of
ZrF, would have to be kept low or the
temperature would have to be maintained
above 700°C to keep solids from
forming. _ . .

NaF-RbF-AIF, Meltlng points of
all the compos1t10ns tested in the
NaF-RbF-AlF, system (about 40) were
substantlally higher than the minimum
melting point observed in the RbF-AlF,
system (525°C).

STUDIES OF COMPLEX FLUORIDE PHASES

Several new complex fluorides have
been prepared to assist in the identi-
fication of various materials found
at the conclusion of static and dynamic
corrosion tests. The complexes of
the alkali fluorides with trivalent
chromium have been studied in detail

PERIOD ENDING DECEMBER 16, 1952

because one such complex, K,NaCr¥,,
had been previously identified as
being present in a loop test. The
complex fluorides synthesized in these
studies have been i1dentified or
characterized by chemical analyses,
x~-ray diffraction patterns, and optical
crystallographic data. Although in
most instances the sealed capsule
technique was used for the preparation
of these samples, 1n some cases welded
tubes similar to those used for static
corrosion tests were employed, with a
concomitant reduction in oxide for-
mation. The density and solubility
of several of the complexes have been
determined, and the data are given
in Table 10.3.

K,CrF -Na,CrF -Li CrF
and L. G. Overholser, Materials
Chemistry Divisien). The identi-
fication of compounds in the system
K,CrF -Na;CrF,-L1,Cr¥F, has been
virtually completed. A constitution
diagram based on x-ray diffraction
data supplied by H. Dunn and interpreted
with help from A. G, H, Andersen is
given in Fig. 10.6. Apparently
Na,CrF, and Na,LiCrF, are the limits
of one solid solutlon region, and
K,NaCrF_, K3Na3(CrP6)2 and KNaLiCrF,
are the limits of another such area.
The corrosion products found in loops
in which NaF-LiF-KF-UF, had been
circulated have been in the region
between K, NaCrF, and K ,Na,(CrF_ ),
except in one case in whlch the product
resembled KNaLiCrF

So0lid Phases 1n NaF-BeFQ»UF4 and
NaF-ZrF, Systems (A. G. H. Andersen,
ANP Division). Work done by Andersen
on the solid phases in NaF-BeF,-UF,
and NaF-ZrF, systems was summarlzed 1n
his final report.(®’ The solid phases
in the NaF-BeF, -UF, system will
probably be studled further when time

1s available.
UF,-ZrF, (V. S.

(B. J. Sturm

Coleman, cC. J.

Barton, Materials Chemistry Division;
T. N. McVay, Consultant, Metallurgy
(B)A G. H. Andersen, The Solid Phases of

Alkali and Uranium Fluoride Systems, Y-F33-3

{Det. 1, 1952},

115
ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 10,3,

PHYSICAL PROPERTIES OF FLUOCOMPLEXES OF CHROMIUM,

IRON, AND NICKEL

 

 

 

 

 

 

 

 

 

COMPOUND OPTICAL DATA MES:::G SOLUBILITY IN WATER ) DENSITY
(o) (g/ml) (o cu’)
K4CrFy Cubic,** n = 1.422 1055 2.5 x 1073 at 290°C | 2.69
NayCrF, Cubic,** n = 1.411 880 4.0 x 10°% ar 250°C
LisCrF, Biaxial,** (=), a = 1.444, 672 3.16
2V = about 40°, a = 1. 464
K,NaCrFy Cubie,** n = 1,422 1006 2.9 x 10°% 3.07
K,LiCrFg Cubic, n = 1.422 952
K,Na,(CrFy) , Cubic, n = 1,418 1000
MNaLiCrF, Cubic, n = 1,418 875
Li,KCrFg Anisotropic 791
LiyNaCrFg Anisotropic
Na,LiCrFg Cubic, n = 1.400 8 30
Rb,CrFg Anisotropic
Cs;CrF Biaxial, (+), avg. n = 1,520%**
K,RbCrF Cubic
K,CsCrFy Cubie, n = 1.457
K FeFg Cubic, n = 1.414
Na,FeF Cubic
K,NaFeF, Cubic, n = 1,414 965 3,19
(solid solution)
K,NiF, . Anisotropic, avg. n = 1,423 9 40
Na,NiF, 805
KNaNiF, 785
*Most of the melting point data were obtained by C. J. Barton, R, J. Sheil, R. E. Traber, and

L. M. Bratcher.

**Data from T. N. McYay,
***Data from G, D. White.

116
Nc]:,_l‘Cr'F6

0 X-RAY PATTERN OBTAINED AT THESE

  

PERIOD ENDING DECEMBER 10, 1952

AN
DWG. 17407

 

 

 

 

 

 

 

 

MPOSITION

COMPO : S SOLID SOLUTION

o
NajLiCrF,
K NO (Cr o
SOLID SOLUTION
Ko NaCrFg 7 LizNaCrfy

A _

 

   

 

 

 

 

 

K3CrFG K'?_LiCrFG

Fig. 10.86.

Division).
heating UF,

The samples prepared by

with ZrF,, with and
- without NaF, show a complex compound
UF,*2ZrF,, which is an orange-red
crystalline solid that is occasionally
found as a reduction product of
ZrF,-bearing fuels. When this material
is heated with NaF at 850°C or when
UF, and ZrF, in proper ratio are
heated with NaF, the products include
UF,, some UF,-2ZrF,, and some NaZrF,.
The latter compound usually contains

LigCrig

L.ial'(CrF'6

The Syste@ Kschfi-Naachfi—LiSCrFe.

some NaUF; in solid solution, presumably
formed by traces of ox1dants in the
system. Since NaF can "steal” ZrF,

from UF -2ZrF,, apparently NaZrFs is a
more stable compound than the UF;'2ZrF,

complex. The 2ZrF, -UF, compound has
been found in loops in which there was
severe reduction of the UF, in the
fuel because of the injection of Nak,
and alsoin capsules in which zirconium
metal or NaK had been added to the
fuel.

117
ANP PROJECT QUARTERLY PROGRESS REPORT

NaF-Zr¥, (P. A. Agron, Materials
Chemistry Division). The x=-ray studies
of the NaF-ZrF, system(?) have been
extended to the region of 75 mole % of
ZrF,. The crystal structures of the
solid phases Na,ZrF, and NaZrF, have
been reported previously.(1?) Studies
of the behavior of the phases that
appear in compositions lying between
these two compounds and of those that
lie beyond the NaZrF, compositions are
being made; considerable additional
study will be required, since the
system 1s an exceedingly complex one.

Other Fluoride Complexes (B. J.
Sturm and .. G. Overholser, Materials
Chemistry Division). It was previously
noted that a preparation corresponding
to K,NaFeF, was thought to be iso-
morphous with K,NaCrF . Further study
has shown that Na,FeF_, and K,Fel,
apparently form solid solutions over a
wide range, probably 1nall proportions.
X-ray patterns for materials corre-
sponding to K, FeF,, K,NaFeF, K, Na(Feky),,
and KNa ,FeF_  differ only by slight
shifts 1n lines.

Fusion of equal wolar proportions
of KHF,, NaHF,, and NiF, -4H,0 at 850°C
resulted ina yellow compound, probably
KNaNiF,. A fusion in the proportions
corresponding to K,Na(NiF,), gave a
mixture of K,NiF, and KNaNiF,. Heating
a mixture of KNiF;+1%H,0 (prepared
from aqueous solution) and NH,F, at
700°C yielded a product that agrees
with the ASTM x-ray diffraction data
for anhydrous KNiF,,

Interaction of chromium metal with
K,NiF, at 850°C for 5 hr resulted
in the formation of K,CrF_ by the
following reaction:

3K,NaF, + 2Cr —> 2K CrF, + 3Ni
This behavior is comparable to that

previously noted for displacement of

Fe from K,NaFeF, by Cr,

(g)L. M. Brateher and C. J. Barten, "Coolant
Development, ' this chapter.

(IO)P. Agron, ANP Quar. Prog. Rep. Sept. 10,
1952, ORML-1375, p. BS.

118

REACTIONS OF FLUORIDE MIXTURES
WITH REDUCING AGENTS

W. R, Grimes
Materials Chemistry Division

L. A. Mann
ANP Division

The reactions of possible ARE fuel
mixtures with reducing agents were
first examined in an effort to evaluate
the damage that would result i1f NaK
were inadvertently admitted to the
fuel circuit during operation. Identi-
of the products 1n these
complex mixtures was shown to bea very
difficult problem. Hecently, however,
it has been demonstrated that addition
of reducing agents to the fluoride
melts decreases the corrosion by these
materials. Since production of UF,
and the consequent precipitation of
uranium at high temperature can result
from excessive addition of reductant,
a careful studyof the fluoride systems
under reducing conditions has been
undertaken.

Reducing Power of Various Additives
(J. C, White, Analytical Chemistry
Division). The total reducing power
of reaction products resulting from
the addition of NaK to such materials
as ZrF,, NalF, and UF, and such mixtures
as Nab-ZrF, -UF, and NaF-ZrF, has
been measured by two methods: (1) oxi-
dation with standard ceric sulfate
solution and (2) hydrogen evolution
from dissolution in mineral acid.
The validity of these procedures has
been verified by using control samples,
such as UF; and Zr. 1In practically
none of the NaK addition tests,
however, has a reducing action eqguiva-
lent to the amount of NaK added been
observed.

This reduction phenomenon has also
been obhserved when UF, has been added
to NaF and ZrF, and treated as in fuel
preparation. One compound, Na,U)F,,
gave nearly the theoretical amount
of UF, added.

Several samples of fuels that had
undergone corrosion testing were

fication
arbitrarily selected and the total
reducing power determined by the
ceric sulfate oxidation procedure,
Since UF, was the only reductant
(with respect to ceric sulfate) known
to be present in these fuels, the
reducing power was calculated as total
uranium and compared with the value
obtained from oxidation with ferric
sulfate solution.

All the values obtained by ceric{IV)
oxidation are higher than those ob-
tained by ferric(III) oxidation.
Since ferric(IIIl) oxidizes only
uranium(IV) in this instance, other
oxidizable species that have not been
identified satisfactorily must be
present in rather appreciable quanti-
ties. ' |

Identificationof Reduction Products
(F. F. Blankenship, D. C. Hoffman,
Materials Chemistry Division; K. J.
Kelly, Pratt and Whitney Aircraft
Corporation; T. N, McVay, Consultant,
Metallurgy Division). HBeduction
products have been obtained by sealing
the fluoride mixture to be tested,
usually either the ARE fuel (50 mole %
NaF, 46 mole % ZrF,, 4 mole % UF,)
or NaZrFS, in capsules of type 316
stainless steel with the desired
quantity of reducing agent and heating
the capsules for 16 hr in the tilting
furnace. 1In these tests, NaK, Zr,
and ZrH2 were added in 0.2, 0.4, and 0.8
equivalents (1 equivalentis the amount
required in theory to reduce the UF,
to UF,). All samples were rocked
with the hot end of the capsule at
800°C and the cold end at 650°C,
After this heating, they were heated
to 800°C while in a vertical position
for2 to 6 hrto encourage sedimentation
of any 1nsoluble material.

Although 1t is not yet possible to
explain all the processes that take
place in these systems, some of the
qualitative results observed may be
summarized as follows:

1. Little change in the over-all
appearance of the NaZrF, samples was

produced with any of the reducing

PERIOD ENDING DECEMBER 10, 1952

agents.  Some NaZrF,, together with
colorless anisotropic crystals of a
continuous range of lower indexes of
refraction, was present, which suggests
possible solid solutions of NairF,,
Na,ZrF,, and NajZrF,. "

2. A gradual change of the ARE
fuel from bright green to a darker
color with vellow to brown to red
portions could be observed as more
reducing agent was added. A brown
(olive-drab) pleochroie crystal
(average index = 1.558) was found 1in
all the 0.2-eq. additions. With 0.4
eq., more of the brown crystal,
togetherwith partly hieachedNaZr(U)Fy
was present, A trace of a red-orange
phase was also found in the 0.4-eq.
NaK addition. With 0.8 eq. of NaK, a
small amount of UF,, more red-orange
phase, some yellow-to-bleached NaZr(U)F,
and a little opaque material (possibly
metal) were found. 1In 0.8-eq. ZrH, or
Zr additions, no UF, was present,
although a yvellow-to-orange phase was
beginning to grow in and separate from
the nearly bleached crystal solution;
the brown crystals were present. The
Zr addition seemed to show more
reduction products, as well as a small
amount of opaque material. In all
the tests, the concentration of the
brown, red-orange, and UF, crystals
appeared to be highest 1n the bottom
portion of the capsule.

3. The NaF—ZrF4—UF4 (50-46 -4 mole
%) fluoride fuel run as a control
sample with these tests appeared
normal, with only a slight trace of a
brown coloration in some crystal
solution fromthe bottom of the capsule.

The orange-red crystalline phase
has been prepared by heating UF; and
ZrF4 (1:2 mole ratio) in sealed
capsules at B850 to 900°C. This
complex, which 1s almost certainly
UF, 2Zr¥F,, appears to be fairly
uniform; however, variations in the
index of refraction indicate that the
UF,-to-ZrF, ratio is not constant.
By heating equimolar mixtures of ZrF,
and UF,, a yellow-brown crystal is

119
ANP PROJECT QUARTERLY PROGRESS REPORT

formed; however,
remains,

When mixtures corresponding to
NaF-UF, -2ZrF, are heated, the red-
orange phase 1s predominant, but free
UF,, NaZrF,, and a trace of a colorless
phase are present. [If preparations
containing NaF + (UF,-2Z:F,) are
‘agitated to ensure equilibrium during
the heating period,
the olive~drab phase,
brown phase, along with NaZrF,
appear. When the composition is
NaF + 2(UF, +2ZrF,), the red-orange
phase predominates, with some green-
to-olive-drab phase growing in it;
some UF;, NaZrF.,, and a trace of
colorless crystals (not ZrOz) appear.

It seems that NaK produces some
free UF, in the ARE fuel, whereas
ZrH, and Zr do not. These phenomena
may have the following explanation:
The (NaK)F formed can remove ZrF,
fromthe UF,*2ZrF,, or similar compound,
with the ultimate formation of free
UF; when it no longer has enough ZrF,
with which to combine. On the other
hand, the ZrF, formed by the ZrH, and
Zr additions will be able to hold more
UF,. (Large encugh additions of
ZrH, or Zr to produce free UF, have
not yet been obtained.) This suggests,
however, that the ARE fuel is in very
delicate balance with regard to
reductants, and on this point alone
fuels with higher ratios of zirconium
to alkali metal might be preferred.

Reaction of Fluoride Mixtures with
NaX (L. A, Mann, J. M, Cisar, F. M,
Grizzell, ANP Division). A series of
four tests was run by the Experimental
Engineering group, in which various
amounts of NaK were allowed to react
with a fluoride fuel mixture to
determine how much NaK canbe tolerated
in the fuel mixture. In these tests,
the NaK was added to NaF-ZrF, -UF,
(46-50-4 mole %) and allowed to react
at high temperatures (1400 to 1500°F),
The liquid portion was then pulled
through a micrometallic filter,
the residue

an excess of UF3

some UF;, much of
and some reddish

and
(the portion that was

120

solid at high temperatures) and the
filtrate were examined. It was found
that the addition of sufficient NakK
would reduce enough UF, in the fuel
mixture to UF; to cause a solid
precipitate at reactor temperatures
and, hence, would be detrimental to
the proper functioning of the ARE.
It was also found, by adding varying
amounts of the NaK to a given amount
of fuel, that a limited addition of
NaK would notcause such precipitation.
The maximum amount that could be
added was between 0.27 and 0.70 eq.
(per equivalent uranium) at 1200 to
1300°F. Additional confirmatory
tests, plus tests with other fuel
mixtures, are scheduled to be performed
by the Materials Chemistry group.
Reaction of Na¥F-ZrF,-UF, with Zrii,
(J. D. Redman, L. G. Overholser,
Materials Chemistry Division). Ad-
ditions of small amounts of ZrH, have
been shown to be of considerable
benefitin decreasing attack on Inconel
by fluoride mixtures. Since production
of UF; in amounts sufficient to
exceed the solubility of this compound
at 1100°F might occur as a result
of such additions,
made

studies have been
to demonstrate whether such
systems are completely liquid at this
temperature.

The apparatus developed for this
purpose is shown in Fig, 10.7. This
apparatus will minimize the possibility
of oxidation of the mixture. The
charge bottle 1s loaded with the
required materials, and the apparatus
1s assembled in a dry box. The
apparatus 1s then rocked, and after an
equilibrium temperature is reached,
the rig 1s inverted to permit the
charge material to pass through the
nickel filter and into the receiver
tube. Thermocouple wells are provided
for temperature control, and the gas
lines permit evacuation of the chamber
or maintenance of a particular atmos-
phere, as desired. The filter may be
removed for examination by disassembling
the apparatus.
 

In these studies, a mixture con-
taining 30 mole % NaF, 46 mole % ZrF,,
and 4 mole % UF, that had been purified
in the standard fashion and stored
in a sealed bottle in a dry box was
treated with varying guantities of
ZrH,. BSamples were equilibrated at
800°C for 4 hr under apositive pressure
of helium and, afrer 2hr equilibration
at 600°C, filtered at the latter
temperature. ‘ |

When ZrH, was added in the range
from 0.2 to 0.7 wt % of the fuel
mixture, no evidenceof solid separating
from the melt was observed. Chemical
analysis of the filtrates in each
case revealed the original uranium
concentration, and microscopic exami-~
nation of material on the filter
revealed no UF;. Although extensive
reduction of uranium to UF; 2ZrF, was
observed in the filtrate, especially
when 0.7% of ZrH, was used, no free
UF, could be detected in the specimens
examined. .

Additional tests will be made to
verify this important point and to

Closed-system

PER 10D ENDING DECEMBER 16, 1952

L2

Batch-Filtration Rig.

determine the maximum ZrH, concen-
tration tolerable without precipi-
tation of the solids. 1t appears,
however, that if the temperature is
kept at 600°C or above, at least 0.7
wt % of ZrH, may be added to the ARE
fuel without difficulty. _
Soelubility of Potassium in the
NaF-KF-LiF Eutectic (H. R. Bronstein,
M. A, Bredig, Chemistry Division).
Additions of alkali metals to fluoride
mixtures have demonstrated a capacity
for supressing the fluoride attack
on container materials. Consequently,
an investigation was undertaken to
determine the amount of potassium
that will remain in equilibrium in
the fluoride eutectic NaF-KF-LiF
(11.5-42.0-46.5 mole %). The need for
separation of the equilibrated liguid
salt and liauid metal phases led
to the design of aball-check apparatus
that consists of a sealed capsule with
two outlets, either of which may be
sealed by the steel ball within the
capsule, The capsule 1s charged with
the fluoride, the metal added, and the

121
ANP PROJECT QUARTERLY PROGRESS REPCRT

mixture intimately mixed. The capsule
is then tilted so that the ball closes
the lower valve and isolates the upper
part of the salt phase, which is then
solidified and analyzed for 1ts
potassium content. The data obtained
are summarized in Table 10.4.

PRODUCTION AND PURIFICATION OF
FLUORIDE MIXTURES

F. F. Blankenship G. J. Nessle

Materials Chemistry Division

Use of the hydrogenation-hydro-
fluorination process previously
described for fuel-sample preparation
has been continued on laboratory and
pilot scales for production of standard
materials for corrosion testing,
physical property evaluation, and
large-component testing.

Construction of the production
equipment to furnish 3000 1b of NaZrF,
for the ARE is in an advanced stage.
Raw materials are being accumulated at
a satisfactory rate,

A modification of the purification
treatment has been adopted for micro-
scale processing of samples containing
enriched uranium for radiation-damage
testing.

Aslight modification of the standard
procedure has been used successfully
to purify mixtures containing 50
mole % NaF, 25 mole % ZrF,, and 25
mole % UF,.

Preparation of NaZrl; by direct
hydrofluorination of Zr0O, in the
presence of NaF has been moderately
successful in simple equipment. Further
study of this process 1s planned,

Laboratory-Scale Fuel Preparation
(R. E. Thoma, C, M. Blood, F. P. Boody,
Materials Chemistry Division). Fourteen
batches of purified molten fluoride
compositions, both coolant and fuel,
have been prepared for loop tests,
measurement of physical properties,
and corrosion testing. These were
treated 1n essentially the same manner
as previously reported, (1)

To explore the unlikely possibility
that hydrogen reduces UF, or Zr¥,,
two of the fuel batches were given an
additional hydrogenation for 1 hr
following the regular procedure. No
evidence of reduced phases in the
products from these runs was shown by
examination with the petrographic
microscope. However, there 1s evidence
that the material so prepared was
slightly less corrosive in small-scale
tests than was the standard fuel. No
explanation for this effect has been
suggested; however, further study of
this behavior is planned.

Modifications of techniques of
assembling and disassembling apparatus
have resulted in less frequent failure
than was experienced at the beginning
of the hydrofluorination program,
Corrosion of the reactors has not yet
been a problem when the standard
treatment 1s applied to relatively
pure fluoride mixtures.

Pilot-Scale Fuel Purification
(G. J. Nessle, J. E. Eorgan, Materials
Chemistry Division). Two apparatus
capable of producing up to 3 kg of

(1D, M. Blood, F. P. Boody, A. J. Weinberger,

and G. J. Nessle, ANP Quar. Prog. Rep. June 10,
19592, ORNL-1294, p. 97.

 

 

 

TABLE 10.4. SOLUBILITY OF POTASSIUM IN NaF-KF-Li¥F
TEMPERATURE OF NO. OF POTASSIUM CONTENT
EXPERIMENT (°C) EXPERIMENTS (mole %)

544 1 3.8
686 1 3.5
920 5 3.7 £ 0.5
10 40 1 3.9

 

 

 

122
purified fuel mixtures andone apparatus
capable of handling up to 30 kg of
the molten materials are 1in use at
present. The small units are es-
sentially duplicatesof the laboratory-
scale models previously described.
The 50-1b equipment, however, was
built mainly from existing equipment
and utilizes an 8-ft transfer line and
a receiver furnace that can be lowered
topermit rapid handlingof the product.
This apparatus has 3/8-in. nickel
tubing in the gas and transfer lines
and 10-in.-dia reactor and receivers.
Swagelok fittings are used on all
tubing connections.

A total of 323 kg of fuel has been
prepared i1n the equipment, and 1t 1is
possible to produce up to 60 kg per
week, 1f necessary.

Each fuel batch 1s sampled as 1t
is transferred under inert gas pressure
to the receiver, Except 1n a few
instances, the concentrations of Fe,
Cr, and Ni have been less than 900,
100, and 900 ppm, respectively. There
appear to benodifferences in chemical
analysis that can be attributed to the
presence or absence of the sintered
nickel filter in the transfer line.

Since there have been some un-
explained differences in corrosion
between various batches of fuel
prepared in this equipment, it may be
necessary to increase the time of
treatment in the 50-1b rig. This
possibilityis being studiedat present.

In general, performance of the
pilot-scale equipment has been quite
satisfactory. The 50-1b rig would be
improved by larger valves in the HF
lines, although heating all such lines
to 100°F has minimized the difficulty
resulting from condensation of HF,
No difficulty with faulty vessels of
nickel or with corrosion of reactors
has been observed. Transfers of
material from the l0-in. reactor are
remarkably complete; material balances
indicate less than 100-g holdup on
batches of 50 pounds.

PERIOD ENDING DECEMBER 10, 1952

Fuel Production Facility (G. J.
Nessle, Materials Chemistry Division).
The 3000 1b of NaF-ZrF, (50-50 mole %)
for the ARE is to be prepared in
250 -1b batches. Equipment for this
purpose 1s being constructed of nickel
from scaled-up designs that are very
similar to those used for the pilot-
scale units. Duplicate units are to
be used, and production of 750 to
1000 1b per week should be possible.
These units are in the final con-
struction stages and should be available
by mid-December.

The equipment i1s instal led in an
area that is convenient for operating
personnel and where safety can be
maintained during the entire cycle.
The process includes storage, transfer,
weighing, and blending of the dry
solid components; treatmentby melting,
fluorinating with hydrogen fluoride,
and reducing with hydrogen; transfer
of the pure liquid melt to the storage

vessel; and storage of the solidified
product. All treatment processes at
molten temperatures are handled

remotely, and the negative-pressure
ventilation over the confined process
equipment is at a rate of more than
one change per minute. The process
equipment from outside manufacturers
and Y-12 shop fabricators has been
delivered and is scheduled for instal-
lation, without delay, upon completion
of the construction work.

The ZrF, for the fuel is low-
hafnium material prepared by the Y-12
Production Division by hydrofluori-
nation of ZrCl,. The ZrCl4, produced
at the Bureau of Mines in Albany,

Oregon, has been delivered, and
conversion to ZrF, is virtually
complete. Table 10.5 shows the

chemical analysis of the sublimed
ZrF, received fromY-12 and stock-piled
for this operation.

The NaF of the purity required is
available in stock, and HF, H,, and
He of the proper purity are available
in gquantity.

123
ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 10.5.

ANALYSIS OF LOW-HAFNIUM Lr¥, FOR PRODUCTION OPERATION

 

 

 

 

 

BATCH COMPOSITION (%) HAFNIUM BORON
NO. 7r F cl C (ppm) (ppm)
B-1 54.8 { 42.8 0.03 0.06 85
B-2 55.2 42.8 0.03 0.12 70

- 43.9 0.01 0. 29 50 1
B-4 43.8 0.01 0. 14 50 0.5
B-5 44.7 0.01 0.07 50 0.5

e Lo .

 

 

 

 

 

 

 

 

Preparation of Hydrofluorinated
Fuel Samples (F. P. Boody, Materials
Chemistry Division). Ten small
samples containing enriched UF, have
been treated in the program for
preparing fuels for the cold critical
test and for radiation-damage tests
in the MTR. The base material was a
mixture of NaF and ZrF, that had been
hydrofluorinated before the enriched
UF, was added. Samples containing
10.8, 16.5, and 35.0 wt % UF, have
been prepared for the radiation-damage
experiments,

The batches of NaF-ZrF,-UF, (46-50-4
mole %) for the cold critical test
were treated in the same manner as
batches containingnormal UF,. However,
the batches required for radiation-
damage tests were small (10 to 20 g),
and the procedure was modified so
that the HF, H,, and He gases were not
bubbled through the melt but were
applied as an atmosphere.

The melt was contained in either
a nickel or platinum crucible 1nside
the nickel reactor. The reactor was
opened and the sample transferred to
a gas-tight glass shipping container
in a dry box that had an inert atmos-
phere that was produced by evacuating
and flushing with dry helium three
times. A small amount of NaK was
included in the sealed glass shipping
container to ensure the quality of the
inert atmosphere.

The formation of a hlack crust on
the surface of several of the samples

124

has caused considerable concern.
This crust has been examined petro-
graphically, spectroscopically,
chemically, and by x-ray diffraction.
The spectroscopicand x-ray diffraction
analyses showed only NaUF; and NaZrF,
to be present in appreciable quantity;

however, the petrographic microscope
indicated small amounts of carbon
and UO0,. Chemical analyses indicated

0.06 to 0.20% carbon and approximately
0.2% U0,.

Hydrofluorination of Zrx0,-NaF¥
Mixtures (C. M. Blood, R. E. Thoma,
Materials Chemistry Division), Efforts
to determine the optimum conditions
for the smooth and complete hydro-
fluorination of NaF-Zr0, mixture to
NaZrF; have continued, The successful
results obtained during the previous
quarter{'?? with the hydrofluorination
of small samples (300 to 400 g)
encouraged attempts to treat larger
batches (2 to 3 kg) without the aid
of a mechanical stirrer. Because of
the high temperatures (700 to 900°C)
involved and the pronounced 1in-~
convenience of mechanical stirring
in the presence of HF at this tempera-
ature, a technigque was sought that
required no more agitation than that
provided by gas bubbling through 6 in.
of melt 4 in., in diameter.

Completion of the reaction by
passing HF through a nickel reactor at

(12)g F. Blankenship, R. E. Thoma, Jr., F. P.

Boody, and C. M, Blood, ANP Quar. Prog. Rep.
Sept. 10, 1952, ORNL-1375, p. 91.
low temperatures (50 to 150 °C}, at
which a liquid reaction medium was
obtained through the agency of the
low-melting-point polyhydrofluorides
of NaF, or at high temperatures
{550 to 850°C), at which liquefaction
resulted from the melting of the
NaZrF,, appeared feasible. The main
problem was to find the optimum
combination of low- and high-tempera-
ture stages. Farlier results indicated
that although essentially stoichi-
ometric conversion could be obtained

at low temperatures, complete neutral-
1zation and removal of water was best

ensured by high-temperature treatment
of the molten salt. Accordingly, all
trials were finished withat least 1 hr
of high- -temperature treatment.

Difficulties associated with the
appearance of insocluble intermediates
were encountered at both low and high
temperatures. Cements, presumably
related to ZrOF,, were developed in
the partly reacted mixtures; rapid
passage of HF or helium was 1neffecti?e
in mixing the cemented portions, and
channeling occurred. '

Successful results were achieved
1f long scaking periods (12 to 16 hr)
atlow temperatures were used, followed
by high-temperature hydrofluorination;
when 25% of the charge material was
the previously prepared NaZrF., the
high-temperature treatment alone was
sufficient. Both petrographic exami-~
nation and x-ray analysis indicate
that even in those samples in which
the material balance i1s 100% there
occurs a trace amount of Na,ZrF,, and
as a result the refractive index or
the x-ray crystal pattern deviates
slightly from the exact standard for
NaZrF, .

The feasibility of making ZrF,
directly from ZrO, and NaF has been
demonstrated from a laboratory stand-
point; the practical application of
theprocess requ1resfurther'englneerlng
development,

PERTOD ENDING DECEMBER 10, 1952

PURIFICATION OF HYDROXIDES

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

The purification of hydroxides has
continued, but at a further decreased

rate, An increased interest 1in
handling alkali metals, especially the
alloy NaK, made it necessary to

condition the vacuum dry box and
perform numerous loadings and un-
loadings, which cut heavily into the
time allotted for hydroxide purifi-
cation. sufficient NaOH has
been purified to maintain an inventory
large enough to fill all requirements
for this material.

However,

The method previously given, ('3¢1%)

which involves the removal of Na,CO,
from a 50 wt % solution of NaOH and
subsequent dehydration at 450°C under
vacuum, was used for all NaOH purifi-
cation., Nine batches averaging
1 1/21b per batch were prepared during
the period. The purified product
continues to meet the specification of

0.1 wt % for both Na,CO, and t,0.

Additional pure KOH has been prepared
by the method previously describedt 1319
namely, the interaction of pure
potassium with water, followed by
dehydration at 450°C under vacuum.
The K,CO; content has ranged from
0.04 to 0.12 wt % and the water
content has been 0.1 wt % or less.
Approximately 2 1b of KOH was prepared
during the period, and the inventory
was thus increased to about 5 pounds.
At present, the demand for pure KOH
is virtually nonexistent, and unless
some unforeseen demand arises 1t 1s
not planned to react more than. an
additional pound of potassium.

Two batchesofSr(Ofl)2 were purlfled
by the method previously described

(13)g. g

Quar.

Ketchen and L.: G. Overholser, ANP
Prog. Rep. June 19, 1952, OBNL-1294, p. 89.
(14)D.lL Cuneo, E. E. Ketchen, D. E. Nicholson,
and L, G. Overholser, ANP Quar. Prog. Rep. March
10, 19532, ORNL-~1227, p. 104, :

125
ANP PROJECT QUARTERLY PROGRESS REPORT

in detail.{'%) This material is not
being used for experimental purposes

(IS)L. G. Overholser, D, E. Nichclson, E. E.
Ketchen, and D. R. Cumeo, ANP Quar. Prog. Rep.
Dec. 10, 1951, ORNL-1170, p. B84.

126

at present; consequently, no further
significant production is anticipated.

No LiOH was purified during the
period. Present requirements are
being met by dehydration of the
commercial monohydrate,
PERYOD ENDING DECEMBER 10, 1952

11. CORROSION RESEARCH

W. R, Grimes, Materials Chemistry Division

W. D. Manly, Metallurgy Division
H. W. Savage, ANP Division

Most of the effort on corrosion was
expended in dynamic tests of the ZrF, -
bearing fluoride melts. These studlea
have been primarily concerned with the
effect of various additives to the melt
on the corrosion behavior of Inconel,
It was found that additions of ZrH,
and NaK to the fuel mixture inhibit
the corrosive attack. Other tests of
fluoride corrosion have been performed
to determine the effect of crevices,
temperature, and pretreatment of the
fluoride and the resistance of various
oxide coatings and ceramic bodies.
Several Inconel loops circulating the
fuel mixture NaF-ZrF, -UF, (46-50-4
mole %) exhibited more severe corrosion
than was previously encountered 1in
such tests; this is believed to be due
to the lower purity of the fluoride
mixtures recently prepared in large-
batch apparatus,

Temperature-dependence tests and
experiments on the effects of various
additions on the corrosion behavior of
the hydroxides were carried out. How-
ever, the maintenance of a hydrogen
atmosphere 1is the only proved means of
controlling hydroxide corrosion.

The work on liquid metal corrosion
was concentrated on the operation of
thermal convection loops of two types
to study the mass transfer properties
of lead. One of the most critical
variables in lead corrosion was found
to be the cleanliness of the lead and
the container., However, even the
purest lead prepared to date cannot be
satisfactorily contained in Inconel at
temperatures above 500°C., The in-
vestigation of the stability of BeO 1n
NaK was continued, but additional ex-
perimentation is required to make an
. evaluation, .

FLUORIDE CORROSION IN STATIC AND
SEESAW TESTS

D, C, Vreeland L. R. Trotter
E. E. Hof{iman J. E. Pope
Metallurgy Division

F. Kertesz C. BR. Croeft
H. J. Buttram R. E. Meadows
Materials Chemistry Division

Oxide Additives. Static tests were
run in Inconel tubes for 100 hr at
816°C with the fluoride mixture
NaF-KF-LiF-UF, (10.9-43.5-44.5-1.1
mole %) plus approx1mately 10% additions
of ferric oxide, nickel oxide, and
chromic oxide. These'tests'were run
to determine whether the presence of
these oxides, which may also be present
on Inconel, would increase the cor-
rosion by fluorides. 1TIn these tests,
the nickel oxide and ferric oxide had
little or no effect on the extent of
corrosion. The addition of chromic
oxide apparently increased corrosion.
The results are summarized in Table
11.1. Similar tests are being run
with oxidized Inconel specimens to
determine whether increased corrosion
can be expected on oxidized Inconel,

Comparison of Liguid- and Vapor-
Phase Corrosion. Sections of the
testing tube used in some of the tests
run with the fluoride mixture NaF-KF-
LiF-UF, (10.9-43.5-44,5-1.1 mole %)
for 100 hr at 816°C were removed, both
from above and from below the bath
level, for metallographic examination.
The corrosive attack noted on all
sections examined would seem to indi-
cate that attack can be expected on
metals exposed to the vapor phase of
the fluoride mixture. The sections of
type 304 stainless steel exposed to
the vapor phase appeared to be attacked

127
ANP PROJECT QUARTERLY PROGRESS REPORT

more severely than those exposed to
the liquid phase. Inconel, type 309
stainless steel, and type 316 stainless
steel appeared to be attacked by the
vapor phase to approximately the same
extent as by the liquid phase of the
molten fluoride salt., Figures 11.1
and 11.2 compare type 304 stainless
steel as exposed to the liquid phase
and to the vapor phase of the fluoride.
Details of these tests are presented
in Table 11.2.

Crevice Corrosion. Some concern had
been expressed about the possibility
of crevice corrosion in materials used
to contain the molten fluorides. In
order to check this possibility, tests
were set up with specially prepared
crevices. Ordinary static corrosion
tubes were employed, and the lower
section of the tube above the bottom

weld was partly crimped so that a
crevice about 1 in. long and approxi-
mately 1/16 in. wide was obtained.
After being exposed for 100 hr at
816°C to the fluoride mixture NaF-KF-
LiF-UF, (10.9-43.5-44.5-1.1 mole %),
the tubes were cut longitudinally
through the partly crimped section and
examined for evidence of accelerated
corrosion. The crimped ends of several
ordinary static corrosion test tubes
were also examined for accelerated
corrosion. Corrosion in these crevices
appeared to be somewhat erratic, with
some sections being unattacked. How-
ever, even in the sections that were
attacked, the depth of penetration did
not exceed that normally expected for
these materials after exposure to
fluorides. Details of these tests are
presented in Table 11.3,

 

 

TABLE 11.1. RESULTS OF STATIC CORROSION TESTS OF INCONEL IN NaF-KF-LiF-UF,
WITH VARIOUS ADDITIVES FOR 100 hr AT 816°C
ADDITIVE METALLOGRAPHIC NOTES

 

10% ferric oxide
10% nickel oxide

10% chromic oxide

 

Attack to 2 mils on specimen and tube
Attack to 1 mil on specimen and tube

Attack to 5 mils on specimen and 2 to 10 mils on tube

 

TABLE 11.2.

RESULTS OF EXAMINATIONS OF VARIOUS MATERIALS EXPOSED TO LIQUID

AND VAPOR PHASES OF NaF-KF-LiF-UF, FOR 100 hr at 816°C

 

 

 

METALLOGRAPHIC NOTES

 

 

MATERXAL }+H————rrorovooo o
LIQUID-PHASE EXPOSURE VAPOR-PHASE EXPOSURE
Inconel Erratic attack, subsurface voids Erratic attack, subsurface voids to
to 1.5 mils 1 mil
Type 304 Erratic attack, intergranular Intergranular attack to 5 mils

stainless steel

Type 309
stainless steel

Subsurface voids to 3 mils

Type 316

stainless steel

 

with some subsurface voids to 1 mil

Intergranular penetration of 1 mil

Subsurface voids to 3 mils

Some slight evidence of attack, less
than 0.5 mil penetration

 

 

128
PERYOD ENDING DECEMBER 16, 1952

 

Fig. 11.1. Type 304 Stainless Steel Exposed to Attack by Ligquid Phase of
NaF-KF-LiF-UF, for 100 hr at 816°C. 250X. | | |

   

P UNCLASSIFIED ©

 

 

   

' . ° & o, ‘ : i . - , . S . o |

T ! - - . £ i - - ;

- Foo . ® . K s 5 . . 3
; i . . | e e .
A i : . 5, ' o . %, P : % . g acu s,
7 s, . S b * . s 4 Ll T . :
. v ‘ . St . : \ \ o
5 . - i . : : :
L - C . { ; e 05

¢ "’““w.v// T
| 'é . - e

 

{;

Fig. 11.2. Type 304 Stainless Steel Exposed to Attack by vapor Phase of
NaF-KF-LiF-UF, for 106 hr at 816°¢C. 250X, |

129
ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 11.3. RESULTS COF EXAMINATION FOR ACCELERATED CREVICE CORROSION IN

 

Inconel (specially prepared crevice)

Type 304 stainless steel (specially

prepared crevice)

Type 316 stainless steel (specially
prepared crevice)

Incone! (regular static test)

Type 304 stainless steel (regular

static test)

Type 309 stainless steel (regular

static test)

 

Carboloy and Stellite Alloys.,
Several Carboloy and Haynes Stellite
alloys that might possibly be used as
valve seat and facing materials have
been subjected to static corrosion
tests with the fluorides. Carboloy
44A and 55A were apparently unattacked
in fluoride mixture NaF-ZrF,-UF, (46-
50-4 mole %). The Stellites tested
were attacked much more severely by
this fluoride mixture than by NaF-KF-
LiF-UF, (10.9-43.5-44.5-1.1 mole %).
These materials were exposed for 100
hr at 816°C in evacuated Inconel tubes.
The test results are presented in
Table 11.4.

Reducing Agents.
with various reducing agents were
exposed in the tilting furnace for
100 hr with 4 cpm between 650 and
800°C., Data were obtained by addition
of 1 wt % of Zr or ZrH, to fuel con-
taining NaF-ZrF, -UF, (50-46-4 mole %).
Chemical analyses of the fluoride,
without and with the additives, indi-
cate that with the additives the
amount of structural elements dissolved
is drastically reduced.
tration

Fluoride samples

The concen-
of nickel in the fuel is con-
siderably increased, whereas the iron
content remains essentially constant;

130

METALLOGRAPHIC NOTES

¥

Subsurface voids to a maximum of 2 mils

Subsurface voids to a maximum of 2 mils

Subsurface voids to a wmaximum of 1 mil

Subsurface voids to a maximum of 2.5 mils

Very little attack, a few subsurface voids to 0.5 mil

Subsurface voids to a maximum of 1.5 mils

the dissolution of chromium, however,
1is reduced to negligible proportions.
It appears that this technique may
alleviate the difficult problem of
selective leaching of chromium from
Inconel and stainless steel. Metal-
lographic examinations of these tubes
indicate that addition of Zr or ZrH,
reduces penetration of the fuel into
the metal.

Tests still in progress indicate
that addition of these agents to the
extent of an 0.8 equivalent of Zr or
ZrH, (based on reduction of UF, to UF;)
causes formation of an orange-brown
compound and that addition of an 0.8
egquivalent of NaK causes some reduction
to UF;. At lower concentrations, that
is, below 0.4 equivalent, no free UF,
1s detected. The range between 0.4
and 0.8 is being investigated,

Examination of Inconel tubes 1in
which varying amounts of NaK were
added to NaF-ZrF,-UF, (50-46-4 mole %)
and NaF-Zr¥F, (50-50 mole %) showed
that in the tubes containing NaF-ZrF, -
UF, there was a nonmetallic deposit
and some pitting at the deposit-metal
inter face. The tubes containing NakF-
ZrF, appeared to be attacked similarly,
but there was somewhat heavier pitting.
Additions of NaK cannot be recommended
at present. '

Fluoride Treatment. Static cor-
- rosion tesfls were made on a special
batch of the fluoride mixture NaF-ZrF,-
UF, (50-46-4 mole %), which was pre-
pared by starting with ZrO, and hydro-
fluorinatirig the mixture to obtain
ZrF,. The tests indicated that this
mixture 1s no more corrosive than
material prepared by the hydrogenation-
hydrofluorination process in which
- ZrF, is used as the startingmaterial.

The effect of a hydrofluorination
treatment for the fluoride mixture

NaF-KF-UF, (46.5-26-27.5 mole %) was

PERIOD ENDING DECEVBER 16, 1952

determined in a tilting-furnace cor-
rosion test in which the hot end was
at 800°C and the cold end at 650°C,
The duration of the test was 100
hours. It was found that Inconel
suffered approximately the same degree
of attack as when exposed to untreated
fluorides of the same composition,
that is, 1 to 6 mils of penetration.
The results of exposure of type 316
stainless steel indicated approximately
the same severity of attack, but
slightly less penetration, as when
untreated material ‘was used. It
appears that the purification treat-
ment should be extended for longer

TABLE 11.4., RESULTS OF TESTS OF SEVERAL CARBOLOY AND HAYNES STELLITE ALLOYS
EXPOSED T0 FLUORIDE MIXTURES FOR 100 hr AT B16°C IN INCONEL TUBES

 

 

 

FLUORIDE -
TER - METALLOGRAPHIC NOTES
MATERIAL BATH NO. | |
Carboloy 44A 27(@) No apparent attack
Carboloy 55A 27 No apparént attack
Carboloy 608 27 General roughening of surface, 1 to 2 mils ‘
Carboloy 779 27 Some surface spalling to a depth of 1 mil, which may have
occurred in grinding :
Carboley 907 27 Some surface spalling to a depth of 1 to 3 mils, which may
have occurred in grinding
Haynes Stellite 3 14® Light attack of what appears to be a carbide phase, 1 mil
in depth
Haynes Stellite 3 27 Selective attack of what appears to be a carbide phase,
4 ' maximum depth 22 wils, minimum 12 mils, average 15 mils
Haynes Stellite 6 14 Selective attack of what appears to be a carbide phase,
{specimen A) maximum depth 6 mils, minimum 2 mils, average 4 mils
Haynes Stellite 6 27 Selective attack of what appears to be a carbide phase,
(specimen A) maximum depth 24 mils, minimum 12 mils, average 15 mils
Haynes Stellite 6 14 Selective attack of what appears to be a carbide phase,
(apecimen B) : maximum depth 3 mils, minimum 1 mil, average 2 mils
Haynes Stellite 6 27 Selective attack of what appears to be a carbide phase,
(specimen B} 4 maximum depth 29 mils, minimum 8 mils, average 1l mils
Haynes Stellite 19 14 Scattered attack to 2 mils in depth
Haynes Stellite 19 27 Selective attack of what appears to be a carbide phase,
maximum depth 29 mils, minimum 10 mils, average 20 mils

 

 

 

(a)

a ) . .
Composition:

(b)Composition:

NaF-ZrF,-UF,, 46-50-4 mole %.

NaF—éfF.LiF;UF4, 10,9-43,5-44.5-1.1 mole %.

131
ANP PROJECT QUARTERLY PROGRESS REPORT

times when high uranium concentrations
are used.

A test was run in the tilting
furnace under standard conditiomns of
time and temperature with NaF-ZrF, -UF,
(50-46-4 mole %) that had been treated
with hydrogen after purification by
the normal hydrogenation-hydrofluori-
nation method. Type 316 stainless
steel showed somewhat less attack when
subjected to the hydrogen-treated
material than when exposed to material
prepared in the standard fashion. The
Inconel tubes were found to be lightly
attacked (0.5- to 1-mil penetration)
when in contact with the hydrofluorin-
ated material, but they evidenced no
attack when exposed to the hydrogen-
treated material. These preliminary
tests indicate that a final hydrogen
treatment of the fuel would be bene-
ficial, Additional tests are scheduled.

Temperature Dependence. Although
the immediate application of the

fluoride fuel NaF-ZrF4-UF4 (50-46-4

mole %) is to be limited to a tempera-
ture of 1500°F, some corrosion tests
under static conditions were performed
at considerably higher temperatures to
determine the temperature-dependence
of the corrosiveness of this fuel. As
the data in Table 11.5 indicate,
penetration of the Inconel by the
fluorides in these 100 hr tests in-
creased slightly, 1f at all, over the
temperature range 800 to 1200°C,

It 1s i1interesting to note that
Inconel seems to lose weight at 1200°C
instead of gaining weight as is usual
at the lower temperatures, No expla-
nation for the behavior athigh temper-
atures 1s suggested,

Ceramic Materials. Two specimens
of beryllium oxide blocks with approxi-
mate dimensions 0.25 by 0,25 by 0,50
in., were exposed to NaF-BeF, (57-43
mole %) for 100 hr (24,000 cycles) in
the tilting furnace with the hot end
(beryllium oxide mechanically held at
hot end) at 800°C, and the cold end at

TABLE 11.5. CORROSION(H?INCGNEI,BYNaFVZrF4-UE1AT TEMPERATURES FROM 880 70 1200°C

 

 

 

 

 

 

 

 

RUN NO TEMPERATURE WEIGHT CHANGE PENETRATION CHEMICAL ANALYSIS (ppm)
' (°C) (mg/dm?/day) (mils) Fe Cr Ni
299(a) 800 +4 0.5 to 1 170 90 20
+8 0.5 to 1 180 1860 45

-1 0.5 to 1 440 2100 20

234(2) 1000 +292 1 to 2 185 2700 20
+9 1 to 4 100 3000 20

+12 1 to 4 90 2600 20

255(%) 1100 0 0.5 to 2 95 980 20
+6 0.5 to 2 100 1000 20

258(¢) 1200 -13 0.5 to 2 200 1030 75
~31 0.5 to 1 290 670 110

 

 

 

 

 

 

 

{a)

Prerun analyses of fuel:

(&)

Prerun analyses of fuel:

{c)

Prerun analyses of fuel:

132

3000 ppm Fe, 980 ppm Cr, 230 ppm Ni.
720 ppm Fe, >20 ppm Cr, 135 ppm Ni.

340 ppm Fe, 220 ppm Cr, 580 ppm Ni.
650°C, One specimen came out with
~clean surfaces and showed a 2.0%
weight loss, whereas the other was
found to have a 4.6% weight gain as a
result of a magnetic oxide layer on
one side. This oxide was probably due
to iron and/or nickel from the tube
walls,

Several ceramic materials with
approximate dimensions 0.25 by 0.25 by
0.50 in. were tested 1n the fluoride
mixture NaF—ZrF4 50-50 mole %) under
static conditions, and the following
information was obtained: '

SPECIMEN WEIGHT CHANGE (%)
SiC i +35
TiO2 - - 17
ZJ."O2 {stabilized) - -51
Zr02 Dissolved
MgO Dissolved
Al ,0, - 317

The weight gain shown by the silicon
carbide may be due to soaking up of
the fluorides; if this is the case,
there is indication that silicon
carbide resists attack by this melt.

FLUORIDE CORROSION IN THERMAL
CONVECTION LOOPS

G. M. Adamson, Metallurgy Divisioni

Several Inconel loops were operated
with NaF-Zr¥F,-UF, (46-50-4 mole %)
under standard conditions, and in every
case the depth of attack was slightly
greater than had been encountered
previously., The fluoride charges for
these loops were all prepared in large
batches. The inhibiting effect of
zirconium hydride was confirmed in a
test with NaF-KF-LiF-UF, (10.9-43.5-
44.5-1.1 mole %). The inhibiting
effect was also observed with ZrF -
bearing fluorides, but 1t was not
quite so effective and caused changes
to take place in the fuel. NaK also
has an inhibiting effect on both these
fluorides but causes even greater
changes to take place. It was noted
that the fluorides possess a self-

PERIOD ENDING DECEMBER 10, 1952

insulating property that makes plugging
a flowing system more difficult than
would be expected, |

The results obtained during this
period with the Inconel loops in which
fluorides were circulated are summarized
in Table 11.6. This table is a con-
tinuation of Table 15 in the previous
report. ?) Table 11.6 should be
referred to in reading the following
discussion because i1t presents the
individual test data not given in the
specific sections. All loop tests,
except one that terminated prematurely
because of a power failure, operated
for 500 hr with a hot leg temperature
of 1500°F,

Mixtures Containing ZrF, .
as yet undetermined reason, the attack
found in the TIncomnel loops in which
fluorides containing ZrF, have been
circulated is increasing, The attack
by these fluorides had been reduced to
a depth of about 5 mils maximum after

500 hr at 1500°F,

For some

but i1n recent tLtests
the attack has been significantly
greater, that is, from 9 to 12 mils.
Although some of the earlier loops
were attacked to a depth of 10 mils,
the attack could be attributed to the
cleaning cycle, since 1t has heen shown
(both with other fluorides(?’ and with
lead) that the cleaning methods pre-
viously used caused an increase in
attack. The new attack is actually as
deep as, or deeper than, the attack in
all the earlier tests,
cleaning variables are i1ncluded.
Figures 11,3, 11.4, and 11.5 show
typical hot leg sections from these
loops. Figures 11.3 and 11.4 permit a
comparison of the earlier and the
recent attack of NaF-ZrF _ -UF, on
Inconel. Figure 11.5 shows a hydrogen-
fired loop of the earlier group. One
other point of interest is that an as
vet unidentified brown material 1is
found in the fluorides from the recent

even when the

 

(I)G. M. Adamson, ANP Quar. Prog. Rep. Sept.
10, 1952, OBNL-1375%, p. 105,

(ypid., p. 104,

133
el

TABLE 11. 6.

CORROSION DATA FROM INCONEL THERMAL CONVECTION LOOPS IN

WERE CIRCULATED FOR 500 hr{e)

AT 1500°F

WHICH FLUORIDE MIXTURES

 

LOoOP
NO.

CLEANING
PAGCEGURE

 

FLUCRIDE

METALLOGRAPHIC NOTES

 

MTXTUR$E

Hot Leg

Cold Leg

CHYMICAL NOTES

 

238

223

241

235

240

239

243

242

244

246

245

247

249

255

256

 

NaK

Nak

Degreased

Degreased

Degreased

Degreased

Degreased

Degreased

Degreased

Degreased

Degreased

Degreased

Degreased

NaF-KF-LiF-UF
10.9-43.5-44.5-1. 1 mole %

NaF-KF-LiF-UF,
10.9-43.5-44.5-1. 1 mole %

NaF-ZrF -UF
46-50-4 mole %
+ 0.5 wg % NaK

NaF-KF-LiF-UF.
10.9-43.5-44.5-1. 1 mole &
+ 0.5 we % Cr

NaF-KF-LiF-UF,
10.9-43.5-44.5-1.1 mole %

NaF-ZrF,-UF
45-50-4 muIe %

NaF-Zrf - UF,
46-50-4 mole %
+0.5 we % Ll

NaF-KF-LiF-UF
10.9-43,5-44.5-1. 1 mole %
+ 0.5 wt % ZrH2

NaF-ZrF‘-UF‘
46-50-4 mole %

NaF-ZzF, - UF,
52-48 mole %

NaF-ZrF, - UF
46-50-4 mole %

NaF-KF-LiF-UF,
10,9-43.5-44.5-1. 1 mole %
+ 0.5 wt % Cr

Hydrafluorinated
NaF-KF-LiF-UFi
10.9-43.5-44.5-1.1 mole %

NaF-KF-LiF-UF,
10.9-43.5-44.5-1.1 mole %

Hydrofiuorinated
NeF-KF-LiF-UF
16.9-43.5-44.5-1, 1 mole %

Moderate general and intergranular
attack, maximum 12 mils, average
7 mils

Typical Inconel attack, 8 to 10
mils; depth slightly deeper in
crevice, but twice usual amount
found!®!

Light pitting and intergranular
attack, maximum 4 mils, average

15 mils

Typical Inconel attack, more grain
boundary preference, maximum 13
mils, average 8 mils

Moderate attack, up to 11 mils,
sample attacked only 3 milsfe)

Moderate te heavy attack, 9 to 17
mils deep, sample attacked only
I miltd?

Widely scattered attack, 1 to
4 mils

Widely scattered attack, 1 mil;
occasional depozits in some grain
boundaries

Moderate Inconel attack, maximum
% mils, average 5 mils

Scattered intergranular attack, up
to B8 mils in top section; pitting,
1 mil, with some areas up to 6
mils in lower part of hot leg

Moderate to heavy Incanel artack,
maximum 12 mils, average § mils

Light to moderate intergranular
attack, 1 to 7 mils deep

Very heavy Inconel attack, average
18 to 20 mils, maximum 30 mils

Moderate intergranular atcack, to
20 miis in bath sections, ne evi-
dence of chrome plate!®!

Maximum penetration 18 miis,
average 10 mils, not so much in
lower section

 

Thin noametallic layer

Thin nonmetaliic deposit

Metallic deposit and needle- like

metallic crystals

Thin nonmetallic surface layer that
follows surface contours

Thin surface deposit

Light, spotty deposit with some
metallic erystals

Very thin surface deposit anrd de-
posits of crystals

Metaliic layer with nonmetallic
particies occluded

No deposit

No depesit

 

Occasicnal nonmetallic particles

Thin layer that appears to be
metallic with some nonmetallic
particles embedded in it

Tightly adhering metallic layer of
L mil, surface rough with some
nonmetallic particles

Heavy, light-colored deposit
throughout leg, varies from 0.25
to 0.75 mil in thickness

Heavy metallie layer with some
attached crystals, up to 0.004 in.
long

 

Fe and U decreased, Cr increased
slightly

Fe and U decreased, Cr increased

Zr and Fe decreased, U, Cr, and
Ni vary

Cr added still in top pot, Fe
and U slightly lower than origi-
nal

U varies widely, Cr increased,
Fe decreased

U increased but varies, Zr de-
creased about 3%, Cr increased,
Fe decreased

U fairly constant but 3% above
original, Zr decreased, Fe
varies, Cr increased

U and F decreased siightly, Cr
and Fe vary but decreased

U increased about 1%, Zr de-
creagsed 2.5%, Fe and Ni de-
creased slightly, Cr increased

Cr and U increased, Fe and Ni
decreased

Cr and U increased, Fe and Ni de-
creased, all high in original
analysis

 

(')Operltion cf each loop was terminated

 

 

(b)Llp Jjoint in hot leg

U¢)Sample in hot Leg.

(‘)Slmpie in top of hot leg

(e} gix-inch chrome-plated saction welded

inta hot leg.

at 500 hr, as scheduled, except loop 235, which was terminsted at 455 hr becavee of « power failure.

LHOdAY SSAYO0Ud ATHALHYNO LOALOMd dNV
PERIOD ENDING DECEMBER 10, 1952

 

Fig- 1i1-3-

NaF-ZrF4~UF4 Prepared in 5-1b Batches.

tests that was not found in
fluorides from the earlier tests.

The major difference in these tests
is that the earlier loops were filled
from small batches of fluorides made
in the laboratory, whereas the later
ones were filled from 50-1b batches
made in a production operation. The
attack now being measured is not
serious, since it 1s comparable to
that normally found with NaF-KF-LiF-
UF, (10.9-43.5-44.5-1.1 mole %), but
the trend may be serious. The actual
fuel material for the reactor will be
produced in 250-1b batches instead of
the 50-1b batches being produced at
present. A joint chemical, experimental
engineering, and metallurgical program
is under way to determine whether the
increase in batch size affects the

the

Hot Leg of Inconel Loop Tested for 500 hr at 816°C with

250X.

corrosion or whether the increase 1n
corrosion is the result of some change
in handling procedure.

Most of the postrun examlnatlons(g)
of the fuels from thermal convection
loops have been confined to petrographic
and x-ray examinations. The NaF-ZrF,
UF, (46-50-4 mole %) fuel circulated 1n
several Inconel and type 316 stainless
steel loops was studied, The ple-
ochroic, brown or olive-drab phase
that has an average refractive index
of 1.556 was first noted in a type 316
stainless steel loop, but it has also
been found in several Inconel loops
containing this mixture, as well as in
ZrF,-bearing fluorides treated with
NaK, Zr, and ZrH,. In these dynamic

(S)Examination of fluoride mixtures were made by
D. C. Hoffman, Materials Chemistry Division.

135
ANP PROJECT QUARTERLY PROGRESS REPORT

 

Fig. 11. 4.

NaF-ZrF4-UF4 Prepared in 50-1b Batches.

corrosion tests, there appears to be
some correlation (which did not exist
in static tests) between the amount of
the olive-drab phase and the extent of
corrosion,

Hydrofluorinated NaF-KF-LiF-UF,.
One other series of tests points to
poor production technique as the cause
of the trouble discussed above. The
attack found in Inconel loops in which
NaF-ZrF, -UF, (46-50-4 mole %) was
circulated had been only one half of
that found with NaF-KF-LiF-UF, (10.9-
43.5-44.5-1.1 mole %). One explanation
for this has been the higher purity of
the ZrF,-bearing fluoride. During
production, the ZrF, -bearing material
is purified with both hydrogen and
hydrogen fluoride, whereas the NaF-KF-
LiF-UF, is not. Two special batches
of NaF-KF-LiF-UF, were prepared by the

136

Hot Leg of Imconel Loop Tested for 500 hr at 816°C with
250X,

group that makes the NaF-ZrF,-UF,, and
the same techniques were used., Un-
expectedly, the metal impurities in the
specially prepared batches were much
higher than in the normal material.
In the two special batches, the metal
impurities were:

BATCH 1 BATCH 2
Chromium, ppm 2200 1280
Iron, ppm 5600 7000
Nickel, ppm 7100 5300

The presence of these impurities has
not yet been explained. Metallographic
examination after a test with the
first batch revealed a wmaximum hot
leg attack of 30 mils, with an average
of 18 mils. This 1s the deepest
attack yet found with any fluoride
loop. A loop 1s now being run to
 

PERIOD ENDING DECEMBER 190,

1952

s 5l . ; N e )
. ¢ * . : ‘r & & S - . ¢
I . ‘ b boraen t P e, MILS
i ; . T _ fg . ) M e .
: ot . L O e T TINGS s N
" ol ey ! . .8 : L e
A T At "J L, e .
. % e i PR et i / L ‘ .
j‘ + ..1 2 v \\ . __-f‘f ; ~ ’ %' W ; - F"}*';“' :v- : vy -
- » » - o ,.--,’, e '?‘W/ : } » - Lt . . 4 - N
e ¢ v o K o ‘ . S « 7 ‘Q,
o ts - . J - . - . ¥ - L
% 3 o . » » . .
o i R N T e B
o " ' » &, LT : ; - . '. -~ . Mm .‘;w" . t to . AQ- . »
. » . - TR L e . fi.- e P ; ‘ ‘-' . P
Fig. 11.5. Hot Leg of Hydrogen-Fired Inconel Loop Tested for 500 hr at

816°C in NaF-ZrF,-UF,. 250X.
determine whether the addition of
iron and nickel metal powders to
NaF-KF-LiF-UF, will cause an increase
in attack. The test with hydro-
fluorinated fuelswill also be repeated.
Corrosion Inhibitors. It was
pointed out in the previous report(*’
that the addition of 0.5% ZrH, caused
a reduction in the hot leg attack on
Inconel by NaF-KF-LiF-UF, (10.9-43.5-
44.5-1.1 mole %) to only 0.5 mil.
The test has now been repeated, and
the attack found in the second loop
was 1 mil., The unidentified material
found in many of the grain boundaries
in the hot leg section of the first
loop was also found here, but only in
very small amounts in widely scattered
areas. A thin adherent layer that
appears to be metallic was found in
the cold leg. The fluorides in the
loop after the run consisted of

(4)9. M. Adémson, op. cit., b. 108.

alternating layers of a colorless
material and the usual green phase.
Petrographically, these two materials
seem to be very similar, or even to
be the same. No reduced or partly
reduced phases were found in the
Yoop itself.

Since optimistic results were ob-
tained with the loops discussed above,
two loops were run with zirconium
hydride added to NaF-ZrF,-UF, (46-50-4
mole %). Operation of both these loops
has been completed, but the results of
metallographic examination are availa-
ble for only the first one. Scattered
attack, with a maximum penetration of
4 mils (average 2 mils), was found.
This attack is definitely less than
that found in the other NaF-ZrF, -UF,
loops run during this period, but the
effect of the zirconium hydride 1s not
so great as was found with the NaF-
KF-LiF-UF, mixture. No grain-boundary

137
ANP PROJECT QUARTERLY PROGRESS REPORT

constituent was found in this loop.
A considerable change had taken place
in the fluorides in both loops, as
evidenced by the unidentified brownish
material found, along with the green
phase, in all sections. No free UF,
was produced, although intermediate
reduction products similar to those
present 1n NaK loops were found.

The static corrosion tests had
indicated optimistic resnlts when
base metals were used as inhibitors.,
Previously reported work had also
shown a reduction in attack when
nonuranium-bearing fluorides were
circulated. Small amounts (about 10
cm®) of NaK were therefore added to
two Inconel loops, one of which
contained NalF-ZrF, -UF, and the other
NaF-KF-LiF-UF,. In these loops, both
the amount and maximum depth of attack
were lower than those in comparable
loops without the NaK. Although the
improvements were not large and could
have resulted from the normal spread
1n results, the fact that the metallic
impurities in the fluorides were so
low would indicate that the im-
provements were real. The NaF-ZrF, -UF,
(46-50-4 mole %) was changed in
appearance and included yellow material
that was present both as clumps and
as a grain-boundary constituent.
This loop also gave considerable
difficulty during startup, and it was
necessary tostrike the legs frequently
with a hammer to wmaintain circulation.
Free UF,, as well as intermediates,
was produced in the fluoride.

the material
being leached from the hot leg, the
addition of chromium to the fuel could
possibly slow down the reaction. It
would be necessary to keep the chromium
concentration low enough to prevent
saturation and deposition in the cold
leg, but high enough to reduce the rate
of solution in the hot leg. The solu-
bility of chromium in the fluorides
has not yet been determined. Two
loops were operated with a chromium

Since chromium 1s

138

addition, but the results are 1in-
conclusive since the chromium did not
mix with the fluorides. The analysais
of the fluorides in both loops showed
a chromium concentration in the various
loop sections that was even lower than
normal, whereas a chromium concen-
tration of over 5% was observed 1in the
fluorides in the expansion pot of one
loop.

Inserted Corrosion Samples. Two
Inconel loops were operated with small,
flat, Inconel samples 1nserted 1in the
top of the hot leg. In a loop operated
with NaF-ZrF,-UF, (46-50-4 mole %),
the wall attack in the hot leg con-
sisted of moderate integranular
and general attack from 9 to 17 mils

deep, which is deeper than normal,
The sample in this loop, however, was
attacked to adepth of only 1 mil. The

second loop, which contained NaF-KF-
LiF-UF, (10.9-43.5-44.5-1.1 mole %),
showed moderate wall attack from 5 to
11 mils deep, which is average attack
for NaF-KF-LiF-UF,. The sample was
attacked lightly to only 3 mils 1n
depth. Two more loops are now circu-
lating with small thin-walled tubes
inserted in the hot leg. Thermocouples
inserted in these tubes show that a
large temperature drop isnot responsi-
ble for the lack of attack on the
samples.

Crevice Corrosion. additional
check on the crevice corrosion found
at the top weld in the thermal loops
and i1n one of the pump loops, a loop
was operated with two crevices fabri-
cated into the hot leg., NaF-KF-LiF-UF,
(10,9-43.5-44.5-1.1 mole %) was circu-
lated in this loop at 1500°F. 1In
the upper section of the hot leg,
the wall attack varied from 4 to 7
mils, whereas in an adjoining crevice,
the attvack was from 6 to 8 mils and
twice as concentrated. At the lower
crevice, the pipe showed only scattered
pits, up to 5 mils deep, whereas
normal attack, up to 6 mils, was found
in the crevice. This loop showed only
a slight increase in depth of attack

As an
in the crevices, but a considerable
increase 1n amount.

Variation in Loop Wall Composition
(J. P, Blakely, Materials Chemistry
" Division, G. M. Adamson, Metallurgy
- Division). Additional drillings have
been made from the inside of the loop
- walls, and steps were taken to assure
that the pipes were not deformed during
the drilling. The chemical analyses
of these samples are shown in Table
11.7. TInconel loop 229, in which
NaF-KF-LiF-UF, (10.9-43.5-44.5-1.1
mole %) was circulated, and Inconel
loop 236, in which NaF-ZrF,-UF, (46-
50-4 mole %) was circulated, were
operated under standard conditions.
The change of analyses with depth from
the surface follows the same general
pattern as for the loops discussed in
the previous report.(s) '

Self-insulating Properties of
Fluorides. An Inconel loop was filled
with NaF-KF-LiF-UF, and allowed to
circulate in the normal manner at
1500°F. After normal circulation had
been established, two helium jets were
directed against an area at the bottom
of the cold leg. Each jet delivered
over 1000 ft3/hr of gas for 18 hr but
did not cause the loop to plug.
Initially, the cold leg temperature
dropped, but the power was increased
and the temperatures gradually returned
to normal. The helium jets were then
replaced by air jets that delivered a
sufficient volume to form a black spot
about 6 in., long. The loop circu-
lated over 1000 hr with these jets
operating, but there was apparently no
plugging caused by the temperature
differences.

Air jets were then turned on a
section of the loop in the cold leg,
and, after a total of 2300 hr, operation
was terminated. The fluoride 1in both
sections appeared normal upon visual
inspection, but in neither section

(S)Ibid.. p. 11D,

PERTOD ENDING DECEMBER 10, 1952

would the fluoride melt under the
conditions normally used for this
operation,

Other Loop Tests. A type 316
stainless steel loop (No. 117) circu-
lated NaF-KF-Li¥F-UF, (10.9-43.5-44.5-
1.1 mole %) for 500 hours. This is
only the second type 316 stainless
steel loop that has not plugged, and
both of these loops were operated under
similar conditions, The cold leg of
this loop was held above 1500°F by
using a hot leg temperature of 1630°F
and wrapping the cold leg with asbestos
tape. No sign of plugging was evident
at 500 hours. At this time, the hot
leg temperature was reduced to 1500°F,
and the asbestos tape was removed so
that the loop would circulate under
standard conditions. The loop then
plugged after 366 hr at the lower
temperature, or a total of 866 hours.
A considerable quantity of metallic
flakes was observed in the fluorides.
The fluoride in the sections of this
loop melted at below 1250°F, and it
all melted.

An Tzett iron loop, which had been
cleaned with dry hydrogen, plugged 1in
46 hr of eirculation of NaF-KF-LiF-UF,
(10,9-43,5-44.5-1.1 mole %) at 1500°F,
The inside surface of the hot leg was
very rough, and considerable reduction
in wall thickness had taken place. A
heavy metallic deposit in the form of
large metallic crystals with some non-
metallic particles embedded i1n them
was found in the cold leg.

Inconel loop 246 circulated NaF-ZrF,
(52-48 mole %) for 500 hr at 1500°F.
Scattered intergranular attack of from
3tp8 mils was found in the top section
of the hot leg, The lower section
showed general pits of 1 mil, with an
occasional patch up to 6 mils deep.
No deposit was found in the cold leg,
This attack is comparable to that found
in the previous tests with fluorides
that did not contain uranium. '

139
0v1

 

 

 

 

 

 

 

 

 

 

TABLE 11.%. ANALYSES OF LOCGP WALL COMPOSITION
DEi;j‘OF COMPOSITION (%)
(mile) Fe Ni Cr Fe/Ni NaF* KF* LiF* ZrF,* | Other** Total
Hot Leg of Loop 229
3 6.73 81.7 .19 0.083 0.26 0.74 0.74 1.34 69.77
6 6.83 80.9 9.38 0.085 0.26 0.44 0.41 1.19 99.41
10 6.99 79.7 11.1 0.088 0.14 0.30 0.27 1.23 99.73
15 6.98 78.1 13.1 0.089 0.11 0.15 0.22 1.40 100.16
20 6.85 77.8 14.2 0.0C88 0.07 0.12 0.16 1.42 160.62
25 6.91 76.8 15.4 9.090 1.47 100.58
30 6.87 16. 4 15.6 0.089 1.66 100.53
External 6.90 76.5 16.0 0.9090 1.36 100.76
Cold Leg of Loop 229
3 8.46 73.8 16.6 0.114 1.45 106. 31
6 8.15 74.6 16.5 0.110 1.64 100.89
10 7.82 4.4 16.17 0.105 1.75 100.67
15 7.63 74.5 16.6 0.102 1.79 100.52
20 7.72 74.4 16.5 0.104 1.64 100.25
External 7.62 74.6 16.7 0.102 1.65 106,57
Hot Leg of Loop 236
3 11.7 74.8 10.3 6.156 0.14 0.92 1.61 99.47
6 9.44 74.6 13.7 0.126 0.04 0.55 1.49 99.82
10 9.02 74.8 14. 38 0.121 1.75 100. 37
15 7.70 74. 4 16.5 0. 43 1.56 100,16
20 7.70 74.1 16. 6 0.104 1.43 99.83
25 7.63 74.3 16.5 0.102 1.56 89.99
30 7.68 74,2 16.6 0.103 1.56 100.04
External 7.68 74.6 16.6 0.103 1.35 100,23

 

 

 

 

 

 

 

 

 

 

 

‘These values caiculated from spectrographic analysis of the metals by assuming them to be present as fluorides,

‘.Total of Ca, Cu, Si, Al, Mg, Mn, and Ti, from spectrographic analysis.

LHGJAM SSAYI08d ATHALYVNO LOACO¥d 4NV
HYDROXIDE CORROSION
Vreeland L. R, Trotter
Hof fman J. E. Pope
Metallurgy Division

F. Kertesz
Materials Chemistry Division

Mo
et O

Corrosion Inhibitors. A survey of
the abstracts contained in ORNL-1291¢8)
indicated that additions of Ca(Q,
Ca(OH)z, NaNO,, NaNO;, and NaClO,
might have some effect in inhibiting
corrosion by hydroxides, CaQ and
Ca(OH), were reported to participate
1n reactions of the following type:

Na2C03 + Ca(()H)Q"'“_“> 2NaQOH + CaCOa,
Na,CO, + CaO +H,0 ~—> NaOH + 3CaCo, .

Reactions for the alkali metal nitrites
and nitrates and sodium chlorate were
not given, but the reduction of cor-
rosion by these additives was stated
to be as much as 90% under certain
conditions, Static tests of type 316
stainless steel in sodium hydroxide
with various additives were run at
816°C for 100 hours., None of these
approximately 5% additions seemed to
have an effect in inhibiting cor-
rosion in these tests. It is believed
that the inhibiting effect mentioned
in the references may take place at
much lower

(G)M. Lee, General Information Concerning

Hydroxides, ORNL-1291 (April 21, 1952).

temperatures than are

PERIOD ENDING DECEMBER 10, 1952

employed in these tests. The test
results are presented in Table 11.8.

Temperature Dependence. A series
of tests for determining the effect of
temperature on the corrosion resistance
of types 304 and 316 stainless steels
exposed to sodium hydroxide for 100 hr
has been completed. Attack on these
materials was not noted until the test
temperature was increased to at least
550°C. The results are tabulated in
Table 11,9, :

Nickel Alloys in NaOH, Static
tests in NaOH of an 80% Ni-20% Mo alloy
and an B0% Ni-20% Fe alloy were run
for 100 hr at 816°C. The 80% Ni-20%
Mo alloy showed subsurface voids to a
depth of 1 to 2 mils, and the 80%
Ni-20% Fe alloy showed 3 to 5 mils of
subsurface voids upon metalloegraphic
examination. |

Compatibility of Be0O in KOH. A
sample of BeO was tested in molten KOH
contained in an Inconel tube with a
hydrogen atmosphere. The hydrogen was
purified with a “Deoxo’ unit., The KOH
was dehydrated, under vacuum, for 56
hr by gradually raising the tempera-
ture to 500°C before introducing the
BeO specimen., During the test, a
constant flow of hydrogen was maintained
through the system, The time of the
test was 100 hr and the temperature
was B800°C. The BeQ was rather severely
attacked, as can be seen from the

 

 

 

 

TABLE 11.8. RESULTS OF STATIC TESTS OF TYPE 316 STAINLESS STEEL
~ IN NaOH WITH VARIOUS ADDITIVES FOR 100 hr AT 816°C

ADDITIVE METALLOGRAPHIC NOTES

5% NaClO3 Unattacked material decreased from 33 to 26 mils

5% NaNO2: Unattacked material decreased from 33 te 25 mils

5% NaNO3 Unattacked material decreased from 33.5 to 20 mils: |
intergranular penetration of 1 to 2 mils beneath uniform
corrosion layer :

5% Ca0 Unattacked material decreased from 33.5 to 26 mils

5% Ca(OH), Unattacked material decreased from 33.5 to 26 mils

 

 

141
ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 11.9,

RESULTS OF TEMPERATURE-DEPENDENCE TESTS O0F TYPES 304 AND

316 STAINLESS STEELS EXPOSED IN NagH FOR 100 HOURS

 

 

TEMPERATURE (°C)

 

MATERI AL METALLOGRAPHIC NOTES
anype 304 stainless steel 350 No attack .
450 No attack
550 Light intergranular penetration
of 1.2 to 1 mal
650 Unattacked material decreased
from 34 to 24 mils
Type 316 stainless steel 350 No attack
450 No attack
- 550 Intergranular attack of 1 to 2
mils

 

 

 

welght and dimensional changes listed
in the following:

BEFORE TEST  AFTER TEST
Length, in. 0.486 0.413
Width, 1in. 0.263 0.207
Thickness, in. 0.269 0.213
Weight, g 1.5254 0.6882

of the
Inconel tube used in this test revealed
a 4 to 10-mil oxide layer.

Nickel in NaOH Under Hydrogen
Atmosphere. Maintenance of a hydrogen
pressure from an external source has
been shown to minimize corrosion and
transfer of nickel in NaOH.
Maintenance of an atmosphere of hydrogen
in components of a high-temperature
reactor would 1ntroduce a number of
problems; however, the prospect of
containing hydroxides in this fashion
has appeared sufficiently promising to
justify systematic study of this
possibility,

In the experiments, sodium hydroxide
containing less than 0.1% Na,CO, and
less than 0.1% H,0 was used, and all
handling of the material was conducted
in a dry box that conld be evacuated
and flushed with pure, dry helium. The

Metallographic examination

mass

142

metal to be tested was heated for 30
min at 800°C in pure flowing hydrogen
and subsequently handled so that the
inner surface of the tube was exposed
only to pure helium before the corrosion
test., The hydrogen used 1in these
experiments was purified by passage
through a “Deoxo’ catalytic mass. The
water was removed by liguid-nitrogen-
cooled traps and by magnesium per-
chlorate,

In one type of test, the movement
of hydroxide was achieved by thermal
convection in a single 1/2-in.-dia
tube inclined at 45 deg and heated to
800°C at the bottom. A temperature
gradient of about 200°C was maintained
between the top and bottom of the
hydroxide that filled the 18-in. tube
to a depth of about 12 inches,

Tubes of nickel were placed in an
outer jacket of stainless steel designed
to contain the furnace assembly so that
the hydrogen can bathe the outside
walls as well as the contents of the
tubes and were exposed to sodium
hydroxide for 48 hours. Figure 11.6
indicates the great difference in the
behavior of nickel 1n NaOH when hydrogen
instead of helium is used in tests of
TEMPERATURE GRADIENT

80G°C — BOTTOM
560°C ~ TOP

TUBE INCLINED 45°

100-hr TEST

" Boo®C

BOTTOM
Fig. 11.6.
this type. Tests conducted in the

helium atmosphere show large deposits
of nickel crystals at the cold end of
the tube, whereas the bottom sections
of the tube are highly polished. When
hydrogen is used, the bottom sections
of the tube show considerable polish,
but virtually no crystal deposit is
visible at the cold end.

When tests are conducted in a
similar fashion but with the hydrogen
(at atmospheric pressure) applied only
to the inside of the tube under test,
the beneficial action isstillobtained.
The bottom section of the tube shows

PERIOD ENDING DECEMBER 10, 1952

   

PHOTO 11862

LIQUID LEVEL~—0u TOP

    
   
    
  

56Q°C

580°C

Standpipe Test of NaOH in Nickel Capsule with Hydrogen Atmosphere.

the typical polished appearance, but
no mass transfer to the cooler portions
is evident,

In another type of test, nickel
tubes fitted with a delivery tube for
maintenance of the proper atmosphere
were exposed in the tilting furnace.
The tubes were rocked four times per
minute, with the hot end maintained at
800°C and the cold end at 650°C, for
100 hours. Results very similar to
those described above were obtained in
these studies, as the specimens shown
in Fig. 11.7 indicate. Since the
hydrogen has ready access to virtually

143
ANP PROJECT QUARTERLY PROGRESS REPORT

800°C

 

14S1H51

t492H54

HYOROGEN ATMOSPHERE
TUBE ONLY DEGREASED

HYDROGEN ATMOSPHERE
TUBE HYBROGEN-~FIRED
BEFORE TEST

Fig. 11.7.

all parts of the tubes during the test,
it is hardly surprising that prior
hydrogenation of the specimens did not
appear to be beneficial.

It is noteworthy that when helium
atmospheres were used in these tests,
the NaOH recovered was brownish-gray
in color and contained about 1500 ppm
of nickel. When the hydrogen atmos-
pheres were used, the snow-white
hydroxide contained 20 to 35 ppm of
nickel.

 

(?)A. des Brasunas, G. P. Smith, D, C, Vreeland,
and F, Kertesz, ANP Quar. Prog. Rep. Sept. 10,
1952, ORNL-1375, p. 119.

144

PHGTO 11881

 

650°C

 

1493H51 147 7H51

HELIUM ATMOSPHERE
TUBE HYDROGEN-FIRED

HYDROGEN ATMOSPHERE
TUBE HYDROGEN-FIRED BEFORE
TEST INTERMITTENT VACUUM
APPLIED

Seesaw Tests of NaQOH in Nickel Capsules After 100 Hours,

LIQUID METAL CORROSION

D, C. Vreeland J. B, Pope
E. E, Hoffman J. V. Cathcart
L. R, Trotter G. P. Swmith, Jr.
G. M. Adamson
Metallurgy Division
L. A. Mann J. M. Cisar

ANP Division

High-Velocity Corrosion by Sodium.
A dynamic test, which has become known
as the spinner test,’? has been used
to check the effect of high-velocity
corrosion and erosion in molten sodium
on several materials. In this test,
relatively high velocities (400 ft/min)
are obtained by rotating tubular
specimens in a pot containing the molten
corrodant under isothermal conditions.
Inconel and types 316, 347, and 446
stainless steels have been tested in
this apparatus. Greater attack than
that obtained in ordinary static cor-
rosion tests was noted onall materials
tested. Tt should be emphasized that,
except in the test of the type 347
stainless steel specimens, these were
actually three-component tests because
the spinner pot was fabricated of type
347 stainless steel.

| The effect of this test on Inconel
spun at 400 ft/min for 72 hr at 816°C
is shown in Fig. 11.8. A 0.5-mil
surface layer, with 0.5 mil of inter-
granular voids beneath it, is apparent.
This surface layer may be a mass
transfer effect caused by the three-
component system used during test.
The analysis of a portion of the

surface layer scraped from the specimen
and

showed 64% nickel, 27.5% chromium,

 
 
    
  

o -
o :
: »
. a?“:
e  ?_ 5
- *?
Fig. 11.8.

Speed of 400 ft/mim. 500X,

PERIOD ENDING DECEMBER 10, 1952

8.3% iron. (The composition of Inconel
is 80% nickel, 15% chromium, and 5%

iron.)

The intergranular attack on type
316 stainless steel 1s shown 1in Fig.
11.9. This attack extended to a depth
of 1 to 2 mils near the front, or
leading edge, of the tubular specimen
and, as would be expected, was less
severe at the trailing edge, where the
depth of attack was 0.5 to 1 mil. The
test was conducted at 816°C for 100 hr
with a speed of 400 ft/min. Type 347
stainless steel suffered intergranular
attack to 2 mils after testing for 74
hr at 816°C with a speed of 400 ft/min.
After 53 hr at 816°C at a speed of
400 ft/min, type 446 stainless steel
showed slight intergranular attack to
a depth of less than 0.5 mil. The
results of these tests are summarized

in Table 11.10,

{ UNCL ASSIFIED
Y8252

1SURFACE LAYER

. ""fiq ATTACHED REGION

 

INCONEL

Inconel from Spinner Test with Sodium at 816°C for 72 hr at a

145
ANP PROJECT QUARTERLY PROGRESS

  
    
 

Fig. 11.9.

for 100 hr at a Speed of 400 ft/min.

TABLE 11.180.

REPORT

| UNCLASSIFIED
. Y-8020

 

Ty, -

 

 

Type 316 Stainless Steel from Spinner Test with Sodium at 816°C

250X,

RESULTS OF SPINNER TESTS RUN AT 400 ft/min AT 816°C

 

 

TIME OF
MATERIAL
A TEST (hr) METALLOGRAPHIC NOTES

Inconel 72 Sur face layer of 0.5 mil with 0.5 mil of inter-
granular voids beneath

Type 316 stainless steel 100 Intergranular attack, 1 to 2 mils deep on lead-
ing edge and 0.5 to 1 mil deep on trailing
edge

Type 347 stainless steel 74 Intergranular attack, to 2 mils

Type 446 stainless steel 53 Slight intergranular attack, less than 0.5 mil

 

 

 

Ceramic Materials in Sodium,. A
series of static corrosion tests of
various ceramic materials received
from the Ceramic laboratory has been
run in sodium at 816°C for 100 hr in
Inconel tubing, The results are
tabulated in Table 11.11 for the
materials that did not disintegrate
during test. Only two of the specimens
were intact after testing (Al,0,; and

146

Spinel). It 1is possible that if the
specimens that broke up during testing
had been thicker (specimens were
approximately 0,020 in, thick), some
of them would have remained intact,
The specimens that disintegrated into
fine powder during the testing included
porcelain (A1-576), mica, titanium
dioxide (Al-192), 5i0,, barium titanate
(A1-T-106), Al1,0, (single crystal), lead
 

 

 

PERYIOD ENDING DECEMBER 10, 1952
TABLE 11. 11, RESULTS OF TESTS OF VARIOUS CERAMIC MATERIALS
"IN SODIUM FOR 100 hr AT 816°C
MATERI AL OBSERVATIONS AFTER TESfl'
Al,0, Three-mil increase in thickness, weight increase
of 0.01 g/in. 2 |
S1C Specimen broken into several pieces, no thick~
ness change ‘
MgO Specimen broken into several pieces, no thick-
ness change
MgO (single crystal) Specimen broken into several pieces, no thick-

ness change

Spinel
g/in. 2

20, (Al 550)

Zircon
ness

BeO

 

 

glass, Zr0, (Hf free), ThO,, 2MgO-Si0,
(fosterite, Al-243), 2Mg0‘2A1203‘FSiO2
(A1-202), aluminum silicate (Al-A),
pyrex glass, zirconium silicate (A]l-
475), HiO,, soda glass, TAM Zr0,, and
MgO- 310, . ;
Lead in. Metal Convection Loops. The
problem of circulating lead has been
revived during this quarter. lLoops
identical to those being used with the
fluorides were first used in this work,
The lead used in all the following
tests was chemical lead that was
further purified by bubbling hydrogen
through 1t.
Lead was first circulated
Inconel loops. The first loop, which
had been cleaned with NaK and then
evacuated while hot, would not circu-
late at all, The second loop was
rinsed with water, after the NaK
cleaning, before being evacuated.
This loop circulated only 2 1/2 hours.
After the lead was drained out, plugs
were found in both loops in the lower

in two

No change in thickness, weight loss of 0.006

Specimen shattered, l-mil increase in thickness

Specimen shattered, l.5~mil increase in thick-

Specimen shattered, 7-mil increase 1n thickness

 

portion of the cold leg. Neither
plug would melt at 1575°F. Besides
the plug in each loop, ‘amass resembling
a vine wound around the inside of the
cold leg and occupied about one-fourth
of the volume. This mass held 1ts
shape where 1t turned into the hot
lJeg., The analyses of these masses
show them to be high in nickel and
lead content; however, the x-ray dif-
fraction lines do not fit those from
any material in this system.

ILead was also circulated 1n a type
430 stainless steel loop that was
rinsed with cold nitric acid, washed
with water, and then evacuated while
hot., This loop circulated 16 1/2
hours. The two cleaning methods were
then combined and used on a type 410
stainless steel loop. This loop was
NaK cleaned, water washed, cold nitric
acid rinsed, water washed, and heated
while under vacuum. Tt circulated for
351 hr at 1500°F before plugging. The
plugs in both these loops again

147
ANP PROJECT QUARTERLY PROGRESS REPORT

appeared to be at the bottom of the
cold leg. They remained when the lead
was meleced out at 1575°F, The plugs
have been submitted for chemical and
x-ray diffraction study.

Lead in fuartz Convection Loops.
The prohibitive amounts of mass trans-
fer that occurred 1n lead-Inconel
thermal convection loops and in static
corrosion tests have prevented the
acquisition of significant data on
this system. In addition, the results
were open toquestion because sufficient
care had not been taken to deoxidize
the lead and because the Inconel tubing
used was covered with a heavy coating
of oxide.

An apparatus built in which
lead could be deoxidized with hydrogen
and then brought into contact with
Inconel specimens without exposure to
an oxidizing atmosphere., This appa-
ratus, shown in Fig., 11.10, consisted
of a thermal convection loop constructed
entirely of 1/2-1in. quartz tubing. Two
6-in. lengths of 7/16-1in. Inconel tubing
were mounted in the hot and cold legs
of the loop. Before being placed in
the loop, the Inconel tubing was
scrubbed with cleaning powder and
rinsed with alcohol and ether. The
lead was deoxidized in a 1 1/2-in, -
dia quartz bulb attached above the
loop and separated from it by a fritted
quartz disk. The lead i1n the de-
oxidation bulb was heated to 425°C,
and hydrogen was allowed to pass
through the fritted quartz disk and
bubble up through the lead. When the
deoxidation was complete, the lead was
forced through the fritted disk and
into the loop. Circulation of ‘the
lead was maintained by the thermal
convection currents thus induced.

was

This procedure assured that the
oxide content of the lead was reduced
to a minimum and, in addition, pre-
vented re-oxidation of the lead during
the loading of the loop. This technique
also offered several other advantages.

148

The use of quartz loops rather than the
conventional metal loops eliminated
many of the difficulties previously
encountered with leaks at welds. The
new loop required only a small amount
of metal tubing, a factor of some
importance for cases in which the
metal to be tested 1s expensive or in
short supply. Since the test specimens
were completely enclosed in the quartz
tubing of the loop, itwas not necessary
to provide a protective atmosphere
around the outside of the specimens,
as would have been the case 1fordinary
metal loops of such reactive metals as
molybdenum or columbium had been the
solid metal under test.

Two tests have been made on Inconel
specimens with 0,035-1in. wall thickness
in the apparatus described above for
circulating lead. Both loops had a
hot leg temperature of 800°C; one had
a cold leg temperature of 400°C and
the other 475°C. The results of the
two tests differed only in deta1l.
Circulation of the lead in one loop
stopped after 125 hr of operation and
in the other after 51 hr because of
formation of plugs in the cold legs.
The depth of attack in the hot leg
loop and 3 mils 1in
the other loop. In addition to the
usual mass transfer, a matrix-like,
corroded layer formed on the Inconel.
This corroded layer is shown in Fig.
11.11, which is a photograph of a cross
section of the Inconel tubing 1in the
hot leg of the loop. The corresponding

was 10 mils in one

cold-leg specimen 1is shown in Fig.

11.12.

In a preliminary test, the primary
purpose of which was to test loop
design, a loop was used that was
similar in design to that shown in
Fig., 11.8, except that 1t was con-
structed of pyrex glass. The hot and
cold leg temperatures were 500 and
365°C, respectively. No mass transfer
or corrosion was noted in either leg
during 96 hr of operation. These
results, together with the lack of
PERIOD ENBING DECEMBER 190, 1952

UNCLASSIFIED
DWG. 17408

TO HYDROGEN AND VACUUM LINES

LEAD
DEOXIDATION

BULB ~————w]
77
iy
LEAD-——m—-m&ééégé
FRITTED
QUARTZ DISK—m=

 

 

 

 

 

12in.

    
  

'PIPS" TO SUPPORT
INCONEL TUBE

 

 

 

 

it

Y,-in. QUARTZ TUBING

 

 

 

Fig. 11.10. Quartz Thermal Convection Loop.

149
ANP PROJECT QUARTERLY PROGRESS REPORT

Ee UNCLASSIFIEDY
T Y8120 . .&
«gfl@@wfi@*;g%

# Wi

el

 

 

F1g.

Inconel Specimen in Hot
(800°C) Leg of Quartz Loop after Circu-

11.11.

lating Lead for 125 Hours. 90X
corrosion in the cold leg of the first
loop (at 400°C) and only very slight
corrosion in the cold leg of the
second loop (at 475°C), led to the
conclusion that the corrosive attack
of liquid lead on Inconel probably
began in the neighborhood of 500°C,

The chemical analysis of the plugs
formed in the guartz loop tests was
only slightly different from the
nominal composition of Inconel. Tt
was therefore concluded that the
liquid lead had virtually no selective
leaching action on the Inconel speci-
mens. It was further concluded that
both severe mass transfer and corrosive
attack will occur in lead-Inconel
thermal convection loops with hot leg
temperatures of 800°C, even when the
lead has been carefully deoxidized
with hydrogen.

Compatibility of Be0 in Nak.‘®’

Dynamic tests of BeO in NaK are being

(S)The chemical analysis of NaK for beryllium is

reported in section 15, “Analytical Studies of
Reactor Materials.”

150

 

A

" UNCLASSIFIE
Y.8122

3 -

¥

rig. 11.12. Inconel Specimen in Cold
(400°€) Legof Quartz Loop After Circu-
lating Lead for 125 Hours. 90X,

conducted by the Experimental Engi-
neering Department. Blocks of BeO
approximately 0.25 by 0.25 by 0.50 in.
were weighed and measured as received
from the shop, heated to a bright red
color, reweighed and remeasured, and
then placed in a test assembly con-
taining eutectic NaK heated to 1400°F.
The BeO was fastened to the end of a
rotating shaft and rotated at a linear
velocity of 0.15 ft/sec. 1In the tests
to date, there has been considerable
BeO weight loss upon exposure for 100
hr or longer at 1400°F., The test

results are summarized in Table 11.12.

Since earlier static tests showed
no instability of the BeO, additional
static and dynamic tests are being
carried out, along with an 1investi-
gation of cleaning, drying, and similar
techniques for obtaining accurate
weighing of the BeO before and after
exposure to NakK.

Sodium in Forced-Convection Loops.
Metallurgical analyses of the figure 8
loop in which sodium was circulated
have been completed. Samples from the
hot, cold, and heat exchanger sections
of the loop were examined. Corrosion
was negligible, as shown in Table 11.13.

FUNDAMENTAL CORROSION RESEARCH

W, R. Grimes
Materials Chemistry Division

W. D. Manly
Metallurgy Division

The basic research needed for
determining corrosion mechanisms has

TABLE 11.12,

PERIOD ENDING DECEMBER 10, 1952

been continued. Examinations have
been made of corrosion products from
dynamic corrosion test by making use
of chemical and physical means to
determine identities of the products,
and studies are continuing on high-
temperature reactions of various
liquids with structural metals, thermal
stability of NiO and NiF,, and reactions
of molten fluorides under applaed
potentials, The preparation of various
complex interaction products of
structural metals with fluorides and
the characterization of these compounds

SUMMARY OF BeO-NaK COMPATIBILITY TESTS

 

 

 

 

 

 

 

 

 

 

 

 

 

RUN AND SAMPLE NO.
1(0) 2 3(5) 4_((:) 5 6(:1',6] 7(C,f) B(d,g) g(f,g)

Initial weight, g 1.5054 | 1.4196 1.4049 1.2287 1.2736 1.8437 25.3124 [ 29.9273
Initia]area,in.2 0.625 | 0.625 | 0.625 0.623 0. 609 0.455 0.625 4. 60 5.45
Initial volume,

in.? 0.0313 | 0.0313 {0.0312 | 0.0307 | 0.0310| 0.0293*)| 0.0378*) | 0.567 | 0.640
Initial density,

g/cm’ 2.93  |2.80 2.80 2.60 | 2.65 2.98 2.72 | 2.85
Approximate

weight of NaK, g| 90 82 82 82 82 82 82 75 75
Duration of

heating, hr 6.4 96 129 203 174 212 212

Final weight, g 1.3320 | 1.1047 1.3653 1.0896 0.9384 1. 4335
Weight loss .

In grams 0.1734 }0.3149 0.0396 0.1391| 0.3352 0.4102

In per cent 11.5 22.2 2.81 11.3 26.3 22.3

Final volume,

in.” 0.0306 |~0.0263|~0.028"> | 0.031¢®
Approximate

volume loss, % Negligible | 15 3¢h? 19(")
Weight loss per

initial unit

area, mg/in.2 280 500 60 230 740 670

(aLFemperature cycled from foom temperature to ()

1500+"F once and BOO to 1500 F five times; one “’Surface removed irregularly, leaving

cornesr chipped off during test.

(b)sample in two pieces at end of run; no other
pieces found,

(C}Hotated 14 hr; static 189 kr.

(d)Porous inner section of BeO block {for
density, see sec. 12, “Metallurgy and Ceramics®™ ).

“*ditches”™ and rounded cormers.

(f)Dense outer section of BeO block (for
density, see sec. 12, “Metallurgy and Ceramics'’).

g)Irreguiar semiwedge shafies.

(h)Calculated for irregularly abaped samples;
values uncertain,

151
ANP PROJECT QUARTERLY PROGRESS REPORT

TABLE 11.13.

ANALYSIS OF CORROSIVE ATTACK BY SODIUM IN FIGURE 8 LOOP

 

 

TEMPERATURE | TIME | VELOCITY ATTACK
(°F) (hr) (ft/sec)
Heat exchanger 1320 2700 Occasional depressions, to 1 mil deep
section 1300 1000 1
Hot section 1350 2700 2 Numerous small depressions, to 0.25 mil deep
1320 1000 1
Cold section 750 2700 No noticeable corrosion or erosion
1100 1000

 

 

 

 

 

were also continued 1in an effort to
aid in the i1dentification of the
corrosion products,

Interaction of Fluorides with
Structural Metals (H. S. Powers, J. D,
Redman, L. G. Overholser, Materials
Chemistry Division). Investigation of
the reaction of chromium and iron with
molten NaF-KF-LiF eutectic has con-
tinued. Special emphasis has been
placed on the behavior of K2NaCrF6 and
K,NaFelF, when contacted with the
eutectic in the temperature range from
600 to 1000°C, The filtrates col-
lected at various temperatures in the
range from 500 to 800°C have been
subjected to chemical analysis, as 1n
the past, and the residues and filtrates
have been examined by x-ray diffraction.

In a number of experiments, 3 wt %
of K,NaCrF_, was added to samples of
the NaF-KF-LiF eutectic. The mixtures
were heated at 800 to 1100°C under an
atmosphere of helium for several hours
and then filtered at various tempera-
tures in the range from 500 to 800°C,
and the residues and filtrates were
examined. Heating of the K,NaCrF, at
temperatures up to 1100°C prior to its
addition to the eutectic did not
affect 1ts solubility, and the temper-
ature of equilibration of this compound
with the eutectic prior to filtration
was unimportant.
ever,

Solubility was, how-
a function of temperature of
filtration. The filtrate collected
at 500°C contained 400 to 500 ppm Cr,

whereas the filtrate obtained at either

152

650 or 800°C contained 2000 to 3000
ppm Cr. Microscopic examination re-
vealed that most of the Cr was present
as K,NaCrF_., which indicates that this
compound 1s appreciably soluble 1in the
NaF-KF-LiF eutectic atelevated temper-
atures, X-ray examination shows that
in some runs, materials other than
K,NaCrF_ have been present, including
K,CrF,, Cr,0,, and KNaLiCrF,

Similar experiments in which Cr was
added to the esutectic as metal 1instead
of as KzNaCrF6 showed that Cr203, Cr,
and K,NaCrF_, were present after the
interaction., Additional products that
have not yet been i1dentified were also
found.

A study of the solubility of K NaFeF
in the NaF-KF-LiF eutectic and the
NaF-KF-LiF-UF, mixture at 600 and
800°C was made. It was found that with
2.5% of K,NaFeF  the solubility in
both fluoride mixtures and at both
temperatures of filtration was 6000 to
7000 ppm. Apparently, all the K,NaFeF,
added was dissolved.

Since the presence of oxidation
products 1in a number of the filtrates
indicated that the digestion and
filtration apparatus in which a blanket
of helium is used does not exclude
oxygen completely, two new types of
reaction-filtration equipment have
been designed and fabricated that
should eliminate oxidative effects.

Air Oxidation of Fuel Mixtures
(R, P. Metcalf, F., F, Blankenship,
Materials Chemistry Division). The
investigation of the oxidation of fuel
mixtures contained in nickel has been
complicated by the corrosion of the
nickel containers. The corrosion has
been studied in a series of experi-
ments.,

In each experiment,
fluoride, in a nickel boat, was placed
in a “combustion tube’’ consisting of a
horizontal 1 1/2-in.-0D stainless
steel tube heated by a tube furnace.
The inner surface of the steel tube
was protected by a sleeve of 5-mil
nickel sheet. The tube was evacuated
and heated to the desired temperature;
dry air was then admitted and passed
through at a rate of about 40 cm?/min
for about 4 hours. The solid products
were examined by x-ray diffraction.
No gaseous products were observed,

In summary, nickel 1s oxidized
severely by dry air when 1in contact
with fused fluorides. The oxidation
1s accompanied by creeping of the melt
over the entire heated surface of the
nickel container. Uranium, when
present in the fluoride mixture, reacts
to form a yellow product that probably
contains oxygen but has not yet been
completely 1dentified.

It is suggested that the oxidation
occurs because the fused fluorides
dissolve the adherent protective scale
of nickel oxide, which forms
absence of the melt.

The formation of the greenish- gray
nickel oxide observed in several of
the experiments appears to be associated
with the presence of moisture in the

salt employed.
EMF Measurements in Fosed Fluorides

(L. E. Tepol, L. G, Overholser,
Materials ChemistryDivision). Previous
sbudies,(g)involving the determination
of the decomposition potential of NiF,
in molten KF, suggested that a thermal
decomposition of NiF, occurred during
the experiments and led to a study of
the decomposition of NiFZ, N10, and
K,NiF, at temperatures from 700 to

(9)L.'E. Topol and L. G. Overholser,
Prog. BRep. Sept. 10, 1952,

5 to 10 g of

in the

 

ANP Quar.
ORNL- 1375, p. 123.

PERIOD ENDING DECEMBER 10, 1952

900°C in a purified helium atmosphere.
The study has been continued and
extended to include the possibility of
decomposition under vacuum. An attempt
has been made to measure the decompo-
sition potential of KF on Ni at 8%90°C
in a nickel container.

Repeated attempts 'to prevent the
decomposition of NiQ and NiF, at 750
to 900°C under a pur1fled helium
atmosphere were unsuccessful, even
when all the rubber tubing was re-
placed with glass or copper tubing.
In some cases, the weight losses were
apparently greater than could be
explained by complete decomposition of
the NiF, or NiO. An attempt is being
made to measure the dissociation of
NiF, and Ni0O in vacuum at various
temperatures. A vacuum system has
been constructed, and a preliminary
experiment showed that no apparent
decomposition of the Ni0 occurred, even
though it had been heated to 1020°C,
This absence of dissociation 1is 1in
agreement with the findings of Johnson
and Marshall,('%) who measured the
vapor pressure of NiO at temperatures
from 1440 to 1566°K. Studies of the
behavior of N10 1in vacuum
tinuing.

are con-

Decomposition potentials have been
determined in KF at 890°C with nickel
electrodes in various vessels. Measure-
ments were made in Norton stabilized-
zirconia crucibles (TZ-5601) and in
Norton thoria crucibles (LT-7010),
although these materials are somewhat
permeable to the KF, Subsequent runs
were made with a nickel cup as both
container and cathode and a 1/16-in.-
dia nickel rod as the anode. 1In every
case except one, a decomposition
potential was observed at 0.7 to 0.9
volt. These values agree with earlier
values®®? and with those determined by
Neumann and Richter(!!) on graphite.
In one run in which higher potentials

(10)H. L. Johnson and A. L. Marshall,
Chen. Soc. 62, 1382 (1940),

(Il)B. Néumann and H., Richter, Z. Elektrochen,.
31, 296 (1925).

 

J. Anm,

153
ANP PROJECT QUARTERLY PROGRESS REPORT

were applied, a change in slope was
noticed at voltages above 2.5, in
addition to the break in the current-
voltage curve at 0.8 volt: but, un-
fortunately, not enough data were taken
at these higher potentials to obtain
an accurate determination of the
decomposition voltage. In the last
experiment, no break in the current-
voltage curve was found at 0.8 volt,
but a definite break occurred at 2.6
volts., In all cases, the anode was
markedly attacked and was surrounded
by a black sludge that was somewhat
magnetic. Possible explanations of
these results are being studied.
Preparation of Trivalent Nickel
Compounds in the Hydroxides (L. D,
Dyer, G. P. Smith, Metallurgy Division).
Lithium nickelate(III) and sodium
rnickelate(III) have been shown to
occur (see previous guarterly reports)
in the corresponding hydroxides under
highly oxidizing conditions by actually

154

preparing large amounts of these com-
pounds in hydroxides, Therefore the
preparation of a presumed potassium
nickelate(1II) was undertaken to show
1ts occurrence under certain circum-
stances in potassium hydroxide,

By using molten potassium peroxide
as the oxidizing agent on nickel,
enough higher valent nickel crystals
were prepared to get a single crystal
for x-ray study. Because of the
apparent instability of the supposed
potassium nickelate, the nickel in the
compound either lost some oxidizing
power when removed from the KOH and
K204 with alcohol, or 1t never was
completely oxidized in the tube. The
separated compound contained only 20%
Ni**", whereas the theoretical value
for KNIO, is about 45% Ni'**,

The experiments are being continued
to find the conditions for obtaining
more and larger crystals and to refine
the separation method.
12.
~W. D. Manly .

PERIOD ENDING DECEMBER 10, 1952

METALLURGY AND CERAMICS

J. M. Warde

Metallurgy Division

Additional studies on the cone-arc
welding process are under way in an
effort to obtain more complete under-
standing of the basic principles of
the process. Resistance welding of
solid fuel elements has been tried,
with some degree of success. After
the proper electrode force was de-
termined, a convenient welding time
was selected, and welds made over a
considerable current range have been
studied. An automatic heliarc welder
has been set up for making longi-
tudinal welds on stainless steels
under very carefully controlled con-
ditions sc that the optimum welding
conditions for the various structural
metals can be determined. The high-
temper ature brazing alloy development
program is continuing, and results of
various corrosion tests in liquid
metals, fused fluoride salts, and
hydroxides, as well as the results
from the various high-temperature
tensile tests, are presented.

A very marked effect of the en-
vironment on the load-carrying abili-
ties of Inconel has been observed in
tests run at 815°C in atmospheres of
air, argon, and hydrogen. The same
effect has been seen to a lesser
extent 1n creep tests of type 316
stainless steel. Preliminary results
obtained at 815°C indicate that the
fluorides affect the load-carrying
~abilities appreciably at the lower
- stress values. _

An attempt is being made to produce
spherical particles of uranium-bearing
alloys approximately 0.010 1in. in
~diameter. These particles will re-
ceive a protective coating of nickel
plating, approximately 0.005 in.,
prior to being used in a pebble-bed
~type of reactor. The methods being
investigated are the following:

1. suspension in refractory powder
and heating above the solidus, '

2. momentary melting by passing
through a high-temperature arc,

3. spraying from a metallizing gun.
Solid-phase bonding studies are being
initiated to determine which materials
are easy to bond and which materials
are difficult to bond. After this
screening operation, the work will be
expanded to include a study of ‘the
production of solid fuel elements by
a hot-pressing operation in which
electroformed screens will be filled
with a mixture of U0, and stainless
steel powder prior to the bonding
step. Studies on the reaction of
columbium with various gaseous atmos-
pheres have been initiated in an
attempt to understand the effect of
these atmospheres so that the me-
chanical properties of columbium and
its alloys can be studied.

Disks in which a layer of U0, 1s
incorporated have been successfully
prepared from Cr-Al,0; cermet compo-
sitions., A ZrC-Fe cermet has been
developed that has possible interest
as a material for pump parts, seals,
and bearing materials. Work 1s being
carried out on SiC-C cermets. A
satisfactory dipping technique has
been developed for the application of
an oxidation-resistant ceramic coating
to nickel parts for use in an aircraft
type of radiator. A ceramic coating
containing 10% boron for possible
application in reactor shielding has
been successfully flame-sprayed on
20-gage stainless steel. |

FABRICATION OF SOLID FUEL IN SPHERES
E. S. Bomar J. H. Coobs

H. Inouve
Metallurgy Division

A geometry for the fuel in the
Supercritical Water Reactor, suggested
by the Pratt & Whitney Alternate
Design group, 1s that of spherical

155
ANP PROJECT QUARTERLY PROGRESS REPORT

particles 0.010 in. in diameter.
After receiving a 0.005-in. protective
nickel plate, the particles would be
packed 1n beds 1n the reactor core,
and cooling water would flow through
the interparticle channels.

Several methods for preparing the

spheres, with the use of both high-
and low-melting-point alloys, were
given a preliminary examination. The

high-melting-point alloy contained
5% uranium in molybdenum, and the
low-melting-point alloy contained 5%
uranium i1n nickel or copper. The
results obtainedbyusing the following
methods of preparation were only
partly successful: (2) suspension in
refractory powder and heating above
the solidus, (2) momentary melting by
passing through a high-temperature
arc, and (3) spraying from a metal-
lizing gun.

Suspension in Refractory Powder,
The U-N1 and U-Cu alloys were swaged
and drawn to 0.010 in. in diameter and
then cut into short segments. The
U-M particles were mixed with Al O,
and heated to 1450°C; one sample was
heated in hydrogen and another in a
vacuum of 3 microns. The U-Cu parti-
cles were spread on a refractory plate
and heated to 1100°C for 5 min in
hydrogen. A surface filwm formed in
all three cases and prevented spheroid-
1zation, even though a liquid phase
was present.

Momentary Melting in a High-
Temperature Arc, A rig for housing
two water-cooled electrodes and
supplying an argon atmosphere was
assembled. Both carbon and tungsten
electrodes have been used. The U-Na
alloy particles, similar to those used
in the first method, were dropped
through an struck between the
electrodes. Alignment proved to be
very critical, but the particles that
passed through the arc zone did melt.
The yi1eld was quite low, however.
Particles that passed through the
carbon arc were porous, whereas those

arc

156

that passed through the tungsten arc
were not.

Similar U-Ni alloy particles were
also dropped through an atomic-hydrogen
arc, and a small number of essentially
spherical, but porous, particles were
formed.

Spraying from a Metallizing Gun,
A length of 0.057-in. -dia U-N1 alloy
wire was fed through a metallizing
and the resulting particles were
collected in water. As with the other
methods, a porous product resulted.
Substitution of argon for the normal
air blast produced what appeared to
be a sound product, with only a
superficial surface film.

During the process of droppaing
segments of wire through the tungsten
arc, 1t was noted that small particles
were ejected from the electrode at
current densities high enough to melt
the electrode tip. The bulk of the
particles ranged in size from 0.010
to 0.020 inch. These results were
duplicated when U-Mo electrodes were
substituted for the tungsten elec-
trodes. The U-Mo electrodes were
prepared by pressing a mixture of the
elemental powders and sintering at
1200°C for 5 hr in a vacuum of 0.5 to
1 micron. The spheroidized particles
did not give a response on exposure
to an alpha counter as did the parent
electrodes and U-Ni particles from an
earlier spraying experiment.

gun,

SOLID PHASE BONDING OF METALS

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

The 1nitial tests on the solid-
phase bonding of metals, performed
essentially as a screening operation,
have been completed. These tests were
undertaken to determine which materials
are easy to bond and which are diffi-
cult to bond. The tests were run on
both sheet stock and sintered-powder
compacts and were performed by stacking
wafers 1n a molybdenum-lined graphite
die. The laminates were hot pressed
for 5 min at 3000 psi at 1150, 1200,
and 1250°C in an argon atmosphere.
Further tests have been run at 1150°C
for 1, 10, and 30 min to determine
the effect of time on the extent of
"bonding. Evaluation of these tests
i1s not complete.

Examination of the first group of
samples revealed little difference in
the degree of bonding at the various
.~ temperatures. As was expected, con-
siderably more diffusien occurred at
- 1250°C than at the lower temperatures,
but bonding was only slightly better.

In general, the variocus couples may
be divided into four groups on the
basis of the extent of bonding: ex-
cellent, interface almost obliterated,
~inseparable; good, interface well
defined, inseparable; fatir, separated
with some difficultyj poor, only
superficial bonding. The couples
studied fall into these categories in
the following manner: '

Excellent

Nickel to Inconel

Nickel to stainless steel

Stainless steel to stainless steel
Iron to iron '

Good

Nickel to iron

Nickel to nickel

Mol ybdenum to nickel foilto molybdenum
Fair . '

Molybdenum to nickel

Meclybdenum to Inconel

Molybdenum to molybdenum
Poor ~

Mol ybdenum to iron

Mol ybdenum to stainless steel _

The self-welding of the molybdenum,
with or without a nickel foil bonding
layer, was poor at 1250°C; so supple-
mentary runs were made at 1400 and
1500°C, with results as indicated
above.

For molybdenum to nickel and mo-
lybdenum to Incomel, there was a
measurable amount of diffusion but
only fair bonding, possibly because
of the formation of a third phase of

PERIOD ENDING DECEMBER 10, 1952

poor strength at the interface. Ex-
tensive diffusion (2 to 3 mils)
occurred between molybdenum and iron
and molybdenum and stainless steel,
but there was even less bonding than
was found between mol ybdenum and
nickel or between molybdenum and
Inconel. - fi

Further work in which electro-
deposited chromium is used as a bonding
layer is being done on the bonding of
molybdenum to the other metals.
Chromium is reported to be completely
miscible with molybdenum and may be
more compatible with the other ma-
terials.

Several small fuel assemblies have
been prepared by using the electro-
formed nickel screen, filling with
mixtures of UD, and stainless steel
powder, and pressing rubberstatically.
This step is followed by cladding both
sides with iron or nickel sheet and
by hot pressing at 1250°C. The as-
semblies are now being examined for
fuel distribution and bond integrity.

" COLUMBIUM RESEARCH

E. S. Bomar H. Inouye
Metallurgy Division

Gaseous Reactions. TInvestigation
of the reaction of gases with columbi-
um was initiated after a literature
survey was made that revealed the
high reactivity of the metal above
400°C and some preliminary high tem-
perature tests made that gave erratic
results., Minute quantities of dis-
solved gases change the mechanical
properties considerably, and hence
these reactions must be controlled.

The primary interest during this
period was in finding a suitable
atmosphere i1n which to conduct creep
tests, An argon atmosphere was
selected, initially, and some of the
observations are presented in Table
12.1. Commercial, tank argon was
purified by drying with a tower of
magnesium perchlorate, followed by

157
ANP PROJECT QUARTERLY PROGRESS REPORT

two towers of titanium sponge at 850°C,
and finally one tower of titanium
sponge at 200°C,

Oxidation in Air. It has been
reported that columbium oxides evapo-
rate between 1200 and 1500°C. Pre-
liminary tests of the stability of
the oxides show that there is a
negligible loss through volatilization
up to 1375°C (see Table 12.2). This
temperature being above that of the
intended investigation, it was decided
to measure the course of oxidation
directly with time by suspending the
sample in the furnace from one arm of
an analytical balance.

The oxidation rates of columbium 1in
dry air have been measured at 600,

800, 1000, and 1200°C. A linear
oxidation rate that becomes more
rapid with increase in temperature 1is
in Table 12.3. The oxidation
tests were made in air flowing at
2 liters/min that was dried by anhy-
drone and a cold trap of acetone and
dry ice mixture prior to entering the
furnace.

At 400°C, columbium undergoes short
periods of rapid oxidation followed
by longer periods of relatively slower
oxidation up to about 20 hours. Beyond
this time, the oxidation rate becomes
linear; that 1s, the weight gain is
proportional to the time of exposure,
The oxide formed on the sample at the
endof 3000 min 1s reasonably adherent.

shown

 

 

 

 

 

 

 

 

 

 

 

TABLE 12.1. COLUMBIUM REACTIONS IN PURIFIED ARGON
TEMPERATURE TIME OF WEIGHT HARDNESS
( °CA TEST INCREASE | INCREASE REMARKS
) (hr) (g/cn?) (VPN)

600 291 0.0005 48 Film of CbO, 4, = 4.21, ductile
800 67 0.0015 106 Black film, not identified
1000 88 0.0026 416 Black film, brittle
1000 89 0.0006 97 Static argon, 5-in. mMercury pressure
1000 72 0.0002 85 Capsulated sample in static argon

TABLE 12.2. STABILITY OF Ch,0, IN AIR (18 hr)
TEMPERATURE WEIGHT CHANGE
o REMARKS
(°cC) (g)
1600 +0.0001 Starting material had a few black
specks of a lower oxide

1100 +0.0001 Oxide completely white

1200 0 Oxide changed to a yellow color

1250 0 Oxide yellow

1300 —0.006 Oxide yellow

1350 =-0.0009 Oxide yellow

1375 -0.0009 Oxide yellow

 

 

 

158
TABLE 12.3. COMPARISON OF GXIDATION

RATES AT END OF 90 MINUTES

 

 

TEMPE RATURE

 

WEIGHT GAIN
("C) (g/cm?)
400 0.00003
600 0.0088
800 0.0465
1000 0.0650
1200 0.0915

 

 

At 600°C and above, the oxide forma-
tion is fairly thick, and some diffi-
culty has been experienced in obtain}
ing a picture because of the spalling
of the oxide within a few seconds
after removal {rom the furnace.
Attempts to preserve the oxide 1in
place have been made by pouring solder
or ““castalite’” plastic around the
specimen. '

CREEP RUPTURE TESTS OF
STRUCTURAL METALS

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

Inconel in Air. Inconel specimens
have been tested 1n an air atmosphere
at stresses ranging from 3000 to 4500
psi at 815°C., The results indicate an
increase in rupture life by a factor
of 10 as compared to tests run in
argon in this stress range. It 1is
thought that subsurface oxygen and
possibly nitrogen may be the elements
influencing the creep rate and pro-
longing the rupture life. Tests are
being designed to investigate the
relative effectsof oxygen and nitrogen
on the creep-rupture properties., It
is postulated at this time that the
creep properties of Inconel in en-
vironments other than air may be
improved by a pretreatment with oxygen
or nitrogen. Experiments will be
conducted to explore this possibility,

PERIOD ENDING DECEMBER 10, 1952

Inconel in Hydrogen. Fine-grained
Inconel specimens have been tested
in a purified dry hydrogen atmosphere
at stresses ranging from 2500 to 7000
psi at 815°C. The results of these
tests as compared with those from
tests made in an argon atmosphere show
a marked decrease in rupture life,
ranging from a 75% decrease at the
lowest stress to a 55% decrease at the
highest stress tested. A similar
series 0f tests 1s being made on
coarse~grained specimens.

Inconel in Molten Fluorides. Tests
of Inconel in the fluoride mixture
Na¥F-ZrF,-UF, {(46-50-4 mole %) are
being made at 815°C at stress levels
from 2500 to 7500 psi. Besults to
date indicate a marked decrease 1in
rupture life that 1s similar in magni-
tude to the rupture life obtained in
tests made 1n a hydrogen atmosphere.,
Figure 12.1 1s a summary of the
available data for Inconel tested at
815°C (1500°F) in the several environ-
ments. {he tentative data for tests
in this fluoride and hydrogen are
based on two and three tests,
spectively,

Type 316 Stainless Steel in Argon,
Specimens of type 316 stainless steel
have been tested at 815°C in argon
at stresses from 5300 to 7300 psi.
A comparison of these data with those
for Inconel tested in argon shows the
type 316 stainless steel to have a
longer rupture life than that of
Inconel by a factor of 10 at the 5000
psi stress level and by a factor of
3 at the 7000 psi stress level. The
data from creep-rupture tests of
type 316 stainless steel tested 1in
air at the Cornell Aeronautical Lab-
oratory, together with a few tests 1in
an air atmosphere at ORNL, do not
indicate any significant difference
between the rupture life of type 316
stalnless steel specimens tested in
argon and the rupture life of those

tested in air.
Type 316 Stainless Steel in Molten

Fluorides. Tests of type 316 stain-

re -

159
091

UNCLASSIFIED
OWG. 17409

!

 

 

 

|
|
|

 

 

 

 

 

|

ot el

o N P
! |~ inco-rounD BAR IN AR !
| N i - L] | 1
| N\ | TN i
C : VNG IS TN e

 

 

 

 

 

 

 

 

 

 

 

|
|
|
—

 

 

 

 

 

 

 

 

 

 

ORN_-0.065-1n, SHEET _L-INCO- 0.0483-in

 

 

NOF-Z!F4“ UF, SHEST IN AIR

!
I
|
|
| |
5 | |
: | ORNL-0.065- in. SHEET i AN | r |
i o N HYDROGEN —— ! oo VN j ‘ | ]
| : e NN NS L

+ORNL- 0.083~in. SHEET IN |
| ARGON !
(WIDE RANGE OF GRAIN
SIZES) D

1
1

i
|
|
|

   
 

 

 

 

 

  
  

 

 

 

 

 

 

STRESS (psi x1000)

 

 

] | -

 

 

3
'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

] : |
B o é | RS
| I R
- I | " SRR
‘ a | | | L | o
| N | | o
0

 

1 10 100 1,000 10,000
TIME  (nn)

Fig. 12.1. Summary of Available Rupture Data for Inconel Tested at 815°C (15008°F) in Air, Argon,
Hydrogen and the fFused Fluoride Mixture NafF-Zrg,-UF, (46-50-4 mole %).

LHOAHY SSHYO0¥d ATHALYVNO 12Af0Md dNV
less steel are being made 1in the
fluoride mixture NaF-KF-LiF-UF,
(10.9~43.5-44.5~-1.1 mole %) at 815°C
at stresses ranging from 5000 to 7500
psi. Preliminary results indicate a
marked decrease in the rupture I1fe.
It will be necessary to complete more
tests before a definite evaluation
can be made of the effect of the
fluorides on the creep properties of
type 316 stainless steel. An anoma-
~ lous behavior has been noted in the
stress range from 6800 to 7300 psi -
the rupture life increases as the
stress is increased. This effect has
also been seen 1n argon tests, al-
though to a lesser degree. Detailed
metallographic studies, together with
x-ray diffraction and vacuum fusion
analyses, will probably be necessary
to determine the structural changes
that promote this behavior.

BRAZING AND WELDING RESEARCH

G. M. Slaughter V. G. Lane
C. E. Schubert
Metallurgy Division

ARE Welding. A manual entitled
‘““Joint Design for Inert -Arc Welded
Vessels’ was prepared as an aid in
construction of the ARE. This manual
describes the preferable basic joint
design for use i1n pressure vessels
that are to be operated at elevated
temperatures in a severely corrosive
environment.,

Cone-Arec Welding. A topical,
comprehensive report of cone-arc
welding research conducted at this
laboratory is being written for publi-
cation in the Welding Journal Research
Supplement. This report will discuss
the basic principles of the cone-arc
process, the actual physical setup of
the apparatus, and the extent of the
research on the welding of thin- walled
tube-to-header joints.

During the research on the cone-arc
welding of the 0.100-in.-0OD stainless
steel tubes with 0.010-in. wall thick-
ness to the 0.125-in. stainless steel

PERIOD ENDING DECEMBER 160, 1952

headers, it was noticed that many
welds exhibited a two-pass appearance,
In order that the mechanism of the
cone-arc welding process could be
better understood, an investigation
was conducted to study this phenomenon.
Figure 12.2a¢ shows a microstructure in
a cone-arc weld that resembles, ‘in
many ways, that resulting from a
multiple-pass butt-weld. There were,
apparently, two stages of melting,
with growth of dendrites across the
interface of the two regions.

This two-pass effect apparently
results from the very rapid melting
of the tube during the initial phases
of the welding process. With the tube
flush against the top header surface,
the intense heat of the welding arc
is at first centered upon the thin-
walled tube. The tube apparently
undergoes melting along 1ts length
for a short distance until the thermal-
insulating capacity of the very fine
tube-to-header spacing is overcome.
After this occurs, much of the heat
is then used in melting the header
material. At this time, conditions
are reached that approach more nearly
the equilibrium state, The lower pass
is then that resulting from the initial
phase of the welding, and the upper
pass, represented by the molten weld
pool, remains throughout the major
portion of the welding cycle. '

Since it would seem that the arc
would play more directly upon the
header sheet at the initial point of
striking i1f the tube were recessed 1in
the hole, this technique was employed
in an attempt to eliminate the two-
pass weld., Figure 12.2b 1s a photo-
micrograph of a cone-arc weld in
which the tube was recessed 0.022 inch,
Complete welding occurred, and the
two-pass effect was eliminated. It
should be remembered that the two-pass
weld is not considered as detrimental
but is merely a characteristic of
joints made under certain conditions,

Relatively long arc distances also
reduce the tendency for this type of

161
ANP PROJECT QUARTERLY PROGRESS REPORT

" UNCLASSIFIEDS
v-7450 K

Fig. 12.2.
plate. (b) Tube recessed 0.022 inch.
joint, As the arc distance increases,

more of the arc beam impinges upon the
header sheet, and the drastic heating
of the tube wall 1s minimized.
Resistance Weldimg. Since 1t 1is
desirable to determine the feasibility
of electrical -resistance spot welding
of stainless steel clad fuel elements
to both themselves and to stainless
steel sheet, an 1nvestigation 1is
being made in which the welding vari-
ables are precisely controlled. The
goal of the preliminary experiments
in this program was merely the produc-
tion of sound, high-strength welds
under conditions that do not materially
harm the core structure of the fuel
elements, Welds were made at several
values of electrode force to obtain
the optimum value, which was considered
to be the minimum force that would
consistently result in crack-free,
nonporous, weld nuggets. DBy using
this optimum electrode force and a
reasonable weld time, experimental
welds were then made over a wide

162

 

Two-Pass Characteristces of Cone-Arc Welds,
Etched with aqua regia.

UNCLASSIFIED
Y5355

 

(a) Tube flush with
194X,

range of welding current. Examination
of metallographic sections of these
welds will enable the determination
of the permissible range of current
over which good welds can be made and
also allow the selection of the most
satisfactory value of welding current.
Any major defects resulting from
welding can also be investigated by
metal lographic sectioning.

A photomicrograph of a sound, high-
strength, spot weld made in 0.030-1in,.
fuel elements containing cores of 50%
U0, and 50% Fe is shown in Fig. 12.3a.
The electrode force used was 800 1b,
the weld time was 5 cycles, and the
welding current was 6300 amp. The
dark etching areas in the cores are
thought to result from the partial
melting of the 1ron portions of the
core metal. The lower limit of the
permissible current range is exhibited
in Fig. 12.3b. The welding conditions
in this case are identical to those
used in Fig. 12.3a except that the
welding current was reduced to 5800
(o)

PERIOD ENDING DECEMBER 10, 1952

ELDED SURFAC

§ b e

 

Fig. 12.3. Spot Wejds of Stainless-steel-¢lad Fuel glements. {a) Wwelding

current §300 amp,
al penetration.

excellent penetration.
Etched with aqua regia.

(b) Wwelding current 5800 amp, margin-
50X.

163
ANP PROJECT QUARTERLY PROGRESS REPORT

amp. The sheet-to-sheet interface
line has been nearly removed; as a
result, a moderate strength “stick
weld” has been produced. It may be
noted that melting begins at the
stainless steel~core interface instead
of at the usual sheet-to-sheet inter-
face.

Tensile-shear tests on clad fuel
elements that were spot welded under
optimum conditions indicate that
relatively high shear strengths can
be obtained. Strengths of 575 1lb per
spot weld were recorded when a 3/16-in.
restricted-dome electrode was used.
Welds in standard stainless sheet of
the same type and thickness would be
somewhat stronger, but this would be
expected because the nugget size would
also undoubtedly be larger.

Avtomatic Heliarc Machine Welding,
A welding program has been originated
to study the operation and welding
characteristics of a G-E Fillerweld
Machine. This machine consists of a
standard inert-gas heliarc torch with
an auvtomatic filler-wire addition
mechanism. When coupled with amachine
welding carriage, this equipment
presents an excellent means for semi-
automatically welding long straight
joints or, when used in conjunction
with an appropriate cam, joints of
irregul ar and intricate shapes. In an
investigation of the type to be con-
ducted 1n this laboratory, 1t 1s
desirable not only to determine abso-
lute values of the welding variables
but also to study the effects of these
variables upon the gquality of the

welds produced. o .
Preliminary work on this investi-

gation first consisted of adapting
the basic equipment to conform with
the needs of the program. Experiments
have been conducted for the purpose of
gaining familiarity with the process
and for determining the ranges of
welding current and travel speed
necessary to produce satisfactory
welds with various rates of filler-
wire addition. All weld beads have

164

been deposited on flat stainless steel
sheet 0.090 in. 1n thickness, but
later the investigation will be
focused upon actual joints in sheet
stock of similar thicknesses,
Fabrication of Heat Exchanger
Units, 1Tt is necessary to develop
techniques for successfully brazing
large, complicated, heat exchanger
test assemblies of a type suitable for
aircraft use., An actual unit was
assembled in this laboratory and
subjected to the Nicrobrazing oper-
ation. Figure 12.4 is a photograph
of the heat exchanger brazed in dry
hydrogen by the canning techniqgue.
As can be seen from the photograph,
the fins are badly warped and distorted
and 1in some cases, even brazed to-
gether. After the leak test with
helium, 1t was found that there were
leaks in the tube-to-fin matrix as

UNCLASSIFIED
01

 

Fig. 12.4. Liguid-te-Air gadiator
Nicrobrazed by Canning Techunigue,
well as in the tube-to-header joints
of the manifold section. The tube-to-
header joints could be repaired by
rebrazing, but the leaks in the body
of the unit are almost impossible to
repair successfully. '

Examination of the completed unit
revealed several probable causes for
the leaks and excessive distortion,
This examination has led to techniques
for overcoming these flaws in future
brazing operations. The tube-to-fin
leaks were thought to be caused by
excessive use of Nicrobraz alloy,
with the resultant dissolving away of
the tube walls.
al so tended to aggravate this situ-
ation by allowing puddling of the
molten brazing alloy. The formation
of “hot spots’ in the body of the
assembly was also thought to be a
prevalent condition. ‘A very rapid
heating rate was also an important
factor in the warpage. ’

Techniques that are planned for
incorporation in further brazing
operations are the use of lesser
quantities of Nicrobraz, the use of
several aspilrators to promote more
even hydrogen flow between the fins,
the build-up of the whole assembly off
the can bottom to overcome the drastic
inittial heating rate, and the use of
a slightly lower brazing temperature.

8razing of Copper to Inconmel. The
need for a satisfactory brazing alloy
for joining copper fins to Inconel
tubing has been emphasized. The
resultant braze should have relatively
high strength at 1500°F and should
also be a diffusion barrier against
copper penetration into the Inconel
during service. It seems likely that
some alloy other than a copper-base
alloy would best fill these require-
ments. Nicrobrazing at 1900°F is very
promising, but the alloy does not
completely melt at this temperature.
Tt 1s hoped, however, that the use of
other alloys very similar in compo-
sition to Nicrobraz but of somewhat

The bad distortion’

PERIOD ENDING DECEMBER 16, 1952

lower melting point will result in a
satisfactory brazed joint.

In view of the possibility that it
might be advisable to use copper-base
alloy for the production of these
tube-to-fin joints, several such
alloys are being investigated. Small
melts of Cu-Be, Cu-5S1, Cu-Sn, and
Cu-Mn alloys have been prepared and
are to be tested for flowability
characteristics, oxidation resistance,
and room- and elevated-temperature
tensile strengths. ' :

EVALUATION TESTS OF BRAZING ALLOYS

G. M. Slaughter V. G. Lane
C. E, Schubert
Metallurgy Division

The results of recent high -tempera-
ture oxidation tests and static cor -
rosion tests in sodium hydroxide,
sodium, and lithium are summarized
in Table 12.4. The high ~temperature
oxidation tests were conducted in
a stream of moist air at 1500°F on T
specimens of Inconel and type 316
stainless steel brazed with the
various alloys. The samples tested
in molten sodium hydroxide at 13500°F
were A nickel T joints brazed with
these alleys. Brazed Inconel and
type 316 stainless steel T joints were
used as specimens for the corrosion
tests in sodium and lithium, The
tensile strength nf several brazed
joints has been examined at room and
1500°F temperatures.

Oxidation of Brazing Alloys. Ex-
amination of metallographic sections
of the high-temperature oxidation
specimens showed the 60% Pd-40% Ni
alloy to be the most oxidation re -
sistant of the brazing alloys tested.
This alloy was unattacked when in
contact with the base metals tested;
the Inconel joint is shown in Fig.
12.5. An example of severe oxidation
is shown in Fig. 12.6, which is the
photomicrograph of a type 316 stainless
steel joint brazed with the 64% Ag-33%

165
991

TABLE 12, 4.

RESULTS OF 104 hr CORROSION

TESTS OF BRAZING ALLOYS AT 1500°F

 

 

BRAZING ALLOY

HIGH- TEMPERATURE
OXIDATION ON
BRAZED INCONEL

HIGH- TEMPERATURE
OXIDATION ON
BRAZED TYPE 316

STATIC CORROSION
ON BRAZED
A NICKEL IN NaOH

STATIC CORROSION

BRAZED INCONEL

STATIC CORROSION
ON
BRAZED TYPE 316
STAINLESS STEEL

 

 

STAINLESS STEEL
In Na In Li In Na In Li
Nicrobraz Slight None Moderate | Slight Severe
60% Mn-40% Ni Severe Moderate Moderate Moderate Severe
60% Pd=-40% Ni None None None Severe Severe Moderate Severe
16. 5% Cr-10.0% Si- Moderate None Fractured Moderate | Severe None* Slight
73.5% Ni during
examination
16, 5% Cr-10,0% S1= Moderate None Fractured Moderate Slight None* Severe
2.5% Mn=71.0% Ni during
examination
75% Ag=-20% Pd-
5% Mn Severe Moderate Severe Sevare Severe Severe Severe
64% Ag~-33% Pd-
3% Mn Slight Severe Moderate Severe Severe Severe Severe

 

 

 

 

 

 

 

 

*Severe cracking present in joint.

LUHOdIH SSHUO0¥d ATHILYVIO L13A704d dNV
PERIOD ENDING DECEMBER 10, 1952

E UNCLASSIFIED

Y7474

 

Fig. 12.5. Inconel Joint Brazed with 0% l’_d~4fl% Ni Alloy After Oxidation
Test for 100 hr at 1500°F. Etched with KCN(NH,),S,0,. 200X.

E UNCLASSIFIED
Y.7AB2

 

Fig. 12.6. Type 316 Stainless Steel Joint Brazed with a §0% Ag-33% Pd-3%
Mn Alloy After Oxidation Test for 100 hr at 1500°F. Etched with agua regia.
200X, | -

167
ANP PROJECT QUARTERLY PROGRESS REPORT

Pd-3% Mn alloy. As a result of this
investigation, it was found that a
brazing alloy in contact with one base
metal may be severely oxidized, whereas
it may be relatively unattacked when
in contact with another base metal,
This might be explained by the fact
that the least oxidation -resistant
constituents of the brazing alloy may
diffuse more readily into certain base
metals than inteo others,

Corrosion of Brazing Alloys by
Sodium Hydroxide. Since A nickel 1is
relatively unattacked by molten sodium
hydroxide at 1500°F, it serves as a
suitable base metal for testing the
corrosion resistance of brazing alloys
in this medium, Examination of a
joint brazed with 60% Pd-40% Ni alloy
and tested in NaOH for 100 hr showed

e
\'.A - ““\ *

B

Fig' 12@ 70

 

that no attack is present, Several
other alloys were attacked, and the
joints brazed with the Cr-Ni-Si-Mn and
the Cr-Ni-Si alloys fractured during
removal from the test capsule.
Corrosion of Brazing Alloys by
Sodium and Lithium. Because brazing
operations may be proposed for as-
sembling fuel plates into subas -
semblies for certailn types of reactors,
it was of interest to study the static
corrosion resistance of brazed joints
in contact with sodium or lithium,
The 60% Pd-40% Ni alloy, which ex-
hibited excellent corrosion resistance
to all other media tested, was severely
attacked by the liquid metals. The
extreme attack of sodium on brazed
Inconel 1is illustrated in Fig. 12.7,
which is the photomicrograph of a

UNCLASS!FEEJ

Sodium.

168

Etched with aqua regia.

Inconel Brazed with 60% Pd-40% Ni After
150 X.

100 hr at 15300°F in
sample tested for 100 hr at 1500°F,
As a check on these results, small
specimens of the pure braze metal
were immersed in the metal baths,
The sample tested in sodium had a
weight loss of over 50%, and the
sample tested in lithium completely
dissolved.

Figure 12.8 is a photomicrograph
of a type 316 stainless steel brazed
with the Cr-Si-Ni-Mn alloy and tested
in sodium for 100 hr at 1500°F. No
attack is exhibited, but the cracking
tendencies of this brazing alloy are
shown very clearly. It was noted that
the cracking of the Cr-Si-Ni-Mn and
Cr-Ni-Si alloys was more severe when
alloyed with stainless steel than when

- alloyed with Inconel.

Tensile Strength of Brazed J01nts.
Results of the room-temperature ten-
sile strength tests of Inconel butt-
joints brazed with Nicrobraz alloy
seem to indicate that a short brazing
time of 10 min or less leads to the

 

Fig. 12.B8.

PERIOCD ENDING DECEMBER 16, 1952

production of weak joints. Samples
held at the brazing temperature for
20 min seem to produce stronger joints.
Figure 12.9 is a photomicrograph of
the room-temperature fracture of a
joint held at the brazing temperature
for 5 minutes. Some Nicrobraz can be
seen in this photograph, but no brittle
fracture is evident. Although the
specimen that was held for 20 min at
the brazing temperature broke near the
brazed joint, no evidence of Nicrobraz
alloy conld be found upon metallo-
graphic sectioning. The greater
alloying of the braze and the base
metal is a result of the increased
diffusion occurring at this longer
time. The increase in tensile strength
of the second bar probably results
from this alloying.

The average of four room-tempera -
ture tensile tests on joints brazed
at short times was 38,300 psi. A
joint brazed for 20 min at the brazing
temperature fractured at 81,700 psi.

 

 

 

Type 316 Stainless Steel Brazed with 16.5% Cr-10.0% Si-2.5%

Mn-71.0% Ni After 100 hr at 1500°F in Sodium. Etched with aqua regia. 150X.

1169
ANP PROJECT QUARTERLY PROGRESS REPORT

 
  

’ 2 ‘
Y
! oo A\ N
¢ j : ’ A ; - : } ‘
v - J {
¥ R < - ¥ ';‘ .
' e } "}_ #
< ! “ { s
X ’ : ' [ *
e ’ . . . # / }
/‘, { . - €
\\ . . .
e -.#) \ . ‘,
‘r ‘ K;r"{ j 7
! /e : ] V
fl’ ¢ /
‘\ P .. J LAl
- ! e N L
: : ; » § { '
R
3 N\ Y ’ ’ .

Fig. 12.9.

Brazing time, 5 min.

The average tensile strength of four
Inconel bars that were heat -treated
similarly was 87,300 psi. This seems
to indicate that if time is allowed
for sufficient diffusion to occur,
joint efficiencies of over 90% can be
obtained. However, a more thorough
investigation of the effect of brazing
time on room-temperature strengths is
being conducted at present,

The tensile strengths of Nicrobrazed
Inconel joints tested at 1500°F were
much more consistent, The average
tensile strength of joints brazed at
short times was 24,400 psi, whereas
the sample brazed for 20 min fractured
at 24,700 psi. Several test bars
failed in the parent metal at a point
definitely not in the diffusion zone
of the brazed joint. The average
elevated-temperature tensile strength
of four Inconel test bars that were
heat treated at the brazing temperature
for 5 min was 23,700 psi. These data
indicate that the Nicrobrazed Inconel
joints at 1500°F are often as strong

170

Etched with aqua regia.

| UNCLASSIFIED
y Y.7835

  

v i -
ey
. i . / ' k *
P L 3
L 3 - 1
. §
- § \’v {—.
2 : ¥ W
/ \ ' *
/‘i -’ » i A 1
P F, - Joe
VEET e |
: "\
{4 R - i
f < . L~ -~ , *
7 : 2 . \"/’ ’ 3
o i i
- f : i
W . ¥ *
v N v i e
’ ™ * ¢ -t
é‘ \ ! £ ot
i < # . \
i L g “ .
3 ‘ - . iy ..t
¢ ayy it
- j /
¥
! /
{ ' -
t h N
] Sy .

« . e -~

Room-Temperature Fracture of a Nicrobrazed Inconel Test Bar.

200X.

as the Inconel base metal, The dif-
fusion of the braze metal while the
specimen 1s at the testing temperature
undoubtedly has some effect upon the
Joint strength at this temperature,

Butt-brazed Inconel test bars
brazed with the 60% Pd-40% Ni alloy
were tested at room and elevated
temperatures. The average room-
temperature tensile strength of these
brazed joints was 87,600 psi, whereas
those tested at 1500°F averaged 21,000
psi. Check bars of Inconel heat
treated for 5 min at 2300°F were also
tested so that the strengths of samples
subjected to the brazing cycle could
be determined. The joint efficiency
at room temperature was found to be
100%, whereas at 1500°F the value
was 93%.

Extremely deleterious effects of
the high-temperature heat treatment
involved in Ni-Pd brazing were not
evident in short-time tensile tests
on Inconel. Inconel 0.252-1n.-dia
tensile bars heat treated for 5 min
at the Nicrobrazing temperature of
2100°F had an average room-temperature
strength of 87,300 psi and an average
1500°F strength of 23,200 psi. Test
bars heat treated for 5 min at the
Ni-Pd brazing temperatures of 2300°F
fractured at an average room-tempera-
ture strength of 84,500 psi and at an
average 1500°F scrength of 22,600 psi.
Thus, the larger grained Inconel has
a tensile strength at room temperature
of only 3.2% less than the finer
grained Inconel; at 1500°F the dif-
ference 1is only 2.6%. Since all
tensile bars 1n the experiments were
machined from the same Inconel rod,
there is indication that heat treat-
ment was the influencing factor,

Melting Point of 60% Pd-40% Ni
Alloy. Although the 60% Pd-40% Ni
brazing alloy has many very desirable
properties, 1its high melting point
might exclude 1ts use in many appli-
cations., In an attempt to lower 1its
melting point by alloying, beryllium
additions of from 0.25 to 4% were
made. A 1% beryllium alloy was found
to have a melting point slightly below
2200°F, but recent tests indicate that
its flowability properties are very
poor. In further attempts to obtain a
lower melting point alloy without
impairment of the corrosion resistance
and physical properties of the basic
alloy, additioms of Al will be made
to the Pd-Ni alloy.

CERAMICS RESEARCH
J. M. Warde, Metallurgy Division

Development of Cermets for Reactor
Components, The Cr-Al,0, cermet
fabrication studies have resulted imn
successful disks 1 in. in diameter and
20 mils thick. Several sandwich disks
were also made that contained a thain
layer of U0, mixed with Cr-Al,0;. A
die 1is now being made to permit
pressing of annular rings of a fuel-
bearing material. In-reactor tests
of these cermets are now being con-
templated. ' ‘

PERIOD ENDING DFCEMBER 10, 1952

A 7ZrC-Fe cermet was fabricated as
a possible material for pump parts,
seals, and bearing materials. This
cermet, along with Kennametal 151-a,
is being tested for corrosion resis-
tance in fluorides, Na, and Pb. These
materials appear to be very promising.
Corrosion data and photomicrographs
will be included in the next report.

Attempts to hot press SiC-81 mix-
tures have not yet been successful.
A vacuum induction furnace is nearly
built that will be used to study
sintering of this and other cermets.

Ceramic Coatings for an Aircraft
Type of Radiator. A satisfactory
dipping technique was developed for
the application of NBS ceramic coatlng
A-418 to provide oxidation resistance
at elevated temperatures to parts
fabricated from nickel for an aircraft
type of radiator. The first composi-
tion of this coating was described
previously.(?’ The radiator parts

-will be coated as soon as they are

available.

Ceramic Coatings for Shielding
Metals. Work was commenced on the
development of a coating contalning
a minimum of 10% by weight of boron
that could be flame-sprayed onto metal
surfaces for use in reactor shielding.
A lead-borate enamel was developed’
that could be successfully applied in
a 20-mil coating to No. 20 gage stain-
less steel by flame-spraying. The
batch composition of this enamel 1s
as follows:

B,0, 35. 0%
PHO 62.5%
ZnO 2. 5%

Further work 1s in progress in an
attempt to coat a mild steel plate
5 ft by 5 fv by 7/8 in. for use in a
shielding experiment. :
Uniformity of Beryllium Oxide
Blocks (.. M. Doney and J. M, Warde,
Metallurgy Division). A study of the
physical structure of the BeO blocks
(1)
1952,

T. N. McVay, ANP Quar. Prog. Rep.

June 10,
OBRNL-1227, p. 152. :

171
ANP PROJECT QUARTERLY PROGRESS REPORT

to be used in the ABE was made to aid
in the interpretation of BeO-NzakK
compatibility test data, These blocks,

fabricated by the Norton Co., were
made by hot-pressing BeO powder. They
are hexagonal blocks 2 1/8 in. on a

side, 3 11/16 in. in diameter across
center of faces, and 6 in. long. They
are fabricated with a center hole
running axially through the block;
the diameter of the hole (depending on
the position of the block 1n the
reactor core) is 1/2, 1 1/8, or 1 3/4
inches. A visual examination was
made of about 30 blocks of the type
having a 1/2-in.-dia hole. These
blaocks had been sawed in two on a
- plane perpendicular to the long axis.
It was observed that in all cases the
core of the blocks had a different
physical structure than did the outer
portion of the same blocks. In order
to study this core structure, .six
blocks that had not been sawed were
soaked 1n fuchsin dye for 12 hours.
The dye 1s most absorbed in the more
porous areas in the block, and Fig.
12.10, a photograph of one of the
dyed specimens, shows the typical
soft core structure noted in all ‘the
dyed specimens. Samples taken from
the inside and outside surfaces of
these blocks showed density variations
from 2.80 to 2.83 at the outside,
which decreased to 2.26 to 2.43 at the
inside. The apparent porosity of the
dense portion of the block i1s practi-

TABLE 12. 5.

 

compared with values up to
23.3% for the soft core. Table 12.5
gives results of density, apparent
porosity, and water absorption measure-
ments for specimens cut from various
portions of the block,

Blocks that have a 1 1/8-in.-dia
central hole are being split longi-
tudinally so that the two halves can
be placed around a cooling pipe. With
the blocks split, a large amount of
inner surface will be expoesed to the
coolant material. This inner surface
is low-density material that 1is
particularly susceptible to NaK cor-
rosion. To determine the extent of
the low density region, a random
sample of these blocks was soaked in

cally zero,

7912

 

Fig. 12.10. bye-treated Bel Block
Showing Porous (Central Section.

DENSITY, POROSITY, AND WATER ABSORPTION MEASUREMENTS
OF VARIOUS SECTIONS OF THE Be0 BLOCK

 

 

T wer weronn, v | oy weionn, 5 | L T RBSORPTION, T SUSPENDED | | ppmecry | FOROSITY,
SQURCE (&) (8) ¥ - D (¥ - D)/D WEIGHT, &* ¥F - 3e» (g/cms) (¥ - D)W - 8)
(%) (g) {%)
Quier edge I 1.3612 1.3609 0 0,881 0.4802 2.83 0
Outer edge 1.8459 1.8458 0 1.187 0.6589 2.80 0
OQuter edge 2.0859 2.0860 0 1.350 0.7359 2.83 Q
Core 1.3957 1.2789 0.-1168 9.13 0.845 0.5497 2.33 21.25
Core 1.3555 1.2661 0.08%4 7.06 0.835 0.5205 2.43 17.18
Core 1.1778 1.0630 0.1098—L 10. 28 0.707 0.4708 2.27 23.32

 

 

 

 

 

 

 

 

*Weight suspended in water,

**Equivalent te volume.
.|

172
fuchsin dye for 12 hr, and specimens
were cut from selected localities in
the block. Figure 12.11 shows the
density and per cent porosity placed
on the block in the location from
which the sample was taken. The dark
shaded areas on the right of the block
indicate the porous portions shown up
by the penetration of the dye. Five
samples of blocks with a 1 3/4-1in.-dia
central hole were soaked in dye and
were also found to have a porous
central core, _ :

A general observation was made
regarding the prevalence of cracks in
all the blocks examined. All the
blocks have at least one crack, and
some blocks contain several cracks.
In some cases these cracks are visible
at the surface of the block, but in
the majority of the blocks the cracks
are confined to the interioer portion.
These cracks are quite narrow and

 

Fig. 12.11.
size ARE BeO Block.

"heating,

PERIOD ENDING DECEMBER 10, 1952

contribute only slightly to the po-
rosity. The presence of one of these
cracks may be observed in Fig. 12.11.

The porous condition developed 1in
the hot-pressed BeO blocks i1is a defect
caused by fabrication. A probable
cause of the defect is that the hot-
pressed die had too tight a fit on the
top and bottom punches around the
central pin, which would cause binding
and therefore uneven pressure appli-
cation. A further factor is faulty
filling of the die, which would cause
bridging of the powder around the pin.
Other factors such as too short time
of applicdation of pressure, too short
time of holding at pressing tempera-
ture, uneven temperature distribution
in the BeO powder because of too rapid
cooling of the die to room
temperature without extracting the
core pin, etc., could also have af-
fected the final density.

PHOTD 5 5675 ¥

pensity and Per Cent Porosity at various Locations in z Full-

173
ANDP PROJECT QUARTERLY PROGRESS REPORT

13. BEAT TRANSFER AND PHYSICAL PROPERTIES RESEARCH

H. F. Poppendiek, Reactor Experimental Engineering Division

The thermal conductivity of the
flnoride mixture NaF-BeF, (57-43 mole
%) was experimentally determined to
be 2.4 Btu/hr-ft-°F over a temperature
range of 452 to 586°C. This result
1s comparable with the thermal con-
ductivities previously obtained for
some of the other fluoride mixtures.
Two new thermal conductivity devices
have been developed and successfully
tested with ordinary liqguids.

The enthalpies and heat capacities
of fuel mixture NaF-KF-ZrF,-UF,
(4.8-50.1-41.3-3.8 mole %) have been
obtained. The heat capacity can be
represented by ¢ = 0,28 + 0.015
cal/cm*°C over theptemperature range

540 to 900°C.

The viscosities and densities of
the fluoride mixtures NaF-ZrF, (50-50
mole %) and NaF-ZrF,-UF, (50-25-25
mole %) have been measured with the
modi fied Brookfield viscometer. The
viscosity of the first mixture ranged
from 15 cp at 610°C to 5 cp at 1030°C.
The viscosity of the second mixture
ranged from 20 cp at 650°C to 11 cp
at 880°C. These results are compared
with the viscosity measurements
previously determined for some of the
other fluoride mixtures. Development
of the capillary viscometer is con-
tinuing.

The vapor pressure of NaZrF, in-
creases from 8 mm Hg at 800°C to 30
mm Hg at 900°C; this behavior is
similar te that previously reported
for the NaF-ZrF,-UF, (50-46-4 mole %)
mixture. FExperimental values of the
vapor pressure of BeF, measured at
BMI agreed with values calculated at
ORNL..

Preliminary experimental heat
transfer coefficients have been
pbtained for the case of turbulently
flowing NaF-KF-LiF mixture in a long
tube. It appears that heat transfer
with this mixture can probably be
described by the usual forced-convection

174

expressions characterizing ordinary
fluids, as was found to be the case
for molten sodium hydroxide.

Several thermal analyses pertaining
to the reflector-moderated reactor
have been conducted. One study was
made concerning the heat generation
within the reflector and shell of the
reactor. Another analysis consisted
of preparing design charts for cooling-
hole distributions in reactor re-
flectors. An attempt to minimize the
weight and volume of air radiators
has been initiated.

A pyrex thermal convection harp
has been designed and constructed for
an experimental inquiry into harp
convection velocities. The experi-
mental data are to be used to check
the validity of analytical methods
used to predict the circulation
velocities.

An investigation has been initiated
in an attempt to derive a heat-momentum
transfer analogy for the annulus 1in
much the same way as it has been
derived for the pipe system.

An experiment has been designed
that will make possible an experimental
study of the heat transfer charac-
teristics of a simple, fuel-circulating
system; a long pipe will be used in
which the flowing fluid is heated
internally by an electric current.
The experimental results will be
compared with the previously obtained
theoretical results.

THERMAL CONDUCTIVITY OF LIQUIDS

W. D. Powers R. M. Burnett
S. J. Claihbhorne W. B. Harrison
Reactor Experimental Engineering
Division

A thermal conductivity device
similar to the one described previ-
ously(!) was used to study the fluoride
Ll e

L. F. Basel and M. Tobias, ANP Quar. Prog.
Rep. Dec. 10, 1950, ORNL-91%, p. 196.

 
mixture NaF-BeF, (57-43 mole %). The
results of three independent sets of
measurements are given in Table 13.1.:

Figure 13.1 shows a plot of the
thermal conductivities of several
fluoride salt mixtures as a function
of uranium concentration. It appears
that the thermal conductivities of
these somewhat similar fluoride
mixtures decrease as the uranium
weight percentages increase.

A new thermal conductivity device
for studying both liguids and solids
has been developed that consists of
an electric heater, a thermal con-
ductivity cell, a heat meter, and a
water cooler. Calculations reveal
that thermal conductivities can be
measured to within +10% of their true

PERIOD ENDING DECEMBER 10, 1952

TABLE 13.1. THERMAL CONDUCTIVITY
OF NaF-BeF,

 

 

THERMAL AVERAGE TEMPERATURE
CONDUCTIVITY TEMPERATURE RANGE
(Btu/hr+ ft- °F) (°c) (°c)
2.3 528 470 to 586
2.5 527 47] to 584
2.4 526 452 ta 500

 

 

 

values. Thermal conductivity measure-
ments of water, obtained with the
heat-meter thermal conductivity device,
agree with literature values to within
6.5%. _ _

A transient method for determining
the thermal conductivity of liquids

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DWG. 47410
30
FLuome-:1 1 ‘
MIXTURE NO.  MIXTURE COMPOSITION
12 :
5s | 2 NaF-KF-UF, (46.5-26.0-275 mole %) __|
' 39 12 NaF -KF-LiF  (11.5-42.0-465 mole %)
m o' 14 NaF -KF~LiF -UF, (109-435-44.5-11 mole %)
* a0 30 NaF-ZrF, -UF, (50-46-4 mole %) _ ]
g :
2 35 NoF-BeF, (57-43 male %)
a
kS
&
= 1.5 :
> ey
5 30
2
0
=z
D
o .
aj "0 \
< ; .
u'_‘ .
L
€I
L \
\\ 2
0.5 e frorean
0 :
) 10 20 20 40 50 60 70 80 90
' URANIUM CONCENTRATION (wt %) ‘
Fig. 13.1. Thermal Conductivity of Fluoride Mixtures as a Function of

Uranium Conecentration.

175
ANP PROJECT QUARTERLY PROGRESS REPORT

has been developed. A thin-walled
metal tube with a small diameter is
immersed in the liguid being studied,
and the tube i1s suddenly heated by a
direct current of electricity. The
transient temperature rise of the
tube 1sa function of the heat capacity
of the tube and the thermal conduc-
tivity, heat capacity, and density of
the liquid. The early part of the
temperature-time function is completely
controlled by the conduction mechanism.
Analyses of the solutions of the
transient conduction equations per-
taining to the system described(?’
show that liquid thermal conductivities
may be determined from the experi-
mental temperature-time measurements.
This transient apparatus has been used
to check the thermal conductivity of
water and glvcerine at room tempera-
ture; the results agree to within
+10% of the values reportedin the lit-
erature.

In addition to the continued study
of fluoride salt mixtures, it 1is
planned to measure the thermal con-
ductivity of molten sodium hydroxide
by oneor more of the methods discussed
above.

HEAT CAPACITY OF LIQUIDS

W. D. Powers G. C. Blalock
Reactor Experimental Engineering
Division

The enthalpy and the heat capacity
of the fuel mixture NaF-KF-ZrF,-UF,
(4.8-50.1-41.3-3.8 mole %) have been
determined by Bunsen i1ce calorimeters
for the temperature range of 540 to
900°C and are given by the equations:
H, (liquid) - Hyo. (solid)

= .15 + 0,287 ,
c, 0.28 + 0.015
where H is in cal/g, T in °C, and c,

in cal/g-°C. The fuel mixture
NaF-ZrF,-UF, (50-46-4 mole %) and the

(2)W. B. Harrison, Transient Methods for
Determining Thermal Conductivity of Liquids,
ORNL CF-52-11-113 (Nov. 1, 1952).

i

 

176

eutectic mixture of NaCH and LiOH are
currently being studied.

VISCOSITIES OF FLUCRIDE MIXTURES

Measurements with the Brookfield
Viscometer (R. F. Redmond, T. N. Jones,
Reactor Experimental Engineering
Division). The viscosities of the
fluoride salt mixtures NaF-ZrF, (50-50
mole %) and NaF-ZrF, -UF, (50-25-25
mole %) have been measured with a
Brookfield viscometer contained in an
inert atmosphere. These results are
plotted in Fig, 13.2, together with
the viscosities of several other
fluoride mixtures. The zirconium-
bearing salts have much higher vis-
cosities than NaF-KF-LiF (11.5-42.0-
46.5 mole %) and NaF-KF-LiF-UF,
(10.9-43.5-44.5-1.1 mole %)}; the
beryllium-bearing fluoride mixture
NaF-BeF, (57.0-43.0 mole %) is also
characterized by higher viscosities.
In general, it appears that an increase
in uranium content yields an increase
in the viscosity.

Capillary Viscometer (F. A. Knox,
N. V. Smith, F. Kertesz, Materials
Chemistry Division). The capillary
viscometer previously described(3+*)
has been used to study fused fluorides
over the temperature range of 600 to
800°C. All the materials studied 1in
this apparatus were purified by the
hydrogenation-hydrofluoerination
procedure. The apparatusin its present
form does not sufficiently protect the
mixtures from exposure to air, and
modifications are being made to the
equipment to improve the gquality of
the inert atmosphere maintained over
the melt.

Bloom, Harrap, and Heymann (%)
determined values for the viscosity
of various chloride mixtures by using

(3)J. M. Cisar, F. A. Knox, F. Kertesz, R. F.

Redmond, and T. N. Jones, ANP Quar. Prog. Rep.
June 10, 1952, OBRNL-1294, p. 146.

4 F. A, Knox, N. V. Smith, and F, Kertesz,
ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL-1375,
p. 145,

(S)H.

Proc. Roy. Soec.

 

Bloom, B. S. Harrap, S. E. Heymann,
(London) 194A, 237 (1948).
PERIOD ENDING DECEMBER 10, 1952

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

e
DWG. 17411
30
20
10
6 \\‘ ‘\ ?\\\
’gnt-; ~ \:‘\/14 \'\\ \
2 /::A ‘.\\ . -
0
=2 4 12 '\\\ >
~ ~
% SN0
> ‘\\
1.._._
o
S 2
9
> "~ — __NaOH
L ~%“"----\_
= ~ -
3 ‘.‘-.M\
S
& : —
< o8 I FLUORIDE -
il MIXTURE NO. MIXTURE COMPOSITION
0.6 12 NGF -KF~LiF (11.5-42.0-465 mole %)
— 14 NaF - KF~LiF - UF, (109-435-44.5-1.1 mole %)
0.4 27 NaF- ZrF, -UF, (46~50-4 mole %)
31 NaF-ZrF, (50-50 mole %)
- 33 NaF-ZrF, -UF, (50-25-25 mole %)
35 NaF ~BeF, (57-43 mole %)
0.2 ‘ ‘ .
800 900 1000 1100 1200 1300 1400 1500
TEMPERATURE (°K)
Fig. 13.2. Vviscosily of SeveralfFIuoride Mixtures as a Function of Temper-
ature. ~

177
ANP PROJECT QUARTERLY PROGRESS REPORT

glass capillaries in apparatus of this
type. The use of glass capillaries
with fluorides, even at temperatures
below 650°C, proved impossible, how-
ever, because the innmer surfaces of
the capillaries were attacked to such
an extent that the original
brations could not be reproduced.
The use of small nickel capillaries
(about 0.08-mm bore) tended to cause
plugging by the mixtures being studied.
Larger capillaries (2.4-mm bore)
minimized this difficulty, but they
were impractical for studying low-
viscosity liquids. With the large
capillaries, Reynold’s numbers below
300 were obtained, and the viscosities
measured —- based on calibrations with
glycerol solutions at room tempera-
ture - are probably reliable as
preliminary estimates. The viscosities
of the several fluoride mixtures thus
determined are in gualitative agree-
ment with those presented in Fig. 13.2.

cali-

DENSITY OF FLUDRIDE MIXTURES

R. F. Redmond T. N. Jones
Reactor Experimental Engineering
Division

Preliminary density measurements
have been made for the fluoride
mixtures NaF-ZrF, (50-50 mole %) and
NaF-ZrF,-UF, (50-25-25 mole %), and
the values are tabulated in Table 13.2.

 

 

 

TABLE 13.2. DENSITIES OF T®¥d FLUORIDE
MIXTURES AT VARIOUS TEMPERATURES
TEMPERATURE DENSITY
FLUORIDE MIXTURE . 3
(°C) (g/cm”)
NaF-ZrF, 592 3.35
(5050 mole %) 690 3.28
762 3.21
862 3.14
NaF-ZrF,-UF, 682 4.13
(50-25-25 mole %) 156 4,0%
820 3.94

 

 

 

178

VAPOR PRESSUREL OF MOLTEN FLUCRIDES
H. E. Moore

Materials Chemistry Division

The vapor pressure of ZrF, above
the pure compound NaF-ZrF, (50-50
mole %) is of interest because this
substance will serve as the fuel
carrier introduced initially in the
ARE. Vapor-pressure data for this
salt were obtained in the temperature
range of 795 to 994°C by the method
originally described by Rodebush and
Dixon, (%) which has been discussed in
previous reports.(’*8) The values
obtained, given in Table 13.3, are
best represented by the equation

-7213
m_ff—-+ 7.635 ,

log P =
where P is in mm Hg and T 1s in °K,
The heat of vaporization is 33 kcal/
mole, and the boiling point, calcu-
lated from the equation, is 1244°C,
The vapor pressure of the NaF-ZrF,
(NaZrFs) i1s very similar to that of
the mixture containing 46 mole % ZrF,,
50 mole % NaF, and 4 mole % ur,,
reported previously.(%) This, of
course, 1s not surprising, since
substitution of 8 mole % of NaUF, for
NaZrF, should not greatly influence
the equilibrium pressure of ZrF,.
Verification of results obtained
by this method has been attempted
during the past quarter by using other
procedures applicable to fused salts
at high temperatures. Agreement among
the various methods has, in general,
been encouraging. Values for the
vapor pressure of ZrF, (36 mm Hg at
768°C and 96 mm Hg at 806°C) were
obtained by using a diaphragm apparatus
similar to that used for vapor pressure

 

(G)W. H. Rodebush and A. L. Dixen, Phys. Rewn,
26, 851 (1925),

R. E.

. Moore and C. J. Barton, ANP Quar,
Prog. Rep. Sept. 10, 1951, ORNL-1154, p. 136.
(B)H. E. Moore and C. J. Barton, ANP Quar,
Prog. Rep., Dec, 10, 1951, OONL-1170, p. 126,
(9)H. E. Moore and C. J. Barton, ANP Quar.
Prog. Rep. Sept. 10, 1952, ORNL-1375, p. 147.
TABLE 13.3. VAPOR PRESSURE OF NaZrF,

 

 

TEMPERATURE OBSERVED PRESSURE
(°c) (mm Hg)
795 8
803 8
807 : 8
835 13
835 14
854 - 16
858 18
868 20
886 28
887 27
923 39
846 47
948 50
964 64
967 63
994 a0

 

 

measurements of UF,.('%) These data
are in fair agreement with the values
(39 and 94 mm Hg) obtained by the
Rodebush method. (%) In addition,
tentative values for this compound,
29 mm Hg at 751°C and 56 mm Hg at
786°C, obtained by using a capillary
“bridge’’ apparatus, (11+12) gpree
fairly well with the values of 26 and
59mm Hg obtained by using the Rodebush
me thod. ' :

In the past quarter, ‘vapor-pressure
data obtained by applying a trans-
piration method to beryllium fluoride
were reported by Battelle Memorial
Institute, ¢*3) A comparison of the
data from Battelle with values calcu-
lated from the equation reported by
this Laboratory('*) is

(ID)K. 0, Johnson, The Vapor Pressure of
firaniun Tetrafluoride, Y-42 (Oct. 20, 1947). :

(11)C. G. Maier, Yapor Pressures of the Common
Metallic Chlorides and a Static Method for High
Temperatures, U.S, Burean of Mines, TP-360.

(12)D. W. Kuhn, A. D. BRyon, and A. A. Palko,
The VYapor Pressures of Zirconiuw Tetrachloride
and Hafniua Tetrachloride, Y-552 (Jan. 17, 1950).

(3¢ A, Sense, M. J. Snyder, and J. ¥. Clegg,
Prog. Rep. Sept. 1952, BMI-772, p. 46, ;

R. E. Moore, ANP Quar. Prog. Rep. June 10,
1952, ORNL-1294, p. 150, '

given 1n

 

PERIOD ENDING DECEMBER 10, 1952

Table 13.4. The two sets of values
are probably in agreement, within the
experimental errors of both methods.

TABLE 13.4., COMPARISON OF VAPOR
PRESSURE OF BeF, BY DIFFERENT METHODS

 

 

 

 

VAPOR PRESSURE
TEMPERATURE (mm Hg)
{(°C) -
BMI Data ORNL Data

778 1.25 2.1
849 7.25 8.8
918 23.5 24
969 51.0 47

 

 

 

CONVECTIVE HEAT TRANSFER 1IN
FLUGRIDE MIXTURE NaF-KF-LiF

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

The construction and instrumentation
of the experimental system for the
determination of convective heat
transfer coefficients for the fluoride
mixture NaF-KF-LiF (11.5-42.0-46.5
mole %) have been completed. This
system is a modification of the ‘one
used for obtaining heat transfer
coefficients in a tube with molten
sodium hydroxide in turbulent flow. (1%
All system components, with the
exception of the test section, have
been constructed of Inconel; A nickel
tubing was used for the test section.
Swagelok connections and “0’" ring
flanges have replaced the all-welded
construction of the sodium hydroxide
system.

The heat loss calibration and five
heat transfer runs with NaF-KF-LaiF
were made. These runs covered a
Reynold’s modulus range of 2000 to
6000. Figure 13.3 presents the data
for NaF-KF-LiF compared with data for
air that were obtained by various

(IS)H. W. Hoffman and J. Lones,

ANP Quar.
Prog. Rep. Sept. 10, 1952, ORNL-1375, p.

148.

179
ANP PROJECT QUARTERLY PROGRESS REPORT

 

Q.01 pemmmmmmmmemmn

 

 

 

 

 

 
      

o GRAETZ (FROM DREXEL AND McADAMS); £/0 = 39; AIR |
o |.OWDERMILK, HUMBLE AND DESMON; £/0 =60, AIR — e ool
’ -~~COLBURN; £/0 AS INDICATED; OILS

| & THIS

lNVESflGAflON L/D =200, NaF ~KF- UF(EUTECflC) |

Llwmqifllgi

UNGLASSIFIED
DWG 17412

 

 

 

 

 

?(FULLY DEVELOPED
T** TURBULENT F%OW)‘*j

MML;ML,@, ,,,,,,,,,,, ______ _____

i
i ’
i
{

 

 

 

 

 

oo b
102 2 5
Re
Fig. 13.3. J-Factor vs. Reynolds Modulus for NaF-¥F-L1iF, Air, and Other

Fiuids.

investigators. {16+ 17.18) The heat
transfer coefficients are correlated
in this figure by the j factor (the
product of the Stanton modulus, h/c¢ G,
and the two-thirds power of the Prandtl
modulus) as a function of the Reynold’s
modulus. It is to be noted that the
transition from laminar flow to fully
developed turbulent flow occurs over
a region extending from Re = 3000 to
Re = 10,000. The preliminary data of
the current investigation fall near
the air datain this transition region.
Further work is to be done at higher
fluid temperatures to extend the data
into the region of fully developed
turbulence.

The experiments with the NaF-KF-ILaF
mixture were prematurely terminated
because of the failure of the nickel
test section. '“Ring’ cracks caused

(16)R, E. Drexel and W. H. McAdams, Nat. Adv.
Comm. Aero., Wartime Report ARR4AF28.

(IT)W. H. Lowdermilk, L. V. Humble, and L. G,
Desmon, KHeasurements of Average Heat Transfer and
Friction Coefficients for Subsonic Flow of Air in
Smooth Tubes at High Surfece and Fluid Tempera-

tures, NACA-102 (1951),

(la)A. P. Colburn, Engineering Bulletin Purdue
University, Vol. 26, No. 1 (Jan. 19, 1942).

 

180

a portion of the test section (5 1/4
in. long and located approximately at
the center of the test section) to fall

out. The failure was not caused by
corrosion but appears to be due to

metal fatigue resulting from the high
temperatures to which the nickel
section was subjected.

A new test section of Incomnel
(L/D = 137) 1s being fabricated.
Although the length-to-diameter ratio
for this tube is less than that for
the nickel tube, 1t 1s still sufficient
to ensure that the entrance length
will be exceeded within the confines
of the test section.(!?)

ANALYSIS OF SPECIFIC REACTOR
HEAT TRANSFER PROBLEMS

W. S. Farmer
Reactor Experimental Engineering
Division
Heat Generation in the Reactor

Reflector. 1In order to arrive at an
estimate of the heat generation to be

 

(19)11 W, Hoffman and J. Lones, ANP Quar.
Rep. Sept., 106, 1952, ORNL-1375, p. 148.

Prog.
expected in the reflector and shell
of the reflector-moderated reactor
(cf., “General Design Studies,’ sec.
3), several radiation source problems
were analyzed. The largest source
of heat in the reflector results from
the absorption of gamma rays generated
in the fuel coolant as a result of
prompt fission and decayof the fission
products. The magnitude of this
source of heat generation was deter-
mined by using three methods of
calculation based on isotropic source
distribution and exponential attenu-
ation. In the first calculation, the
gamma -ray source was assumed to be
distributed uniformly in a spherical
annuilus, and the same absorption
coefficient was assigned to both the
reflector and fuel coolant. 1In the
second method, the source was distrib-
uted uniformly in an infinite slab
of thickness equal to the annulus,
and the true absorption coefficients
of the fuel coolant, shell, and
reflector were used in evaluating the
heat generation from gamma-ray absorp-
tion at any point in the reflector.
The resulting heat generation was
corrected for geometry by a plane-to-
sphere transformation, The third,
and probably the best, approximation
to the true heat generation in the
reflector was obtained by solving for
the current of radiation from the
surface of the fuel cooclant by using
a spherical annulus geometry. Thisg
cirrent was then assumed to enter the
face of a composite slab made up of
the shell and reflector that surround
the fuel coolant. The resulting heat
generation was then transformed from
a slab geometry to a spherical geometry.

A secondary source of heat gener-
ation in the reactor reflector and
shell arises from the absorption of
gamma rays generated by neutron capture
in the reflector and shell. 1In order
to evaluate this source of heat gener-
ation, a general solution was deter-
mined for the case of a nonuniform
source of radiation in a slab geometry

PERIOD ENDING DECEMBER 10, 1952

in which the mneutron flux could be
approximated by a power series or
exponential expression,(

In order to simplify the com-
putation of the maximum temperature
rise in the reflector resulting from
heat generation, general design charts
that give this temperature rise for
various values of heat generation,
coolant-hole diameter, thermal con-
ductivity, and coeolant-hole spacing
were prepared for several pertinent
materials. (?1) :

Analysis of Fluid-to-Air Radiators.
The design of the air radiator for
transferring the heat generated in the
reactor coolant to the air propulsion
stream flowing through the turbojet
engines has been analyzed in an effort
to minimize air radiator welght and
size. An analysis of the radiator
volume, weight, and pressure drop was
made for fixed conditions of power
output, mass velocity per frontal
area, and thermal conditions for a
design in which plane-plate, finned
radiators of stainless steel and
nickel are used. In arriving at a
solution, the heat transfer coeffi-
cients for the radiator were deter-
mined on the basis of laminar flow
through flat ducts, since the Reynold’s
number for flow through the radiation
over the pertinent range of conditions
was less than 2000, Multiple calcu-
lations were made for a range of values
of tube diameter, fin thickness, fin
spacing, and tube pitch. On the basis
of this evaluation, the optimum
radiator design appears to be that
with tube diameters between 1/8 and
1/4 in., 20 to 30 fins per inch, a fin
thickness of 0.010 in., and a patch
of three times the tube diameter.

A review of the literature covering
the thecoretical and experimental work

 

(20)W. S. Farmer, Heat Generagtion in ¢ Slab for
o Non«Uniform Source of Radiation, ORNL CF-52-9-202
(Sepr. 4, 1952). '
(21} . . . ,
W. 85, Farmer, Cooling Hole Distribution for
Rz;tc)tor Reflectors, ORNL CF-52-9-201 (Sept. 3,
1952), :

181
ANP PROJECT QUARTERLY PROGRESS REPORT

on air radiations and simulated flat-
duct systems is being made. A solution
is sought that will suitably correlate
the nonisothermal heat transfer
results of air radiator tests being
conducted by the ANP Division,

NATURAL CONVECTION IN CONFINED SPACES
AND THERMAL LOOP SYSTEMS

D. C, Hamilton F. E. Lynch
L. D. Palmer
Reactor Experimental Engineering
Division

Temperature profiles have been
measured in the flatnflate natural-
convection apparatus(??) over the
practical range of variables attainable
with the present equipment. The data
will be presented in a forthcoming
ORNL report and will be analyzed in
relation to the theory previously
developed. An example of a typical
experimental temperature profile is
given in Fig., 13.4.

The annulus, natural-convection
apparatus is ready to be operated,

(22)0. C. Hamilton asd F, E. Lynch, ANP Quar.
Prog. Rep. June 10, 1952, ORNL-129%4%, p. 158,

 

   

  
 
 

UNCLASSIFIED
DWG. 1743
1.0 -~ m—fl——u——w
0-8 _— - e . oo e ——
CONDUCTION CURVE — —_
{NO CONVEfiHON) T
06 e e e — e — ,.

 
 

*gi)fii .

WITH CONVECTION~__

    

0.4

TEMPERATURE FUNCTION, &

Q.2

 

   

 

   

0 — e e
LEFT CENTERILINE RIGHT
walLL WALL
Fig. 13. 4. Typical Temperature Pro-

file with Natural Convection.

182

except for construction of the two
thermocouple assemblies. After the
flat-plate data have been analyzed,
the annulus experiments will be
stacrted. The test section shown in
Fig., 13.5 was described in detail
previously. (?3) The test section has
been assembled and installed in a
20-in,-ID can, and the can will he
sealed and maintained at a slight
vacuum to prevent mercury vapor from
leeking into the room during operation.

In the pyrex model of a thermal
harp, shown in Fig. 13.6, the tube has
an inside diameter of 16 mm and the
over-all height of the apparatus is
30 inches. The temperature differ-
ential between the hot and cold legs
is provided by running hot and cold
water through the condenser tubes.
The natural-convection velocity of
the water in the inner tube is measured
by observing the motion of small
particles. The data now being obtained
with this apparatus will provide a
means of checking the validity of
analytical methods of predicting harp
velocities.

TURBULENT CONVECTION IN ANNULIZ

W, B. Harrison J. 0. Bradfute
Reactor Experimental Engineering
Division

Relatively little satisfactory
information on turbulent-flow, forced-
convection, heat transfer in anrulus
systems 1s available in the litera-
ture. (2%)  However, a significant
number of momentum transfer studies
have been reported, (257 and investi-
gation has been initiated in an attempt
to derive a heat-momentum transfer
analogy for the annulus in muech the

 

(23)p. C. Hamilton and F. E. Lynch, ANP Caar.
Prog. Rep. Sept. 10, 1952, ORNL- 1375, p. 149.

(24)H. C. Claiborne, A Review of the Literature
on Heut Transfer in Annuli and Noncircular Ducts
for Ordinary Fluids and Liquid Metals, ORNL
CF-52-8-166,

25]H. C. Claiborne, Critical Review of the
Literature on Pressure Drop in Noncircular Ducts

and Annuli, OBNL-1248 {(May 6, 1952).
PERIOD ENDING DECEMBER 10, 1952

; UNCLASSIFIED
5 J DWG_ 17414

     
    
  
   
   
    
     

UPPER ELEGTRODE = CENTER THERMOGOUF’LE ASSEMBLY

 

77 '
= I!fi“‘% ——WALL THERMOCOUPLE ASSEMBLY
Y  (8%in. LONG)

5
II
|

   

i i
i i
B

. i

3

i 4
K ;
i ;

'l i il

b i

i

f 2

ik -
i N

]

COOLANT QUT COOLANT QUT

——— TEST SECTION ASSEMBLY
(%-in. 1.D., O187-in. WALL , 16 %g-in. LONG,

INSIDE "SURFACE ELECTRICALLY INSULATED
WITH ENAMEL)

~_ |

~~— COOLANT JACKET
(1%g-in. 1.D., 0.125-in. WALL , 1535 -in. LONG)

 

 

 

 

 

 

 

 

e
b
r

i

  

COOLANT N

 

 

 

fFig. 13.5. Test Section Detail of Annulus Natural-Convection Apparatus.

183
ANP PROJECT QUARTERLY PROGRESS REPORT

 

 

-

UNCL ASSIFIED

5
e

R

 

 

=

o

ey
S

L

TR

hE
SR

P

Pyrex Thermal Convection Loop.

13- 69

Fig.

184
same way as it has been done for the
pipe system. The velocity profiles,
as well as the eddy diffusivity
profiles obtained from radial shear
stress and velocity gradient data, are
required in such an analysis. Thus,
the first step in such a study must
be the development of a generalized
velocity profile for the annulus.
Several possible ways of generalizing
experimental annulus velocity data are
currently being tried; the experi-
mental data of Knudsen and Katz, (?%)
Mikrjukov, (?27) and Rothfus, Monrad,
and Senecal, (%) are being utilized
in this study. :

CIRCULATING-FUEL HEAT TRANSFER
H. F. Poppendiek G. Winn
Reactor Experimental Engineering
Division
Mathematical studies of circulating-
fuel heat transfer in long pipes have

 

(26)J. G. Knudsen and D. L. Katz, Proceedings
of the Midwestern Conference on Fluid Dynamices,
First Conference, Moy 12-13, 1950, p., 175.

{27)V. Mikrjukov, J. Tech. Phys.

(U.8.5.R.) 4,
961 (1937). S

(2B)H. R. ‘Rothfus, C. C. Monrad, and B, E.
Senecal, Ind. Eng. Chea. 42, 2511 (1950},

PERIODfENDING DECEMBER 10, 1952

been previously presented. (?2:30) 4
modest experimental study has been
initiated in an attempt to compare
the theoretical and experimental
behavior of this volume heat source
system. A sulfuric acid solution: is
to be circulated in a closed loop by
means of aglass pump. An a-c electric
current is to be passed through the
acid flowing in a long plastic pipe
(1/4 in. in inside diameter and 2-ft
long) located in the flow system. The
current flow will yield a volume heat
source within the acid. Acid flow
rates, pipe wall temperatures, and
mixed-mean acid temperatures into and
out of the test section will be
measured. Test section and heat
exchanger heat balances will also be
made. The temperature differences 1in
this system will lie within the range
necessary to yield only minor changes
in electrical resistivity of the acid,
so an almost uniform volume heat source
exists in the system., It is proposed
to study both laminar and turbulent
flow regimes.

(ZQ)H. F, Poppendiek and L. D. Palmer, Forced
Convection Heat Transfer in o Pipe System with
Volume Heat Sources Within. the Fluids, Y-F30-3

(Nov. 20, 1951},

(SO)H. F, Poppendiek and L. D. Palmer, Forced

Convection Heat Transfer in Pipes with Volume
Heat)Sources Within the Fluids, OBNL-1395 (Dec. 2,
1952).

 

185
ANP PROJECT QUARTERLY PROGRESS REPORT

14.
D. S. Billington,

RADIATION DAMAGE

Solid State Division

A, J. Miller, ANP Division

Radiation-damage studies of ma-
terials exposed in the ORNL graphite
reactor, the LITR, and the 86-in.
cyclotron have continued. A sample of
a zirconium-bearing fused fluoride
fuel has been irradiated with a high
flux 1n the MTR, but examination of
the irradiated capsule and its con-
tents has, not yet been made. A sodium-
filled loop was operated in the LITR
at about 1200°F, but it was shut down
after seven days because of a flow
stoppage.

Additional irradiations of fuel
have been carried out in the LITR at
ARE fission rates for extended periods
of time. High power dissipations, in
the range that would occur in an air-
craft reactor, were achieved for long
periods of time in fuels by using 22-
Mev protons from the cyclotron, and
there was some indication, but no
positive evidence, of radiation damage
or radiation-induced corrosion.

In-reactor cantilever creep measure-
ments on Inconel in an air atmosphere
were continued in both the graphite
reactor and the LITR. From these more
recent measurements, 1t appears that
the radiation has only a small effect
on the creep strength. Examination of
some of the older test rigs showed
that the decrease in creep strength
previously reported could possibly
have been due todefective extensometer
devices. Apparatus is being con-
structed for creep measurements 7in the

MTR.

These radiation damage studies are
described in the following. Additional
information is contained in the Solid
State Division Quarterly Progress
Report for the Period Ending November
10, 1952.

186

TIRRADIATION OF FUSED MATERIALS

G. W. Keilholtz D, F. Weeks
J. G. Morgan M. T. Robinson
H. E. Robertson D. D, Davies
C. C. Webster A. Richt
P. R, Klein W, J. Sturm
M. J. Feldman

Solid State Division

R. J. Jones R. L. Knight

Electromagnetic Research Division

B, W. Kinyon
OBNL Engineering Division

The installation in the MTR of the
testing facility for fused salts was
completed, The guide tube extends
from a 1 3/8-in, hole in a beryllium
“A’ piece through a specially designed
flange in the north-tank access hole.
In the first run, a l-in. column of
salt was contained in a 0,1-1in,-1ID
Inconel capsule. The capsule tempera-
ture was maintained at 1500°F by means
of a controlled flow of cooling air.
The flux at the capsule position was
determined by means ofcobalt wire with
the MTR operating at 19 and at 25
megawatts. The estimated thermal flux
for the first capsule run at 30 mega-
watts was 2,1 x 104,

The capsule contained a fuel with
the composition NaF-ZrF -UF, (50.0-
46.2-3.8 mole %). It was maintained
at 1500°F for 116 hr at full flux, and
there was an estimated power dissipation
of 1900 watts/cm® in the fused salt,
This approaches the power density
expected 1in an aircraft reactor,
Examinations of the irradiated capsule
and the salt are now being made, In
addition, two capsules containing the
fuel have been irradiated in the LITR
at 140 watts/cm® for 565 hr, and two
more have been placed in the LITR for

1000-hr tests. Examinations of the
irradiated materials are now being
made. ‘

Preparations are nearing com-
pletion for melting-point checks and
fission-fragment-activity determi-
nations on all specimens of reactor-
irradiated fuel. Equipment has been
constructed for rocking a capsule so
that the fuel will cycle through a
thermal gradient; this equipment will
be used when a horizontal beam hole in
the LITR is available for the experi-
ments. :

Inconel capsules containing fuels
NaF-KF-Z¢F -UF, (4.8-50.1-41.3-3.8
moele %) and NaPV-ZrF, -UF, (46-50-4
mole %) were bombarded with 22-Mev
protons from the 86-in. cyclotron.
Power densities of several thousand
watts per cubic centimeter of irradi-
ated fuel were used for periods up to
92 hours, In evaluating the effects
of theirradiation, only metallographic
examinations of the Inconel could be
considered reliable, since the irradi-
ated volume of the capsule containing
the fuel was large enough to cause a
threefold dilution of dissolved
Inconel components., For each experi-
ment, listed in Table 14.1, there was

PERIOD ENDING DECEMBER 10, 1952

a furnace-heated control that showed
0.5-m11 paits. Tt can be seen that
there is some indication of attack on
the Inconel at the higher power
densities, : '

IN-REACTOR CIRCULATING LOOPS

0. Sisman R. M. Carroll

W. W, Parkinson C. D. Bauman

J. B. Trice C. Ellis

A, S. Olson W, E. Brundage
M. T. Morgan F. M. Blacksher

Solid State Division

An Inconel loop containing sodium
was operated in the LITR at a tempera-
ture in the region of 1200°F for seven
days. The loop was shut down and re-
moved because of a flow stoppage; the
reason for the stoppage is now being
determined.

A prototype pump for use in an MTR
fluoride loop 1s being constructed.
Fabrication of other portions of the
loop is being delayed until information
about the MIR capsule tests i1s obtained
and further evaluwation can be made of
the experiment planned. |

TABLE 14.1. CYCLOTRON IRRADIATIONS OF FUELS

 

 

FUEL TRRADIATION POWER DISSIPATION INCONEL PITS
‘ " TIME (hr) (watts/cm®) (mil)
NaF-KF-ZrF ,-UF, 3 2400 0.5
4.5 2300 0.5
7 4100 1.5
8 1150 0.5
NaF-ZrF,-UF, 5.6 2800 0.5
10.6 2200 0.5
14.3 2000 0.5
15. 4 2900 0.5
46. 4 3300 2.0
92 1700 0.5

 

 

 

 

187
ANP PROJECT QUARTERLY PROGRESS REPORT

CREEP UNDER JTRRADIATION

W, W, Davis J., C, Wilson
J. C, Zukas
Solid State Division

Additional cantilever creep tests
on Inconel with the standard ARE 2-hr,
1700°F anneal were carried out at
1500°F in air atmospheres in the
graphite reactor and the LITR. As
shown in Figs., 14.1 and 14.2, there
are only small differences between the
bench and in-reactor tests, The upward
deflection after 400 hr in the curve
for the 3000-psi test in the ORNL
graphite reactor is probably due to
faulty operation of the temperature
controller. The results are to be
checked by postirradiation measure-
ments of the creep deflections on a
remotely controlled profilometer.

Postirradiation measurements were
made on specimens annealed at 1650°F
and irradiated in the graphite reactor
at 1500°F., Their creep curves were
shown in a previous report¢?!) as curve
A, 1500 psi, and curve C, 2000 psa,
The profilometer measurements indicated
that the large extensions which were
measured were in error and that the
actual extensions were not much
greater than those occurring in the
bench tests. Several tests reported
last quarter as being 1n progress were
negated by faulty microformers,

The design for the bellows-loaded
tensile crecep test in the MIR is near-
ing completion, and most of the com-
ponents have been fabricated
partly assembled for test.

(l)J. C. Wilson, J. C, Zukas, and W, W, Davis,

ANP Quar. Prog. Rep. June 10, 1952, ORNL-1294,
p. 164,

and

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SSD-A-499
DWG.1TO3T7A
006 |- A-—BENCH TEST
O—HOLE HB-3 LITR. FAST FLUX, > 10'?
3 ® —HOLE 15 ORNL GRAPHITE REACTOR. FAST FLUX, 4 x10'°
& ANNEALED 2hr AT 1700°F M&——A—’M
&
l...
- B\
0.02 Cr__{}fl——-—-Cf”ff"‘p==55£::‘;~—“JD 441’——““-4‘
0 200 400 600
TIME (hr)
Fig. 14.1. Cantilever Creep Tests om Inconel in Air at 1300°F and 1500 psi.

188
PERIOD ENDING DECEMBER 10, 1952

S
55 D-A-500
OWG. 170 38A

 

014

A\ —BENCH TEST _ /

042 |- O ——HOLE HB-3 LITR.FAST FLUX,>10'?
® —HOLE 15 ORNL GRAPHITE REACTOR. FAST FLUX, 4 x!

 

O‘IO

 

 

 

0.10

ANNEALED 2hr AT 1700°F

 

 

 

 

 

 

 

 

 

= _ &/AA
5(106 _ {r?QZ//”’
>
ul : 7
0.04 e J‘% :
& |
0.02 N e

 

 

 

 

 

 

 

 

 

0 . 200 400 600
TIME (hr)

Fig. 14.2. Cantilever Creep Tests on Inconel in Air at 1500%‘ and 3000 psi.

189
ANP PROJECT QUARTERLY PROGRESS REPOCRT

15. ANALYTICAL STUDIES OF REACTOR MATERIALS
C. D. Susano
Analytical Chemistry Division
C. R. Baldock

Stable Isotope Research and Production Division

J. M. Warde
Metallurgy Division

An empirical volumetric method for
the determination of zirconium 1n
reactor fuels has been developed. The
method 1s based on the measurement of
the basicity produced by the reaction
between zirconium hydroxide and excess
potassium fluoride. An 1indirect
volumetric method for the determina-
tion of zirconium, which depends on
precipitating zirconium as the slightly
soluble benzoate or m-nitrobenzoate
and subsequently determining the
organic acid by titration, proved
unsatisfactory because of the variable
composition of the precipitate. At-
tempts to determine zirconium by
titrating zirconium m-nitrobenzoate in
nonaqueous media with perchloric acid
were not successful. The use of
mandelic acid as a precipitant for
zirconium in the presence of a large
excess of uranium was shown to be
superior to phenylarsonic acid,
m-nitrobenzoic acid, and cainnamic
acid.

The stability of the chromium
complex with diphenylcarbazide was
increased by development of the complex
in 0.2 N perchloric acid. The separa-
tion of traces of aluminum 1n reactor
fuels by means of anionic-exchange
resins was studied; preliminary results
indicate that this technique is feasi-
ble. A bibliography of recent litera-
ture on bromine trifluoride as a
reagent for the determination of oxygen
in metallic oxides has been started.
An improved apparatus for the determi-
nation of trace amounts of water in
solids was designed and fabricated for
measuring the electrical conductivity
of anhydrous hydrogen fluoride to
determine 1ts water concentration.

190

In addition to the chemical analy-
ses, various compounds of 1nterest
were examined by spectrographic or
petrographic techniques. The useful-
ness of the spectrograph has been
enhanced by the construction of a
source that is good for temperatures
approaching 2850°C. The petrographic
studies have led to the classification
of many complex compounds that are not
found 1n the literature.

The work of the ANP Analytical
Laboratory service group consisted
chiefly of analyses of beryllium- and
zirconium-bearing fluoride mixtures.
A review of the methods used and a
summary of the services performed are
included.

CHEMICAL ANALYSIS OF REACTOR FUELS

J. C. White J. E. Lee, Jr.
C. M. Bovyd W. J. Boss
C. K. Talbott

Analytical Chemistry Division

The research and developmental work
on methods for the analysis of reactor
fuels and their components has been
of a diversified nature during thais
quarter. The primary interest has
been in the development of a volumetric
method for the determination of zir-
conium. Other problems that have been
studied include: the colorimetric
determination of chromium; the de-
termination of oxygen 1n metallic
oxides; the separation and determina-
tion of traces of aluminum 1n reactor
fuels; the determination of water 1in
components of reactor fuels; the
design of apparatus that will permit
the measurement of the conductivity
of liquid hydrofluoric acid so that
1its water content can be determined;
the 1dentification of the products
that are produced by the reaction of
- NaK and components of the fuels; the
adaptation of a method for the de-
termination of Be0 in NaK; and the
adaptation of a method for the de-
termination of UOQ, in UF,. A discus-
sion of these problems 1s included in

the following paragraphs.
Zirconium. An empirical volumetric

method for the determination of zir-
conium has been developed. The method
15 based on the measurement of the
basicity produced by the reaction
between zirconium hydroxide and excess
potassium fluoride, which was discussed
in a previous report.(!) TFree acid in
the starting solution in excess of
0.5 N produced results approximately
2% high. This interference can be
overcome by dilution; a practical
limitation must be 1imposed, however,
because of the difficulty encountered
in determining the end point. The use
of acid media other than sulfuric acid
was also explored. Only perchlorate
systems were definitely unsatisfactory
because of excessive hydrolysis at
high acidity.

Another approach to the development
of a volumetric method for zirconium
was to precipitate the hydroxide from
a mineral acid solution with excess
ammonium hydroxide, centrifuge, wash,
and dissolve in standard acid; the
excess of standard acid was determined
by titration. This method 1s satis-
factory in an academic sense, but 1is
hardly practical because of the copious
washing required to free the hydrous
oxide from basic salts and excess
precipitant. j

The guantitative precipitation of
zirconium by benzoic acid and nitro-
benzoic acids has been reported.¢2?:3)

{1)

 

R. Rowan, Jr., J. C. White, C. M. Boyd,

W. J. Ross, and C, K. Talbott, ANP Quar. Prog.
Rep. Sept. 10, 1952, ORNL-1375, p. 159.
(E)C. Lakshman Rao, M. Venkataramaniah, and

Bh. S5, V. RAaghava Rao, J. Sci,
(India) 108, No. 7, 152-4 (1951},

M. Venkataramaniah and Bh. S. V. Raghava Rao,
Z. anal. Chem. 133, 24B-31 (1651).

Ind., Resecoarch

PERTOD ENDING DECEMBER 10, 1952

The reported method was used as a
basis for formulating a volumetric
method in which the precipitate 1is
dissolved in dilute acid, and the
liberated organic acid 1is extracted
with ethyl ether and determined by
titration in methyl alcohol-benzene
solution with sodium methylate by
using thymol blue as the indicator.
However, a precipitate of varying
composition is formed, and attempts
to establish an empirical relationship
were unsuccessful.

Further study of the reported
gravimetric methods showed low results
in all cases, the extent being de-
pendent upon the mnature of the medium
of precipitation. For example, 1in a
sulfate system, no precipitate 18
formed with benzoic acid because of
the strength of the zirconium sulfate
complex in solutions that are more
acid than pH 3 to 4. A slight pre-
cipitate was noted when m-nitrobenzoic
acid was used as the precipitant.
Nitric or hydrochloric acid solutions
are recommended for this procedure.

An attempt was made to titrate
zirconium m-nitrobenzoate with per-
chloric acid in ethylene glycol~
isopropyl alcohol solutien. Zirconium
1s not sufficiently basic in this
medium, however, to give a distinct
break in the potentiometric titration
made by using a glass electrode as
the indicator electrode and a calomel
electrode as the reference electrode.

Since mandelic acid gives a pre-
cipitate of definite composition with
zirconium, this reagent will be
studied for application to volumetric
procedures. '

The effect of large excesses of
uranium on the gravimetric determina-
tion of zircomium was studied briefly
in an attempt to select the precipitant
least affected. Four reagents, phenyl-
arsonic acid, mandelic acad, m-nitro-
benzoic acid, and cinnamic acid, were
tested on solutions of zirconium
containing a fivefold excess of
uranium. . The number of determinations

191
ANP PROJECT QUARTERLY PROGRESS REPORT

made was i1nsufficient to permit a
statistical study; however, the re-
sults clearly 1i1ndicated that the
mandelic acid method was superior 1in
the presence of uranium. Phenylarsonic
acid gave slightly high results;
m-nitrobenzoic acid, slightly low
results. Cipnamic acid 1s unsatis-
factory for this purpose; the average
of three determinations in the uranium-
containing solution was 15% lower than
the amount taken.

chromium. The chromium-diphenyl-
carbazide complex 1s usually developed
in 0.2 N H,S0,, a medium said to
produce maximum stability of the
complex, Thirty minutes is considered
the maximum period before serious
fading occurs. The use of 0.2 N
perchloric acid solution increases the
period of color stability to nearly
60 min, regardless of the oxidant
used. Acid concentrations greater
than 0.2 N are harmful; for example,
0.6 N prevents color formation en-
tirely.

Aluminum. The determination of
traces of aluminum can be accomplished
by means of “aluminon’*) reagent;
however, complete separation from a
number of interfering cations 1is
necessary. JIn order to apply this
method to reactor fuels, aluminum must
be separated from zirconium and 1iron.
A chemical separation based om con-
trolled-acidity precipitation with
*“oxidine” was unsuccessful. Separation
by 10on exchange 1s currently being
studied. One procedure is to retain
the zirconium and uranium as sulfate
complexes on an anionic exchanger that
passes aluminum. An alternate pro-
cedure 1s to retain both aluminum and
zirconium as fluoride complexes and
separate them by elution, as reported
by Freund and Miner. (%) Results on

(¢)C. J. BRodden (Editor-in-Chief), Analytical
Chemistry of the Manhattan Project, p. 387,
McGraw-Hili, New York (1950).

(S)H. Freund and F, J. Miner, Determination of

Aluminum in Zirconium Utilizing Ton Exchange
Separation, paper presented at Northwest Regional
Meeting of American Chemical Society, June 20-21,
1952, Oregon State College, Corvallis, Oregon.

192

these tests are inconclusive, but
promising.

Oxygen. The preparation of a
bibliography of recent literature on
bromine trifluoride has been started
in connection with the design and

fabrication of experimental equipment

for work with this reagent. Of special
interest are the physical data, the
inorganic oxide fluorimations, and

the material corrosion information.

Water. Developmental work on the
me thod for determining traces of water
in solids 1n a finely divided state
was completed with the successful
testing of an apparatus that has re-
duced the blank {(that is, before
addition of the samples) to approxi-
mately 15 mg of water., This method
has been reported in more detail 1in
previous reports(?!? and a formal
report will be i1ssued in the near
future.

The concentration of water 1in
anhydrous hydrogen fluoride can be
conveniently determined by measuring
1ts electrical conductivity. A pro-
cedure(®’) has been developed and used
for determining the water content of
the HF used as a cover gas in fuel
preparation. The two cylinders tested
thus far showed 0.5 and 0.35 wt %

water.

DETERMINATION OF BERYLLIUM IN NaK

J. C. White J. E. Lee, Jr.
C. M. Bovyd W. J. Ross
C. K. Talbott

Analytical Chemistry Division

A method for the determination of
beryllium in NaK has been adapted for
evaluating the compatibility of DBeQ,
as a moderator, with NaK. Beryllium
is determined, 1n both the soluble
and insoluble phases following dis-
solution of the NaK, by colorimetric
measurement of the complex formed with

 

p-nitrobenzenecazo-orcinal.(7) This
(6) . . B
(7f. A. Westbrook, private communication.

C. J. BRoddes, op. cit., P. 354,
method can be utilized to determine
0.2 ug per milliliter of solution. A
method was developed for this dis-
solution in which NaK was submerged in
a high-boiling-point inert hydrocarbon
(2,2,5-trimethylheptane) and reacted
slowly by drop-wise addition of methyl
alcohol. This procedure provides a
rapid and safe method for dissolving
alkali metal eutectics.

SPECTROGRAPHIC ANALYSIS

C. R. Baldock
Stable Isotope Research and
Production Davision

For some time it has been recognized
that an extremely high-temperature
ion-source oven was needed for the
investigation of such refractory
materials as the oxides and other
compounds of uranium. A direct-heating
type of oven has been constructed that
makes possible heating of materials to
a temperature that approaches the
melting point of tantalum (2850°C).
The oven 1s small, contains no thermal
insulating material, and 1s thereby
free of the need for troublesome
outgassing. The heater i1s mounted on
a subassembly that i1is easily attached
to and removed from the i1on source;
thus, 1t 1s easy to change samples.
Because of the favorable sample loca-
tion, considerable improvement 1in the
ion-generating efficiency has been ob-
tained, compared with that of previous
equipment. Quite satisfactory analy-
ses can be performed with much less
than 1 mg of sample. |

Uranium Oxides. Three oxides of
uranium, UO,, UOQ,, and U,0,, have
been i1nvestigated 1in considerable
detail. The presence of these materi-
als as 1mpurities in aircraft fuels
or in other uranium compounds can now
be detected with high precision. :

Nickel Fluoride. In an effort to
explain the existence of some non-
stoichiometric compounds of nickel and
fluorine, several tests were made on
NiF, material. The prepared NiF, was

PERIOD ENDING DECEMBER 10, 1952

carried to charge exhaustion 1n an
older socurce oven that reached a tem-
perature of 750°C. The residue of
this test was subsequently examined
at 1100°C 1in the new type of oven
heater, and i1t proved to be nickel
oxide (NiO). Evidence from this
subsequent test proved that above
1100°C the NiO dissociates by thermal
decomposition to nickel and oxygen.
It was of interest that the oxygen
came off as O,. : :

PETROGRAPHIC EXAMINATION OF FLUORIDES

G, D. White, Metallurgy Division
T. N. McVay, Consultant

Routine petrographic examinations
of some 600 samples of fluoride mix-
tures were made in comnection with
fuel investigations. The products of
the ZrQ, -HF reaction and the NaF-ZrOF,
reaction have been examined in some
detail. |

Zr0,-HF. A petrographic analysis
was made of the products of the reac-
tion between Zr(Q, and 48% HF, reacted
at room temperature; two phases were
observed. Chemical analysis gave
compositions that were approximately
those of hydrates of ZrF,. The first
of these phases 1s uniaxial, which
indicates that it is either tetragonal
or hexagenal. The index of refraction
parallel to the ¢ axis is 1.504, ‘and
parallel to the other axes 1s 1.528.
The second phase 1s biaxial negative,
which indicates that it 1s either
triclinic, monoclinic, or orthorhombic.
The crystals are tabular and have a
fibrous structure; cleavage is good.
The birefringence is low; the maximum
interference color 1s a first-order
red. The refractive indexes range
from 1.42 to 1.47 and thus indicate a
variation in the water content of the
material. ' :

NaF-ZrOF,. Sodium fluoride was
reacted with zirconium oxyfluoride
which had been prepared in the labora-
tory. Well-developed, but small,

193
ANP PROJECT QUARTERLY PROGRESS REPORT

ZrO, crystals were formed, and there
was some NazZrF, in the 2Na¥F.ZrOF,
mixture. There was also a very small
amount of another phase, probably NaF,
that had a low index of refraction.
These products would be expected to
form if the following reaction took
place:

2(2NaF'ZrOF2)"““"> NayZrF, + Zr0, + NaF.

The ZrO, crystals, together with
what is thought to be Na,ZrF,, were
formed in the NaF-ZrOF, mixture. This
would be expected from the following
reaction:

2{NaF.ZrOF,) —> Na,ZrFg t Zr0, .

OPTICAL PROPERTIES OF SOME FLUORIDE
COMPOUNDS

T. N. McVay

Consultant, Metallurgy Division

The compounds examined were pre-
pared by V. Coleman of the Materials
Chemistry Division. The x-ray pattern
of the Na,UF, was analyzed by B. S.
Borie, Jr., of the Metallurgy Division.

Na ,UFg. The Na,UF, had the fol-
lowing properties: hexagonal; uniaxial
negative; color, green; O = 1,495 *
0.003; birefringence about (.005.

K;UF,. The crystal system of K;UF,
was not determined, but the material
is orthorhombic, monoclinic, or tri-
clinic, with the following properties:
biaxial negative; 2V about 70 deg;
color, light blue; refractive index of
about 1.414, with low birefringence;
pleochroic, Z = light blue, X = color-
less,

Na ,Zr¥,. The crystal system of
Na,ZrF, was not determined, but the
material 1s either tetragonal or
hexagonal, with the following proper-
ties: wniaxial negative; O = 1,386
0.003; colorless; birefringence low,
not greater than 0.005.

2ZrF,-UF,;. The crystal system of
272rF,.UF,; was not determined, but the
material 1s orthorhembic, monoclinic,
or triclinic, with the following
properties: biaxial; 2V about 90 deg;
color, deep orange-red; refractive

194

index of about 1.576,
fringence.

with low bire-

CHEMICAL ANALYSES FOR UG, IN UF,

J. C. White J. E. Lee, Jr.
C. M. Boyd W. J. Ross

C. K. Talbott
Analytical Chemistry Division

The * ammonium oxalate insolubles”

procedure(®? for the determination of
U0, in UF, was applied successfully to
the determination of UO, in UF,. The
trifluoride is more resistant than the
tetrafluoride to complexing with
oxalate 1on and requires a reflux
period of about 8 hr for a 1- to 2-g
sample. It is postuvlated that the
mechanism of the reaction involves the
slow oxidatian of UF, to UF, and its
subsequent complex formation with
oxalate 1on.

SERVICE CHEMICAL ANALYSES

H. P. House I.. J. Brady
J. R. Lund
Analytical Chemistry Division

In order to determine zirconium 1in
the corrosion test samples to which
titanium had been added either as the
metal or the hydride, 1t was necessary
to use mandelic acid as the precipi-
tating agent, since titanium 1s pre-
cipitated along with zirconium when
phenylarsonic acid is employed.

Recent spectrographic analysis of
the zirconium oxide residues from
ignition of the phenylarsonic acid
precipitates has shown that varying
amounts of arsenic remain with the
zirconium. If conditions for complete
volatilization of the arsenic durang
ignition cannot be established, the
mandelic acid method will be used to
replace the present method for de-
termining zirconiuim.

In order to study the reactions
that take place when NaK 1s introduced
into molten zirconium fuel, a pro-
cedure that i1ncluded measuring the
hydrogen evolved during dissolution

(E)Ibid. » P. 81,

 
of the sample im acid was employed.
It was hoped that this method would
give an accurate measure of the UF,
formed during the reaction with NaK.
However, the results indicate that
other compounds or possibly metals are
formed that evolve hydrogen, since
more hydrogen than can be attributed
to the formation of UF, is often
evolved. The isolation and subsequent
determination of UF; in samples of
this type have not been accomplished,
and further work will be necessary to
develop an adequate method.

There was an increase 1n the number
of samples of beryllium-sodium~uranium
fluoride eutectic analyzed during this
period. Requests for the estimation
of the concentration of the major
constituents of the eutectic made
necessary the adaptation of methods
for determining uranium and berylliium.
Since the uranium concentration was
relatively high (15 to 25%) and beryl-
lium caused no interference, the use
of the Jones reductor with a final
oxidation titration with ceric sulfate
proved adequate for determining this
element. An estimate of the beryllium
content was made by precipitating the
uranium and beryllium together with
ammonium hydroxide, igniting the pre-
cipitates to the oxides, and subtract-
ing the weight of the uranium oxide,
as previously determined by titration,
from the weight of the combined
oxides. The weight of beryllium oxide
is thus obtained by difference. A
separation step, which includes re-
duction of uranium with zinc amalgam
and precipitation with cupferron,
serves to remove the uranium prior to
precipitation of the beryllium with
ammonia; the ignited beryllium oxide
is then free from contamination with
uranium. This method of analysis has
been employed recently for samples_of
beryllium eutectic.

One series of samples of lead was
tested for minor impurities by apply-
ing colorimetric methods after removal
of the lead by electrolytic deposition.

PERIOD ENDING DECEMBER 106, 1952

Samples of alkali hydroxides de-
rived from corrosion tests were ana-
lyzed for iron, chromium, and nickel.
The range of comncentraticn of impuri-
ties in these samples was great enough
to require the use of colorimetric and
gravimetric methods. Since the
material was not homogeneous and could
not be ground to fine particle size,
the entire sample was dissolved to
ensure that the aliquot tested was
representative of the batch.

A few purified alkali hydroxide
samples were tested for carbonate
impurity by reacting the carbon dioxide
evolved with standard barium hydroxide
solution. The excess barium hydroxide
was then titrated with standard hydro-
chloric acuid.

A number of Sodlum samples was
analyzed for the General Electric
Company. Each sample was tested for
sodium monoxide content, and selected
samples of the series were analyzed
for minor 1impurities.

A greater diversity of miscellaneous
samples was tested during the quarter,
but further description of the samples
or details of the methods will not be
attempted in this report.

Of the total of 905 samples ana-
lyzed during the quarter, 617 were
submitted by the Reactor Chemistry
group, 167 by the ANP Experimental
Engineering group, 68 by the ANP
Critical Experiments group, 34 by the
General Electric Company, 17 by the
Flectromagnetic Research group, and
2 by the Ceramic laboratory. The
backlog of samples was reduced during
the period by approximately 50%, as
indicated in the summary, Table 15.1.

 

 

TABLE 15.1. BACKLOG SUMMARY
Samples on hand, 8-10-52 211
Number of samples received 80 2
Total number of samples 1013
Number of samples reported 905
Backlog as of 11-10-52 108

195
 
SUMMARY AND

The list of reports that has been
issued by the Project during the last
quarter includes 14 formal reports and
46 informal documents {(not including
internal documents) on all phases of
- ANP research at ORNL (sec. 16).

A directoryof the research projects

INTRODUCT ION

of the ANP Division of ORNL 1is given
in section 17. Also included are the
Laboratory’s subcontracts to its ANP
Project, as well as the research
projects being performed by ORNL for
the ANP programs of other organi-
zations. |

 

16. LIST OF REPORTS ISSUED

REPORT NO. TITLE OF REPORT AUTHOR(S) DATE ISSUED
I. Experimental Epgineering and Design

Y-F17-19 Performance and Endurance of a Nicrobrazed G. D. Whitman 8-11-52
Stainless Steel Sodium-To-Air Radiator Core
Element

ORNL.-1215 Heat Transfer and Pressure Loss in Tube G. H. Cohen 8-12-52
Bundles for High Performance Heat Exchengers A. P. Fraas
and Fuel Elements M. E. LaVerne

Y-F17-24 Model DA ARE Centrifugal Pump H. W. Savage 8-29-52

Y-F17-25 Dynamic Test of Compatibility of BeO with L. A. Mann 9.2-52
NaK Under Temperature Cycling

Y-F15-11 Investigation of the Fluid Flow Pattern in a R. E. Ball 9-4-52
Model of the “Fireball” BReactor

ORNL - 1287 Preliminary Investigation of a Circulating- R. W, Schroeder {(to be issued)
Fuel Reactor System '

ORNL~ 1330 Heat Exchanger Design Charts A. P. Fraas (to be issued)

Y-F17-28 Valve and Pump Packings for High Tempera- H. R. Johnson 9-15-52
ature Fluoride Mixtures

CF.52-11-156 Report of ARE Design Review Committee T, E. Cole 11-19-52.

Y-918 Potential Lubricants to Meet Extreme Temper- E. P. Carter - 16-29-52"
ature, Pressure and Radiation Demands

IY¥. BReactor Physics

Y-F10-111 Some Solutions of Age Equatibn R. Osborn 11-17+52

Y-F10-112 The Temperature Dependence of a Cross Section R. Osborn 11.17-52
Exhibiting a Resonance ' :

Y-B4-58 Literature Survey of Danger Coefficient Data E. P. Carter 7-11-52

_ I1I. Shielding

CF-52-5-1 Neutron and Gamma Dose Distribution Beyond C. E. Clifford 8-14-52

Part 2 Shield, Parc 2 '

CF-52-T-71 Gamma Ray Spectral Measurements with the F. €. Maienschein 8-8-52
Divided Shield Mockup, Part 11

CF-52-7-83 Gamma Measurements on the GE Reactor Ducts C. E, Clifford

8§~16-52
C. L. Storrs, Jr. :

199
ANP PROJECT QUARTERLY PROGRESS REPORT

CF-52-8-120

Y-F30-8

CF-52-9-99
CF-52-9-145

CF-52-9-161

CF-52-11-143

ORNL-1217

ORNL- 1438

CF-52-8-38

CF-52-10-9

CF-52-11-124

Y-889

CF-52-8-163

CF-52-8-212

CF-52-9-201

CF-52-9-202

CF-52-11-72

CF-52-11-103
CF-52~11-104
CF-52-11-105

CF-52-11-106

CF-52-11-113

ORNL-1342

ORNL.-1395

200

Aircraft Shield Analysis

Radiation Through the Control Rod Pene-
trations of the ABE Shield

Neutron to Gamma Ray Ratio Experiments

Biological Experimentation in Support of
ANP Project

The Development of Lead-Impregnated Moulding
and Modeling Materials for Shielding of
Penetrating Radiations

Basic Shielding Research

Neutron Transmission Through Air Filled
Ducts

Determination of the Power of the Bulk
Shielding Reactor,

Gamma-Ray Spectral Measurements with the
Divided Shield Mockup, Part ITI1

Summary of Lid Tank Neutron Dosimeter
Measurements in Pure Water

Gamma Ray Spectral Measurements with the
Divided Shield Mockup, Part IV

>

.. Wilson
L. Walker

fi. L. F, Enlund

J. L. Meem
J. Furth
E. P. Blizard

E. P. Blizard

. E. Clifford

Johnson
. Meem

&

. Maienschein

V. Blosser

s

IV. Heat Transfer and Physical Properties Research

Selected Physical Properties of Mercury in
the Temperature Range 100 to 1000°C. A
Literature Search

Measurement of the Thermal Conductivity of
Flinak

Physical Property Charts for Some Reactor
Fuels, Coolants, and Miscellaneous
Materials

Cooling Hole Distribution for Reactor
Reflectors

HHeat Generation in a Slab for a Non-Uniform
Scurce of Radiation

Measurement cf the Thermal Conductivity of
Fluoride Mixture No. 33

Heat Capacity of Fused Salt Mixture No. 21
Heat Capacity of LiOH

Density Measurements of Fuel Salts No. 31
an

Viscosity Measurements of Fuel Salts No. 31
and 33

Transient Methods for Determining Thermal
Conductivity of Liquids

Enthalpies and Heat Capacities of Stainless
Steel (316) and Zirconium

Forced Convection Heat Transfer in Pipes
with Volume Heat Sources

F.

L.
S.

Sachs

Cooper
J. Caliborne

ANP Physical
Properties

Group, BEEDivision
W,

W,

i

= 0P A7 F =

S. Farmer

S. Farmer

Gy

Claiborne

Powers
Powers

Redmond
Jones

Redmond
Jones

W Zm =2W T O

Harrison

F. Redmond

Lones

. Maienschein

8-20-52

9-.25-52

9-18-52
9-24-52

9-26-52

11-19-52
11-11-52

(to be issued)

8-8-52

10-1-52

11-17-52

T-10-52

8-26-52

8-28-52

9-3-52

9-4-52

11-11-52

11-15-52
11-15-52
11-15-52

11-15-52

11-1-52

{to be issued)

F. Poppendiek (to be issued)

D. Palmer
CF-52-11-85
ORNL-1433

ORNL-1372

CF-52-10-187

MM-19¢ 1)

CF-52-8-151

CF-52-9-5

CF-52-9-59

CF-52.9-125

CF-52-11-7

CF-52-11-12

CF-52-11-13

CF-52-11-116

CF-52-11-146

ORNL-1354

Y-F33-3
MM-48¢ 1)
Y-F26-44

ORNL-1375

ORNL- 1407

PERIOD

V. Radiation Damage

Effect of Irradiation on BeQO

Preliminary Title List for Blbllography on

Hadiation Effects in Solid Materials
Neutron Spectra from Fluoral Activation

Xenon Problem - ARE

VI. Metallurgy and Ceramics

Metallographlc Examination of Nicrobrazed
Stainless Steel Sodium-Air (No. 2) Heat
Exchanger Fpllow1ng Failure During Test

Unusual Diffusion and Corrosion Phenomena

Observed in Liquid Metal and Molten Caustic

Corrosion Tests

Memorandum on Reflector-Moderator Matrix
Subsurface Void Formation in Metals During
High Temperature Corrosion Tests

Cermet Program

Fabrication of Spherical Particles
Structure of Norton’s Hot Pressed Beryllium
Oxide Blocks :

Fuel Element Program

Mass Transfer in Dynamic Lead-Inconel
Systems

Structure of the BeO Block w1th the 1-1/8
1n. Central Hole

A Compilation of Data on Crucibles Used in
Calcining, Sintering, etc.

VI1I. Miscellaneous

The Solid Phases of Alkali and Uranium
Fluoride Systems

Meteorology and Cllmabology in the 7500
Area

ANP Information Meeting of November 25,
1952 :

Aircraft Nuclear Propulsion Project
Quarterly Progress Eort for Period
Ending September 10,

Aircraft Reactor Experiment anards Summary

Report

 

(I)Nu-her assigned by ANP Réporta Dffice.

ENDING DECEMBER 10,

G. W. Keilholtz

R. . Cleland

J. B, Trice

W, A, Brooksfiank

E. E. Hoffman

A. deS. Brasunas

J. B. Johnson
G. D. White

A. deS. Brasunas

J. R. Johnson

. S. Bomar
. Inouye

v 13

L. M. Doney

J. M. Warde
J. V. Cathcart

L. M. Doney

M., Schwartz

A. G. H. Anderson

U. 8. Weathef
Bureau

W. B. Cottrell

¥. B. Cottreil

J. H. Buck
W, B. Cottrell

1952

11-13-52
11-14-52

{to be issusd)

10-22-52

8-18-52
8-29-59

9-2-52
9-5-52

9-23-52

11-1-52
11-1-52

11-3-52
11-24-52

11-17-52

(to be issued)

10-1-52
10-52
11-15-52

11-5-52

11-24-52

201
ANP PROJECT QUARTERLY PROGRESS REPORT

I'

A,

202

17.

REACTOR AND COMPONENT DESIGN

Aircraft Reactor Design

1.

Do W B

Reflector-Moderated Reactor Studies

Fluid-Fuel Flow Studies

Gamma Heating of the Moderator
Nuclear Ramjet Design Studies
Design Consultants

ARE Reactor Design

.

SO 1 O

.Core and Pressure Shell
Fluid Circuit Design

Pressure and Flow Instrumentation

Structural Analysis

Thermodynamic and Hydrodynamic Analysis

Remote-Handling Fquipment
Shielding and Off-Gas System
Electrical Power Circuits

ARE Control Studies

1.
2.

3.

High-Temperature Fission Chamber
Control System Design

Control Studies on Simulator

ARE Building Facility

1.

Internal Design

ARE Installation and Operation

1.

2.
3.
4.

Instrumentation
Coordination Installation and Design
Expeditors

Control Equipment

Reactor Physics

1.
2.

3.
4.
5.

Analysis of Critical Experiments

Fermi-Age and Boltzmann Equation on Computing
Machines

Kinetics of Circulating-Fuel Reactors
Computation Techniques for ARE-Type Reactors

Computation Techniques for Reflector-Moderated
Reactors

Critical Experiments

1.

2.

ABE Critical Assembly

Reflector-Moderated Beactor Critical Assembly

9704-1

9704-1
9704-1
9704-1
9704-1

9201-3
62061-3

9201-3
9201-3

9201-3
9204-1

9201-3
9201-3
1000

4500
4500

4500

1000

7503
7503
7503
7503

9704-1

2068
9704-1
9704-1

9704-1

9213

9213

DIRECTORY OF ACTIVE ANP RESEARCH PROJECTS AT ORNL

Fraas, lLaVerne
Bussard

Abernathy
Fox, Longyear
Fraas, Stumpf

Wislicenus, JHU
Haines, Wyld, BMI
Chambers,

Hemphill

Cristy, Jackson,
Fckerd, Scott

Hluchan, Affel,
Williams
Maxwell, Walker

Lubarsky, Green-
street

Hutto
Enlund
Walker

Hanauer

Epler, Bates, Ruble,

reen, Mann

Stone, Oakes, Mann

Browning

Hluchan, Affel
Wischhusen, Watts
Perrin, West

Mann

Mills

Edmouson, Coveyou
Prohammer, Frgen
Bengston

Frgen

Callihan, Zimmerman,
Williams, Scott,
Keen

(Same as above)
i.

IX.

A,

PERIOD ENDING DECEMBER 10, 1952

Seal and Pump Development

1. Mechanical Pumps for High-Temperature Fluids

2. High-Temperature Seals for Rotary” Shafts

3. Rocking-Channel Sealless Pump
4. Seals for NaOH Systems

Valve Development

1. Valves for High-Temperature Fluid Systems

Heat Exchanger and Radiator Development

1. High-Temperature Fluid-to-Air Heat Exchange
Test : :

2. Fluoride-to-Liquid Metal Heat Exchange Test

3. Heat Exchanger Fabrication
4. Boeing Turbojet with Na Radiator’

5. Radiator and Heat Exchanger Design Studies

Instrumentation
1. Pressure-Measuring Devices for High-Temperature
Fluid Systems ' :

2. Flow-Measuring Devices for High-Temperature
Fluid Systems

3., Liquid-Level Indicator and Control for High-
Temperature Fluids

High-Temperature Strain Gage
5. High-Temperature Calibration Gage

SHIELDING RESEARCH

Cross-Section Measurements

1. Neutron Velocity Selector

2. Analysis for He in Irradiated Be
3. Total Cross Sections of N** and 0'® (GE)

4. Elastic-Scattering Differential Cross Section

of N'* and 016 (GE)

5. Total Cross Sections of Li%, Li7, Be®, B!?,
Bl2’ C12 .

6. Fission Cross Sections

7. Cross Sections of Be, C, W, Cu
8. Inelastic-Scattering Energy Levels
9. FEffective Removal Cross Sections

Shielding Measurements

1. Divided-Shield Mockup Tests (GE)

9201-3

9201-3

BMI

- BMI

- 9201-3

9201-3

£ 9201-3

9201-3

- %201-3

9201-3

9201-3

9201-3

9201-3

BLH
Cornell

~Aero. Lab.

4500
3005

3026
5500

3500

3500

5500

3001
5500

3001

3010

McDonald, Cobb,
Whitman, Huntley,
Grindell

McDonald, Tunnel,
Ward, Smith, Msason

Dayton
Simmons, Allen

Tunnell, Ward,
MeDonald

Whitman,
-Fraas,

Salmon,
LLaVerne

Salmon, Fraas,
‘Longyear

Fraas

Fraas, Crocker (GE),
Potter (GE) ‘

Fraas, Bailey,
(Tulane Univ.)

Peterson,

Southern, Taylor
Engberg

McDonald, Smith
Taylor, Southern
Trumme 1

Southern

Pawlicki, Smith,
‘Thur low

Parker

Willard, Bair,
Johnson

Johnson, Bair,
Willard, Fowler
Johnson, Willard,
Bair :
Lawphere, Willard
Clifford, Flynn,

Blosser
Willard, Bair,
Kington
Blizard, Flynn,

and crew

Meem and crew

203
ANP

II1.

204

PROJECT QUARTERLY PROGRESS REPORT

Bulk Shielding Reactor Power Calibration
Bulk Shielding Reactor Operation

Heat Belease per Fission

Air Duct Tests - Large Mockups (GE)
Thermal Shield Measurements (DuPont)
Streaming of Neutrons Through Metals

Air-Scattering Experiments

O @~ O LN b N

Primate Exposures to Radiation (School
Aviation Medicine, ORNL, Wright Field,

Convair)

Shielding Theory and Caleculations

1. Survey BReport on Shielding
2. Shielding Section for Reactor Technology

3. Correlation of Bulk Shielding Facility and
Lid Tank Data

4, Calculations on Air-Scattering Experiments in
Bulk Shielding Facility

5. Divided-Shield Theory and Design
Air Duct Theory (GE)
7. Shielding Section for Reactor Handbook

8. Consultation on Radiation Hazards (GE)
9. Shielding Consultant

Shielding Instruments

Gamma Scintillation Spectrometer
Neutron Dosimeter Development

Proton Recoil Spectrometer for Neutrons
He? Counter for Neutrons

LiT Crystals for Neutrons

NN s WO RO

Neutron Spectroscopy with Photographic Plates

MATERIALS RESEARCH

Ligquid Fuel Chemistry

1. Phase Equilibrium Studies of Fluorides

2. Preparation of Standard Fuel Samples
3. Special Methods of Fuel Purification

4. Complex Fluorides of Structural Metals

5. Thermodynamic Stability and Electrochemical
Properties of Fuel Mixtures

6. Hydrolysis and Oxidation of Fuel Mixtures

7. Stability of Slurries of UO3; in NaOH

8. Phase Fquilibria Among Silicates, Borates, etc.

9, Fuel Mixtures Containing Hydrides

10. Chemical Literature Searches

11. Solution of Metals in Their Halides

3010
3010
3010
3001
3001
3001
3010

3010

4500
4500

4500
4500

NDA
3001
4500

2001
Cornell

Univ.

3010
3016
3010
3010
3010
3006

9733-3

9733-3
9733-3

9733-3

9733-3
9733-3

BMI
MHI
9704-1
4500

Johnson, McCammon
Leslie

Meem

Flynn and crew
Flynn and crew
Flynn and crew
Hungerford, Blizard

Meem and Crew

Blizard
Blizard

Blizard

Blizard, Simon,
Charpie

Goldstein
Simon, Clifford

Blizard, Hungerford,
Simon, Ritchie, Meen,
Lansing, Cochran,
Maienschein, Burnette

Morgan
Bethe

Maienschein

Blosser, Hurst, Glass
Cochran, Henry
Cochran

Maienschein, Schenck

Johnson, Haydon

Barton, Bratcher,
Traber, Snell

Nessle, Forgan

Grimes, Blankenship,
Nessle, Blood

Overholser, Sturm

Overholser, Topol
Blankenship, Metcalf
Patterson

Crooks

Banus

Lee

Bredig, Johnson,
Bronstein
12.

13.
14.

15.

16.

17.
18.
19.

PERIOD

Preparation and Properties of Complex
luorides of Structural Elements

Reactions of Fluorides and Metal Oxides
Prgparation, Phase Behavior, and Reaction of
U
3 :
Preparation of Specimens for Radiation Damage

Phase Behavior of Fuels with Added Reducing
Agents

NaK-Fuel Reaction Tests
Free Energy of Fluorides _
Ligquid Fuel Chemistry Consultants

B. Liguid-Moderator Chemistry

Preparatton and Evalwation of Pure Hydroxides

Electrochemical Behavior of Metal Oxides in
Molten Hydroxides

Moderator Systems Containing Hydrides
Hydroxide-Metal Systems

C. Corrosion by Ligquid Metals

9.
10.
11.
12.
13,

Static Corrosion Tests in Liquid Metals and
their Alloys

Dynamic Corrosion Research in Convection Loops
Effect of Crystal Orientation on Corrosion

Effect of Carbides on Liquid Metal Corrosion
Mass Transfer in Molten Metals

Diffusion of Molten Media into Soiid Metals
Structure of Liquid Pb and Bi

Alloys, Mixtures, and Combustion of Liquid
Sodium

Protective Coating§ for Corrosion Resistance
Mass Transfer and Corrosion Inhibitor Studies
Handling of Liquid Metal Samples

Compatibility of BeO and NaK

Compatibility of Materials at High Temperatures

D. Corrosion by Fluorides

1.
2,

Static Corrosion of Metals and Alleys in
Fluoride Salts

Static Corrosion Tests in Fluoride Salts

Fluoride Corrosion in Small-Scale Dynamic
Systems

Dynamic Corrosion Tests of Fluoride Salts

Reactioq of Metals with Fluorides and
Contaminants

E%uilibr@a Between Electropositive and
ransition Metals in Halide Melts:

Corrosion by Fluorides in Stand Pipes

ENDING DECEMBEB 10, 1952

9733-3
9733-3
9733-3

8733-2
9733-2

9201-3
$9201-3

9733-3

1 9733-2

4500

2000
9201-3
2000

2000

2000
9201-3

2000
2000

2000
2000
9201-3
9201-3
9201-3
2000

2000

9766

9766
2000

6201-3
9733-3
3550

9766

Overholser, Sturm
Blankenship, Heoffman

Coleman, Hoffman,
Kelly, Barton

Boody |

Overholser, Redman,
Blankenship, Hoffman,
Kelly :

Mann, Cisar

Mann, Petersen

Gibb, Tufts College
Hill, Duke Univ.
Weekes, Texas ASM
Carter

Overholser, Ketchen

Bolomey, Cuneo
Banus

Bredig, Johnson,
Bronstein

Vreeland, Hoffwman
Adamson

Smith, Cathcart,
Bridges

Vreeland, Hof fman

Cathcart,
Adamson

Smith, Cathcart
Smi th

Smith, Hall
Vreeland
Adamson

Ke tchen

Mann, Cisar
Bomar, Vreeland

Vreeland, Day,
Hof fman

Kertesz, Buttram
Smith, Meadows,
Croft

Kertesz, Buttram,
Croft, Smith, Meadows,
Vreeland, Nicholson,
Hoffman, Trotter

Adamson, Beber

Overholser, Redman,
Powers, Sturm

Bredig, Johnson,
Bronstein

Kevrtegz, Buttram,
Croft

205
ANP PROJECT QUARTERLY PROGRESS REPORT

G.

206

Corrosion by Hydroxides

1.

2,

3.

4,

5.

Static Corrosion of Metals and Alloys in
Hydroxides

Mass Transfer in Mclten Hydroxides
Physical Chemistry of the Hydroxide Corrosion
Phenomenon

Static and Dynamic Corrosion by Hydroxides

Static Corrosion by Hydroxides

Physical Properties of Materials

e W B2

W oo~ 3 oLn
e e x e s

Density of Liquids

Viscosity of Liquids

Thermal Conductivity of Seolids
Thermal Conductivity of Liguids

Specific Heat of Solids and Liquids
Viscosity of Fluoride Fuel Mixtures
Vapor Pressure of Fluoride Fuels
Vapor Pressure of BeF,

Density of BeO Blocks

Evaluation of Materials for High Thermal
Conductivity Suitable for Radiator Fin

Heat Transfer

1.

6.

Heat Transfer Coefficients of Fluoride and
Hydroxide Systems
Heat and Momentum Transfer in Convection Loop

Free Convection in Liquid Fuel Elements

Heat Transfer in Circulating-Fuel Systems
Forced Convection in Annuli

Radiator Analysis

Fluoride Handling

5.

6.

Fuel Production
Design of Special Fluoride Handling Equipment

Small-Scale Handling of High-Temperature
Liquids

Preparation (Experimental) of NaZrF

Preparation of Enriched Fuel for Radiation
Studies

Fluoride Sampling Techniques

Liguid Metal Handling

1.
2.
3.

Sampling Techniques
NaK Distillation Fquipment
NaK Handling

Dynamic Liguid Loops

1,
2.

Operation of Convection Loops
Operation of Figure-Eight Loops

2000

2000

2000
9766

BM1

9204-1
9201-4
8204-1
9204-1

9201-4
9766
9733-2
BMI
9766
2000

5204-1
9204-1

9204-1
9204-1

920]1-4

9204-1

9201-3
9733-3

9733-2

9733-2

9733-3
9201-3

9201-3
9201-3
9201-3

9201-3
9201-3

Vreeland, Day,
Hof fman

Vreeland, Smith,
Cathcart

Cathcart, Smith

Kertesz, Croft,
Buttram, Smith

Jaffee, Craighead

Jones, Redmond
Jones, Redmond
Powers, Burnett

Claiborne, Powers,
Burnett

Powers, Blalock
Kertesz, Knox
Barton, Moore
Patterson, Clegg
Doney

Bomar, Slaughter,
Oliver

Hoffman, lLones

Hamilton, Palwmer,
Lynch, Poppendiek

Hamilton, Lyrnch

Poppendiek, Winn,
Palmer

Bradfute, Harrison,
Poppendiek

Farmer

Nessle, Eorgan

Grimes, Blankenship,
Nessle

Blood, Thoma,
Weinberger

Blankenship and crew

Baody, Blankenship
Mann, Blakely

Mann, Blakely
Mann, Cisar

Mann

Adamsan
Coughlen
3.
4,

5.

PERIOD ENDING DECEMBER 10,

UOQ,«NaOH Slurry Loop

Operation of Thermal Convection Loops

Fluid Circuit Mockups

K. Materials Analysis and'lnspection Me thods

1.

N

O o =1
. e

10.
11.
12,

13.
14,

15.

16.

17.

18.
19.

Determlnatlon of Metalliec Corroqlon Products
in Proposed Reactor Fuels

Analysis of Beactor Fuels

Determination of Reduced Species in Reactor
Fuels Following Reaction with NaK

Determination of Oxides in Reactor Fuels
Determination of Water in Hydrogen Fluoride

Determination of Efficiency of Removal of
NaK by Distillation from ARE

High-Temperature Mass Spectrometry
X-Ray Study of Complex Fluorides
Petrographic Examination of Fuels
Chemical Methods of Fluid Handling
Metallographic Examination

Identification of Compounds in Solidified
Fuels

Preparation of Tested Specimens for Examination

Tdentification of Corrosion Products from
Dynamlc Loops

Assembly and Interpretation of Corrosion Data
from Dynamic Loop Tests

Determination of Combined Oxygen in Fluoride
Mixtures

Metallurgical anmlnab1on of Englneerlng
Parts

Analyses of Na and NaK for Na,0
Ydentification of Corrosion Products

L. Radiation Damage

1.

Liquid Compound Irradiation in LITR

Liguid Compound Irradiations in Cyclotron
Liquid Compound Irradiations 1in MTR

Ligquid Metal Corrosion in ORNL Graphite
Reactor loops

Stress Corrosion and Creep in LITR Liquid
Metal Loops

Creep of Metals in OBNL Graphite Reactor
and LITR

Thermal! Conductivity of Metals in ORNL
Graphite Reactor and LITR

Neutron Spectrum of LITR
Radiation~-Damage Consultants

BMIY
2000

9201-3

9733-4
9733-4

9733-4
9733-4
9733-4

9201-3
9735
3550
9766
9201-3
2000
9733-2

9733-2

9733.3
2000

9733-3

9201-3
9201-3

9201-3
2000

3005

9201-3
NRTS

3001

3005

3001
3025

3001

3005

3005

1952
Simons
Cathcart, Bridges,
Smith
Mann

White, Talbott
White, Ross

White, Talbott
White, Lee
Whive, Talbott

White, lLee

Baldock

Agron

McVay, White

Mann, Blakely

Gray, Krouse, Roeche

Barton, Hoffman
Kelly

Meadows, Didlake

Hof fman, Blankenship,
Smith, Borie, Dyer

Adamson, Blankenship,
Smith, Vreeland,
Blakely

Vhite, Lee

Adamson, Vreeland,
Gray ,

Mann
Dyer, Borie

Keilholtz, Morgans
Webster, Kinyon,
Davies, Feldman,
Bobinseon :

Sturm, Jones, Binder

Keilholtz, ¥lein,
Morgan, Bacht,
Robertson

Sisman, Bauman,
Carroll, Brundage,
Blacksher, Parkinson,
Ellis, Dlsen, James,
Mor gan

Sisman, etc.,
(as above)

Wilson, Zukas,
Davis

Cohen, Templeton
Sisman, Trice

Ruark, JHU,
Smith, Cornell Univ.

207
ANP PROJECT QUARTERLY PROGRESS REPORT

208

Strength of Materials

1,
2,

3.

Creep Tests in Fluoride Fuels

Creep and Stress-Rupture Tests of Metals in
Vacuum and in Fluid Media

High-Temperature Cyclic Tensile Tests
Tube Burst Tests
Tube Burst Tests

Relaxation Tests of Reactor Materials

Creep Tests of Reactor Materials (GF)
Fvaluation Testing for Other Development
Groups

Designing and Evaluating New Testing Methods
Fatigue Testing of Reactor Materials

Metals Fabrication Methods

SY O B WM

-3

10.
11.
12,
13.
14.
15.

New

1.
2.
3.

Welding Techniques for ARE Parts

Brazing Techniques for ARE Parts

Mol ybdenum Welding Research

Molybdenum Welding Besearch

Resistance Welding for Mo and Clad Metals

Welds in the Presence of Various Corrosion
Media

Nondestructive Testing of Tube-to-Header Welds

Basic Evaluation of Weld-Metal Deposits in
Thick Plates

Evaluation of the Cone-Arc Welding Technique

Development of High-Temperature Brazing
Alloys

Evaluation of the High-Temperature Brazing
Alloys

Brazing of Testing Components for Evaluatioen

Application of Resistance-Welding Methods
to ANP Materials

Welding Comsultant

Fabrication Procedures for Nickel and Its
Alloys

Metals Development

Me and Cb Alloy Studies
Heat Treatment of Metals
Alloy Development of New Container Materials

Solid Fuel Element Fabrication

1.
2.
3.

Solid Fuel Element Fabrication
Diffusion-Corrosion in Solid Fuel Flements

Determination of the Engineering Properties
of Solid Fuel Elements

Electroforming Fuel-Tube-to-Header
Configurations

Electroplating Mo and Cb

Carbonyl Plating of Mo and Cb

9201-3
2000

2000
9201-3
2000

2000
2000
2000

2000
2000

2000
2000
BMI
MIT
RPI

2000
2000

2000

20600

Wall-
Colmonoy

2000

2000

2000

International

2000
2000
2000

2000
2000
2000

Gerity
Mich.

Gerity
Mich.

2000

Adamson
Oliver, Douglas,
Weaver
Oliver, Woods
Adamson
Oliver, Woods,
Weaver
Oliver, Woods,
Weaver
Olivexr, Woods,
Weaver
Oliver, Woods,
Reber
Oliver, Beber
Oliver, Reber
Slaughter

Housley, Slaughter
Parke

Wulff

Nippes, Savage

Vreeland, Slaughter
Slaughter

Slaughter, Gray
Slaughter
Peaslee

Slaughter

Slaughter, Petersen,
Fraas

Slaughter
Patriarca

Nickel Company

Incuye
Coobs

Bomar,

Inouye

Bomar, Coobs

Bomar, Coobs

Bomar, Coobs, Woods,
Oliver, Douglas

Graaf

Graaf

Bomar
PERIOD ENDING DECEMBER 10, 1952
7. Fuel Spherical Particle Production {P&W)
8. 1Inspection Methods for Fuel Elements 2600 Bomar, Levy
9. Carbonyl Plating Commonwealth Eng. Corp.
1¢. Production of High-Temperature Screens Gfigifiy Graaf
- ich.
Q. Ceramics and Metals Ceramics
1. Metal Cladding for BeO Gerity Graaf
‘Mich,

2. Control Rod Development 2000 Bomar, Coobs
3. Hot Pressing of Be#ring Materials 2000 Bomar, Coobs
4., High-Temperature Firing of Uranium Oxide to

Produce Selective Powder Sizes 2000 Bomar, Coobs
5. Development of Cermets for Reactor Components 9766 J%hnfon, Shevlin,

: _ aylor

6, Ceramic Coatings for ANP Radiator: 9766 White, Griffin
7. Ceramic Valve Parts for Liguid Metals and oSy Shevlin, Johnson,

Fluorides Taylor
8. Production of Control Rod Samples :(GE) for ~ :

Testing 2000 Bomar, Coobs
9. Ceramic Reflector 9766 Johnson, White
10, Ceramic Coatings for Shielding 9766 Yhite, Shevlin

OSU
11. High-Temperature Ceramic Coatings H?tpoint,
: - Inc.

1Vv. TECHNICAL ADMINISTRATION OF AIRCRAFT NUCLEAR PROPULSION PROJECTY

AT OAK RIDGE

PROJECT DIRECTOR

 

*
Dual Capacity.

NATIONAL LABORATORY

R. C. Briant*
ASSOCIATE DIRECTOR FOR ARE J. H. Buck*
ASSISTANT DIRECTOR FOR COORDINATION A. J. Miller*
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 11 AB,C,D
Reactor Physics W. K. Ergen* 1 E
Critical Experiments A. D. Callihan I G
Nuclear Measurements A. H. Snell I1 A
STAFF ASSISTANT FOR RADIATION DAMAGE A. J. Miller*
Radiation Damage D. §. Billington I11 L
STAFF ASSISTANT FOR GENEBAL DESIGN A. P. Fraas*
General Design A. P. Fraas* I A
STAFF ASSISTANT FOR ARE J. H. Buck*
ABE Operations E. S. Bettis 1 C,D,E
ARE Design R. W. Schroeder 1 B,D
Experimental Engineering. H. W. Savage I §r§,}.§: IfI C,D,

209
STAFF ASSISTANT FOR ENGINEERING RESEARCH
Heat Transfer and Physical Properties
Ceramics

STAFF ASSISTANT FOR METALLURGY
Metallurgy

STAFF ASSISTANT FOR CHEMISTRY
Chemistry
Chemical Analyses

 

*Dual Capacity.

210

n== == -

o®® ST Em oD

Briant*
Poppendick
Warde

Manly*
Manly*

Grimes*
Grimes*

Susano

I11
111

11

III
III

F,G

c,D,E,J,K,M,N,0,P,Q

AB,D,E,F . HK
K

*i