ORNL-1771.txt

MARATIN MARIETTA ENERGY SYSTEMS LIBRARIES

3 445k 0250997 O

e
-

OLLECTION

 
ORNL-1771
T his document consists of 197 pages.

Copy? 7 of 254 copies. Series A,

Contract Neo. W-7405-eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT
For Period Ending September 10, 1954
W. H. Jordan, Director
S. J. Cromer, Co-Director
R. I. Strough, Associate Director

A, J. Miller, Assistant Director
A. W, Savolainen, Editor

DATE ISSUED

OCY 25 1984

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MARTIN MARIETTA ENERGY SYSTEMS LISRARIES
A Division of Union Carbide and Carbon Corporation

OAK RIDGE NATIONAL LABOQRATORY |
i
Pos?'foice Box P |
Ouk Ridge, Tennessee 3 445k DESD:l:l? 0

CARBIDE AND CARBON CHEMICALS COMPANY

 
1. G. M. Adamson
2. R. G. Affel

3. C. R, Baldock
4, C. J. Barton
5. E. S. Bettis

6. D. S. Billingten
7. F. F. Blankenship
8, E. P. Blizard
9. G. E. Boyd

10, M. A. Bredig
11. F. R. Bruce
12. A. D. Callihan
13. D. W. Cardwell
14, C. E. Center
15. R, A, Charpie
16. G. H. Clewett
17. C, E. Clitford
18. W. B. Cottrell
19. R. G. Cochran
20. D. D. Cowen
21. S. Cromer

22, F. L. Culler
23. L. B. Emlet (K-25)
24, A, P, Fraas
25. W. R, Grimes
26, E. E. Hoffman
27. A. Hollaender
28. A. S. Householder
29. J. T. Howe

30. R. W. Johnson
31, W. H. Jordan
32. G. W. Keilholtz
33. C. P. Keim

34, M. T, Kelley
35. F. Kertesz

346. E. M. King

37. J. A, Lane

38, C. E. Larson
39. R.S. Livingston
40. R. N. Lyon

41. W. D. Manly

 

ORNL-1771
Progress

INTERNAL DISTR}BUTION

42,
43.
44,
45,
46,
47.
48.
49,
50.

51,
52.

53.

54,

55.
56.
57.
58,

59.

60.
61.

62.
63.

64.
65.
66.
67.
68.
69.

70,
71,
72.

73.
74.
75-84.
85.
86-90.

91.

92.
93.
94,
95.97.

 

. A, Munn

. B. McDonald

L.. Meem

. J. Miller

. Z. Morgan

. J. Murphy

P. Murray (Y-12)

. J. Nessle

. Patriarca

. Fi Poppendiek

. M. Reyling

. W. Savage

. W, Savolginen

. D. Shipley

. Sisman

. P, Smith

. P. Smith (consultant)

. H. Snell

. L. Storrs

. D. Susano

A. Swartout

. H, Taylor

B, Trice

. R. Van Artsdalen

. C. VonderLage

M. Warde

. M, Weinberg

C. White

G. D, Whitman

E. P. Wigner (consultant)

G. C, Williams

J. C. Wilson

C. E. Winters

X-10 Document Reference Library (Y- ]2)
Biology Library

L aboratory Records Department
Laboratory Records, ORNL R.C.
Health Physics Library
Metallurgy Library .
Reactor Experimental Engineering L|brdry
Central Research Library ~

&>&mm&m&fin>rOom>IvItO%mK>&ér
98.

99.

100.
101.
102.
103.
104-115.
116.
117.
118-122.
123.

124.
125.
126-128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138-142.
143.
144,

145.
146-148.
149.
150-157.
158.
159,
160.
161-162.
163-166.
167-168.
169-170.
171,
172.
173.
174.
175-178.
179.
180-181.
182.
183.
184-185.
186-188.
189.
190-196.

 

EXTERNAL DISTRIBUTION

Air Force Engineering Office, Oak Ridge

Air Force Plant Representative, Burbank

Air Force Plant Representative, Seattle

Air Force Plant Representative, Wood-Ridge

American Machine and Foundry Company

ANP Project Office, Fort Worth

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

Armed Forces Special Weapons Project, Washington (Gertrude Camp)
Atomic Energy Commission, Washington (Lt. Col. T. A, Redfield)
Babcock and Wilcox Company

Battelle Memorial Institute

Bendix Aviation Corporation

Brookhaven National Laboratory

Bureau of Aeronautics (Grant)

Bureau of Ships

Chicago Patent Group

Chief of Naval Research

Commeonwealth Edison Company

Convair, San Diego (C. H. Helms)

Curtiss-Wright Corporation, Wright Aeronautical Division (K. Campbell)
Department of the Navy — Op-362

Detroit Edison Company

duPont Company, Augusta

duPont Company, Wilmington

Duquesne Light Company

Foster Wheeler Corporation

General Electric Company, ANPD

General Electric Company, APS

General Electric Company, Richland

Glen L. Martin Company (T. F. Nagey)

Hanford Operations Office

lowa State College

Kirtland Air Force Base

Knolls Atomic Power Laboratory

Lockland Area Office

Los Alames Scientific Laboratory

Materials Laboratory (WADC) (Col. P. L. Hill)

Nuclear Me'm!s, Inc.

Monsanto Chemical Company

Mound Laboratory

National Adyisory Committee for Aeronautics, Cleveland (A. Silverstein)
National Advisory Committee for Aeronautics, Washington
Naval Research Laboratory

Newport News Shipbuilding and Dry Dock Company

New York Operations Office

North American Aviation, Inc.

Nuclear Development Associates, Inc.

Patent Branch, Washington

Phillips Petroleum Company (NRTS)

 
197-198.
199-208.
209-210.
211.
212.
213.
214.
215.
216-217.
218-219.
220.
221-226.
227-238.
239-253.
254.

 

Powerplant Laboratory (WADC) (A. M. Nelson)

Pratt and Whitney Aircraft Division (Fox Project)

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

San Francisco Field Office

Sylvania Electric Products, Inc.

Tennessee Valley Authority (Dean)

USAF Headquarters

U. 5. Naval Radiological Defense Laboratory

University of California Radiation Laboratory, Berkeley
University of California Radiation Laboratory, Livermore
Walter Kidde Nuclear Laborotories, Inc.

Westinghouse Electric. Corporation

Wright Air Development Center (WCSNS, Col. John R. Hood, Jr.)
Technical Information Service, Oak Ridge

Division of Research and Medicine, AEC, ORO

 
vi

ORNL-528
ORNL.-629
ORNL-768
ORNL-858
ORNL-919
ANP-60
ANP-65
ORNL-1154
ORNL-1170
ORNI.-1227
ORNL.-1294
ORNL.-1375
ORNL.-1439
ORNL-1515
ORNL.-1556
ORNL.-1609
ORNL-1649
ORNL-1692
ORNL-1729

 

Reports previously issuved in this series are as follows:

Period Ending November 30, 1949
Period Ending February 28, 1950
Period Ending May 31, 1950
Period Ending August 31, 1950
Period Ending December 10, 1950
Period Ending March 10, 1951
Period Ending June 10, 195]
Period Ending September 10, 1951
Period Ending December 10, 1951
Period Ending March 10, 1952
Period Ending June 10, 1952
Period Ending September 10, 1952
Period Ending December 10, 1952
Period Ending March 10, 1953
Period Ending June 10, 1953
Period Ending September 10, 1953
Pericd Ending December 10, 1953
Period Ending March 10, 1954
Period Ending June 10, 1954

 
 

FOREWORD

This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL records the
technical progress of the research on Circulating-Fuel Reactors and all other ANP research at
the Laboratory under its Contract W-7405-eng-26. The report is divided into three major parts: |.
Reacter Theory, Componenf Design and Testing, and Construction, Il. Materials Research, and
[l. Shielding Research.

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

The effort on circulating-fuel reactors was, until recently, centered upon the Aircraft Reactor
Experiment. The equipment for this reactor experiment has been assembled, and the current
status of the experiment is summarized in Section 1 of Part |.

The design, construction, and operation of the Circulating-Fuel Reactor Experiment (CFRE),
with the cooperation of the Pratt & Whitney Aircraft Division, are now the specific long-range
objectives. The CFRE is to be a power plant system that will include a 60-Mw circulating-fuel
reflector-moderated reactor and an adequate heat disposal system. Operation of the system will

be for the purpose of determining the feasibility and the problems associated with the design,

construction, and operation of a high-powered reflector-moderated aircraft reactor system. The

design work, as well as the supporting research on materials and problems peculiar to the CFRE

(previously included in the subject sections), is now reported as a subsection of Section 2,

¥

““Circulating-Fuel Reflector-Moderated Reactor.’

The ANP research, in addition to that for the Circulating-Fuel Reactor Experiment, folls

into three general categories: studies of aircraft-size circulating-fuel reactors, materials problems

associated with advanced reactor designs, and studies of shields for nuclear aircraft. These"

phases of research are covered in Parts |, H, and lll, respectively, of this report.

 

vii
 

 
 

CONTENTS

FOREWORDHOII.II..II. ...... # 8% & % & B A E & % 2 B A& S B 0% E 8 o8 & & & ¥ & & 5 2 8 F F &5 8 TN Vii
SUMMARY & o0 vvvn s, e DI B

1.

PART I. REACTOR THEQORY, COMPONENT DESIGN AND TESTING, AND CONSTRUCTION

CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT ..... e et e e e e 9
The Experimental Reactor System . .o v it e ittt iinnnnstanans e s n e e s 9
Characteristics of the Fuel and the Sodium Systems During the Water Tests ... ... o0y, 10
Preparations for Loadingthe ARE .. .. it i ittt i i it i s e onanso 1
Fuel-Concentrate injection Nozzle , . v . .. e e s s e e s s e e e 12
ARE Unloading Experiment. . « o v i v v oo s osanoosnenssassosnnse C et e e ceea. 12
ARE Fuel Recovery and Reprocessing « v v v e v st s sunouvasnsnsasanaasasnssssnnas 12
ARE Pumpss e v v s v o v us te et et ar e £ s ee b ias s s st A s et a0aa st e N an 14
Fabrication and Testing. « « v v s v v v s esosassnes C e e L i e e e s E s e 14
Seal-Gas Pressure-Balancing System . .o v v v ittt ittt e i i ittt it e 14
Radiation Damage to Pump Drives . v v v o v v v e v v oo e N e e e e 14
Radiation Damage to Shaft Seal . . .. .. i n e e et e r st e e s ettt e e b e 15
MOtor TSt v v m v o v v v o a s o s s o asnsconososoesonesssosssonsesssss P s e 15
Reactor System Component Test Loop. v v vt vt i v e ie i it e s neesataessanononssan 16
Operation of Loop. . ... e et e s e e e s s e n e s e 16
Examination of Hairpin Tube. « v v v vttt ittt v enasnoosaressonnososcnonasss 16
Reactor Physics v v v v v v v e v e e et s e st s et e e e h s ae e 16
REFLECTOR-MODERATED REACTOR ,..... T i e s s e 17
Designof the CFRE . .. .. ittt ittt ininans C e i e e s s s e e 17
CFRE Component Development Projects + o v v v et it i snn ottt etonnnennoanasenon 8
Reactor Physics. o v v s s v os s aisrnsnnotisontansnsoneraassssnnsosssansnsss 24
Reactor Calculations v v v v a v v v v sn v nn e Canre f e i a e s e s i ettt e e 24
Beryllium Thermal Stress Test . . v i it ittt ittt nssaaras s eena et eana 24
Beryllium-Inconel-Sodium Systems . . ..o v vttt i i i Cra s e e s e 29
Beryllium-Inconel-Sodium Compatibility Tests. « . ¢ . . . & b e h s e s s Ch e e 29
Mass Transfer Tests in Thermal-Convection Loops ... .. .. Ceeaas s h s e e e e 30
Mass Transfer Tests in Forced-Circulation Loops + v v v v s st vt evonnnvan e sonnans . 32
EXPERIMENTAL REACTOR ENGINEERING 4 v v e v vt v s v s v nsnnsantnoosnsansnanss 34
In-Pile _oop Component Development . . ..o v v v v v v e b e e e s e e s e as s ceen. M4
Horizontal-Shaft Sump Pump . 4 v ottt i ittt st v s s s e nsaanonas Car e e 34
Vertical-Shaft Centrifugal Sump Pump « v v v v s v i it i it i i st ii et i s i e e o s as 37
Hydraulic Motor Pump Drives + v v vvii i i oo e h e s n e e s eea e ae A es 38
Forced-Circulation Corrosion Loops v v v v v v o0 v v e au s s e ae e et e e e e 38
Inconel Loops « v vt vttt et s m e s santsasesoncssnsosenesnssenss L e e e e e 38
Dissimilar Metal Loops « v v v e e v v v e e e o e e e cenacssee s e 41
Gas-Furnace Heat-Source Development . ... oo vovvi v Cee e et e b e e 41
Study of the Cavitation Phenomenon .« o v v v vt it i i ittt e et osnasaronansneses 41
Expansion and Improvement of Thermal-Convection Loop Testing Facilities . .. ..o 0 v v vn v 42
CRITICAL EXPERIMENTS . . v v i et v as L et e e v e e e st e e s 44
Reflector-Moderated Reactor « v v o v v v s v s s s v s s nanessnnsonsoe s e e e s e R .
Supercritical-Water Reactor. « v o v v v vt v s eaeeneesi et e s s et e e .. 47

 
 

PART Il. MATERIALS RESEARCH

5. CHEMISTRY OF MOLTEN MATERIALS .. ...... e R X |
Solid Phase Studies in the NaF-ZrF -UF , System .o oivvin ooy e M
Visual Cbservation of Melting Tflmperatur&s in the NaF-UF Sys'rem .................. . 35
Phase Relationships in UF -Bearing Systems. ... ... Gt e a e e es s esnnes e ¢ haanes 56

UF3iinZrFA-BeqrinQSystems.. ......... ettt e e e e e I
UF, in NaF-KF-LiF Mixtures «oo0vvve it e e e Y
UF, in NaF-RbF-LiF Mixtures . ....... .o i e e crr e 58
UF, in the Individual Alkali-Metal Fluorides .« v vvcviiv v i, 58
UF; in Binary Alkali-Metal Fluoride Systems .. ..o v, P 60
Purification of Rubidium Fluoride . ... oo i iv i ettt 60
Chemical Reactions in Molten Salis + . ... .. ... et et chee e verrseens 60
Chemical Equilibria in Fused Salts v v v v i i ittt ittt st anoss oo ssssaeas 60
Stability of Chromium Compounds in Molten Fluerides ........... < X
Reduction of NiF, by H, in NaF-ZrF,  Systems. . ..o vvviv i e N«
Reduction of FeF, by H, in NaF-ZrF, Systems ..o uvn. et e e e 64
Preparation of Various Fluorides v v v v vttt v it i it it i st et i nansaoens 67
Fundamental Chemistry of Fused Salts v v v v i v v i vt vnenneasnsnrnesrnoesnssnsess O
EME MeaSUrements « o v v v o s v oo vt o o vt oot s aoosussasennsesaneanennsssssess 67
Solubility of Xenon inMolten Salts. . . v v v v i v vt it e it e i ittt e i e i e 70
X-Ray Diffraction Studies in Salt Systems . v iviiiiiiriornnsersnsnnsesssssaess 70
Physical Chemistry. v o v o v v v v oo n o e et et e et e e e 72
Production of Purified Molten Fluorides . . o i vttt i ittt v in it en et oo s e 72
Use of Zirconium Metal as a Scavenging Agent + v v v v v v v v e e C e e e 72
Purification of NaF-ZrF, Mixtures by Electrolysis « oo vvvviviiiivacnen P e e /3
Preparation of UF3wBenring Fuels oo i v oo i n ot h e e e e Gt e e e 77
Alkali-Metal Fluoride Processing Facility. « v v v v a0 v v b e e e e seeeee. 78
Production Facility. v v v s v v v s v i v e i e oo et e e et e e 79
In-Pife Loop Loading v v v vt vt v v vnoennnns e e et et anes 19
Chemistry of Alkali-Metal Hydroxides . . . v v v v v v v v et e e e s ceresees 80
Purification. « v v v vt i vt i e et e e an et e e n s e r e e . 80
Reaction of Sodium Hydroxide withMetals . . ..o v oo i it i e ceseeras. 80

8. CORROSION RESEARCH v v it i ittt s i e nnse oo C et a et 81

Static and Seesow Corrosion Tests . v v v v v v e v ot v s v o s et e e e e e 82
Brazing Alloys in NaF-Zrt ,-UF  and in Sodium ... ... Ce e T - 7.
Special Stellite Heats in NaF ZrF -UF, and in Sodium. .. oo e T - <
HasfelloyRinVariousMedio........... ............. C e ne e e st e 85
Incone!l in Molten Rubidium., v v oo v o e v vt Ce e e et et e e e e ae e 86
Carburization of Inconel by Sodium. « v v v v v v v i v v v v v e P -1
Special Tar-Impregnated and Fired Graphite . v v . . .. .. L
Ceramics in Various Media oo ittt vn e I

Fluoride Corrosion of Inconel in Thermal Convection Loops ... ... Cen e e e e 95
Effect of UF, in Z¢F ,-Base Fuels . ... ... f e et e e s e e 95
Effect of UF, in Alkali-Metal Base Fuels .. ..o vvvvv vy e e e ceeraaeaae. 96
Effect of Zirconium Hydride Additions to Fuel .. . oo v v v ittt it iiainon 98
Effect of Uranium Concentration « v v v v v v v v o v s oen oo e e 98
Effect of Inconel Grain Size . ... ... C e st s e rn e es e st e e s e e s e e e .. 98

Fluoride Corrosion of Hastelloy B in The rmc:l Convechon lLoops v vii v e vieeanes 98

Lithium in Type 316 Stainless Steel « v v i vt it it it i i i et ettt ettt a e 99

Fundamental Corresion Research o v o v v v v v i v v i v i v o s Gt s e 100

 
 

Mass Transfer in Liquid Lead . . o v v e v i n s v v v e

 

100

Flammability of Sodium A”oys e s st e s e s et . e h e et e e . . 102
Thermodynamics of Alkali-Metal Hydrox:des..........,....................... 102
Chemical Studies of Corrosion. « v v v v v v ettt vt i e erenan Cie v aneaseses 108
Effect of Temperature on Corrosion of Inconel ond Type 316 Stainless Steel . . .. ... vee... 108
Corrosion by Fission Products . ..o i i i en st st ennnesons s s 109
Controlled-Velocity Corrosion Testing Apparatus . .. . . e e e erenas 109
Reaction Between Graphite and Fluoride Melt. . v o v . i i e vt e it nnenenenanasaa. 110
Lithium Fluoride Castings ...... e s r e e e e P 110
. METALLURGY ... .. iiiinineenan, e e s s e senavaranas 111
Stress-Rupture Tests of Inconel . . . .. .. v v v v co ceee . bee coes 112
High-Conductivity-Fin Sodium-to- Alr Rudsator s a e e s . veeraeseass 116
Investigations of FinMaterials + « v v v v s e i in i it i s nennesrennonooenanssnseess 117
Development of Brazing Alloys for Use in Fabricating Radiators . v v v 4. .. cha e 118
Radiator Fabrication .+ . . . . G r e e e e “a e e a e e e . 120
Nozzles for the Gas-Fired qumd Mefal Heoter System .« . v v s O (0]
Special Materials Fabrication Research. . ..... ch s e I vaw e 121
Stainless-Steei-Clad Molybdenum and Columbium. e h et st e e e e 121
tnconel-Type Alloys + v vv v v e u v v e e e e e s e e e s e 122
Nickel-Molybdenum-Base Alloys .+ . v v e v vt vt o vt netevnonnnnaen e e et 122
Duplex Tubing o v o vt e sttt a s oo noneeesanosanotonsnonsnsossnennssesss 123
Sigma-Phase Alloys . . . v ittt i vt ittt ttoncanonseenonss e 124
Boron Carbide Shielding. v v v 0 v s v v v v v v v vt e . Cee s e e e 124
Tubular Fuel Elements & . o it i i it ittt it i it ittt et et entnesasanan 124
Metallographic Examination of a Fluoride-to-Sodium Heat Exchenger . ... ... ... caeees 124
. HEAT TRANSFER AND PHYSICAL PROPERTIES . .o e . ‘e 127
Physical Properties Measurements « v v o v v v v st o st v neaasens e RPN e e 127
Heat Capacity v v o v oo v o st v i s s s o nnoessonaosassssnassssrensnnse . 127
Density and Viscosity v o v o v st st s vn oo tasessnsonnesosssnss . A VY
Thermal Conductivity & v s o v e v it s v i e st s o s v aononsroorsorosssesssanosanes 129
Electrical Conductivity . oo v v v et a s e e e . . e . . 129
VapPOr PressSure o o v v o o o s s 0 s 60t asiesorenonssenses c. . e h e e . . 129
Fused-Salt Heat Transfer v « v v v o vt v v s e st tnonesaonssontosnoasonsssannoasss 130
Transient Boiling Research . ...... C e s e et e e e 130
Fluid Flow Studies for Circulating-Fuel Reactors . .. ........ ettt et 131
Heat Transfer Studies for Circulating-Fuel Reactors .. v v v v v vt o vt s e v sennannasas 131
. RADIATION DAMAGE + v v v v s v vt s i et st st et o et ensonnsnnesanonansness . 134
MTR Static Corrosion TestS « v v s evesssasnssssssnsnossosssssesarenaasssanes 134
Fission Product Corrosion Study + o v cu e v ve v on v Ces se e e e creeaes 135
Facilities for Handling Irradiated Capsules & o v v v v v v s v i e it s vt e s s st ansnonnasons 135
Analysis of lrradioted Fluoride Fuels for Uroniom . . v oo o0 . caes . . . . 136
High-Temperature Check Valve Tests o v v v v v v v v Pean e s e e . 137
Miniature In-Pile Loop — Bench Test v v v o s vt o i entntnntatossasososnnssssases 137
Life Tests on an RPM Meter, Bearings, and a Small Electric Motor Under lrradiation. . . . . s ees 137
Removal of Xenon from Fused Fluorides « v v v v s v v 0o e e e e e sa s s e aaaas 140
CoF UF Imadiation, v u v v s ss v i st it o s i i nnans v o ea s ... 140
LITR Horrzonmf Beam-Hole Fluoride-Fuel Loop s « v v v v v v v e v v s s et nonnenns e ea 141
ORNL Graphite Reactor Sodium-Inconel Loop ...... e s e eas 142
Creep and Stress- Corrosnon Tests v o v nn et tanenan e e . . ere. 142

xi
 

Remote Metallography v v i v i v o v i ittt te it et s e s ne st o onnasnnsstsanassssnas 144

Fission-Fragment Annealing Studies . v v v v s v v v v et e s e e e 145
High-Temperature, Short-Time, Grain-Growth Characteristics of Inconel. .. v v oo v v v i v v in 145
BNL Neutron Spectrum ~ Radiation Damage Study « v o0 v i it it i it ittt i e e 145
MTR Neutron-Flux Spectra — Radiation Damage Study . v v v v v s v v st v st st i v c oo 147
10. ANALYTICAL STUDIES OF REACTOR MATERIALS ... i ittt ittt e s s annnas 148
Analytical Chemistry of Reactor Materials o v v v v i i it e i i i it e it i s i i e st enaan 148
Determination of Oxygen in Fluoride Fuels. . . v o i ittt it i i et it e v i s i i s aenn s 148
Oxidation-Reduction Titrations inFused Salts . . . oo v v s i i it i i i i e i en e 150
Polarographic Studies in Fused Ammonium Formate + v v v v v i v vt vt i i i st i e v nnnas 150
Conversion of UF ; and UF, to the Respective Chlorides with BCly. .o vu v v vvvovonsy 151
Solubility of Tri- and Tetravalent Uranium Fluorides in Fused NaAICl, v oo vvv i onn e, 151
Oxidation of UF 5 with Oxygen . v v v v v v vi i vns et e s s e ceee. 152
Determination of Lithium, Potassium, Rubidium, and Fluoride lon
in NaF-KF(RbF)-LiF-Base Fuels . v v v v v v v v it i v v v vt v a e C e e e e e 152
Oil Contamination in ARE Helium., v« o v v v i it i i i s it it e s e s s e s veanane 154
Petrographic Investigations of Fluoride Fuels. i i e v i vt ittt s e vt v o esans 154
Summary of Service Analyses. ¢ v v v i ittt i i et c i fe st e s assuae e 154
PART I, SHIELDING RESEARCH
11, SHIELDING ANALY SIS Lo ittt ittt i it et a st a st st oot aossoannsnans 157
Slant Penetration of Composite Slab Shields by Gamma Rays ... ... e et e e 157
Air Scattering of Neutrons. v v v v v vt v vt et v i et osnaasn et e et 158
Single Anisotropic Air Scattering in the Presence of the Ground (Shielded Detector) .. ... .. 158
Single Isotropic Air Scattering in the Presence of the Ground (Unshielded Detector) . ... ... 158
Formulas for Multiple Scattering in a Uniform Medium o « o v oo v v v v o e v C e e 159
Ground Scattering of NeUtrons 4 v v v v v v it i st it ittt st s e nonsossassnoens 162
Focusing of Radiation in a Cylindrical Crew Compartment « v v v v v v v v v v s Cea e e 163
12. LID TANK SHIELDING FACILITY 4 it it ettt vt c s s s aaton s snssaossnonasasas 164
Reflector-Moderated Reactor and Shield Mockup Tests v v v v v i i i it i i et i e i e e en o 164
Effective Removal Cross Sectionof Carbon & v v v v vt it i i it it it i i e s i e st e s s 164
GE-ANP Helical Air Duct Experimentation « « v v « v v e s v s o s nnonssensettnennenss 166
13, BULK SHIELDING FACILITY & v v ittt i e ivaean st astnntoosooenassns P e 168
Reactor Radiations Through Slabs of Graphite. s v v v v v e v vt e i e v i st n e e bene e e 168
Reactor AIr Glow v v v v v v oo v oo s oo oot tnnsnnsennsesnssssotensossnsansessses 170
Fuel Activation Method for Power Determination of the ARE . .o v v v v v i v v v v i v i v an 173
14, TOWER SHIELDING FACILITY & v v vttt ittt vt s e st s aotoossossoasnesnasssns 175
Fast-Neutron Ground and Air Scattering Measurements v v v v v v vt v v v oo v o s e e a e e 175
Calorimetric Reactor Power Determination « v v v v v s v v e v v vt c o st ton oot tonasnsas .. 175
GE-ANP R-1 Divided-Shield Mockup Tests ¢ « v v v v v v e e st e s a oo osonasosansons .. 178
PART V. APPENDIXES
15. LIST OF REPORTS ISSUED DURING THE QUARTER v it it i s v v s v st s s snnsononos 183
ORGANIZATION CHART OF THE AIRCRAFT NUCLEAR PROPULSION PROJECT .......... 185

x11

 
 

ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART |. REACTOR THEORY, COMPONENT
DESIGN AND TESTING, AND CONSTRUCTION

Circulating-Fuel Aircratt Resctor Experiment

The water tests of the fuel and the sodium cir-
cuits of the ARE system at room temperature were
completed (Sec. 1). The sodium circuit was pres-
sure-filled with water from the sodium fill tanks,
while the fuel system was vacuum-filled to ensure
the elimination of gas pockets in the fuel system.
Both systems were found to function properly, and
the filling, circulating, and draining operations
were effected with a minimum of difficulty. After
the water was drained, the system was dried by
heating it to approximately 600°F. The electrical
heating system was found to be satisfactory in
the check thus afforded.

The final system completion work is now under
way, that is, removal of the sodium system reactor
bypass, completion of the fuel-enrichment system
installation, completion of thermocouple and insu-
iation installation, and other minor modifications.
When this work is completed,
charged to the system and the high-temperature
checkout phase of the experiment will be initiated.

The neutron source was put into the reactor, and
the nuclear checked out.
Also, mechanical checks were made of the per-
formance of the safety and control rods. The
building electric and helium systems were made
ready to accommodate the loading facilities, and
final arrongements were completed for attaching
the fuel-sampling equipment.

Radiation damage experiments indicated the de-
sirability of providing gamma shielding at several
points where elastomers were used (belts, dia-
phragms, etc.}. Where shielding was impractical,
the composition diaphragms were replaced with

sodium will be

Cinsfrumentation was

metal diaphragms.

The results of operation of the reactor system
component test loop at K-25 are encouraging in
that the loop has now been operated without major
difficulties for more than 1800 hr. None of the
minor difficulties encountered would indicate that
serious problems might arise in operation of the

ARE.

Reflector-Moderoted Reactor

The program for the development and construc-
tien of the Circulating-Fuel Reactor Experiment
(CFRE) has been outlined, .and many of the de-
velopment projects are under way. Tentative
design data have been compiled and o flow sheet
has been prepared (Sec. 2). The first priority
development project, the test of beryllium in con-
tact with sodium and Inconel under thermal stress,
has been completed. The results of this test were
needed in the determination of the dmcunf of
poison to be expected in the reflecter. The test
indicated that beryllium will not crack under the
thermal stresses involved in the temperature range
1000 to 1300°F. Since corrosion and mass transfer,
as well as thermal stress, will be important in the
beryllium-lnconei-sodium system, many static and
dynamic tests under varieus conditions have been
made. There is considerable evidence to indicate
satisfactory compatibility in the beryllium-Inconel-
sodium system at temperatures up to 1200°F.

The temperature coefficient of reactivity for the
CFRE was computed on the UNIVAC and, for rapid

- temperature changes, was found to be -3.5 x

10-°/°F. The critical mass computed for the
rhombicuboctahedral critical assembly was rede-
termined because of errors found in the originaol
data. The redetermined value agreed closely with
the experimental value, but, since the criticol
mass is not very sensitive to errors in detail,
further evaluation of the agreement must await
additional experimental results,

Experimental Reactor Engineering

The emphasis in the engineering work is now
on development of components for an in-pile loop
for insertion in a horizontal beam hole of the MTR
and the design ond construction of forced-circu-
lation corrosion testing loops (Sec. 3). The in-pile
loop for insertion in the MTR is o joint ORNL and
Pratt & Whitney Aircraft Division project. It is
to circulate proposed fuel mixtures in the high-flux
of the MTR so that the extent of radiation damage
to materials of construction and the effect of radi-
ation on the fuel can be determined. Two types
of pumps have been developed for in-pile use: a

L |
ANP QUARTERLY PROGRESS REPORT

vertical-shaft centrifugal sump pump for instal-
lation external to the reactor shield and o hori-
zontal-shaft sump pump for insertion inside a beam
hole. A turbine-type impeller is being considered
for the horizontal-shaft pump because it would
have the advantage that both the inlet and dis-
charge could be at the bottom. Hydraulic motors
of suitably small dimensions have been found to
be satisfactory drives for these pumps.

Two series of Inconel forced-circulation cor-
rosion loops for circulating fluoride mixtures are
being developed to meet the following require-
ments: (1) a Reynolds number of 10,000 with
temperature gradients of 100, 200, and 300°F and
(2) a temperature gradient of 200°F with Reynolds
numbers of 800, 3,000, and 15,000. The maximum
fluid temperature is to be 1500°F,

A study is under way of the cavitation phe-
nomenon associated with operating liquid metal
systems at elevated temperatures, high flow rates,
and high pump speeds. A correlation of fluid-flow-
noise infensity with pressure data was noted.

The number of stations available for convection-
loop testing was increased from 18 to 31 and the
basic design of the loops was simplified. Various
means of heating the loops and of making operation
of them more automatic are being studied.

Critical Experiments

The first step of the present critical experiment
program was the construction of a small two-region
reflector-moderated reactor to provide experimental
data on a system of simple geometry and materials
for use in checking the calculational methods
being used (Sec. 4). The core consists of alter-
nate sheets of enriched uranium metal and Teflon
and is surrounded by a beryllium reflector. The
uranium loading can be varied, within the specified
dimensions, to make the system critical. The
assembly was looded as prescribed by the multi-
group calculations but was not critical. However,
when the calculations had been corrected to take
into account errors in the original data, a new
attempt to achieve criticality with the prescribed
loading was made. The corrected prescribed
loading was 20.9 to 22.75 |b of U??® and the
experimental loading was 24.35 1b of U235, A
larger critical assembly of the same shape is to
be constructed that will consist of three regions,
with the beryllium island and the reflector sepa-
rated by the fuel annulus. A further check on the
calculational methods will be obtained.

2 oo

 

PART H. MATERIALS RESEARCH
Chemistry of Molten Materials

Studies of the fluoride systems of interest as
reactor fuels were continued, with particular em-
being given to systems in which the
uranium-bearing component is the less corrosive
UF, or a mixture of UF, and UF, rather than UF,
alone (Sec. 5). Recent attempts to correlate the
anticipated reduction of UF, in the UF -bearing
melts with wet chemical analysis for UF, and
UF, and results of petrographic examination show
some surprising anomalies. When UF, dissolved
in LiF, in NaF--ZrF4-UF4 mixtures, or in NaF-LiF
mixtures is treated under flowing hydrogen at
800°C with excess uranium metal, 90% or more
of the UF, is reduced to UF,. However, when
this technique is applied to UF, in NaF-KF-LiF
mixtures, the reduction is only 50% complete at
800°C and, perhaps, 75% complete at 600°C.
Petragraphic examinations of the specimens reveal
no complex compounds of tetravalent uranium; it
is possible that the UF, is “hidden’” in solid
solutions or in complex UF4-UF3 compounds in
which it is not at present recognizable.

Solid phase studies of the NaF-ZrF -UF, system
were initiated following completion of studies of
the NaF-ZrF, system, and to date no ternary com-
pounds have been discovered. A tentative equi-
librium diagram was prepared.

phasis

A methed for large-scale purification of rubidium
fluoride has been developed that can be used if
material sufficiently free from cesium cannot be
obtained from commercial sources. Fundamental
studies of the reduction of NiFF, and FeF, by H,
in NaF-ZrF, systems were made as a means of
determining possible improvements in purification
techniques. Also, methods for preparing simple
structural metal fluorides were studied.

Large quantities of puritied ZrF ,-base fluorides
were prepared for engineering tests at ORNL and
elsewhere, and the demand for purified fluorides
of other types is rapidly increasing. Therefore
preparation and purification methods have been
studied intensively in an effort to lower production
time and costs. Developments indicate that the

price of purified NaZrF, and NoF-ZrF ,-UF, mix-

tures may be halved in the next few months.

Corrosion Research

The static and seesaw corrosion testing fa-
cilities were used for further studies of brazing

-
o
alloys, special Stellite heats, Hastelloy R, In-
conel, graphite, and various ceramics in sodlum
flucride fuel mixtures, and other mediums (Sec. 6).
The brazing alloy 67% Ni-13% Ge~11% Cr-6%
Si—-2% Fe~1% Mn was found to have good corrosion
resistance in fluoride fuels and fair corrosion
resistance in sodium and therefore will be useful
for the fabrication of many reactor components.

In the thermal-convection loop studies, UF,-
bearing fuels were tested in Inconel. Hot-leg
attack is not found in Inconel loops in which ZrF -
base fluoride mixtures with the uranium as UF,
are circulated. A deposit is, however, found on
the hot-leg surface. Only preliminary information
is available, but it appears that neither attack
nor a hot-leg layer is found with alkali-metal-base
fluoride mixtures containing UF,. Mixtures of
UF, and UF, result in a reduction in attack from
that found with only UF,, but some attack is
present, and in high-uranium-content systems the
attack may be significant,

Several Hastelloy B loops have now been suc-
cessfully operated in both the as-received and the
over-aged conditions. In both coses @ consider-
able increase in hardness occurs during operation.
With ZrF, -base mixtures containing UF,, very
little attack is found, even after 1000 hr.

Thermal-convection loop tests of molten lithium
in type 316 stainless steel were operated for 1000
hr. There were no signs of plug formation, and
only a small amount of mass transferred material
was found in one loop. Alloys of 45% Cr—55% Co,
Ni-Mo alloys, and the Fe-Cr-base stainless steels
have been shown to be more resistant to corrosion
and mass transfer in liquid lead than are the pure
metals. Their resistance to mass transfer can
probobly be related to the formation of interme-
tallic compounds. |

Metallurgy

Creep and stress-rupture testing by the tube-
burst method has been studied intensively (Sec. 7).
In the tube-burst tests, a tube that is closed at
cne end is stressed with an internal gas pressure.
The stress pattern introduced into the specimen
in this fest approaches the stress pattern that will
be found in ANP-type reactors. Apparatus for the
tests has been constructed, and a theoretical
analysis has been made with which o check en the
experimental results can be obtained.

In the investigation of high fhermql-conductwlty

PERIOD ENDING SEPTEMBER 10, 1954

fins for sodium-to-air radiators, stress-rupture and
creep tests were made on copper fins with various
types of cladding at stress levels between 500 and
2000 psi at 1500°F. The tests show that for o
1000-hr exposure in air, stresses greater than 500
psi and less than 1000 psi are tolerable; that s,
in this stress range there is no indication of
brittleness in the core or oxidation of the core
due to cladding failure of type-310 stainless-steel-
clad copper fins. From the over-all considerations
of melting point, oxidation resistance, dilution of
fin and tube wall, formation of low-melting eu-
tectics, and flowability, it was found that Coast
Metals alloy 52 was the best brazing alley for
use in the construction of radiators with high-
conductivity fins. A sodium-to-air radiator with
6 in. of type-430 stainless-steel-clad copper high-
conductivity fins was fabricated by use of a combi-
nation heliarc welding and brazing procedure.

Packed-rod nozzle assemblies were fabricated
for the 100-kw gas-fired liquid-metal-heater system,
and work was started on the formation of duplex
tubing. An attempt is being made to prepare tubing
that will have good corrosion resistance on the
inner surface and oxidation resistance on the outer
surface.

Attempts are being made to find new alloys in
the nickel-molybdenum system that will have better
high-temperature strength and fluoride corrosion
resistance than Inconel has. Hastelloy B satisfies
these requirements, but it has poor fabrication
properties and oxidation resistance; it also loses
its ductility in the temperature range of interest
for application in high-temperature circulating-fuel
reactors. lInvestigations are under way to find a
suitable melting practice and heating treatment
that will increase the ductility of Hastelloy B in
the temperature range of interest.

Heot Transfer and Physical Properties

The enthalpies and heat capacities of NaF-ZrF -
UFJ (65-15-20 mole %) and of LEF-N0F~UF4 (57.6-
38.4-4.0 mole %) were determined (Sec. 8), The
thermal conductivity of NaF-KF-UF, (46.5-26.0-
27.5 mole %) was found to be 0.7 Btu/hr-f’rz(ol:/ff)
that for KF-LiF-NaF-UF , (43.5-44.5-10.9-1.1 mole
%) was 2.0, and that for LiF-KF—UFd (48.0-48.0-
4.0 mole %) was 1.4. A new electrical conduc-
tivity device has been constructed and has been
successfully checked with molten salts of known
conductivity.
ANP QUARTERLY PROGRESS REPORT

A device for studying the rates of growth of
tube-wall deposits has been successfully tested
with a simple heat transfer salt. Also, a hydro-
dynamic flow system for studying the reflector-
moderated reactor flow structure has been tested.
A mathematical study of the temperature siructure
in converging and in diverging channel systems
that duct fluids with volumetric heat sources and
a study of wall cooling reguirements in circulating-
fuel reactors were made.

Rudiation Damoge

Additional irradiations of Inconel capsuies con-
taining fluoride mixtures were carried out in the
MTR (Sec. 9).
UF,-bearing fue! has been examined, and it shows
no corrosion, in confrast to that found previously
with UF  -bearing fuels.

Inspection of the LITR in-pile loop, which failed
prior to startup, disclosed that the failure was
caused by ¢ bregk in the weld connecting the pump
discharge nipple to thz fuel line. Design revisions
and refabrication of some of the parts are in prog-
ress. Developmental work continued on a smaller
loop for operation in an L.ITR A-piece.

Only one capsule containing a

Detailed exomination of an Inconzsl loop which
circulated sodium at high temperature in the ORNL
Graphite Reactor showed no evidence of radiation-
induced corrosion.
were corried out in the hot cells on irradiated fue!
plates for Frait & Whitney Aircraft Division, and
studies were made on annealing-out of fission-
fragment dumage. Work continued on examination
of wire and multiple-plate-type units for GE-ANP,

Metallogrophic examinations

Anclytical Studies of Reactor Materials

The primary analytical problem continugs to be
the seporation ond determination of trivalent and
tetravalent uranium in hoth NqF-ZrFS and NaF-
KF-LiF-bose fuels (Sec. 10).
tentiometric titration of UFJ
with metallic

A successful po-
in molten Naldrf
zirconium was gaccomplished by
means of polarized platinum g¢lectrodes. The sclu-
bility of UF, in NaAICl, was determined to be
18 mg/g ot 200°C, in contrast to a solubility of
UF, of less than 1 mg/g.
NaAIClA is expected to dissolve tetravalent ura-

Therzsfore molten

nium selectively from the fuels.
Calibration measurements have been completed
on the apparatus for the determination of oxygen

as metallic oxidas in reactor fuels. The reaction

involves the hydrofluorination of the oxide and
measurement of the increase in conductivity of
liquid HF as o function of the water formed. In-
vestigations are being made of the oxidation of
Ut', and of UF, with oxygen at slevated tempera-
tures,

An improved method for the determination of
lithium in NaF-KF-LiF-base fuels was developed.
Also, studies were made of the solubilities of
potassium, rubidiym, and cesium tetraphenyl-
borates in various organic solvents to ascertain
differential solubilities.

PART ill, SHIELDING RESEARCH
Shialding Analysis

Application of the Monte Carlo method to the
calculotion of gamma-ray penetration of crew
shie!d sides has been worked out, and the method
appears to be quite satisfactory for this type of
problem (Sec. 11). Considerable progress has been
made on understanding the Tower Shielding Fa-
cility (TSF) mecsurements of ground and air
scattering. Expressions have been derived which
describe the effect of ground interference with the
air-scattered flux, and thus it is now possible to
estimate ground-scaitered radiation both «t the
TSF and in on airplane at landing and ftakeoff,
Calculations have been set up for evaluating
multiply scattered radiation in air. The values
obtained will be of considerable importance in
studies of the highly asymmetric but light shields
of several current designs.

Lid York Shislding Facility

Preparations for a second series of shielding
tests for the reflector-modercted reactor ot the
Lid Tank Shielding Facility (LTSF) have included
the construction of a large tank which will hold
all the components of the mockups and the irradi-
ation of the UF C,F,  solution which may be
used to simulate the reactor fue! in mockup ex-
periments (Sec. 12). Measurements of the removal
cross section of carbon made in a continuous
have in a value of
o, = 0.750 barn. This is to be compared with the
previous value of 0.81 1 0.05 barn measured behind
a solid slab of graphite. Thermal-neutron flux
measurements have been mode bevond two con-

carbon  medium resulted

figurations of GE-ANP helical air ducts, a single
duct and a triangular array of three ducts. Meas-

urements beyond g 35-duct arroy will begin soon.
‘Bulk Shielding Facility

The use of a graphite reflector as a shieid com-
ponent has been investigated at the Bulk Shielding
Facility (BSF) with measurements of attenugtion
through various thicknesses of the material (Sec.
13). The fast-neutron spectrum of the BSF reactor
from 1.3 1o 10 Mev through 1 ft of graphite was
also measured. Removal cross section values for
the carbon were determined to be 0.82 barn for a
1-ft slab, 0.84 barn for a 2-ft slab, and 0.80 barn
for a 3-ft slab.

An experiment has been performed at the BSF
to provide an experimental basis for future esti-
mates of the amount of visible light around o
nuclear-powered aircraft. Measurements in an air-
filled tube placed against the reactor have indi-
cated that the maximum glow will occur at!a
pressure correspondmg to an altitude of ubout
30,000 ft.

in a proposed method for the determination of
the power of the ARE, the relative octivity of the

PERIOD ENDING SEPTEMBER 10, 1954

fuel samples exposed in the ARE ond in the known
flux of ancther reactor is meosured. The method
has been tested at the BSF,

Tower Shielding Facility

Measurements of the ground and goir scattering
of neutrons have been made at the Tower Shielding
Facility (TSF) (Sec. 14). A preliminary anolysis
indicates that the contribution at the maximum
altitudes (around 200 ft) is between 2 and 5% of
the total scattered neutrons for differential ex-
periments.

A new procedure for a calorimetric determination
of the power of the TSF reactor has been devised
which is based on the relationship between the
rate of temperature rise in the water of the reactor
tank and the reactor power. The resuits of three
experiments were consistent to within 1%. The
next series of tests at the TSF will be on the

GE-ANP R-1 divided-shield mockup.
Part |

REACTOR THEORY, COMPONENT DESIGN

AND TESTING, AND CONSTRUCTION
1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT

E. S. Bettis

J. L. Meem

Aircraft Reactor Engineering Division

THE EXPERIMENTAL REACTOR SYSTEM

A general revision was found to be necessary
in the ARE ‘in that shadow shielding had to be
installed, at several points where elastomers were
used (belts, diaphragms, etc.),to provide protection
from gamma fields that would: have produced pro-
hibitive degrees of radiation damage. At some
points, where shielding was impractical, the compo-
sition diaphragms were replaced by metal dia-
phragms.

The entire ARE system has now been completely
checked out as far as rocometemperature preoper-
ational tests are concerned. Both the fuel and the
sodium circuits have been operated simultaneously
with water as the circulated liquid in each circuit.
The moderator volume of the reactor was bypassed
for these tests to keep water away from the be-
ryllium oxide blocks.

The sodium circuit was pressure-filled with
water from the sodium fill tanks, while the fuel
system was vacuum-filled to ensure the elimination
of gas pockets in the fuel system. Both systems
were found to function quite satisfactorily, and the
filling, circulating, and draining operations were
effected with a minimum of difficulty.

During these operational shakedown tests, it was
found that the gas lines to the sump tanks could
not be kept pressure tight because the Swagelok
fittings developed leaks. All these joints have
therefore been silver-soldered to correct this
situation,  The pumps performed satisfactorily,
and useful system curves for actual future operation
were obtained.

After the water tests of the system had been
completed, it was necessary to remove oll water
from the system. This necessitated heating the
entire system to o temperature of approximately
600°F. This operation provided the first check out
of the electrical heating system.
performance was gratifying.  Such troubles as
arose were corrected rather easily, and, in general,
the heating of the system was effected with very
little ditficulty.

A CO, cold trap was incorporated in the off-gas
line while the system was being heated to collect
all moisture that was driven from the system. Dry

Here again the

air was admitted to the fill tanks, and both the
fuel and the sodium systems were flushed with
this dry air, the air leaving the system through the
CO, cold trap. When the dew point of the exit air
from this cold frap reached ~30°F, the system was
considered to be dry and the electrical heat was
turned off.

Befare the heat was removed from the system,
however, the temperature on the thermal barrier
doors around the heat exchangers was raised to
approximately 1000°F and operation of the doors
was checked., Even though the doors had been
reworked to eliminate sticking, it was found that
the doors were still binding. When the barrier
doors were removed, it could be seen that the side
guides whichrun the height of the door were binding
in the guide ways of the door frames. This binding
was resulting from the bowing of the guide rails
caused by the thermal gradient which exists in the
door. The door does not require these side guides,
since the door housing provides ample guides for
the normal functioning of the doors. Therefore the
side guide rails are being removed so that there
will be no further binding of the barrier doors.

It was also noted while the system was hot that a
considerable volume of kerosene was being driven
out of the beryllium oxide moderator blocks by the
high tempercture. The kerosene had been absorbed
by the blocks during the cutting operations. To
remove the kerosene, a vacuum was pulled on the
moderator volume of the reactor. A CO, cold trap
was inserted in the vacuum line to the moderator,
and while the reactor wos maintained at a temper-
atyre of approximately 450°F, the vacuum pump was
run continvously. This operation was continued
until no more condensate of the kerosene distillation
was being collected in the cold trap. Approxi-
mately 2 gal of kerosene was removed by this
procedure. |t is possible — in fact, fairly certain -
that some residue was left in the moderator blocks.
It was the consensus of the chemists that this
would not have an adverse effect on the sodium
coolant, and therefore nothing further is to be done
about the small amount of tarry residue remaining
in the reactor moderator blocks, ’
ANP QUARTERLY PROGRESS REPORT

After the system had been thoroughly dried, all
heat was turnad off and the final system completion
The major work in this category
consists in removing the sodium system reactor
bypass, removing the flanged filter pots in the
sodium purification system, and completing the
Ther“
mocouple installation also has to be completed,
as well as thermal insulation of the fuel pump
bowls.

work was bequn.

enrichment system tie-in fo the fuel circuit,

The decision to remove the flanged filter pots
and o substitute weld-sealed filter pots was mads
It wos decided at that time,
however, to postpone this substifution until after
the woter test. The original filters were useful in
cleaning up the water used to test the operation of
the sedium circuit, Completion of the fue! enrich-
ment system in no way affected the water test, and
hence this work was postponed in the interest of
the woter test as expeditiously as
The fuel injection system is to be
woter-tested independently of the fuel system.

Insulation of the fuel pump bowls was postponed

several months ago.

completing
possible.

to allow more room for access to the pumps during
the water tests. Since the pumps have riow been
checked out, the insulation job will be completed,

The system should be ready for charging with
sodivm early in September, and the high-tempera-
ture check out phase of the experiment will be
initiated. While the sodium checks arc being run,
the hot.gas leck test of the fuel system will be
performed. This test involves loading the fuel
system with a mixture of helium and krypton at a
pressure of about 15 psi. While this gas is being
circulated at 1300°F by the fuel pump the annulus
circuits will be monitored by mass spectrographic
methods for the presence of krypton in the annulus.

The nsutron source was put in the reactor, and
the nuclear instrumentation was checked out. Al
three (the twe regular plus the spare) fission
chambers were checked, and the count-rate vs
chamber-voltage curves were plotted. The data
obtained provided the necessary information for
establishing the operational plateou for
chamber. Also, mechanical checks were made of
the performance of the safety and control rods.

A modification of the rod-cooling circuit was
required in order to minimize the heat loss from the
center of the reacter.
lation of propertioning orifices to get coivect
helivm flow around ths fission chambers and the
safety rods.

each

This riecessitated instale

10

Final modifications of the sodium and the fuel
loading systems were completed, and the building
electric ond helium systems were made ready to
accommodate the loading facilities when they are
brought into the building. Final arrangements
have been completed for atiaching the fuel-sampling
system. [his sampling system will not be left
connected when the power run is initiated.

The engingering prints for the entire ARE instal-
lation are now up to date, and the electrical prints,
in particular, were used successfully in checking
out the many heater circuits involved in the experi-
ment,

CHARACTERISTICS OF THE FUEL AND THE
SODIUM SYSTEMS DURING WATER TESTS

The characteristics of both the fuel and the
sodium systems while circulating water have been
determined. Upon removal of the fuel heat ex-
changer bypass loop, a glass rotameter was tempo-
rarily installed in the fuel circuit between the
reactor outlet and the heat exchangers. A direct
calibration of the high-temperature fue! rotameter
against the glass rotameter was obtained with
water as the fluid. Conversion of water flow to
fuel flow! is made by

1/2
PF - P/2
92 pfz
9y P — Pf] ’
pf'l
where
g, = watsr flow (gpm) ,
g, = fuel flow (gpm) ,
p;, = water density (g/cm?) ,
1
Pr, = fue! density (g/cm3) ,

pp = float density (g/em®) .

During the period of water operation, data on
pump speed vs flow rate were cbtained (Fig. 1.1).
Also, the pressure head from the pump suction to
the reactor inlet was measured (Fig. 1.2). The
system characteristics, as shown in these two
curves, should be the same during operation with
fluorides.

 

]A. L.. Southern, Discussion of the Rotameter Used in
the ARE Fuel Circuit, ARE Filas.
ORNL—-LR--OWG 3087

 

 

2000 -

 

 

 

 

 

FUMP SPEED trom)
O
<
|
|
\
|

 

 

 

 

 

 

 

0 40 20 30 40 50 80 70 20 S0
FLOW RATE (gom)

Fig. 1.1. Data on Pump Speed vs Flow Rate
Obtained During Water Test of ARE Fuel System.
Sl

ORNL~LR--BWG 3088

 

 

 

 

 

 

 

 

 

 

 

100 s T T g s
[ ’.,
°
L B0 e e — e e ;/,’,,,,, -
¥ /'
o 1 *
L<1IJ a0 L "/ . b
i v
Ll 1
o »
Ty &0 e e o
g L
Ld '/‘
& 1 o«
20 e - fi-;.i"’("""” """"""""
|
[ eeeeeemeenedooes I R U O A S B
O 10 20 30 40 50 50 [49] 80

FLOW RATE (gpm)

Fig. 1.2, Data on Pressure Head from the Pump
Suction to the Reactor Inlet vs the Flow Rate
During Water Test of ARE Fuel System.

By using a calibrated orifice installed tempo-
rarily in the sodium bypass line around the reactor,
it was possible to measure the flow rate while
operating the sodium system with water. The
pump speed was recorded and alsc the pressure
head from the pump suction to the reactor outlet.
The system characteristics with water, shown in
Figs. 1.3 and 1.4, should be valid while operating
with sodium as the fluid. During operation with
sodium, the flow rate in each of the two loops will
be measured with an electromagnetic flowmeter.
The sum of the two flow rates should equal the
total flow as obtained from the pump speed.

PERIOD ENDING SEPTEMBER 10, 1954

g

=ED {rpm)

PUMP Sk

 

 

 

 

 

 

0 20 40 &0 2O tC0 120 140 160 180
FLOW RATE (gpm)
Fig. 1.3. Data on Pump Speed vs Flow Rate

Obtained During Water Test of ARE Sodium System,

CRNL -LLR -DWG 309
e T S L

 

 

 

 

 

 

 

4 8 r T T T T T T T T T T T T T "T = -
£ 1 e
2 |
l |
T |
w 24 S s
£ | .
T 7 i
8 ?,,,,, ‘.} e ”,,,,,,,,;“/t FUU i . — ,,w‘
LT J
0L ‘ e L_ ......................... N
o 20 40) 60 80 100 120 140 160 1380

FLOW RATE {gpm)

Fig., 1.4. Data on Pressure Head from the Pbmp
Suction to the Reactor Outlet vs the Flow Rate
During Water Test of the ARE Sodium System,

PREPARATIONS FOR LOADING THE ARE
G. J. Nessle

Mafericls Chemistry Division

All the preliminary preparations for loading the
barren and enriched fluoride mixtures into the ARE
have been completed. The necessary control
panels have been assembled, and furnaces, heating
units, transfer lines, and various complementary
equipment are now ready for installation. The
exact location of the various operations has been
determined.. However, because of the large amount
of construction work which has been carried out in

il
ANP QUARTERLY PROGRESS REPORT

the ARE building around these locations, none of
the parts necessary for the loading and sampling
operations have yet been installed. Testing of the
equipment ofter installation and before the filling
operation should require approximately ore week.
It is anticipated that all personnel to be invelved
with the loading and sampling operation will be
made thoroughly familiar with the equipment and
its operation during this testing period.

FUEL-CONCENTRATE INJECTION NOZ/LE
W, G. Cobb W. R. Huntley

Aircraft Reactor Engineering Division

Tests were made on a resistance-heated fuel-
concentrate nozzle to dectermine the
effectiveness of the heating in the region where
the nozzle will enter the relatively cold pump tank
cover plate, These tests showed that a temperature
of 1300°F could be maintained in the injsction
tube by using about 140 amp from o high-current
transformer.  This nozzle test was considered
satisfactory to meet the operating reguirements for
the ARE enrichment system,

injection

ARE UNLOADING EXPERIMENT
J. Y. Estabrook

Aircroft Reactor Engineering Division

The mockup of the ARE unloading apparatus was
tested twice. In the first test the storage tank
was filled with 750 b of NaZiF,, which was
successfully unloaded in 90-1b increments into
aluminum cans. In two cases, however, the bottoms
of the aluminum cons melted and the fluoride
mixture {ecked out.

Modifications made in the apparatus before the
second test included improvements in the pouring
indicators, and the pressure
In addition, small omounts

spout, the level
medasuring equipment.
of the fluoride mixture were frozen in the bottoms
of the aluminum cans to minimize the possibility
of melting these containers.

For the second test the apporatus was filled
with 500 1b of NaF.-ZrF, (50-50 mole %), and this
material was fransferred at 1200 to 1250°F to the
aluminum cans in 70~ to %0:1b increments without
difficulty, Attempis to transfer at 1100 to 1150°F,
however, led to occasional freezing at the pouring
spout and indicated that unless much more reliable
temperoture control could be effected, operation
at this low temperature could hardly be considered
to be feasible.

12

ARE FUEL RECOVERY AND REFPROCESSING

D. E. Ferguson G. L. Cathers
M. R. Bennett

Chemical Technology Division

A fluoride-volatility method of processing ARE
fuel, hased on elemental fluorination of the molten
Naf-ZrF ,-UF,, has been investigated because of
its attractiveness from the stondpoint of cost,
inventory of fissionable material, woste disposal,

In preliminary tests2e3

and safety of operation.
than 99.95% of the uranium tetrafluoride
contained in the fuel was converted to the volatile
hexafluoride with o gross beta decontamination
factor of 100 to 300. Resublimation of this uranium
hexafluoride resulted in an over-all decontamination
factor of 4000 to 5000,
excess fluorine UF ( is soluble in molten NaF-ZrF ,
the resublimed product con be dissolved, probably

more

Since in the absence of

. . 4
according fo the reaction”

UF ¢ + NaF > NaF.UF  ~—> NaFUF , + F,

The fuel from an aircraft reactor would be re-
processed in three steps: recovery of the uranium
from the used fuel by fluorination, separation of
the uranium hexafluoride from excess fluorine by
trapping the uranium salt in a cold trap and re-
subliming the UF [ from it, and fabrication of a new
reactor fused salt by absorbing the partially decon-
taminated UF ( in molten NaFF-ZrF . Since uranium-
zirconium and other alloys can be dissolved at
reasonoble rates by hydrofluorination in a fused
salt of the MaF-ZrF ,-KF type,> this also provides
a msthod of processing
fuels.

A decertaminaiion foctor of 4000 to 5000 is
sufficient if the fue! is to be recycled directly

heterogeneous reactor

to a fused salt reactor, but, if the uranium is to
be returned to o diffusion plant or to a heter-
ogeneous reacior, more decontamination is needed.
In further experimenial werk, o decontamination

factor of approximately 20,000 was obtained by

scrubbing the fission-product fluorides from the

2G. I. Cathers, Recovery and Decontamination of
Uranium from Fused Fluoride Fuels by Fluorination,
ORNL-1709 (June 9, 1954),

3¢, N. Browder, G. l. Cathers, D. E. Ferguson, and
E. O. Numni, ANP Quar, Prog. Rep. Mar. 10, 1954,
ORNL-1692, p 42,

4H. Martin, A. Albers, and H, V. Dust, Z anorg, w
allgem, Chem. 265, 128 (1951),

SR. . Leuze and C. E. Schilling, personal com-
municaticn.
uranium hexafluoride with a molten NaF-ZrF | salt
bath (at 650°C) in which the UF, is insoluble in
the presence of excess fluorine. The uranium loss
in the scrubbing step was less than 0.001%. Frac-
tional distillation of the UF  from this step, which
would be performed in order to obtain complete
decontamination, could be carried out in only lightly
shieided equipment.

Two experiments on scrubbing UF , with molten
NaF-ZrF, were carried out. In one, an acetone-dry
ice cold trap was used between the fluorination and
scrub steps; the uranium hexafluoride was cone
densed {and thus separated from excess fluorine)
in the cold trap and was then resublimed from it.
The uranium was recovered from the scrub salt by

TABLE L1

PERIOD ENDING SEPTEMBER 10, 1954

refluorination, In the second experiment no cold
trap was used, but the decontamination obtained
was the same (Table 1.1). it is thought that
resublimation of the salt-scrubbed product would
achieve additional decontamination.

Synethetic ARE fuel was prepared for the two
runs by dissolving a 224-day-irradiated, 30-day-
cooled, 6-g miniature vranium sfug in NaF-ZrF
(56-44 mole %) at 650°C by passing HF through
the molten bath. The final NaF-ZrF (-UF , mixture
had a compesition of 54.0-42.4-3.6 mole % and an
activity of 5 x 10° beta counts per minute per
milligram of uranium. The dissolution and subse-
quent fluorination were carried out in the same

nickel reactor, In each run the scrub salt was

DECONTAMINATION OBTAINED BY MOLTEN SALT SCRUBBING OF UF,

IN THE FLUORIDE VOLATILITY PROCESS

Synthetic ARE fuel; beta activity of 3 X ]06

counts per minute per milli-

gram of uranium; fivorinated at 650°C with elemental fluorine gas at flow

rate of 66 ml/min

 

 

 

 

 

 

 

Run A; Fluorination period of 3 hy; UF6 plus excess fluorine pussed di-
rectly to scrub salt from the flusrination step
Run B: Fluorination period of 5hr; UF& condensed in cold trap andresub-
limed from it prior to scrubking and recovered from scrub by re-
flucrinagtion
Beta Decontamination Factors
Activity o RwmA . RumB
Scrub Qver-all Scrub Cverall
[
Gross 2 % 104 2 % 10%
Ru 6 750 1.4 430
Zr 3 >7 x 10t 9 1 x 10°
Nb 4 700 8 2 % 10
TRE 90 8 % 108 3 2 x 10¢
Cs 130 >3 x 10
S¢ 20 >3 x 109
Bq >8 x 10°
Uranium Losses (% of Initial Chorge)
In fluorinatien salt 1.6 3 x 10"3
In resublimation trap Q.74
In scrub salt 0,23 i3
*Calculated onthe gross Ru, Zr, Mb, and TRE activity rather then on total activity, which included much due to 5.13]-

This does not materially offect comparison with previous gress beta decontaminatien data,

13
ANP QUARTERLY PROGRESS REPORT

67 g of an NaF-ZrF, mixture of the same compo-
sition as that used in preparing the initial ARE
fuel. Both fluorination and scrub operations were
carried out at 650°C,

The scrub decontamination factors
Table 1.1 were calculated on the basis of the
activity remaining in the scrub salt, The gross
beta decontamination factor of 20,000 obtained in
both runs is about fivefold better than the 4000 to
5000 obtained previously. The decontamination
factors for individual activities
uncertain because of the high amount of
the product from the short-cooled slugs.

The conversion of UF, to UF, in Run A was
relatively poor, with 1.6% remaining in the fluori-
nation salt. This was due to a fluorination period
of only 3 hr. A longer fluorination period of 5 hr
lowered the loss to 0.003%. The same time was
used in refluorinating the scrub salt of Run B to
recover 98.5 wt % of the initial uranium, Since
less than 10~ 3% of the uranium passed through the
scrub bath in the absence of fluorine, quantitative
reconversion of UF , to UF , by the reaction

UF6 + NaF —> NaF .UF ( ——> NaF.UFA +F,

given in

are somewhat

|]3] in

is indicated.
ARE PUMPS

Fabrication and Testing

W. G. Cobb A, G, Grindell

Aircraft Reactor Engineering Division

Five of the six rotary elements fabricated for
ARE pumps have passed all acceptance tests.
The sixth rotary element is being used in the hot
shokedown test stand to test welded-fabricated

Inconel impellers and for other miscellaneous
testing.
Unsuccessful attempts were made by ORNL

and by an outside vendor to lap the bronze bearing
wear surfaces of the upper oil seals for the shaft
to an acceptable flatness (three wave bands of
helium light) with a water soluble scouring agent,
Bon-Ami. However, acceptable flatness was
attained with conventional lapping compounds.

Four ARE pumps with welded-fabricated impellers
have been installed in the ARE.
used were four of the six welded-fabricated im-
pellers that hove passed all acceptance tests.
One of the impellers is being used in the pump in
the reactor system component test loop at K-25
and one is being held as a spare.

The impellers

14

Seal-Gas Pressure-Balancing System

W. G. Cobb

Aircraft Reactor Engineering Division

A seal-gaos pressure-balancing system was de-
veloped and has been installed in the ARE pumps.
This system automatically maintains agaspressure
in the lubricant chamber around the lower me-
chanical seal that is slightly higher (]/2 to 2 psi)
than the pressure in the pump tank volume in order
to impede the outward leakage of vapors into the
lubricating-oil system. The equipment involved
consists of a differential supply valve, a differ-
ential vent valve, and the requisite pipe con-
nections. In addition to providing a master-slave
relationship for maintaining the pressure differ-
ential across the seal, the system has two charde-
teristics which are advantageous to ARE pump
operation.  The first advantage is that it will
provide automatic bleed-off of excessive oil system
pressures created by the gas which will be evolved
from irradiation of the lubricating oil. The second
advantage is that it will maintain a greater (1’2- to
2-psi) pressure across the lower seal even when
subatmospheric pressures are introduced to the
pump tank, for example, when the ARE fuel system
is being vacuum filled.

Radiation Damage to Pump Drives

W. G. Cobb A. G. Grindell

Aircraft Reactor Engineering Division

Some alternatives to the usual V-belt pump-

power transmission system were investigated
in an attempt to find a system that would not be
affected by radiation. The alternatives studied in-
cluded direct in-line drives, angle-gear-box drives,
silent-chain drives, and improved V-belt drives.
The use of direct drives would have required
modification of the concrete shielding above the
pump pits, and the use of the angle-gear-box drive
would have required development of suitable lubri-
cant seals. The silent-chain drive tested did not
have the required mechanical reliability.®

Two improved V-belts, a Browning B-105 Super-
grip belt (rayon cords in a neoprene matrix) and a
Browning B-103 Grip belt (cotton cords in a rubber
matrix), were irradiated, under operating tension,
through a short length of the belt by a Co%? source
to total doses of 10% and 107 rep, respectively.

 

6A. G. Grindell, Morse Silent Chain Drives on ARE
Pump Hot Shakedown Test Stand, ORNL CF-54-7-166
(Juty 16, 1954),
After irradiation each belt was used in the cold
shakedown test stand to drive a pump rotary ele-
ment at 1600 rpm with no pump load. After 90 hr
of operation, the B-105 belt showed no visible
damage and was returned to ARE personnel. After
75 hr of operation, the B-103 belt showed no visible
damage and was transferred to the pump in the
reactor component test loop pump to transmit 3 hp
at 800 rpm. This belt, which has now been in
service for more than 900 hr, has increased in
length approximately 1 in. Failure of the belt
does not appear to be imminent, and pump operation
has not been affected. An idler pulley is used to
maintain tension,

On the basis of these tests, it was decided to
use the cotton-corded, rubber-matrix V-belts on
the fuel and the sodium pumps in the ARE. As an
additional sofeguard, the belt on the fuel pump
drive will be shielded with lead.

Radiation Damage to Shaft Seal
W, G. Cobb J. M, Trummel

Aircraft Reactor Engineering Division
W. W. Parkinson
Solid State Division

The rotary subassembly of the ARE pump con-
tains six Buna-N O-ring static seals which will
receive radiation during reactor operation. To get
information on the expected life of these seals,
an O-ring sealing arrangement similar to two of
the pump seals was constructed and was placed in
the LITR flux. The test assembly included six
radial O-ring seals and three gasket O-ring seals.
One side of the seals was supplied with oil at
75-psig pressure, and the other side was vented to
the reactor off-gas system through bubblers and a
catch tank so that it was possible to detect any
appreciable leakage by the seals. Since the seals

PERIOD ENDING SEPTEMBER 10, 1954

were vented in four groups, it wos possible to get
some information on variations in seal life. Differ-
ences in the following might produce such vari-
ations: (1) flux density over test assembly,
(2) twisting and rolling of the rings during assembly
for testing, (3) the O-rings, and (4) the machined
parts.

The gomma radiation intensity at the test lo-
cation in the LITR was estimated by extrapolation
from data taken on the Bulk Shielding Facility.”
The neutron flux was negligible. The resulis are
reported in Table 1.2, which gives time of irradi-
ation to first observed leakage and the total esti-
mated dose to first leakage. Examination of the
O-rings aofter irradiation revealed substantial
hardening. No appreciable elastic recovery of the
rings occurred upon removal from the container.

The O-rings to be used in the ARE will not be
in sohigh a flux as that used for these experiments,
and they will be under much lower pressures,
Since it is not essential in the ARE application
that the O-rings retain their resiliency, the resuits
of these experiments do not have immediate signifi-
cance. It is evident from these tests, however,
that O-rings cannot be used successfully in future
high-power reactor applications.

Motor Test
W. G. Cobb A. G. Grindell

Aircraft Reactor Engineering Division

A test of an ARE-type d-c motor operating in dry
helium at 130 to 140°F was: terminated after 4000
hr of uneventful operation. Examination of the
silver-impregnated graphite brushes used during
the test indicated that total wear was limited to

 

"H. E. Hungerford, The Skyshine Experiments at the
?gusli) Shielding Reactor, ORNL-1611, p 24 (July 2,

TABLE 1.2, RESULTS OF RADIATION DAMAGE TESTS OF RUBBER O-RING SEALS

 

Time to Leakage Dose to Leakage

 

Two radial ring sedals on bottom piston
Two radial ring seals on middle piston
One radial ring on top piston

Three gasket seal rings

Uj) {rep)

128 5 x 168
167 6.7 % 108
259 1 x 107
259 1 % 107

15
ANP QUARTERLY PROGRESS REPORT

approximately ]/“5 in. It is concluded that no
problem will be encountered in the use of similarly
equipped motors in the ARE,

REACTOR SYSTEM COMPONENT TEST LOOP
Cperation of Loop
W. G, Cobb A. G. Grindell

Aircraft Reactor Engineering Division

D. B. Trauger G. A. Kuipers
Technical Division, K-25

Operation of the reactor system component test
loop® at K-25 was temporarily halted after 1047 hr
of operation to remove one ARE core hairpin tube
for metallographic examination, Shutdown, tube
removal, tube replacement, and re-startup were
accomplished under a complete helium blanket to
prevent air from contaminating the fluoride mixture
or the inside of the Inconel system. An irradiated
V-belt that was installed, as described above,
after 875 hr of system operation has operated for
more than 900 hr of the approximately 1800 hr
of system operation. Plugging of gas line con-
nections to the pump tank by ZrF ,-vapor condensate
has occurred periodically. In some instances the
plugs have been removed by heating the gas line
involved and blowing the vaporized ZrF, back
into the pump tank. No completely satisfactory
solution to this problem has been found to date;
the wvse of o lorge-capacity vapor trap is being
studied as a possible solution.

 

8H. W. Savage et al., ANP Quar. Prog, Rep. [fune 10,
1954, ORNL-1729, p 14,

16

Exominotion of Hairpin Tube

G. M. Adamson R. S. Crouse
Metallurgy Division

Metallographic examination of the hairpin tube
showed normal subsurface void attack. The maxi-
mum penetration was to a depth of 5 mils. There-
fore it appears that corrosion from impurities will
not be a serious problem in the ARE. Since this
loop was not precleaned, the corrosive conditions
should have been worse than those in the ARE.
No information on the amount of corrosion that
may be expected from mass transfer could be
obtained because the loop is being operated isc-
thermally,

REACTOR PHYSICS

¥W. K. Ergen M. E. LaVerne
R. R. Coveyou C. B. Mills
Aircraft Reactor Engineering Division
R. Bate C. S. Burtnette
United States Air Force

The Eyewash code, developed by the ORNL
Mathematics Panel for use on the UNIVAC, yielded
the following results for the ARE: critical mass,
30 Ib of U235; U235 required to yield the 4% excess
reactivity needed for operation, 35 Ib; k& . for
40 1b of U%3%, 1.07; critical mass without controls,
safety, or instrument apparatus, 17 Ib (this empha-
sizes the high cost of these materials); AR/(AM/M)=
0.232; temperature coefficient applicable to slow
temperature changes and due to expansion of the
whole reactor and to thermal base effects,
~4.3 x 1073/°F; reactivity value of U233 in tube
bends at reactor ends, 0.006'in & _,..
PERIOD ENDING SEPTEMBER 10, 1954

2. REFLECTOR-MODERATED REACTOR
A. P. Fraas

Aircraft Reactor Engineering Division

Four years of work on the ANP Project ot ORNL.
have led to the belief that the circulating-fluoride-
fuel reactor with reflector moderation and a spheri-
cal-shefl heat exchanger can be developed into
an aircraft power plant of exceptionally high per-
formance.!  This view has been supported ‘in
recent months by the results of studies by USAF
contractors. - The question has now become how
best to bridge the gap between the ARE and a
prototype aircraft power plant.

It was proposed about a year ago that a 50-Mw
reactor be built so that the feasibility of con-
structing and operating a circulating-fuel reflector-
moderated reactor could be investigated and its
performance characteristics, particularly with ref-
erence to control, shielding, heat transfer, and fluid
flow, could be determined. Preliminary estimates
indicated that a power of at least 25 Mw would
be required to disclose the more important control
characteristics. Other studies indicated that an
output of 50 Mw would be required for the lowest
powered aircraft likely to be tactically useful (for
radar picket. ships, patrol bombers, etc.), while
an output of 150 to 200 Mw would be required (with
chemical-fuel augmentation)} to power a strategic
bomber. Preliminary reactor test unit designs and
cost estimates indicated that the time and cost
involved would be roughly propertional to the rated
output of the reactor (largely because the size
and cost of the heat exchangers, pumps, and heat-
dump equipment vary directly with reactor power).
After much analysis and discussion it was decided
that 60 Mw represented a good compromise for the
power rating -and that such a reactor, to be called
the Circulating-Fuel Reactor Experiment (CFRE),
should be built by ORNL, with the aid of the
Pratt & Whitney Aircraft Division. This test unit
is to embody the basic features and proportions
that have made the circulating-fuel reactor at-
tractive for aircraft use, but the details of the
pumps, radiators, and cuxiliary equipment will not
have to meet aircraft requirements of size, weight,
etc. An operating life of 500 hr, of which a sub-
stantial portion should be at or near 60 Mw, was

 

TA. P. Fraas and A. W. Savelainen, ORNL Aircruft
Nuclear Power Plant Designs, ORNL-1721 (in press).

considered to be a desirable goal. The information
gained from this project should serve to provide
a sound basis for the design of the full-scale
aircraft reactor.

As has been the case with the ARE, the CFRE
will require a major supporting effort on the part
of various. groups in the ORNL organization to
obtain much essential information not yet avail-
able. The program on fuel chemistry, for example,
will continue along the lines it has followed. The
results of research on both the basic chemistry
and the in-pile and out-of-pile corrosion tests are
obviously vital to the project. Metallurgical re-
search on promising alloys, welding and brazing
techniques, strength properties at various tempera-
tures and in various ambients, and related work
will help greatly to increase the reliability, and
possibly the operating temperature, of the system.
Tests under way ot the Tower Shielding Facility
coupled with further Lid Tank Shielding Facility
tests will provide a more complete basis for the
aircraft shield design. Multigroup calculations
coupled with critical experiments should provide
additional information on the static physics of
the reactor. Much test work in experimental engi-
neering will be required to validate key features
of the design.

DESIGN OF THE CFRE
A. P. Fraas R. W. Bussard

Aircraft Reactor Engineering Division

The design of the CFRE is still preliminary,
but some tentative data can be presented at ‘this
As presently conceived, the CFRE will
include a reactor, heat-exchanger, pressure-sheli,
and shield assembly, as described in previcus
reports.?*3 A quasi-unit shield structurally and
functionally similar to one suitable for an aircraft
will probably be used,? although the structure will

stage.

 

2A. P. Fraoas, ANP Quar. Prog. Rep. Mar. 106, 1953,
ORNL-1515,.p 61.

3R. W. Bussard and A. P. Fraas, ANP Quur. Prog.
Rep. Dec. 10, 1953, ORNL-1649, p 31.

1g. p. Blizard and H. Goldstein {eds), Report of the
%353) Summer Shielding Session, ORMNML-1575 (June 11,

54). '

17
ANP QUARTERLY PROGRESS REPORT

be simplified if simplification proves to be ex-
pedient. l arge blowers coupled to banks of NaK-
to-air radiators will be used to provide heat dumps
that will be simpler, more reliable, and easier to
control than turbojet engines. While conventional
heat exchangers like those of the ARE might be
used, experience in fobricating and testing tube-
and-plate finned radiators® indicates that aircraft-
type radiators should cost little more and that they
will make o much neater, more compact instal-
lation.  Further, they will effect a manyfold re-
duction in the NaK heoldup in the system and will
give a thermal inertia and a NaK transit time es-
sentially the same as those for a full-scale aircraft
power plant. The fabricating and operating ex-
perience obtained should also prove to be quite
valuable, |t appears that d-c motors will be the
most practical means for driving the various pumps
in the system, but the blowers will probably be
driven by a-c motors, with control being effected
by the use of air bypass valves and/or shutters.
Key data for the CFRE are presented in Tables 2.1
and 2.2, and a flow sheet for the system is given
in Fig. 2.1. The major components are shown
approximately to scale on the flow sheet, although
their arrangement relative to each other is purely
schematic, The more important plumbing and
wiring connections expected to be made through
the reactor chomber wall and to quxiliary equipment

are listed in Tgbles 2.3 and 2.4,

CFRE COMPONENT DEVELOPMENT PROJECTS
The projected design effort for the CFRE has

been tied closely to o comprehensive series of
component development projects. A list of these
projects as currently conceived is given in Table
2.5. 1t is fully expected that the need for ad-
ditional projects will develop as the program
evolves, but those listed should provide the most
important information required.

The first project was completed recently, and
the results are presented below; work on some of
the other projects is well under way. The test
of beryllium in contact with sodium and Inconel
under thermal stress was given high priority be-
cause the results were needed in the determination
of the amount of poison to be expected in the

reflector. With the amount of poison in the re-

 

SW. S. Farmer et al., Preliminary Oesign and Per-
formanc)e of Sodium-to-Air Radiators, ORNL-1509 (Aug.
3, 1953).

18

flector known, the reactor critical mass, the core
size, the fuel concentration, etc. could be de-
termined. In the second test, which is well under
way, the hydrodynamics of the pump-volute, ple-
num-chamber, and core system are being studied.
Work on the same basic problem is also being done
by the Heat Transfer and Physical Properties
Group and by the Pratt & Whitney Aircraft Di-
vision. It is hoped that the different approaches
of the three groups will lead to a thoroughly sound
solution. The objective of the third project is to
produce a plastic model of a pump and header-tank
arrangement that will remove xenon from the fuel,
Data on the solubility of xenon in the fuel and on
the rate of production of xenon in the fuel are
being obtained by the Fuel Chemistry and Radi-
ation Damage groups; the data thus obtained will
provide a measure of the efficiency of xenon re-
moval by the pump, An in-pile test of a full-scale
pump may be required eventually. Xenon removal
by the pump is very important, from the control
standpoint; in addition, if it can be effected as
planned, the bulk of the other volatile fission
products will also be removed. This will dispose
of a major potential hazard in the event of a
reactor failure and should simplify the reactor
installation design problem.

The information given in Table 2.5 for most
of the other component development projects is
largely self-explanatory. To expedite the work,
some of the projects can be undertaken by other
organizations; for example, another organization
might undertake the entire job of designing, de-
veloping, and fabricating the fill-and-drain system,
the fuel addition equipment, the control rod drive
mechanism, or the sample heat exchanger tube

bundle.

The last project listed, the Zero Power Unit
(ZPU), deserves special mention. Experience with
the ARE has demonstrated that many problems will
arise in the fabrication of the CFRE and that there
will be o number of doubtful items, especially
welds, that could hardly be trusted in a high-power
reactor. This seems particularly true of the re-
actor core, heat exchanger, and pressure shell
assembly, If it is accepted that the first attempt
at fabricating the uwnit will leave much to be de-
sired, planning to use it as a hot mockup seems
to be more sensible than spending much time and
money trying to rework and patch it. It can be

endurance-tested ot temperature without developing
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 2.1. TENTATIVE SYSTEM DATA FOR THE tFRE FLUID CI‘RCUITS

 

Cirevit

 

General

Fuel ‘Na NaK

 

Fuel-to-NaK heat exchanger

Temperature drop (or rise), OF

Pressure drop, psi

Flow rate, cfs :
Velocifyifhmugh tube mairix; fps

Reynolds number ;

Over-oll heat transfer coefficient, Btu/hr-ft2- °F

Fuel-to-NaK temperature difference, °F

Moderator cooling system

Temperature drop (or rise) in heat exchanger, CF

Sodium-NakK temperature difference, °F
Pressure drop in heat exchanger, psi
F’ressureg drop in reflector, psi
Pressure drop in island, psi

Pressure drop in pressure shell, psi
Temperature rise in reflector, °F
Temperature rise in island, °F
Temperoture rise in pressure shell, °F
Power generated in reflecfor,I kw
Power generated in island, kw

Power generated in pressure shell, kw
Flow rate through reflector, cfs

Flow rate through island and pressure shell, cfs

NaK cooling system

Airflow through NaK radiators, cfm
Radiator air pressure drop, in. H20
Blower power required, hp (total)

Blower power required, hp (per fan}

Total radiator inlet face ared, 1%

Shield cooling system .

Power generated in lead layer, kw

Power generated in water layer, kw

Pump dato,

Head, ft

Fliow, gpm

Specific speed
Suction épecific speed
impeller speed, rpm
Drive, hp

Total fluid volumes, ft°

400 f 400
35 | 40
2.7 | 12.6
8.0 = 36.4
4,600

3,150

95

100 100
43

36
32

100

72
28

2,040

600

210
1.35
0.53

150,000
24

720

180

72

132
<4

50 250 280
600 430 2,600
3,750 : :
13,200

2,700

25 16 200

4.3 1.0 16

 

19
ANP QUARTERLY PROGRESS REPORT

TABLE 2.2, REFLECTOR-MODERATED REACTOR POWER PLANT SYSTEM DATA

 

 

 

 

 

 

200- Mw
Full-5cale 60-Mw CFRE
Aircraft

NaK in radiator cores, b

Header drums 128 40

Radiator tubes 128 40

Outlet lines 800 240

Inlet lines 200 270

Intermediate heat exchanger 450 135
NaoK flowing in circuits, 1b 2300 680
NaK in header tanks, |b 500 150
Total NaK in complete power plant, b 2800 840
NaK flow rate through radiator circuits (for a 400°F AT), cfs 40 12.6
NaK transit time through system, sec 1.25 1.25
NaK transit time through intermediate heat exchanger, sec 0.23 0.23
NaoK transit time through radiator tubes, sec 0.067 0.067
Fuel in core circuit (flowing), #3 6 3.9
Fuel in header tank, 3 1 0.4
Fuel in core, f13 1.8 1.8
Fuel flow rote, cfs 12.0 2.7
Fuel circuit transit time, sec 0.5 0.7
Therma! capacities of power plant, Btu/°F

Radiator cores 600 180

Linres and pumps 250 75

NaK (flowing) 600 180

Intermediate heat exchanger 250 75

Total for NaK systems 1700 510
Fuel 300 210
Total for power plant, Bru/°F 2000 720

 

nuclear power. Even if a leak should develop in
one of the welds, there would be no serious con-
tamination problem. Upon completion of the en-
durance test, it should not be difficult to make a
thorough inspection of the parts —~ something that
could hardly be done after sustained operation at
high nuclear power.

At first thought, the ZPU might appear to delay
the program, but closer inspection indicates that
it may actually save a substantial amount of time,
because efforts to get it into operation can proceed
at full speed without the inevitable hazards, re-

20

inhibitions, and hesitations that would
accompany high nuclear power operation. Work on
the CFRE can proceed at o more deliberate pace,
and the mistakes disclosed by work on the ZPU
can be avoided. The degree to which the ZPU
will be a mockup of the CFRE will depend upon
the problems that develop in the course of the
component test work. It is hoped that a 10-Mw
gas-fired furnace can be used, in place of one of
the two CFRE heat dumps, in making some heat
transfer tests on the fluoride-to-NaK heat ex-
changer and in obtaining a temperature gradient

straints,
o
-
~

\F
4

!

TN

X

PREHEAT RLANES
PREFIEAT RURNEN

MODRRATOR

BOOO cfm 9:0WER CRNL LR~ 2G 309

T TN
: : b : ! : Vo d ]
! } : ) ! ' } -

 

 

E * o PREFFEAT DOORS — -~

     
 
 

 
  
 
   

    

 

{ WMAIN Nok-TO- AR
§ REDIATOR

 

1z

Tl 75000 ofm

2 ok Fuwe | RN <7
{ g NaK PUMP ] N i
b, SEWER WATER Mo | \\ / |
' Y !
L L
LUBRICATION - G
Ul:fi\\((:f?b%l\ GiL , MalN NG

       
  
   
    
  

 

 

 

- OREREAT ELURNER

BURNER -~
GAS ~

PREHDAT

 

   

FoElL-T0-Nak

K S S

 

 

HEAT EXCHANGER
] .
!
ey
: \ /

 

N
{
i
\,
i
T

 

 

)

 

 

]

 

 

SEWER WATER MAIN
Fe VENT

 

75000 cfm
HLOWER

 

FLIWER

He VENT

 

 

 

 

 

N\
SR AND DRAIN
TANK - FUZL

 

i ]
\____,//“\ Pl AND DRAIN TANK - o

 

 

//\‘ 4
/1 [
- u

Fig. 2.1. CFRE Flow Sheet,

PS6L 0L ¥39WIIdIS ONIGNT QOyId
ANP QUARTERLY PROGRESS REPORT

TABLE 2.3. WIRING, PIPING, AND TUBING CONNECTIONS THROUGH REACTOR CHAMBER WALL

 

Thermocouples

Duplicate thermocouples to oll pipes leaving shield 52
Duplicate thermocouples to main pipes entering shield 8
Single thermocouples to oil inlet and outlet lines to pumps 8
Na and fuel temperatures 12
Pressure shell remperatures 12
Gamma-shield shell temperatures 12
Shield-water temperatures 12
Fuel and Na heut-dump system temperatures 24
Pump drive motor temperatures 8
Control rod drive temperatures 4
Miscellaneous support structure and leak warning temperctures 12

Total 164

Miscellaneous Instrument Wires

Muclear instruments 20

Pump speed 8

Power Wiring

Pump drive motors (4) 12
Control rod actuatar (1) 3
L.ights and receptacles 2
Fuel addition machine 1

Total 18

Tubing

Ma and fuel dump valve actuators™ (]{4 in. OD) 10
Expansion-tank pressure®* (]/4 in. OD) 2
Liquid-level gages* (]{1 in, OD) 2
Main NaK lines (5 in. QD) 4
Moderator-cooling NaK lines (2 in. OD) 4
Gamma-shield cooling water (1 in. OD) 2
Pump cooling and lubricating eil (]/5 ins OD) 4
Xenon vent ]
Atmosphere vent 1
Helium supply 2

T otal 32

 

*These tubes can be replaced by wires.

22
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 2.4. PRINCIPAL WIRING, PIPING, AND TUBING CONNECTIONS TO AUXILIARY EQUIPMENT

 

Thermocouples

Duplicate thermocouples to NaK monifold at outlets of all radiator cores 32

Dupllcrate thermocouples to main NaK pipes erflermg radiator cores 8
Single thermocouples to mlet and outlet lines to oil coolers 6
Rediater air inlet and outlef temperatures : f : 12
Shield_wuter inlet and oui‘lret temperatures . 4
NakK dump system temperatures ' ; 12
NaK pump drive motor temperatures , - ' 4

NaK pump temperatures

Miscellaneous support structure and leak warning temperatures ' 12

Total 28

Miscellaneous lnstrument Wires

Nucledr instruments _ : 20
NakK flow meters 6
NaK pump speed 4
Qil expansion tank level md:cotors 3
Shleld_wafer expansion tank level indicaters 3

Total 36

Power Wiring

NaK pbmp drive motors (2) 6
Oil pump drive motors (3) : 6
Shield:water pump drive motor (1) 2
Main blower drive motors (4) 8
Moderator-cooling system blower drive motor (1) 2
Preheat burner drive motors {3) é

Total 30

- Tubing

NaK dump-valve actuators (] in. OD) ‘ , ; 10
Na and fuel dump-valve octuators (] in. OD) * : i0
Expansion-tank pressures (/ ine OD) : f 2
NaK liquid-level gages ( Cin. OD) 2
Main NaK lines (5 in. OD) 4
Moderator-cooling NaK lires (2 in. OD) 4
Gamma-shield cooling water (1 in. OD) 2
Pump cooling and lubricating oil (16 in. OD) 7
Xenon vent to buried charcoal bed _ 1
Helium supply to expansicn tanks : 5

Total 37

 

23
in the ZPU system. While this arrangement would
hardly represent normal operation, it would simu-
late the one-engine-out condition rather well, It
is not yet clear just what will be done, but every
effort will be made to get as much information as
possible from the ZPU, consistent with the money
and the time available.

REACTOR PHYSICS

W. K. Ergen M. E. LaVerne
R. R. Coveyou C. B. Mills

Aircraft Reactor Engineering Division

R. Bate C. S. Burtnetie
United States Air Force

The three-group, three-region code for reactor
statics calculations® on the ORACLE has been
completed. The temperature effects for the CFRE
were computed on the UNIVAC for three reactors:
a 50-Mw reactor with Incone! coolant tubes in the
reflector, a reactor without these tubes, and a
reactor in which the !Inconel in the core shells was
replaced by columbium clad with 0.010 in. of
The temperature coefficient for rapid
temperature changes turned out to be about -3.5 x
1073 /°F, and the temperature coefficient for slow
effects is —2.6 x 1072 /°F for the heavily poisoned
reactor with the Inconel coolant tubes, but it in-
creases to —3.3 x 107°/°F and -3.8 x 1073 /°F,
respectively, for the reactor without the lnconel
coolant tubes and the reactor with the Inconel-clad
columbium core shells.

Inconel.

REACTOR CALCULATIONS
M. E. LaVYerne

Aircraft Reactor Engineering Division

C. S. Burtnette
United States Air Force

Results of the parametric reactor study done on
the UNIVAC ot the AEC Computing Facility in
New York have been published,”*® and the Curtiss-
Wright Corp. has completed, under contract to the
United States Air Force, a series of multigroup,
multiregion calculations for the ORNL-ANP Proj-

 

5w, K. Ergen, ANP Quar. Prog. Rep. June 10, 1954,
ORNML-1729, p 32.

7C. S. Burtnette, M. E. LaVerne, and C. B. Mills,
Reflector-Moderated-Reactor Design Parameter Study,
Part 1, Effects of Reactor Proportions, ORNI. CF-54-
7-5 (to be publishad).

BM. E. LaVerne and C. S. Burinette, ANP Quar. Prog.
Rep. June 10, 1954, ORNL-1729, p 32.

24

ect. A report on the results of the latter calcu-
lations is being prepared and will be issued by the
Curtiss-Wright Corp.

A redetermination of the critical mass for the
(CA-19)
was found to be necessary because of an error in
carbon transport cross-section scaling, a differ-

rhombicuboctahedron critical assembly

ence in reftlector geometry from that assumed for
the original calculations, and an actual Teflon
density lower than that used in the calculations.
In addition, a small correction {about 1% of critical
mass) was made for the presence of the aluminum
lattice supporting the reactor and for the aluminum
control red sheaths.

The eigenvalues of the difference-equation ap-
proximation fo the age-diffusion differential equa-
tion vary with lattice spacing. Calculations were
therefore made with three different spacings in
order to evaluate this effect. A foil of 0.004 in.
thickness was used in the calculagtions. The
resuvlts, corrected as mentioned in the preceding
paragraph, are given in the following:

Space Interval, Ar Calculated Critical Mass

(em) (Ib)

0.457 22.75
0.914 21.47
1.828 20.91

The data are plotted in Fig. 2.2. A Lagrangian
extrapolation to a zero lattice spacing gives a
critical mass estimate of 24.70 Ib. The assembly
became critical at a loading of 24.35 b of U233,

It is recognized that the agreement between
calculation and experiment may well be fortuitous,
since the critical mass is not very sensitive to
errors in detail. Further evaluation of agreement
(or lack thereof) must await additional experi-
mental results. Calculations are now being made
for the next critical assemblies, that is, the three-
region assemblies consisting of a beryllium island
and reflector with a Teflon and uranium-foil fuel
annulus. The fuel annulus will be surrounded by
core shells; aluminum core shells will be used for
one assembly and Inconel for the next.

BERYLLIUM THERMAL STRESS TEST
R. W. Bussard R, E. MacPherson

Aircraft Reactor Engineering Division

One of the key questions in the design of the
CFRE, as mentioned above, has been that of the
TABLE 2.5. SUMMARY OF CFRE DEVELOPMENT PROJECTS

 

 

o Estimated
ject
roies Project Title Objective Number of References Status September 1, 1954
Number
Man Hours
1 Beryllium Thermal Stress investigation of beryllium under severe therma! stress 2280 ORNL-1515, -1517, and -1721 First test completed; no cracks
conditions . appeared in beryllium speci-
men
2 Core Hydrodynamics Investigation of flow separation, velocity distri- 3880 ORNL-1515, -1692, -1701; YF-15-11; Memo, Tests under way with water
bution, and perverse particle dwell time Fraas to file, B/4/54; Memo, Wislicenus
to Briant, 6/19/53
3 Xenon Removal Pump (plastic Development of attitude-insensitive pump and header 3080 ORNL-1649, ORNL CF-54+5-1; Memo, Tests under way with water
model) tank geometry to remove xenon and degas fuel Fraas to file, 8/4/54
4 Xenon Removal Pump (hot test Performance and cavitation tests with water, 5200 ORNL-1649; Memo, Froas to file, 8/4/54 Design completed; fabricotion
unit) sodium, ond fuel; endurance test under way
5 Model-T Component Test Check of fabricability and endurance test of fuel 8920 Pump, header-tank, and core-
{isothermal) pump, plenum chamber, and core shell assembly shell design held pending
results of projects Z and 3
6 Model-T Component Test (100-kw Check for flow separation; estimate of full-scale 4400 ORNI_ CF-54-2-35 Same as above
heat input to core} boundary layer thickness
7 Second Fuel-to-NaK Heat Exchanger Performance and endurance test at full-scale pres- 9280 ORNL-1215, -1330, -1509, 1729 (p 28); Design of about one half of
{4000 kw, 1000-kw input) sures, pressure drops, and temperature differentials CF-54-1-155; Memo, Ahern to Fraas, components completed; heat
6/12/54 exchanger being fabricated
8 Incone! Therma! Stress Investigation of effects of severe thermal stress 2680 Design of test rig nearly
cycling on cracking and distortion completed
9 Leak Choracteristics of Tube<to- investigation of possibility of self-plugging of weld 760 ORNL-1215, ORNL CF-54-1-155 Parts being fabricated
Header Welds cracks between fuel and NaK
10 Fill-and-Drain System Performance test with water, Nak, and fuel 2280 © Detail design completed for
one proposal
11 Control Drive Mechanism Check of fabricability; shckedown and endurance 3200 Preliminary loyouts completed
tests for several proposals
12 Fuel Addition System Check of fabricability; shakedown and endurance 3200 Preliminary layouts completed
tests for several proposals
13 Wood and Plastic Reoctor Mockup Check of fabricability and test demonstration 260 Fabrication under way
(]/2 scale)
14 Sample Heat Exchanger Tube Check of fabricability and cost determination 22490 Detail design completed for
Bundle one proposal
15 Zero Power Unit Check of fabricobility; shakedown and endurance Preliminary design under way

tests

 

25
SR
ORNL—LR—DWG 3092

 

 

 

 

27 e : |

26 e
. ® MULTIGROUP CALCULATIONS
- 254; 1 —— A |AGRANGIAN EXTRAPOLATION
od
> o\ ® EXPERIMENTAL VALUE (CA—19)
E 24 + \VL ‘ : : | | |
= N FOIL THICKNESS = 0.004 in.
w) \\ |
n 23 . ~ - . / | i} i
< n
=
Ja2 | N I
S l
= a

20 !

| |
‘ }
19 ~E
O 02 04 06 08 10 412 14 16 48 2.0

Fig. 2.2. Effect of Space Interval Size on Com-
puted Critical Mass.

effects of severe thermal stresses in beryllium at
temperatures between 1000 and 1300°F.% 1% There
has been serious concern as to whether cracking
or distortion in the beryllium would prove to be a
major problem. Therefore a test on the beryllium
block shown in Fig. 2.3 was devised to investigate
the effects of power (and hence thermal) cycling.
The test loop used is shown in Fig. 2.4. Heat
was generated in the beryllium block by simple
electrical-resistance heating with a current of
about 15,000 amp through the beryllium block.
Sodium flowed from an electromagnetic pump to
the block inlet header, upward through the lower

 

%R. W. Bussard et al., The Moderator Cooling System
for the Reflector-Moderated Reactor, ORNL-1517 (Jan.
22, 1954).

10g, A, Field, Temperature Gradient ond Thermal
Stresses in Heat Generating Bodies, ORNL CF-54-5-196
(May 21, 1954).

UNCLASSIFIED
PHOTO 22390

 

 

 

BERYLLIUM

 

Fig. 2.3. Beryllium Test Block.

27
ANP QUARTERLY PROGRESS REPORT

RADIATOR ———-___

  
  
   
    

ELECTROMAGNETIC
FLOWMETER

SURGE TANK

 

BYPASS FILTER

 

 

LUG COOLANT LINES

UNCLASSIFIED
ORNL-LR-DWG 3093

250-kva TRANSFORMER

    
  

  

 

 

 

|

 

 

 

 
 

I|

I L
m ' ["~—AIR-COOLED COPPER BUS BAR

—_
X
N\
X

 
  

CONTACT
LUG

 

 

 

 

L
Na INLET HEADER

CANNED BERYLLIUM BLOCK

Fig. 2.4. Beryllium Thermal Stress Test Apparatus.

tube bank to the corresponding holes in the be-
ryllium block, through the upper tube back into
the outlet header, out to the radiator section, and
back to the pump. Banks of tubes were used to
duct the sodium to and from the block to minimize
the flow of electrical current through the sodium
inlet and outlet header tanks. A bypass filter
arrangement was provided, as well as bypass
cooling flow to the sodium-filled lugs connected
to the transformer bus bars.

The high-power-density volume heat source
coupled with transverse sodium flow through the
drilled holes gave a high thermal gradient around
the holes. During the course of the test the power
density was cycled regularly by alternating the

28

operating conditions as shown in Table 2.6, which
also compares the operating conditions for the
thermal stress test with those for the CFRE,

The test included an initial 100-hr period at a
constant high power density followed by power
cycling at the rate of one cycle per day. The
changes from one power level to the other were,
in all cases, performed at a fairly uniform rate in
a l-min interval. The test was concluded after
1000 hr of operation, including 36 cycles. The
beryllium sample is now being examined by the
Metallurgy Division for dimensional stability and
evidence of mass transfer, corrosion, or erosion.
Visual inspection of the sample after the test
indicated that no distortion, cracking, or erosion
of the beryllium had taken place.
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 2.6, COMPARISON OF OPERATING CONDITIONS FOR BERYLLIUM
THERMAL STRESS APPARATUS AND CFRE

 

Beryllium Block

 

CFRE
Power level High Low 60 Mw
Peak power density in berylliom, w/cm’ 40 14 70
Average power density in baryllium, w/cm® 0 14 1.5
Cooling passage diameter, in. 0.25 0.25 0.25
Cooling passage spacing (minimum), in. 1.5 1.5 1.0
Calculated thermal stress (no plastic flow), psi 108,000 38,000 42,000
Sodium inlet temperature to beryllium, °F 1,040 1,060 - 1,050 ’
Sodium outlet temperature from beryllium, °F 1,100 1,080 1,150
Average sedium temperature through beryllium, O F 1,070 1,070 1,100
Maximum thermocouple reading in beryllium, ©F . 1,200 1,110
Calculated maximum beryllia;m temperature, ° F . 1,280 1,150 | 1,230
Time at power level, hr/doy 8 16
Heating .currenf, amp 14,300 8,450

 

 

BERYLLIUM-INCONEL-50DIUM SYSTEMS

The compatibility of sodium, beryllium, and In-
conel has been studied by using whirligig, seesaw,
thermal-convection-loop, and = forced-circulation-
loop tests. The results obtained to date indicate
that it will be possible to use beryllium unclad
in the CFRE if the temperature of the sodium-
beryllium-Inconel region (that is, the reflector and
the island) is kept below 1200°F. The main effect
observed thus for in the tests has been the type
of mass transfer usually found when dissimilar
metals are used in a system. In this case the
mass transfer has been evidenced by the alloying
of beryllium with the Inconel walls in regions
where the beryllium and Inconel were close to-
gether, with only stagnant sodium separating them.

Beryllium-lhconel-Sodium Compatibility Tests

E. E. Hoffman W. H. Cook
C. R. Brooks C. F. Leitten
Metallurgy Division

Nine tests of the compatibility of beryllium,
sodium, and inconel in dynamic systems have been
completed (Table 2.7). In the whirligig tests o
circular loop of Inconel tubing with a berylllum
insert was pdrfiy filled with sodium which was
circulated at o velocity of 10 fps. For four of

 

these tests there was no temperature differential,
but for one test there was a differential of 140°F.
These tests were of 300-hr duration. Two thermal-
convection loops of Inconel with beryllium inserts
were tested for 1000 hr with sodium circulating
at a relatively low velocity (approximately 2 to
10 fpm). In addition, seesaw tests were performed
in which beryllium specimens were retained in the
hot zones of osciilating Inconel tubes partly filled
with sodium which cvrculoted at a velocity of

3 fpm.

Macroscopic examination failed to show attack
on the Inconel, but two specimens showed de-
posits. Black layers were formed in the annular
space between the beryllium insert and the en-
closing Inconel sleeve in the whirligig and thermal-
convection-loop tests. The layer was very ad-
herent and made removal of the insert difficult;
consequently, the weight change data obtained
were not considered as being relmb!e.

A picture of this layer from © whirligig ioop
operated at 1300°F is shown in Fig. 2.5. Micro-
scopic examination showed that the maximum depth
of attack on the inside surface of the Inconel did
not exceed 0.5 mil. Chemical analysis of the
sodium bath reveaied less than 0.04% berylliom
in the sodium. In all tests, beryllium was detected

29
ANP QUARTERLY PROGRESS REPORT

 

TABLE 2.7. RESULTS OF BERYLLIUM-SODIUM-INCONEL COMPATIBILITY TESTS

 

Concentration of

 

 

Test T Time Temperature Be on Surface Q .
est IYPE b CE) of Inconel Tube emarks
(g/em?)

Whirligig 300 1100 (isothermal) 53ta 7.9 Two dark deposits on Inconel analyzed high in
heryllivm; macroscopically, beryllium showed
fine pitted appearance

300 1200 (isothermal) 0.05 No visible deposits on Inconel; beryllium smooth
but discolored

300 1300 (isothermal) 0.07 No visible deposits on lnconel; beryllium showed
dark discontinuous deposit covering surface

300 1400 (isothermal) 39.6 to 127.0 No visible deposits on Inconel; beryllium surface
pitted and partly covered with small deposits

270  Hot zone, 1200 loose, adherent gray deposit found in cold zone,

Cold zone, 1060 analyzed high in sodium; beryllium smooth but

discolored

Thermal- 1000 Hot zone, 1150 Inconel appeared smooth and showed no visible

convection Cold zone, 990 deposits; beryllium smooth but discolored
loop
1000  Hot zone, 1300 Inconel appeared smooth and showed no visible
Cold zone, 1130 deposits; beryllium specimen smooth and light
gray in color

Seesaw 300 Hot zone, 1400 Hot zone, 14.0 Beryllium specimen covered with black flaky

Cold zone, 710 Middle, 9.9 deposit; surface of Inconel tube discolored
Cold zone, 20.6
300 Hot zone, 1100 Hot zone, 0.07 Similar to above

on the Inconel tube wall surface.
of the Inconel surfaces indicated that the beryllium

Cold zone, 740

X-ray analyses

Middle, 0.02
Cold zone, 0.01

the inside surface, but attack up to 5 mils was
observed on the ocutside surface which was in

was present as the metal.

In the thermal-convection loops, a deposit was
formed between the contacting ends of the be-
ryllium insert and the Incone! tube. X-ray anal-
yses of these deposits and those found in the
whirligig loops showed beryllium, Inconel, and
lines very similar to sodium bicarbonate. The
beryllium specimens from the seesaw tests were
covered with a black, flaky deposit, which has
been identified by x-ray anaclysis as beryllium

oxide and beryllivim,  Again, lines similar to
sodium bicarbonate were observed.
Microscopic examination of the beryllium speci-

men from a whirligig test showed little attack on

30

contact with the relatively stagnant sodium in the

0.005-in. annular space (Fig. 2.6).

Mass Transfer Tests in Thermal-Convection
Loops

G. M. Adamson
Metallurgy Division

A series of thermal-convection loops is being
operated to determine the effect of dissimilar metal
transfer in the beryllium-Inconel-sodium
These studies are being carried ouf in

mass
system.
Inconel loops with short beryllium inserts in the
hot legs; sodium is the circulated fluid, The
 

Figa 2-5-

PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
Y-12811

S

INCH

 

 

 

Inside Surface of Inconel Sleeve Enclosing Beryllium Insert Exposed in Whirligig Test to

Sodium for 300 hr at 1300°F, Etched with glyceria regia. 1000X.

TABLE 2.8. EFFECT OF CIRCULATING SODIUM AT YARIOUS TEMPERATURES FOR 500 he IN
INCONEL THERMAL-CONVECTION LOOPS WiTH BERYLLIUM SPECIMENS IN THE HOT LEGS

 

Metallographic Examination

 

 

 

Temperature N

°F) Beryllium Specimen inconel Hot Leg

1500 Outside surface of insert showed holes to a depth Deposit opposite the beryllium
of 14 mils; inside surface slightly rough insert

1300 Intergranular penetrations to a depth of 2 mils on Mo attack or deposit
outside surface; no attack on inside surface

1100 No layer or attack No attack or deposit

200 No layer or attack No attack or deposit

 

results reported in Table 2.8 are from tests oper-
ated for 500 hr. Other loops are now in operation
that wifl be run for longer periods. k
Photomicrographs of the outsides of the be-
ryllium inserts are shown in Fig. 2.7. Black,
flaky material was found in the cold legs of the

loops operated ot 1100, 1300, and 1500°F and

in the hot leg of the loop operated at 1500°F. The
deposit material has not yet been identified by
spectrographic or diffraction studies. The deposit
found on the Inconel opposite the beryllium insert
in the loop operated at 1500°F is shown in Fig.
2.8; it has been identified as a nickel-beryllium
alloy.

31
ANP QUARTERLY PROGRESS REPORT

 

NCLASSIFIED
¥-12994

 

 

Fig. 2.6. Beryllium Specimen After Whirligig Test for 300 br ot 1300°F. (a) Surface exposed to flowing

sodium,
and the Inconel sleeve.

Mass Transfer in Forced-Circulation-Loop Tests

L. A, Mann

Aircraft Reactor Engineering Division

F. A. Anderson'!

University of Mississippi

A test unit!? for investigating mass transfer of
beryllium to Incone! in flowing sodium was oper-
ated for 1006 hr under conditions nearly the same
as those proposed for the reflector-moderated re-
The approximate conditions of operation,
with minor fluctuations, were as follows: sodium
flow rate, 3 gpm; beryllium temperafure, 1200°F;
temperature, 1245°F (sodium entering
economizer); minimum flowing sodium temperature,
900°F (re-entering economizer from cooling sec-
tion); Reynolds number ot minor diameter (0.25 in.)

actor.

Maxiimum

nSummer Resesarch Participant.
120 A. Mann, ANP Quar. Prog. Rep. Jure 10, 1954,
ORNL-1729, p 22.

32

(b} Surface exposed to relatively stagnant sodium in the annular space betwaen the specimen

Etched with oxalic acid. 250X,

of beryllium insert, 155,000; minimum Reynolds
number (at coldest point in economizer annulus),
61,000.

Upon termination of the test, the unit was cooled
rapidly to room temperature to freeze the sodium.
Samples of both the sodium and the Inconel! were
then taken from each of 24 locations in the unit.
The sodium samples were analyzed chemically for
beryllium, und none was found. Metallurgical ex-
amination of the beryllium piece and the Inconel
samples indicated that the beryllium-lnconel-
sodium system should give no trouble from cor-
rosion or transfer under the conditions
presently proposed for the reflector-moderated re-

1200°F, Reynolds

Mmass

actor (temperature not over
number not over 150,000).

The dbove-described test unit was essentially
duplicated, and o second test was started with
maximum and minimum temperatures in the flowing
sodium of 1300°F {at the beryllium) and 1000°F,

respectively. A power failure terminated this test
B UNCLASSITIED
T-583939

) TRERD
§. t-5%az

it s

UNOLASSIFIED
. Y-goat

 

 

 

Fig. 2.7. Surfaces of Berylliuminserts in Inconel
Thermal-Convection Loops After Exposure to
Flowing Sodium for 500 hr at (z) 1100°F, (%)
1300°F, (c) 1500°F. Unetched. 250X. Reduced
36%. '

PERIOD ENDING SEPTEMBER 10, 1954

after 268 hr of operation. Sodium and Inconel
samples and the beryllium insert were removed for
analysis and examination, but results have not yet
been received. A third unit has been built and is
intended to operafe for 1000 hr.

M onduassierenl
- r.s0s2 |

WS

 

 

 

Fig. 2.8. Layer on Inconel Opposite Beryllium
Insert in Thermal-Convection Loop That Circu-
lated Sodium at 1500°F for 500 hr. Unetched.
250X. Reduced 36%. :

33
ANP QUARTERLY PROGRESS REPORT

3. EXPERIMENTAL REACTOR ENGINEERING
H. W. Savage

Aircraft Reactor Engineering Division

The emphasis in the engineering work has shified
to research and component development for the
CFRE and for in-pile loops, since most of the
preoperational ARE tests have been completed.
Twe types of puinps have been developed for in-
pile loop use, The vertical-shaft centrifugal sump
pump, which would be installed outside the reoctor
shield and thus would require auxiliary shielding,
is now being fabricated in sufficient quantity to
meet the demands of the in-pile loop program and
the forced-circulation corrosion testing program,
The small (4-in.-OD)}, air-driven, horizontal-shaft
sump pump being developed for insertion in a
reactor beam hole was tested with NaF-ZiF, at
1350°F.  Some difficulty was encountered with
initial priming, but operation was otherwise satis-
factory. A new, small pump that has the required
small holdup volume is being designed. This
pump, which will use o turbine-type impeller, has
the advantage that both the inlet and discharge
can be at the bottom. Hydraulic drive motors of
suitably small dimensions have been found to be
satisfoctory drives for these pumps.

Additional work has been done on the development
of forced-circulation corrosion-testing loops for
obtaining information on the corrosion of Inconel
in high-velocity turbulent fluoride mixtures with
large temperature differentials in the system. Two
series of loops are being constructed to meet the
following requirements: a Reynolds number of
10,000 with temperature gradients of 100, 200, and
300°F and o temperature grodient of 200°F with
Reynolds numbers of 800, 3,000, and 15,000. The
maximum fluid temperature is to be 1500°F. A
forced-circulation loop is also being developed for
testing combinations of structural metals in contact
with high-velocity turbulent liquid metals under
high temperature differentials,

Several
burners for use with the proposed gas-furnace heat
source for high-temperature reactor mockup tests.
Also, a study is under way of the cavitation phe-
nomenon associated with operating liquid metal
systems at elevated temperatures, high flow rates,

exploratory tests were made of gas

'D, F. Salmon, ANP Quar. Prog. Rep. June 10, 1954,
ORNL-1729, p 19.

 

34

and high pump speeds. A correlation of fluid-flow-
noise intensity with pressure data was noted.

The number of stations available for convection
loop testing has been increased from 18 to 31 so
that many more long-term tests (2000 hr or longer)
and intensive tests of special materials can be
The basic design of the convection loops
has been simplified, and various means of heating
the loops and of making operation of them more
gutomdatic are being studied.

made.

IN-PILE LOOP COMPONENT DEVELOPMENT

¥W. B, McDonald

Aircratt Reactor Engineering Division

An in-pile loop for insertion in the MTR is to be
designed, constructed, and operoted as a joint
ORNL and Pratt & Whitney Aircraft Division
project. The loop is being developed for circu.
lating proposed fuel mixtures so that the extent
of radiation damage to materials of construction and
the effect of radiation on the fuel can be de-
termined, Preliminary ORNL work on this project
is concerned with the design and development of
components of a loop to operate in o horizontal
beam hole. Further developmental work was also
done on the vertical-shaft centrifugal pump for use
with both in-pile and out-of-pile forced-circu-
lation loops.

Horizontal-Shaft Sump Pump
D. F. Salmon J. A, Conlin

Aircraft Reactor Engineering Division

The air-driven horizontal-shaft sump pump de-
scribed previously! was operated with NaF-ZrF ,
at 1350°F, |t produced a 70-psi head at 6000 rpm
and a flow rate of 1.5 gpm. Initial difficulty in
priming the pump was solved by momentarily
flooding the impel!ler labyrinth. This pump, which
was built as a pilot model for checking design
principles, hot pump performance, and reliability,
has too large a volume holdup for in-pile use and
will now be used as a laboratory pump.

A new pump has been desighed that has the
required small pump holdup volume. For priming,
this design incorporates a boffle in the sump tank
that will permit remote, momentary flooding of the
impeller eye and the labyrinth behind the impeller.
A turbine-type impeller is used instead of the
conventional centrifugal impeller.

it has been experimentally determined, with
water, that the turbine type of impeller will prime
itself if the inlet and outlet are placed at the
bottom of the pump and the fluid is maintained at
a level about half way between the bottom of the
impeller and the impeller center. The horizontal
sump pump principle is used, with the sump having
a minimum volume that acts merely as an expansion
chamber and a reservoir to replace and catch the
fluid leaking past the labyrinth seal. Excessive
impeller end clearances are necessary for this
type of pump because of the severe temperature
gradients present. However, estimated available
pump heads (25 ft at 4500 rpm and 1.5 gpm} are
greater than the estimated requirements. Hydraulic
efficiency of the pump will be low but performance
will be reliable.

Further work has been done on centrifugally
sealed and frozen-fluoride-sealed pumps, and it
has been established that the centrifugally sealed
pump requires too greata volume of fluid for sealing
for it to be considered for in-pile use and that the
leakage rate of the frozen-fluoride-sealed pump is
excessive. However, the centrifugally sealed
pump, which will be useful for other applications,
has been modified to include a system that makes
the pump self-priming and deaerating. By con-
trolling the seal level and using a proper venting

 

__PRESSURE EQUALIZATION LINE
L RARGE ~ PUMP HOUSING
VENT —- CISCHARGE J
‘“] TN
PROBE. . ) 4 s

 

 

HOLE - -aa_ \

 

 

 

 

 

 

 

 

 

  

  

   

 

 

 

 

 

 

 

 

 

(I e { B2 [0 —T J\{ - - SHAFT \

PERIOD ENDING SEPTEMBER 10, 1954

sequence, the pump can be easily stopped and
restarted. A pilot mode! of the pump, similar to
the pump previously described,? has been designed
that incorporates these features (Fig. 3.1} and an
integral sump tank connected by a passage to the
seal cavity.

In operation, the system would be filled tc the
full level. The seal cavity and sump tank would
then be pressurized through the gas inlet tube to
the predetermined pressure necessary to obtain a
positive pump inlet pressure during operation, and
the pump would be started. The centrifugal force
would cause the liquid in the seal cavity to form a
rotating annulus. The annulus of fluid would
develop a small pressure and thus pump fluid into
the discharge tube, through the loop, and back to
the pump. The small bieed hole from the pump
discharge to the sump tank would permit deaeration
by bypussing aerated fluid from the pump into the
sump. The bypassed fluid would be replaced by
clear fluid from the seal. When the pump was
primed and deaerated, a continuous flow of fluid
would pass from the bleed hole to the sump to
remove any accumulation of gas. Throughout this
operation, fluid entering the loop from the seal to
either fill the loop or replace the bleed flow would
be replaced by fluid from the sump, and sufficient
fluid would thus be in the seal for sealing. There-
fore there is a lower limit on the run level in the

 

2J. A, Conlin, ANP Quar, Prog. Rep., June 10, 1954,
ORNL-1729, p 18.

UNCLASSIFIED
QRMIL~LR-DWG 3094

LGAS INLET Ol OUTLET -
™~

      

—

A

 

 

      
    

 

 

%
SOl INLET

OlL. DRAIN

Fig. 3.1. Centrifugally Sealed Sump Pump,

35
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 3095

lj?‘s'm e SIR Y 0%

_.—ROTARY SEAL UNION

  
 
 
  
 
   

 

 

 

 

_|-SPARK PLUG PROBE |

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

|
£
\—‘\N
M
’ o
i
COOLING OIL TO ANNULUS |
|
PRESSURIZING GAS LINE ‘
e f— !
i
i
LIQUID LEVEL - -
VOLUME -REDUCING .
ATTACHMENT (OPTIONAL) — fm
INLET |
ANTIFLOWBACK SEAL \ !
IMPELLER VANE = I
. |
Y3,-in HOLE |
DRILLED IN SHAFT [

DISCHARGE

Fig. 3.2. VYertical-Shaft Centrifugal Sump Pump (Model LFB),

36
sump in that it must be sufficiently high at all
times to provide the necessary flow into the seal.

To stop the unit, the sump tank would be vented
at the same time the pump was stopped, but the
gas pressure would be maintained unti! draining
had been completed. This would cause a flow of
gas along the shaft into the seal cavity and would
blow out any fluid that might enter along the shoft.
I+ would also maintain a seal cavity pressure
somewhat greater than the sump tank pressure and
would force the fluid out of the seal into the sump.
If this procedure were not followed, the sealing
annulus of fluid would, upon collapsing when the
pump was stopped, momentarily fill the seal cavity
above the shaft level and cause fluid to enter
along the shaft,

Vertical-Shaft Centrifugal Sump Pump
D. R. Ward

Aircraft Reactor Engineering Division

Developmental work has been continued on the
vertical-shaft cenirifugal sump pump (Model LFB)
for use with forced-circulation loops, both in-pile
and out. This is a vertical-shaft cenirifugal sump
pump with a side inlet and a bottom discharge
(Fig. 3.2). A Graphitar face seal prevents the
pressurizing gas from escaping along the shaft.
Circulating spindle oil serves to lubricate this
face seal and to carry away heat from the seal
region.

Typical performance curves for this pump with
water are shown in Fig. 3.3. In o recent test the
pump was run for 864 hr at 6000 rpm while pumping
1400°F sodium at about 3 gpm. Typical oil leakage
past the shaft face seal was under 0.1 cm3/hr.

Developmental work during the past eight months
has resulted in the following improvements in this
pump: the elimination of the tendency for the pump
to introduce gas into the liquid stream at flow
rates up to 6 gpm; an improved anti-flowback seal
to prevent recirculation of the liquid within the
pump; an improved means for cooling the face-seal
region of the pump; and a reduction of pump volume,
which is desirable when the pump is used for
in-pile loop tests.

The main difficulty, that of gas entrainment, was
primarily caused by vortexing and the passage of
gas downward along the rotating shaft. The ad-

PERIOD ENDING SEPTEMBER 10, 1954

UNCILLASSIFIED
ORNL-LR-OWG 3096

 

48

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[
’_'--...,__‘_. i
\*\ ,,,,,,,
*\ 6000 rpm
T —— ~g —
32 b
&
\4
—_ S ——
o - e e -~ 5000
7 .\J 5 1 rpm
2 e
< pa - - e
T
s e
—_
2 Ay
! \.\ 4000 rpm
o f—1 e
.;'!.fiv-.....‘;_;i.; ---------- — I !
\’\?N 3000 rpm
8 ..................... — “\+;fi.__ ,,,,,,,,,,,,,
0
0 2 4 8 8
FLow {gpm)
Fig. 3.3. Performance Curves for Vertical-

Shaft Centrifugal Sump Pump (Model LFB).

dition of vertical baffles adjacent to the shaft
merely broke up the large vortex into several
smaller ones, which were just os objectionable.
The problem is solved in the present arrangement
by permitting liquid from the high-pressure region
of the pump to be bypassed up through a hole in
the shaft to an annular space surrounding the
shaft.  This liquid, in turn, forms a seal and
prevents the gas from following its former path.

Acceptance tests for these pumps include oper-
ating each pump with water to check the pumping
characteristics and freedom from gas entrainment
and pumping sodium at T1400°F for 72 hr at 6000
rpm with acceptable oil leakage past the face seal.
This hot test also serves to clean oxide film from
the wetted pump parts, as well as to reveal any
mechanical faults that might be present. Five of
these pumps are being built, and four similar
pumps {Model LFA} are now in use.

37
ANP QUARTERLY PROGRESS REPORT

Hydrauvlic Motor Pump Drives
4. A. Conlin

Aircraft Reactor Engineering Division

Two types of hydraulic motors have been found
to be suitable for use os in-pile pump drives. The
first, a gear-type motor rated at 1.75 shp at 3500
rpm, was operated at 6000 rpm for 500 hr under a
light load (about 0.08 shp) that was obtained by
Upon
completion of the test the motor was disassembled
and was found to be in good condition. The rubber
lip shaft seal had worn a slight groove in the
shaft, and o carbon-like deposit which had ac-
cumulated on the shaft just outside the seal caused
the seal to stick slightly after it had been idle for
a prolonged period. However, throughout the test,
sea! leakage was nregligible. The only other indi-
cation of weoar was a slight scoring on the gear
sides adjocent to the shaft end of the motor.

The second motor, an axial-piston type rated at
3.8 shp at 6060 rpm continuous duty, was run for
100 hr at 6000 rpm. Operation was completely
satisfactory, and, since this motor is rated at
6060 rpm, further testing was judged to be un-
necessary.

The axialepiston motor was chosen for in-pile
service, since it is considered to he the more
reliable unit. Throughout these tests, hydraulic
oil with a viscosity of 155 SSU (seconds Saybolt
universal) at 100°F and a viscosity index of 100
It should be noted that during initial
testing of the gear-type motor, oil with a viscosity
of 290 to 325 SSU at 100°F was used, and, as a
result, excessive overheating of the motor occurred
to the extent that the shaft turned blue. Despite
this, the motor completed the test satisfactorily.

coupling the motor to @ small air blower.

was used,

FORCED-CIRCULATION CORROSION LOOPS
Incone! Loops

W, C. Tunnell
Aircroft Reactor Engineering Division
W. K. Stair, Consultant
J. F. Bailey, Consultant
University of Tennessee

The previously described, small, Inconel, forced-
circulation loop designed to have a temperature
differential of 135°F has been operated for a total
of 141 hr with NoF.ZrF UF  (50-46-4 mole %)
at ain apparent mass flow rate of 400 to 450 Btu/hr
with a temperature differential up to 300°F.% The

38

fluid velocity in the loop normally was maintained
at about 1.7 fps at a temperature gradient of 165°F.
Termination of the test resulted from a short in
the motor leads and subsequent rupture of the
heated section of the loop. The test facility,
shown in Fig. 3.4, is being rebuilt and is approxi-
mately 80% complete. The final design conditions
are presented in Fig. 3.5. Reports which give a
detailed presentation of the design and an analysis
of the initial operation are being prepared.?s3

In the design of the loop for electrical-resistance
heating, a resistivity value® of 98 gohm-cm was
used for Inconel in computing the necessary length
of the resistonce-heated tube. However, data
obtained during the loop operation indicated that
the actual resistivity of the Inconel was about
70 to 80 pohm-cm. To resolve the discrepancy, a
duplicate of the heuating section of the loop was
filled with NaF-ZrF .UF, (50-46-4 mole %) and
alternately heated and
potential ond current measurements were made.
Sixty-two measurements were made in the tempera-
ture range of 800 to 1600°F, and the average
resistivity of the Inconel was found to be 75
pohm-cm,

cooled while accurate

A program is under way for the design, con-
struction, and operation of two series of forcad-
circulation loops for studying the effect of temper-
ature gradient and fluid velocity on the phenomenon
of mass transfer in Inconel systems containing
fluoride mixtures at elevated temperatures.” The
requirements of the two series of loops are a
Reynolds number of 10,000 with temperature gradi-
ents of 100, 200, and 300°F and o temperature
gradient of 200°F ond Reynolds numbers of 800,
3,000, and 15,000. The maximum fluid tempercture
is specified as 1500°F, and the Reynolds numbers
are to be evaluated at that temperature. The loops
are to be consiructed of Inconel tubing. The
maximum tube-wall temperature is to be 1700°F,
and the surface-to-volume rotio is to be held
essentially the same for each loop.

3W, C. Tunnell, W. K. Stair, and J. F. Bailey, ANP
Quar. Prog. Rep. June 10, 1954, ORNL-1729, p 21,

‘w. K. Stair, The Design of a Small Forced Circu-
lation Cosrrosion Loop (to be published).

 

SW. C. Tunnell, Operation Report of the First Small
ATV Study Loop (to be publishedf.

6lnterrw.fionol Nickel Co., Properties of Some Metals
and Alloys.

7W. D. Manly, High Flow Velocity and figh Temper-
ature Gradient Loops, ORNL CF-54-3-193 {Mor. 18,
1954).
PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
FHCOTO 22350

é
i

 

Fig. 3.4. Inconel Forced-Circulation Corrosion Loop.

39
ANP QUARTERLY PROGRESS REPORT

TEMPERATURE DIFFERENTIAL, 159.2°F
REYNOLDS NO. AT 1300°F, 3730
REYNOLDS NO. AT 1459°F, 4540
FLOW, O.B1gpm

VELOCITY, 5.29 fps

MASS FLOW RATE, 0.361 Ib/sec
LENGTH OF HEATING SECTION, 5.43 ft
LENGTH OF COOLING SECTION, 5.86 ft

ORNL-~LR -DWG 3097

 

 

 

 

 

 

 

 

    
 

1300°F POWER
POWER, 17.6 kw 1300°F TERMINAL
CURRENT, 400 amp \
COOLING
VOLTAGE, 44 SECTION—n T

Fig- 3.5.
Temperature Differential of Approximately 160°F,

A summary of the design dimensions and expected
operating conditions for each of the six loops is
presented in Table 3.1, which also indicates the
current status of each loop. Three methods of
heating were studied: electrical-resistance, gas-
fired furnace, and liquid bath. The liquid bath
was discounted because it would require pres-
surization of the liquid metal an excessively long
heat exchanger,and a large investment. A gas-
fired furnace was considered, but lack of operating
experience and the high heat flux necessary for
the loops indicated the use of electrical heating,
with which considerable experience has been ob-
tained. A gas-fired furnace is being studied as a
possible heat source for a large number of cor-
rosion loops. Preliminary calculations of the
pressure drop through the loops indicate that the

40

   
   
 
 
    

   

HEATING SECTION

0.30-in. OD,
0.025-in. WALL
INCONEL TUBING

1459.2°F

Configuration and Design Conditions for Inconel Forced-Circulation Loop to Operate at a

Model LFA pump® can be used, and thus it will
not be necessary to develop a pump.

For purposes of preliminary design the following
assumptions were made: (1) The specific resis-
tivity of molten NaF-ZrF4-UF4 (50-46-4 mole %)
is so high that the resistance of the electrically
heated Inconel tube may be taken as the parallel
circuit resistance.” (2} Heat transfer to the
fluoride mixture may be calculated by relations
established for water.'® (3) A thermal loss equal

 

8W. B. McDonald et al., ANP Quar. Prog. Rep. Mar,
16, 1954, ORNL-1692, p 15.

9 ; .
A, L. Southern, personal communication.

]OD. F. Salmeon, Turbulent Heat Transfer from a
Moliten Fluoride Salt Mixture to NaK in a Double Tube
Heat Exchanger, ORNL-1716 {to be issued).
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 3.1. PRINCIPAL FEATURES AND STATUS OF FORCED-CIRCULATION CORROSION LOOPS

Loop Number

 

 

 

3 4 5 6
Reynolds number at 1500°F 10,000 10,000 10,000 200 3000 15,000
Temperature differential, °F 100 200 300 200 200 200
Power input, kw 27.7 54.2 80.0 6.8 25.5 §6.%
Heating section length, ft* 8.08 8.08 8.08 6 22.3 8.08
Cooling section length, ft* 15.64 15.64 15.64 1.4 4 15.64
Heat exchanger tength, ft** 18.75 18.75 18.75 None None 18.75
Pressure drop, psi 35.12 36.54 47.5 G.06 0.57 115.5
Status Detail Detail Construction  Detail Preliminory  Construction
design design and as- design design 20% com-
complete complete sembly 90% complete; plete
80% com- complete detail de-
plete sign 20%
complete

*0.5-in-0D, 0.020-in.~wall tubing.

 

**1.0-in.-0D, 0.049-in~~wall tubing used for outer tube of concentric-tube heat exchanger.

to 10% of the total heat transfer exists at the
heater section and at the heat exchanger.

Dissimilar<Metal Loops

L. A, Mann

Aircroft Reactor Engineering Division

A loop for testing combinations of structural
metals in contact with high-velocity turbulent
liquid metals under high temperature differentials
is being developed. If combinations of materials
are found that are satisfactory for use in liquid
metal systems, greater flexibility and ease of
construction may be attained. An optimized design,
patterned after that used for beryllium-Inconel-
sodium mass transfer tests {cf. Sec. 2, “'Reflector-
Moderated Reactor’’), has been developed and is
to be used first for testing combinations of Inconel
and type 316 stainless steel in sodium.

GAS-FURNACE HEAT-SOURCE DEVELOPMENT
L. A. Mann

Aircraft Reactor Engineering Division

L. F. Roy!'!
University of Mississippi

Several exploratory tests were made of com-
merciai and ORNL-designed-and-fabricated burners
for use with the gas-furnace heat source described
previously.'? The heat release intensity desired

is on the order of 3000 Btu/ft3.sec, which is of the
same order as that attained in jet aircraft burners.

The maximum rate of heat release obtained in
the tests, 170 Btu/ft%, was obtained with the
burner designed by A. P. Fraas. The results of
these exploratory tests are presented in Table 3.2.
Although the rate of heot release contemplated in
the original design (about 3000 Btu/ft3.sec) was
not approached in these tests, it appears that
rates higher than 170 Btu/ft3.sec can be attained
by securing more rapid mixing of air and gas,
preheating the combustion-air, and providing a
positive pressure in the combustion chamber.
Additional tests are to be made.

STUDY OF THE CAVITATION PHENOMENON

W. G. Cobb A. G. Grindell
J. M. Trummel
Aircrgft Reactor Engineering Division
G. F. Wislicenus, Consultant

One of the problems associated with operating
liquid metal systems at elevated temperatures,
high flow rates, high pump speeds, and minimum
container weight is that of local boiling or cavi-
tation. Because there islittle information available

Hfiummer Research Participant.

T2A. P. Fraas, R. W. Bussard, R. E. McPherson, ANF
Cuar. Prog, Rep. June 10, 1954, ORNL-1729, p 23,

4}
ANP QUARTERLY PROGRESS REPORT

TABLE 3.2, EXPLORATORY TESTS OF GAS BURNERS

 

 

 

Furnace
Burner Gos Rate Rate of Heat 5 Temperature™*
{cfm) Release* {Btu/sec-ft”} (°F)
ORNL 2.1 {muax) 170 2000
Fisher, laboratory type 3.5 160
National N-4 1.2 20
Two-in. pipe with jet 1.3 20 2900

Flame did not fill
combustion space on

which rate was based

Rate based on esti-

mated flume dimensions

Maximum attainable

attainable, except where indicated.

on the cavitation phenomenon in forced-circulation
high-temperature sodium systems, a preliminary

study was initiated,

Tests were run on the ARE hot pump test facility
in which pressures at the venturi throat could be
varied by varying the pressure of the blanketing
gas. Thus it was possible to vary the pressures
without changing the pump speed, the throttle
valve setting, or the temperature. Data were taken
during eight runs at temperatures in the range
1200 to 1500°F. Measurements were made of pump
discharge pressure, venturi throat pressure, and
pump suction pressure as the pump suction pres-
sure was varied, Cavitation was found to begin
when the venturi throat pressure ceased to decrease
uniformly with decreasing pump suction pressure.
Also, as cavitation began, the loop pressure drop
increased and the flow rate decreased. Observation
of fluid flow noise revecled a coirelation of noise
intensity with pressure data.

The pressure at which cavitation occurred cannot
be exactly ascertained, since the pressure at the
middle of the throat of the venturi is obviously
not the lowest pressure in the venturi. In water
cavitation, bubbles form at the start of the flow
expansion, and, in the particular venturi used, the
rather sharp radius at the transition from the throat
to the diffusion cone probably coused flow sepo-
The minimum pressures cbserved at the
throat were within 1 psi of the vapor pressure of

ration.

sodium,

42

Instrumentation for measurement of subaimos-

pheric pressures was required for these high-
temperature (up to 1500°F) sodium-cavitation
studies. A Moore null-balance bellows-type pres-

sure transmitter and a Moore Model 60-N *‘nullmatic’’
pilot valve modified to vent the pilot valve to «
vacuum source proved to be satisfactory for this
application. When the venting atmosphere or sink
pressure is reduced, the unit is capable of meas-
uring correspondingly process pressures.
Pressure measurements were made over the range
of 20 to 1.6 psia, with the sodium-filled transmitter
maintained at about 650°F. The zero shift of the
pressure transmitter ranged from +0.3 to 0.2 in.
Hg over the range.

A second check of the zero shift was made as
the transmitter temperature was varied from 800 1o
470°F with a constant system pressure of approxi-
mately 16.2 psia. The zero shift then observed was
+0.1 to -0.2 in, Hg. In general, the performance
ot this pressure-measuring device was quite satis-
factory; a more highly refined application of this
device might lead to even smaller zero shift. A
report covering this modification is being prepared.

lower

EXPANSION AND IMPROVEMENT OF THERMAL-
CONVECTION LOOP TESTING FACILITIES

E. M. Lees

Aircraft Reactor tngineering Division

The number of power panels installed for sup-
plying thermal-convection loops has been expanded
from 18 to 31 so that tests of long duration and
extensive tests of ‘‘new’’ materials can be made
rapidly. Materials such as the Hastelloy alloys,
molybdenum, and nickel-molybdenum alloys are
to be tested to determine the effects of time,
temperature, temperature difference, loop pre-
cleaning methods, fluid pretreatment methods,
corrosion inhibitors, etc.

Samples of newly designed heating equipment
have been ordered for study, with the expectation
of replacing the present method of installing

 

PERIOD ENDING SEPTEMBER 10, 1954

custom-tailored heating units and insulation on
each loop. Split-tube furnaces that would remain
in place from test to test may possibly be used.
The possibilities of automatizing operation of the
loops and of providing safeguards in case of
shutdowns because of power failures or other
emergencies are also being studied. The loops
have been redesigned so that they require fewer
welds, and the sharp angles of liquid flow have
been eliminated. Figure 3.6 shows the old and
the improved designs.

B UNCLASSIFIES
PHOTH 22432

Fig. 3.6. Thermal-Convection Loops. (a) Old design. (4) New design. (c) New design with insulation.

43
ANP QUARTERLY PROGRESS REPORT

4. CRITICAL EXPERIMENTS
A, D. Callihan

Physics Division

REFLECTOR-MODERATOR REACTOR

D. Scott B. L. Greenstreet

ARE Division
J. J. Lynn

Physics Division

J. S, Crudele J. W. Noaks
Pratt and Whitney Aircraft Division

The first step of the present critical experiment
program was the construction of a small two-region
reflector-moderated reactor to provide experimental
data on a system of simple geometry and materials
for use in checking the present calculational
methods.’ The assembly is approximately spheri-
cal, 41 in. in outside diameter, and has a center, or
core region, that is about 15 in. in diameter.? The
core consists of alternate sheets of enriched
uranium metal (4 mils thick) and Teflon, (CFz)n'
and is surrounded by a beryllium reflector. It is
possible to vary the uranium loading, within the
specified dimensions, in order to make the system
critical, The Teflon and the uranium have been cut
to the various shapes necessary to form the sub-
assemblies required to conveniently construct the
core. A summary of the materials in the first criti-
cal assembly is given in Table 4.1. The aluminum,
which is a necessary structural material, is present
in the core as a vertical double sheet at the mid-
plane and in the reflector as fairly well-distributed
sheets and rods. The coating material, which con-
tained 31.8 wt % C, 61.9 wt % F, 5.5 wt % Cr, and
0.83 wt % H, was applied in a thin layer to the
uranivm metal to reduce oxidation. The Scotch
cellophane tape was used to hold the core sub-
assemblies together. A view of the mid-plane of
the reactor as initially assembled is shown in Fig.
4.1. The small rectangular holes below and to the
right of the core are the aluminum guide channels
for the control and safety rods.

1

 

1D, Scott and B. L. Greenstreet, ANP Quar. Prog.
Rep. Mar. 10, 1954, ORNL-1692, p 45.

2A l . . - » -

more complete description is given in Reflector-
Moderated Critical Assembly Experimental Program by
D. Scott and B, L. Greenstreet, ORNL CF-54-4-53
(April 8, 1954).

44

An earlier assembly was loaded with 17.4 Ib of
U233, as prescribed by the multigroup calculations,
and was not critical.
that these calculations were in error because of
a numerical error in one of the ¢ross sections and
that the predicted critical mass range should have
been 20.9 to 22.75 Ib of U?33, The system was
made critical with 24.35 Ib of U?35 and had an
excess reactivity of approximately 7 cents, with
about 20% of the mid-plane aluminum plates omitted.
It has since been shown that the mounting of this
plate resulted in a reduction in reactivity of ap-
proximately 10 cents, and therefore the system as
designed and loaded with 24.35 |b was slightly
subcritical. Sufficient excess reactivity for sub-
sequent experiments has been gained by the addi-
tion of a cylindrical annulus of beryllium reflector
]7]/2 in. long and 27/8 in. wide with its axis perpen-
dicular to the mid-plane. This addition gave an
increase in reactivity of 65 cents, The extra
beryllium at the top of the reactor was removed
during flux measurements.

It was discovered, however,

TABLE 4.1. COMPOSITION OF CORE AND
REFLECTOR OF FIRST REFLECTOR-MODERATED
REACTOR CRITICAL ASSEMBLY

Core
Volume 1.05 #°
Average radius 7.48 in,
Weight of
Teflon* 58.464 kg (128.62 Ib)
Uranium 11.878 kg
y23s*” 11.067 kg (24.348 Ib)
Uranium coating material 0.112 kg
Scotch tape 0.085 kg
Aluminum 0.92 kg (2.02 Ib)
Reflector
Volume 21.24
Minimum thickness 12.9  in.
Weight of
Beryllium 1102.5 kg (2425.7 Ib)
Aluminum 16.8 kg (37.06 Ib)

*Density: 1.96 g/cm3 of care,
**Density: 0.372 g/cma of core.
PERIOD ENDING SEPTEMBER 10, 1954

PHOT 0-22373

N

     

: ]
] i
! |

Fig. 4.1. First Reflector-Moderated Reactor Critical Assembly,

45
ANP QUARTERLY PROGRESS REPORT

Measurements of neutron flux have been made in
a vertical direction along a radius by using indium
foils with and without cadmium covers. The foils
are 10-mil aluminum-indium alloy, 5/16 in. in diame-
ter, having an effective indium thickness of 3 x
107 in., and the cadmium covers are 20 mils thick.
The results of these measurements are given in
Fig. 4.2, The values of the indium-cadmium fraction3
and the indium activation produced by neutrons
having energies less than the cadmium cut-off are
also plotted. A relotive power, or fission-rate,

 

3The cadmium fraction by definition is the ratio of the
difference between the bare and cadmium-covered acti-
vations to the bare activation.

distribution was made through the core along the
radius that was used for the indium flux measure-
ments by observing the activity of the fission
fragments retained in aluminum foils placed in

contact with uranium.

Similar data, presented in

Fig. 4.3, were obtainedwith the uranium and alumi-

num enclosed in cadmium covers.

The average

uranium-cadmium fraction for the core is 0.33. The
fission rates were measured on opposite sides of
three vuranium foils near the reflector, and the

results are plotted,
higher on the side toward the beryllium.
Measurements

In each case the rate was

will be made of the reactivity

coefficients, as functions of the radius, of a few

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

o8

06

0.4

02

CADMIUM FRACTION

 

 

S
CRNL-LR-DWG 3017
ey if_!fi{
—*]L . I
J REFLECTOR —
i BARE INDIUM ACTIVATION MINUS
‘ /CADMIUM—COVERED INDIUM  ACTIVATION
7 ¢ |
S 25| — | : e
5 |
£
5 BARE INDIUM ACTIVATION
o
: 20 _’#;,,fi)_, ‘ ] A_fl.._.l_f._ ]
= )
S d é ’ CADMIUM FRACTION
- i
g )
‘5 N A -‘ N . . P *_1
a N N\
.
@, ‘ e !
10 7, T - T
7 45‘7 T \
" ’
/ \
)
8 b — L) —_— . [ o . . o Y D
[ @ - 1
} | \ \‘
CADMIUM-COVERED INDIUM ACTIVATION \
O
0 | | ‘ ® =g Q
0 2 4 6 8 10 12 14 16 18 20 22
RADIAL DISTANCE FROM CENTER OF CORE (in)
Fig. 4.2, Radial Neutron Distribution at Mid-Plane of First Reflector-Moderated Reactor Critical
Assembly.

46
PERIOD ENDING SEPTEMBER 10, 1954

CRNL - L_I!-J DWG: 3018

 

 

 

 

 

 

 

 

 

 
 
    

 

 

 

 

 

 

 

 

 

40 1
o
|
; |
. | J
‘2)2 L ——— —_— .- ______,,‘,________________|___ —_ _ e e
BARE ALUMINUM FOIL AGTIVATION —--
~ |
@ |
E !
5 e 00 (R C—— —
5
E
>
|,__.
>
g 16 e Lo e . . ]
<< BARE MINUS CADMIUM- GOVERED ALUMINUM FOIL AGCTIVATION -
1o
CADMIUM - COVERED ALUMINUM Q1L ACTIvATION - 2
Bl o AN TRV ~ Nl L 108 =
é &
L i | 08 &
" - -
& _ 7 : % 04 2
. - - g
CADMIUM FRACTION-. e oz 2
| , 3
: T Q
o 0
0 1 2 3 5 6 7 a

RADIAL DISTANCE FROM CENTER OF GURVE {in.}

Fig. 4.3. Power Distribution of Mid-Plane of Reactor.

The paired points on the bare aluminum foil

activation curve at 6.4, 6.9, and 7.1 in. were obtained from foils placed on opposite sides of the same ura-

nium foil.

metals, such as Inconel, nickel, and cadmium.
Upon completion of the experiments on this as-
sembly, it is planned to construct a larger reactor
of the some shape that will consist of three regions,
with the beryllium island and reflector separated
by the fuel annulus. The diameters of the beryilium
islond and the fuel annulus will be about 10.4 and
20 in,, respectively. Initially, the assembly will
have no structural materials, such as Inconel core
shells, and it will provide a further check of the
calculational methods,

SUPERCRITICAL-WATER REACTOR

J. S, Grudele J. W, Noaks
Pratt and Whitney Aircraft Division
The Supercritical-Water Reactor (SCWR) critical
assembly was described previously* as consisting
of an aqueous solution of enriched UC,F, con-

4. L. Zimmérmcn, J. 5. Crudele, and J. W, Nocks,
ANP Quar. Prog. Rep. Mar. 10, 1954, ORNL-1692, p 45.

 

tained in stainless steel tubes distributed in on
organic liquid in a pattern designed 1o give a uni-
form radial thermal-neutron flux. The liquid, Furfural
(C5H402), which also serves as a neutron reflector
on the lateral surface of the cylindrical tube bundle,
has a hydrogen density opproximately the same as
that of water under the temperature and pressure
conditions designed for the SCWR. It simulates
the nuclear properties of supercritical water as
well as can be determined without further knowledge
of the effect of binding energies on diffusion and
slowing down lengths, Stainless steel was inserted
in the fuel tubes to represent the reactor structure.

The loading of the critical assembly was calcu-
lated by Pratt & Whitney Aircraft Division person-
nel,® by use of a four-group diffusion theory method,
to be 15.34 kg of U?3% and 261 kg of stainless
steel in o 76.2-cm equilateral right cylinder with a

 

3Brivate cammunication from Geoarge Chase, Fox
Project, Prott & Whitney Aircraft Division.

47
ANP QUARTERLY PROGRESS REPORT

side reflector 6.5 cm thick. The calculated radial
fuel distribution is plotted in Fig. 4.4 as the smooth
curve and is compared to the stepwise distribution
achieved when aoll tubes are identically loaded,

In the experimental study, a quantity of UO,F,
aquecus solution, containing uranium enriched to
93.14% in U?3%, was distributed among the tubes
at a U23% concentration of 0.505 g/cm®. The
number of tubes required for criticality was meas-
sured as a function of the Furfural height as in-
creasing quontities of stainless steel were inserted
ond as the solution was diluted. In this manner a
stepwise opproach was made to a loading of uniform

0.30 \

concentration which would be critical at the de-
signed linear dimensions of both fuel solution and
Furfural and which would contain the prescribed
mass of stainless steel. These conditions have
not been achieved because the last dilution was
overestimated and a large part of the solution is at
a concentration somewhat lower than required for
criticality, a deficiency compensated for by locating
about 50 tubes of higher concentration fuel in one
peripheral section.
in Table 4.2.

Some neutron-flux measurements have been made
by using effectively thin (3 x 107% in.) indium foils

The configuration is described

S

ORNL-LR-DWG 3019

 

 

0.28 |—

0.26

ACTOR

0.24

 

 

 

RE

-

0.22

 

 

0.20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FRACTION OF TOTAL FUEL X 102/UN{T AREA O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

042 e S
0AQ fmmrmmrserer e B
00B b b o
006 |- i
0.04 .
0 5 10 15 20 25
RADIUS (cm)

Fig. 4.4. Rodial Fuel Distribution in Supercritical-Water Reactor.

48
with and without 0.02-in.-thick cadmium covers,
The data obtained along a radius 36.7 cm from the
bottom of the core and somewhat removed from the
high-solution-density perturbation ore shown in
Fig. 4.5. Similar data from a longitudinal traverse
located 2.4 c¢m from the cylinder axis are given in
Fig. 4.6. Although these results are preliminary,
it is believed that the apparent nonuniformity of
the radial thermal flux is greater than the experi-
A few measures show the
U233 1o also decrease from

mental uncertainty,
radial importance of
the center,

In one experiment to evaluate the Furfural re-
flector, annular sheets of aluminum were inserted
adjacent to the wall of the reactor tank, thereby
reducing the reflector thickness from 10.0 em to
3.6 cm. The effect of the aluminum, indicated by
the critical height of the Furfural, was a slight
increase in the reactivity.

PERIOD ENDING SEPTEMBER 10, 1954

TABLE 4.2, CONFIGURATION OF SCWR
CRITICAL EXPERIMENT

 

 

Experimental Design
Number of tubes™ 215.0 215.0
Height of UO,F, salution, cm 71.2 76:2
Height of Furfural, em 69.5 76.2
Mass of UP33 | g** 11.48 15.34
Thickness of reflector, em 10.0 6.5
Mass of stainless steel, kg 220.4 261.0

 

*Expressed as equivalent number of tubes 1 in. in
diometer; the array contains 193 which are 1 in. in
diameter, 8 which are 3/4 in., and 25 which are 1/2, in,

**Of this loading, 86.8% was in a solution having o
U%33 concentration of 0.196 g/cma, and 13.2% was in
one with a U3 concentration of 0.505 g/cm3.

-

ORNL-LR-OWG 3020

 

 

 

 

 

 

TY

RELATIVE ACTIVA

 

 

 

 

S

 

[

 

 

 

 

 

CADMIUM FRACTICN

 

 

 

 

 

5

 

DISTANCE FROM REACTOR CENTER (cm)

Fig. 4.5. Radial Neutron Distribution in Supercritical-Water Reactor Critical Assembly,

49
RELATIVE ACTIVITY

50

28

26

24

22

20

iy

mn

O

 

ORNL-LR—-DWG 3021

 

 

 

 

 

 
   

  

 

 

   

 

 

    
 
 

 

 

 

 

 

 

        

 

 

 

 

 

[T e e _[ 77777 : ! I - [ !
[ . ; & i ! ‘ |
,,4:r7, _ i e e bl I — e 1 - . ,,,r ________ |
! | |
|
I | ‘ ‘ ‘ |
| 4 i _ ) o o o —1. - .
| INDIUM ACTIVATION |
(R I L L ‘ | .
—— ‘e ‘ e - L - —} 40
] | |
Lo i L L | l | \ 0.9
e - e e
L - CADMIUM - COVERED | x| ‘
| F‘*’J— — TTINDIUM ACTIVATION..  © ! T g ‘ 0.8
1] ' i
L x | s v I 0.7
e e R L 1 o
| O ! i T 1
13 0 I 0 5
SO - R -4 06 =
= s | | 5
| g | ‘ g
@ | [ x
i — e — — 0.5 &
| ; . | \ s
* B i i \ ga
1| ; ’ e ‘ ’r T T E
| | i
i | T e A ! <
o e i T | —Jo3"”
[ : ! =" CADMIUM FRACTION ) ! g ‘
! ’ . : | i |
0 b S B | '
| | [ INDIUM ACTIVATION MINUS S —1 92
| i ! ! CADMIUM - COVERED INDIUM ACTIVATION ™ |
. N . [ ey T T . | oa
1— i ; [ ! |
| | ! i w‘
. - 1 I b L 0
0 5 10 15 20 25 30 35 40 45 50 55 80 85 70 75 1o

DISTANCE FROM REACTOR TANK FLOOR (cm)

Fig. 4.6. Longitudinal Meutron Distribution in Supercritical-Water Reactor Critical Assembly,
Part 11

MATERIALS RESEARCH
5. CHEMISTRY OF MOLTEN MATERIALS

W. R. Grimes
Materials Chemistry Division

Studies of the NaF-ZrF -UF , system were initi-
ated after the solid phase studies in the NaF-ZrF,
system were completed, ond to date no ternary com-
pounds have been discovered in the NaF-ZrF -UF,
system, A tenfative equilibrium diagrom has been
prepared. The apparatus for visual observation of
melfing temperatures was used to observe eleven
NaF-UF, mixtures in the composition range 20 to
50 mole % UF ,, and a partial phase diagram of the
system was prepared. A new apparatus has been
constructed that will permit some manipulations to
be carried out with the fused salts in an inert
atmosphere.

Recent afttempts to correlate the anticipated re-
duction of UF, in the preparation of UF,-bearing
melts with wet chemical analysis for UF, and UF
and results of petrographic examination show some
surprising anomalies. When UF , dissolved in LiF,
in NaF-ZrF , mixtures, or in NaF LiF mixtures is
treated under flowing hydrogen at BO0°C with ex-
90% or more of the UF, is
reduced to UF,. However, when this techmque is
applied to UF, in NaF-KF-LiF mixtures, the re-
duction is only 50% complete at 800°C and, perhaps,
75% complete at 600°C. Petrographic exominations
of the specimens reveal no complex compounds of
tetravalent uranium; it is possible that the UF  is
“*hidden’’in solid solutions or in complex UF ;-UF,
compounds in which it is not at present recogniz-
able,

A method for the large-scale purification of
rubidium fluoride has been developed. All the
rubidium fluoride obtained from commetcial sup-
pliers to date has contained considerably more than
the specified quantity of cesium compounds, and
therefore almost all of it has been returned to the
vendors for further processing. The purification
method developed can be used at a reasonable cost

cess uranium metal,

on a large scale if material sufficiently free from
cesium cannot be obtained from commercial sources,

Values for the equilibrium constant for the re-
action

Cro 4 2UF, o= 2UF,; + GF,
in molten NaZrF hove been re-examined by the use

of various ratios of UF:;/UF4 with chromium metal
in the charge., From the data obtained when this

initial ratio is less than 3, average values of
5 % 10-4 at 600°C and 6 x 10~% at 800°C can be
computed. At ratios larger than 3 the values in-
crease regularly; this rise is probably due to the
extreme difficulty in filtration and analysis of
samples containing 1 to 20 ppm of Cr**,  The
values at low UF /UF ratios agree quite weli
with the correspondmg vciues of 42 x 104
600°C and 4.1 x 1074 at 800°C obtained prevuously
when UF, and chromium metal were the charge
material.

Fundamental studies were made of the reduction
of NiF, and FeF, by H, in NaF-ZrF, systems as
a means of determining possible improvements in
purification techniques. Also, methods for pre-
poring simple fluorides were
studied. Additional measurements of decomposition
potentials of KCI and of various chlorides in molten
KCI at 850°C were made.

Preliminary measurements of the solubility of
xenon in KNO,-NaNO, eutectic (66 mole % NaNO )
show values of 8.5 x ]0 =8 and 10-7 mole/cm? at
280 and 360°C, respectively, af 1 atm xenon pres-
The all-glass apparatus used has been re-

structural metal

sure,
placed with o nickel and glass combination, and
measurements of xenon solubility in molten fluo-
rides are being made.

Lorge quantities of purified ZrF -base fluorides
are being prepared for engineering tests at ORNL
and elsewhere, and the demand for purified fluo-
rides of other types for possible reactor fuel appli-
cation is increasing rapidly. Consequently, an
increasing fraction of the effort is devoted to pro-
duction or to research in direct support of produc-
tion functions.

Since the consumption of NaF-ZrF, and NaF-
ZrF ,-UF, mixtures is expected to reach 10,000 Ib
during the fiscal year, it is important to decrease
the cost of production of this material. The HF-H,
processing currently used is adequate from ali
points of view, but processing times of nearly
100 hr per batch are now required with the rela-
tively poor raw materials available. The sub-
stitution of hafnium-free ZrF, and commercially
available NaF .ZrF, for the impure ZrF, now
used should afford « considerable saving. The

53
ANP QUARTERLY PROGRESS REPORT

proposed use of zirconium metal as a scavenger (to
replace most of the hydrogen processing) has shown
promise on a small scale, and rapid electrolytic
deposition of iron and nickel from the ZrF -base
mixtures appears to be feasible. It is anticipated
thai the price of purified NaZrF . and NaF-ZrF -UF

mixtures may be halved in the next few months.

SOLID PHASE STUDIES IN THE NanZrF4-UF4
SYSTEM

C. J. Barton
R. E. Moore R. E. Thoma

Materials Chemistry Division

G. D. White, Metallurgy Division
H. Insley, Consultont

Studies of the NaF-ZrF ,-UF, system were initi-
ated after the solid phase studies in the NaF-ZrF
binay system were completed. The solid phases
present in both the slowly cooled and the quenched
were studied by x-ray diffraction and
The study of
the NaF-UF, system was suspended, except for
some visual-observation experiments, and will be
resumed when time permits.,

Visual-observation sxperiments (cf. section below
on ‘‘Visual Observation of Melting Temperatures in
the NaF-UF ; System"’), together with petrographic
studies of slowly cooled NoF—UF4 compositions,
demonstrated that the compound previously desig-
nated 1+?:3 as NaUF . is a congruently melting com-
pound Nagll F, | that is anologous to the NagZrgF,

somples
petrographic analysis techniques.

41
Guenching experiments with fernary

compositions on the join Na U F,-Naglr F
show that this join comprises a completely (or
nearly completely) miscible series, with liquidus

compound.*

temperatures descending from 722°C at the uranium
compound to 520°C at the zirconium compound.
Earlier thermal analysis daota and filiration ex-
periments>*® demonsirated that there is another
series of complete solid solutions along the join

T4 R, Grimes ¢1 al., ANP Quan Prog. Rep. Man. 10,
1951, ANP-60, p 129.

2\, H. Zachaoriasen, J. Am. Chem, Soc. 70, 2147 (1948).

3¢. A Kraus, Phase Diagrams of Some Complex Salts
of Uraniwm with Halides of the Alkali and Alkaline Earth
Metals, M-251 (July 1, 1943).

4R, E. Thoma, et al., ANP Quar. Prog. Rep. June 10,
1954, ORNL-1729, Fig. 4.1, p 41.

5S¢ J. Barton, 5. A. Boyer, and R. J. Sheil, ANP Quar.
Prog. Rep. Dec. 10, 1953, ORNL.-1649, pp 50 and 55.

6C. J. Barton and R. J. Sheil, ANP Quar. Prog. Rep.
Mar. 10, 1954, ORNL.-1692, p 54,

54

Na,UF _-NayZrF ,, with the high liquidus tempera-
ture {BgOOC) at the zirconium compound and the low
liquidus temperature (629°C) at the uronium com-
pound. In order to study equilibrium relationships
in the region between these two solid solution joins,
a series of nine compositions equally spaced along
the Na,ZrF -Na UF . join was prepared, Phase
analyses on completely crystallized preparations
of these compositions indicate that Na,ZrF, and
Na,UF, do not exist in equilibrium with each other
and with liquid. Instead, the join between Na, ZrF,
and Na,UF crosses the two three-phase triangles
having their common base on the NagZrF -Na U F
line and their opexes at Na,ZrF, and Na,UF,,
Quenching experiments are being made for estab-
lishing liquidus relationships and boundary curves
in this area,

Four ternary mixtures were prepared in order to
obtain a preliminary indication of equilibrium rele-
tionships in other parts of the system. The com-
positions of these mixtures are given in Table 5.1,
Identification, by x-ray diffraction and petrogrophic
techniques, of the phases in completely crystal-
lized preparations of the mixtures indicated that the
compound  Na.Zr, F .o occupies a small primary
phase area contiguous to the primary phase fields
of the NagZr F,,-NajU.F, | and the ZrF UF,
solid solution series. It also appears that the
primary phase field of the ZrF -UF , solid sclution
series is configuous to the primary phase field of
the NagZr,F,-NagUgF,, solid solution series
along the greater part of the boundary curve that
enters the ternary system at the Nagl,F -UF,
eutectic (42 mole % NaF-58 mole % UF ,, melting
point 680°C), follows approximately the 50 mole %
NaF join for most of its course, and leaves the
ternary system at the NagZrgF ,-Na Zr F o eutec-
tic (approximately 50 mole % NaF--50 mole %
ZrfF ,, melting point 510°C). The mixtures listed
in Table 5.1 will be used for quenching experiments

TABLE 5.1. COMPOSITIONS OF FOUR MIXTURES
IN THE NqF-ZrF4rUF4 SYSTEM

 

 

 

Sample _ Composition {mole %)
Designation NgF ZrF4 UFA
12 41 55 4
T3 42 48 10
T4 48 39 13

 
in the near future to establish solidus and liguidus
relationships.

No fernary compounds have ‘been discovered, to
date, in the NaF-ZrF -UF, system. A tentative
equilibrium diagram, which shows the general relo-
tionships believed to exist in this complex system,
is presented in Fig. 5.1,

 

7R. J. Sheil and C. J. Barton, ANP Quar. Prog. Rep.
June 10, 1954, ORNL-1729, p 42.

162

PERIOD ENDING SEPTEMBER 10, 1954

VISUAL OBSERVATION OF MELTING
TEMPERATURES IN THE NuF-UF4 SYSTEM

M. S. Grim
Materials Chemistry Division

The apparatus for visual observation of melting
temperatures, described in the previous quarterly
report,’ was used to observe eleven NdF“UF4
mixtures in the composition range 20 to 50 mole %

UF,.

Liquidus temperatures obtained with this

ORNL ~ LR~ DWG 2821

1000

LF, 1035°C
975 YALUES ARE
TEMPERATURES IN °C
950
825
200
8?5
850 e

NugUgFyy

 

800
L.— 825
-85
—875

- 900

 

NgF 999°C

 

Z(F, 910°C

Fig. 5.1. Tentative Equilibrium Diagrom for the NaF-ZrF ,-UF, System.

55
ANP QUARTERLY PROGRESS REPORT

apparatus, both from visual observation of crystal-
lizotion and from thermal effects recorded during the
course of the visual observations, were generally
a little higher than those indicated by the published
diagram for this system.! The data are shown on
the partial phase diagram for the NaF-UF , system
in Fig. 5.2, A solid circle indicates the tempera-
ture at which the composition oppeared to be com-
pletely solidified. This point could not be de-
termined with any degree of certuinty with some
compositions, and the disappearance of the liquid
phase was not always accompanied by a noticeable
thermal effect. The liquidus temperatures shown in
Fig. 5.2 seem fo indicate that NajUF, melts con-
gruently at 629°C, that is, ot a temperature only a
few degrees higher than the sutectic temperatures
on both sides of the compound, The data alse
support the results of petrographic studies which
indicate that the compound in the 50 mole % UF,
region has the approximate composition Na U F,,
rather than NaUF .

A new apparatus has been constructed that will
permit some manipulations to be carried out with
the fused salts in an inert utmosphere. |t consists
essentially of a small dry box, with 5/B—in.-'i“hick

ORNL*! R-0OWG 2922

700 -

MPERATURE {°C)
»
w
o

 

 

 

Ll |
e |
\
|
600 : : : : 1‘
w !
5 \
D

o I

o -

LL’\ = uj,r
2 @ .
om = |
= 3 ‘
. =z !
B e e cemmme D e o e e

15 20 25 30 35 40 45 50

UFg (mole %}

Fig. 5.2. Partial Pt ase Diagram of the NaF.UF
System.

56

Lucite side walls and top, attached to a stainless
steel bottom plate which has a 6-in. length of
2-in.-dia stainless steel pipe welded ontfo it to
serve as a furnace core, The pipe is surrounded
by resistance coils and insulating tape. The box is
equipped with face plates for covering the glove
holes so that the box can be partly evacuated

before it is filled with inert gas,

PHASE RELATIONSHIPS IN UF3-BEARING
SYSTEMS

C. 1. Barton R. J. Sheil
l.. M, Bratcher A, B, Wilkerson

Materials Chemistry Division

G. D. White, Metallurgy Division
T. N. McVay, Consultant

It was assumed in the previous studies® of solu-
bility relationships for UF, in various fluoride
mixtures that in the presence of excess metallic
uranium or zirconium the uranium present in the
fused salt solution was trivalent. This assumption
seemed justified, since, for the reaction

3 Lilo

7UF, + 4U°— UF,

Brewer's tabulation® of standard free energies of
formation yields

AF® = =16 keal .

In addition, petragraphic exomination of the melts
obtained revealed the presence of red materiols
tentatively identified as complex compounds of
UF,, with green compounds of UF, showing only
in traces, if ot all. Since these observations were
those anticipated and since the petrographic ex-
aminations had previously been shown to be very
sensitive in detecting compounds of trivalent ure-
nium in systems which had been slightly reduced,
there seemed to be little reason fo believe that the
results of such examination might be unreliable,

During the past quarter, however, considerable evi-
dence has been accumulated by the most accurate
wet-chemical methods yet devised for determining
U3Y and U*" in fluoride mixtures which shows
that the reduction of UF, by excess uranivin metal
is not complete at B00°C. While the reduction in

 

BG. M. Watson and C. M. Blood, ANP Quar. Ffrog. Rep.
June 10, 1954, ORNL-1729 p 51.

?1.. Brewer et al., Thermodynamic Properties and Equi-
libria at High Temperatures of Uranium Halides, Oxides,
Nitrides, and Carbides, MDDC-1543 (Sept. 20, 1945, rev.
Apr. 1, 1947).
NaF-ZrF, mixtures and in LiF appears to be 90%,
or more, complete, in such mixtures as NaF-KF-LiF
only about 50% of the UF, is reduced. Toward the
end of the quarter it became evident that severd
of the materials such as “3KF-2UF ;" previously
believed to contain only U®¥ regularly contained
large and varying quantities of U**. Furthermore,
the crystals containing variable quantities of U4*
are distinguishable as such only with great diffi-
culty, if at all, by petrographic and x-ray diffraction
examination, Consequently, the previously reported
data on the solubility of UF; in various systems
must be reinterpreted,

The reasons for the incomplete reduction of the
tetravalent uranium are not yet completely under-
stood, and it is not possible at present to define
the extent to which the reduction will proceed at
various temperatures aond in the various solvents,
Accordingly, the significance of much of the ma-
terial presented below is not completely known,

UF, in ZrF -Bearing Systems

The solubility of UF, in NaF-ZrF ; mixtures was

previously shown® to increase with increasing
temperature and with increasing Zrf, concentration
of the solvent over the range 47 to 57 mole % ZrF ,,
Since the publication of that information it has been
shown that the reduction of UF, in ZrF -bearing
melts by excess uranium metal is slightly less than
90% complete at 800°C, The temperature de-
pendence of the reduction is not yet known for
this system, but it is likely that the UF, is more
completely reduced at lower temperatures,

An examination of the NaF-ZrF -UF; system has
been ottempted by thermal analysis, with petro-
graphic examination of the resulting solid phases,
In these studies, UF:‘4 and excess uranium metal
are added to the desired NaF-ZrF ; mixiure before
the sample is heated, and the somple is stirred
constantly while in the molten state.

the reaction

Therefore

YUF, + hUe+— UF,

might be expected to reach its equilibrium value at
any temperature above the melting point. Conse-
quently, in contrast to experiments in which the
metallic uranium is removed by filtration af high
temperatures, the solid products in slowly cooled
melts might be expected to be nearly completely
reduced.

These studies indicate that UF, is the primary
phase in systems in which the NaF-to-ZrF , ratic is

PERIOD ENDING SEPTEMBER 10, 1954

less than 5 to 6 and that Na,U,F, appears as the
primary phase in systems in which the NaF-to-Z¢F
ratio is about 15, It also appears that Na,Zr F .,
crystals may contain small quantities of UF, in
solid solution. The solubility data obtained from
these ond previous studies give little reason to
expect that UF, can be dissolved in NaF-ZrF,
mixtures in sufficient amounts to provide fuel for
reflector-moderated reactors,

UF, in NaF-KF.LiF Mixtures
The solubility of UF, in the NaF-KF-LiF eutectic

was stated previously'? to be equivalent to at
least 15 wt % ot temperatures as low as 525°C,
Subsequent careful examination of this system has
revealed that when UF, and an excess of uranium
metal are added to the purified NaF-KF-LiF mixture
the dissolved uranium species aggregate at least
22% total uranium in the mixture. However, it is
obvious that only 40 to 45% of the soluble uranium
is present as UF; at 800°C, while 55 to 60% may be
trivalent at 600°C, Further study of the system will
be necessary before these values can be determined
more accurately,

Thermal analysis date have been obtained for
several mixtures which were prepared from UF , and
the NaF-KF-LiF eutectic and then treated with an
excess of uranium metal. The data obtained, as
shown in Table 5.2, are in agreement with the data
obtained from filtration studies which showed high
uranium concentrations at low temperatures. Petro-
graphic examination of these materialsreveoled that
at low uranium concentrations a red phase (refractive
index, about 1.44), which is probably I'(::,)UF:‘S and
which may contain UF ,, is predominant. At high

 

105, M. Watson and C. M. Blood, ANP Quar. Prog. Rep.
June 10, 1954, ORNL-1729, p 53.

TABLE 5.2. THERMAL ANALYSIS DATA FOR
UF3-BEARING NaF-KF-LiF MIXTURES

 

Theoretical Composition

 

 

(mole %)* U° Used Thermal Effects
(% of theory) Q)
NaF KF LiF U F3
10.8 39.5 43.7 6.0 200 475,455
16.6 38.6 428 8.0 110 510,460,455
10.¥ 37.0 40.9 120 110 520,455,445
9.6 353 39.1 140 130 565,475
8.8 322 357 23.3 200 570,490,470

 

*Bgsed on complete reaction of UF, with ue,

57
ANP QUARTERLY PROGRESS REPORT

uranium concentrations an olive-drab cubic phase
(refractive index, 1.46) and a blue biaxial phase
(refractive index, 1.544) also appear. There is
evidence that the blue material is a solid solution
of Na,U,Fqs-K,U,Fg; whether it dissolves UF, is
not known.

UF3 in NaF-RhF.LiF Mixtures

The previously reported!! data on the solubility
of UF, in the NaF-RbF-LiF eutectic composition
were based on the assumption that all the uranium
in the filtered specimens was UF ;. For more.recent
studies, the solubilities were determined by using
200-g samples of melt, mechanical agitation during
the equilibration period, and 20% excess uranium
for the reduction step. Samples of the equilibrium
mixture were withdrawn with a filter stick contain-
ing sintered nicke!l as the filter medium. The
samples were then analyzed for trivalent uranium
and for total uranium. The results of the analyses
are presented in Table 5.3.

 

VIR, J. Sheil and C. J. Barton, ANP Quar, Prog. Rep.
June 10, 1954, ORNL-1729, p 53,

The significance of the analyses for UF, is not
entirely clear, Petrographic examination of the
solidified filtrates indicated o reddish-brown ma-
terial, presumed to be RbyUF ., as the only colored
phase, While it is possible that this material could
accommodate UF, in solid solution, it deoes not
appear likely that as much as 75% of the total
uranium present could be hidden in this fashion,

The good agreement, for some of the mixtures
studied, between the theorstical uranium content
and the measured total uranium content indicates
that the reaction of UF, with uranium metal pro-
ceeded nearly to completion. Additional data on
this system will be needed before an interpretation
of the results can be presented.

UF3 in the Individual Alkali-Metal Fluorides

Filtered specimens from preparations in which
UF, dissolved in LiF was treated at 825 to 850°C
with excess uranium metal have shown that more
than 90% of the dissolved uranium is trivalent and
that it crystallizes as UF;.  Accordingly, it is
believed that the behavior observed by thermal
analysis of this system is that of LiF with pure,

TABLE 5.3. SOLUBILITY OF UF, IN NoF-RbF-LiF MIXTURES AT VARIOUS TEMFERATURES

 

Theoretical Composition

 

 

 

Theoretical Filtration Filtrate Composition
(mole %) U3* Content Temperature (wt %)
Naf RbF LiF UF, (wt %) e udt Total U
14.1 37.6 42.3 6.0 19.3 793 2.92 15.2
700 1.41 16.9
645 4.54 16.7
600 1.17 14.7
550 10.7
495 5.26
9.4 47.0 37.6 6.0 17.7 750 5.51 17.0
675 9.18 16.9
650 3.22 18.0
625 9.95 16.4
600 No sample*
5.6 48.9 39.5 6.0 17.5 800 5.07 16.8
750 4,02 17.0
700 4.07 17.3
650 4.76 16.4
625 4,86 13.1
600 6.35

 

* Filtration was attempted, but the filter apparently clogged at this temperature ond the sample could not be ob-

tained.

58
or nearly pure, UF,. Recent studies of LiF-UF,
mixtures have confirmed the belief!2 that no com-
pounds form and that these materials comprise a
simple eutectic system. A partial phase diagram
for this simple system is shown as Fig. 5.3.

The NaF-UF, system has been studied more
thoroughly than any other binary alkali-metal
fluoride—UF, system. The best thermal analysis
data obtained for this system (data obtained with
those mixtures which showed the least evidence of
contamination of the melt by tetravalent uranium)
are shown in Fig., 5.4, The earlier data, 13:14 ob-
tained by mixing NaF with previously synthesized
UF; in sealed capsules, are shown by open circles.
The more recent data, shown by solid circles, were

 

12, M. Bratcher et al., ANP Quar, Prog. Rep. June 10,
1954, ORNL-1729, p 43.

13y, s, Coleman and W. C. Whitley, ANP Quar. Prog.
Rep. Sept. 10, 1952, ORNL-1375, p 79.

My c. Whitley, V. 5 Coleman, and C. J. Barton, ANP
Quar. Prog. Rep. Dec. 10, 1952, ORNL-1439, p 109.

ORNL—-LR—DW( 2923

 

1000
Q50
350

BOO [

 

 

 

 

TEMPERATURE (°C)
—
o
C

 

 

 

 

 

 

 

 

 

 

?OO ..... e e L -
B5O e bt T
o
oo J) AN R - —Lv ----- o
1o J5 ESCUUNIUN NONIIURUUE SO0 SN SN ]
Q
500 L | - ]
0 10 20 30 40 50 80 TC
UFy (mole %)
Fig. 5.3. Tentative Partial Phase Diagram of

the LiF-UF, System,

PERIOD ENDING SEPTEMBER 10, 1954

obtained by reducing NaF-UF, mixtures with ex-
cess uranium in open crucubles protected by an
atmosphere of helium. Some data, also obtoined
during the past quarter, were obtained by a method
which represents a compromise between the two
previously mentioned methods, Mixtures were pre-
pared in open stirrer-equipped crucibles by combin-
ing dry NaF with freshly prepared!? UF,. These
data are indicated in Fig. 5.4 by the solid triangles.
The agreement between the data obtained by the
various methods is considered fo be gratilying, in
view of the difficulties associated with the ease of
oxidation of U*" to U4*, The identification of the
compound is based chiefly upon the results of
petrographic examination of slowly cooled melts,
The possibility that the compound might be Na ,UF,
rather than Na,U,F, and the pOSSIbI]Ify thof
appreciable quantities of UF, may be “hidden’’ in
these materials have not been completely ruled out
by these studies. Very few compositions contain-
ing more than 50 mole % UF, have been prepared,
and the data in this part of the system are probably
of little value,

ORNL~( RADW(" 2924

 

M!XTURFS PREPARED N %m_co CAFTSUL.ES
* MXTURES PREPARED IN OPEN CRUCIBLES :
IN A HELIUM ATMOSFHERE T
MIXTURES PREPARED N OPEN
STIRRER - EQUIPPED CRUCIBLES ..

  
       
   

Q

OO - I

 

TG0 [

TEMPERATURE °C}

 

 

 

 

 

BO0 o=

o0 - - I T

 

 

UF 3 {mole ¥)

Fig. 5.4. Tentative Partial Phase Diagram of
the NuF-UFs System.

Phase equilibrium studies of the KF-UF, and
RbF-UFB systems are not so far advanced. |t
appears that the most probable compositions for
the compounds observed are K ,UF,, K,U,F,, ‘and
Rb,UF, ond that the KF-K UF, eutechc melts at
about 720 C at upproxumately 15 mole % UF,, It

 

lsPrepmed'by W. C. Whitley, July 1954,

59
ANP QUARTERLY PROGRESS REPORT

seems that in both systems considercble UF, is
present along with the UF, in a manner such that
it can escape detection by petrographic examination,

UF, in Binary Alkali-Metal Fluoride Systems

A number of thermal analyses followed by petro-
graphic examination of the solid products have
been performed with UF, in each of the NaF-LiF,
NaF-KF, and LiF-KF eutectic mixiures. These
specimens were prepared, in each case, by reducing
the desired UF ,~alkali-metal fluoride mixture with
10% excess metallic uronium in nickel crucibles
blanketed with inert gas and fitted with stirrers of
nickel.

In the NaF-LiF system the lowest liquidus tem-
perature observed was at about 630°C at 10 mole %
UF; ond 36 mole % NaF, In all cases studied (up
to 30 mole % UF ;) only the free alkali-metal fluo-
rides ond the NaF-UF ; complex compound (Najl,F o)
were observed.

In the NaF-KF-UF; system the red compound
K3UF, appears as the only colored species in
mixtures containing up to 10 mole % UF,; the
optical properties are slightly different from those
of the compound crystallized fromNaF-free mixtures
and may indicate slight solid solution of NaF in
the crystals. Mixtures containing 20 to 40 mole %
UF; show increasing quantities of a blue phase
with optical properties which suggest that it is a
solid solution of NajU,Fg and K U, 1. Liquidus
temperatures in this system seem to be higher than
700°C in all cases.

Solutions of UF, in the LiF-KF binary system
show liquidus temperatures in the range 485 to
600°C. Petrographic examination of thesematerials
shows free LiF in all specimens, K,UF, in those
containing less than 10 mole % UF,, and K U, Fy
in those containing much above 10 mole % UF,,

Chemical analyses were not performed on any of
these materials, but it is likely thot considerable
UF , was present in these specimens, especially at
the higher temperotures. Additional studies of
these potentially valuable systems will be made,

PURIFICATION CF RUBIDIUM FLUORIDE

C. J. Barton, Materiols Chemisiry Division
D, L.. Stockton, ORSORT

All the rubidium fluoride cbtained from comimercial
suppliers to date has contained considerably more
than the specified quantity of cesium compounds,
and therefore almost all of it has been retumed to

60

the vendors for further processing. However, the
immediate need for small amounts of pure material
for phase equilibrium studies and for rubidium metadl
production has necessitated purification of some
material at ORNL.

Experiments '¢ have shown the feasibility of
separating rubidium from cesium with the ion ex-
change medium Amberlite IR-105, This resin was,
accordingly, used in all the experiments conducted,
The resin column consisted of a 3-in.,-dia glass
pipe containing a 6-ft bed of the resin. The column
was loaded for each run with about 350 g of crude
RbF known to contain about 20% CsF. Elution of
the rubidium was accomplished in about 35 liters of
0.5 M (NH ) ,CO,4 solution; the purity of the product
was not appreciably affected by elution rates in
the range 30 to 60 ml/min.

Elution of cesium from the column, which was
observed by use of Cs'37, required 65 to 70 liters
of the (NH,),CO, solution; elution was not greatly
improved by use of a 1.5 M solution. From a total
of 1775 g of crude RbF charged in five different
runs, somewhat more than 1 kg of RbF containing
less than 1% CsF was obtained. The KF content
(0.6% KF) was not appreciably affected by this
treatment.
be conducted on a large scale at reasonable cost if

It is apparent that this purification con

material sufficiently free from cesium cannot be
obtained from commercial sources.

CHEMICAL REACTIONS IN MOLTEN SALTS

F. F. Blankenship L.. G. Overholser
W. R, Grimes

Materials Chemistry Division

Chemical Equilibric in Fused Salts
J. D. Redman C, F. Weaver

Materials Chemistry Division

The apparatus and techniques for experimental

determination of edquilibrium constants for the
q

reactions
Ce® + 2U'r:4 _ CrF2 + 2UF3
and

Fe® + 2UF, == FeF, + 2UF,

in molten NaZrFs were described previously, V7 and

 

161, R. Higgins, Chemical Technology Division, ORNL,
work to be published.

17 6. Overholser, J. D. Redman, and C. F. Weaver,
ANP Quar. Prog. Rep. Mar. 10, 1954, ORNL.-1692, p 56.
values were given for the “‘equilibrium constants'’

obtained with UF, and the metal as starting ma-
terials. During the past quarter, the values have
been checked with various ratios of UF4 to UF, as
starting materials in the NaZrF . solution. The
apparatus and details of the experimental technique
were the same as in previous experiments. The
results of the recent series of experiments in which
2 g of hydrogen-fired {(1200°C}) Cr® and various
UF ;-to-UF; rotios were used are shown in Table
5.4. In the table

c?. . .C
UF3 CrF2

C2
UF4

with the concentrations expressed as mole fractions
by using the formula weight -NaZrF5 = 2 moles,
It was necessary to calculate the final UF_ con-
centration, since accurate analyses for U3 in the
presence of Crtt could not be made; any error
arising from this calculation would be important
only at low UF; concentrations,

The data show that at both 600 and 800°C the

PERIOD ENDING SEPTEMBER 10, 1954

values of K _ remain constant at low UF 5-to-UF
ratios but increase rapidly as this ratio rises above
3. This rapid increase in K may be due to in-
complete removal by filtration of small amounts of
chromium metal or to uncertainties in analysis of
these frace quantities. Since, as shown above, the
UF ;-to-UF, ratio in NaZrF . solution in equilibrium
with uranium metal is about 9 at 800°C, it is not
likely that disproportionation of the UF, is re-
sponsible for the increase in K .

In any case, the agreement between the meon of
four values of the constant ot each temperature
with previously published values is extremely good.
Agreement among the various initial Uf""?;i'o:«U[':4
ratios seems to show that a state of equilibrium is
reached in these experiments,

When equilibrium data for the reaction

Fe® 4+ 2UF4F~—"_——"'32UF3 + FeF2

were obtained by the addition of UF,, UF, and
FeF, to NaZrF, in the same way as for the chro-
mium reaction, the results obtained were not so
reproducible.  The data obtained are shown in
Table 5.5, The mean of the values obtained at

TABLE 5.4, EQUILIBRIUM DATA FOR THE REACTION Cr® + 2UF4V_—-..—:“£ CrF, + 2UF, IN MOLTEN NaZrFg

 

Experimental Conditions

Equilibrium®*

 

Concentration

 

Tem([.:ecr;ture ( IUF jkAdc:ed | UF, Added of Crtt K.>**
moles/kg of melt) (moles/kg of melt) (ppm)

600 0.358 2080 3.0 x 1074
0.267 0.078 950 5.1 x 10~4
0.178 0.158 150 2.7 x 10~4
0.089 0.238 35 4.4 x 104
0.035 0.287 50 5.7 x 103
0.029 0.318 40 1.1x 1072

800 0.360 2660 7.4 x 10—%
0.280 0.081 1125 6.1 x 10™4
0.194 0.165 210 3.3 x 1074
0.111 0.244 55 4.2 x 10~4
0.064 0.295 30 1.8 x 1073
0.032 0.325 35 10-3

7.9 x

 

*Yalues shown are mean of closely agreeing volues from duplicate experiments. Blank centained 200 ppm of cett,
**K, values previously obtained: ot 600°C, K, =4.2x 10~4,;

Keqcalculmed from ~AF°= RT aneqby using Brewer’s values for AF®: at 6OOOC,KGq = 1; at 800°C, K, =

at 800°C, K_= 4.1 10~4,

q

0.]-

61
ANP QUARTERLY PROGRESS REPORT

TABLE 5.5. EQUILIBRIUM DATA FOR THE REACTION Fe° + ZUF“;“, ZUF3 + FeF, IN MOLTEN NaZrFs

 

Experimental Conditions

Concentration of

 

 

Temperature UF, Added UF, Added FeF, Added Fe'™ in Filtrate Ky

(°C) (moles/kg of melt)  (moles/kg of melt)  (moles/kg of melt) (ppm)

600 0.322 0.053 0.0319 620 1.2 x 106
0.322 0.053 0.0346 735 1.2 x 107¢
0.326 0.083 0.0347 660 9.3 x 1070
0.326 0.083 0.0347 655 9.3 x 107¢
0.321 0.036 0.0168 995 2.1 x 1077
0.321 0.036 0.0168 940 1.9 x 1077

800 0.242 0.113 0.206 8600 1.2 x 1073
0.242 0.113 0.103 3210 3.2 x 1072
0.242 0.113 0.103 3270 3.7 x 1073
0.277 0.077 0.103 4110 2.7 x 1073
0.277 0.077 0.103 4050 2.2 x 1077
0.0318 0.322 0.213 3110 3.4 x 106
0.0318 0.322 0.210 3300 2.5 x 1073
0.103 0.271 0.210 4530 1.1 x 1079
0.103 0.271 0.207 4645 3.7 x 1077
0,207 0.152 0.213 7940 1.3 x 1072
0.245 0.115 0.053 495 4.2 x 10-6
0.245 0.115 0.053 435 2.8 x 1076
0.245 0.115 0.053 265 1.3 x 1076

 

*Values previously reported: at 600°C, K = 1.2 x 10“5; at 800°C, K,=1.5x 10-8.

Keq calculated from —-AF° = RT In K, by using Brewer’s values for AF°:
at 600°C, K, = 2x 107% at 800°C, K = L4x 1074

600°C (1 x 10=7) agrees with the previously re-
ported value of 1.5 x 10~3, although agreement
among the individual determinations was not en-
couraging. The values obtained at 800°C, however,
do not agree well with the value of 1.5 x 10~
reported when UF, and ferrous metal were used as
starting materials.

In on ottempt to explain this discrepancy, the
experiments in which UF | and ferrous metal were
used as reactants at 800°C were repeated. The
values obtained, os shown in Table 5.6, yielded a
mean of 0.9 x 1078 and ogreed more closely with
the previously reported values. It appears possible
that the reoction of ferrous metal with UF, had
not reached equilibrium in the reaction time allowed.

62

TABLE 5.6. EQUILIBRIUM DATA FOR THE
REACTION OF FERROUS METAL WITH
UF, iN MOLTEN NoZrF, AT 800°C

UF4 added: 0.360 mole/kg of melt

 

Concentration™

 

of Fettin Filtrate K,
(ppm)
520 1.7 x 1078
390 5.4 x 1077
330 2.8 x 1077
480 1.1 x 1076

 

*Blank of 100 ppm to be subtracted from determined
values in calculations.
Stability of Chromium Compounds in
Molten Fluorides

J. D. Redman C. F. Weaver
Materials Chemistry Division

It was shown in previous experiments that CrfF,
or some complex compound of divalent chromium
was the corrosion product in the ZrF4ubuse fluoride
melts of interest., However, when NoF-KF-LiF
mixtures, with and without UF . were tested in
equipment prepared from olloys containing chromium,
the reaction products found were complex com-
pounds of CrF . Since films that were apparently
NaK,CrF . or similar materials have been observed
in some experiments, attempts have been made to
observe the stability ond solubility of CrF, and
CrF, in typical salt mixtures of the two fypes.

In the temperature range 600 to 800°C, CrF, is
not stable in NaZrF. solution in contact with
equipment of nickel. Examination of filtered speci-
mens shows that the reaction

ECrFB + Ni&=—= NEF2 + 2CrF2
The solubility
of CrF, in the NaZrF g solvent seems fo exceed
6000 parts of Cr per million parts of salt at 600°C;
solubility ot 800°C seems to be at least 15,000 ppm
and may be considerably higher,

In the NaF-KF-LiF system it appears that CrF,
is not stable.
action

 

proceeds to essential completion,

Whether the disproportionation re~

3CrF2 ==y 4 ZCrF3

or some other mechonism is involved is not evident
from the data available, It appears from preliminary
evidence, however, that the solubility of Cr3% is
several thousand parts per million at 600°C,

Reduction of NiF, by H, in NaF.ZrF, Systems

C. M, Blood H. A, Friedman
G. M, Watson
Materials Chemistry Division

The use of hydrogen as a reducing agent in the
removal of oxidizing impurities from fluoride melts
has been routine for almost two years. One of the
principal impurities removed is the Ni** introduced
as a result of treating the melts with hydrogen
fluoride in nickel containers at 800°C. Since
hydrogen was first employed in this fashion, kinetic
studies have been under way in order to supply data
on which improved fluoride production procedures

PERIOD ENDING SEPTEMBER 10, 1954

could be based, The Ni*" was found to reduce so
rapidly thot the controlling step was the rate of
removal of hydrogen fluoride from the melt; hence
the chemistry of the process presents no problem of
consequence from an engineering standpoint. With
the completion of measurements af 800°C during the
pust quarter, these studies have been discontinued,

The experimental method used for the measure-
ments at 800°C was the same as that used in
previously described experiments ot 600 and
700°C.'8  Figure 5.5 shows a comparison of the
effect of temperature on the rates of reduction for
four concentrations. The slopes of the curves show
that the apparent activation energy for the rate de-
termining step is about 7800 cal. This energy of
activation is in accord with the hypothesis that
the reaction is heterogeneocus and is catalyzed by
nickel surfaces.

 

186G, M. Watson, C. M. Bload, F. F. Blonkenship, ANFP
Quar, Prog. Rep. Mar. 10, 1954, ORNL-1692, p 62.

UNCLASSIFIED
ORMIL-LR—OWSG 2525

TEMPERATURE

5 X 1072

2 X402

1072 [

 

5 %4073

 

 

EGUIVALENTS OF NEF2 REDUCED PER liter OF B FASSED

2Rt

 

 

 

 

 

L] | ]

10~3 e
GODIRB Q0017 00016 0.0015 0004 GO 00012

A

Fig. 5.5. Effect of Temperature and Concentra-
tion on the Rate of Reduction of NiF, by H, in
Molten NoF-ZrF, (53-47 mole %).

63
ANP QUARTERLY PROGRESS REPORT

The data presented here were all obtcined at
roughly the same gas flow rate (100 mi/min).
However, since the rate of reduction is dependent
on the flow rate and on the geomeiry of the appo-
ratus, the results must be considered merely as
representative and are subject to some voriation
with conditions; they apply to 3-kg batches in
4.in,-dia pots with hydrogen bubbles rising from
the notched end of a ]/4-in. vertical dip tube im-
mersed to a depth of 5 in.

Reduction of FeF, by H, in NaF-ZrF ; Systems
C. M. Blood G. M, Watson

Materials Chemistry Division

Since the reduction of FeF, with hydrogen is the
step in purification of NaZrk,
mixtures for reactor use aond testing, an extensive
study of the equilibrium constant of this reaction

has been attempted.

FeF, + Hye===2Fe®  2HF

time-consuming
The reaction

is relatively easy to study, since the hydrogen
fluoride concentration in hydrogen can be accu-
rately determined, and, after filiration of the melt,
accurate determinations of FeF, in the melt can
be obtained. Since at these elevated temperatures
the activities of hydrogen and hydrogen fluoride
can be assumed to be equal to their partial pressures
in the gas phase, then

where Cp_ e is expressed in mole fraction, and

from the relationships

eq

and

 

the activity coefficient for FeF, at low concentra-
tions in NaZrk
equilibrium
estimated by two different methods:

method and on equilibration method.

should be easily determined. The
constant for the reaction has been
a dynamic

Dyncmic Methed. In preliminagry experiments in
which H, was passed through molten NaF-ZrF,
{53-47 mole %) contaminated with known quantities

64

of FeF, or CrF,, the HF concentration of the
effluent H, was studied as a function of the H
flow rate. The dota obtained are plotted in Fig. 5.6,
in which the ratio of HF concentration at a given
flow rate to that found at a standard flow rate (210
ml/min) is plotted against flow rate. As antici-
pated, the ratio drops at high flow rates; however,
the ratic shows no sign of leveling off at low flow
rates, as would be expected. Although the curve
shows no justification for the practice, the lowest
flow rate found to be practicable (9 ml/min) was
considered to be equivalent to a zero flow rate
(thot is, the flow rate at which the HF concen-
tration could be expected to be at equilibrium),
and experimental values found in subsequent ex-
periments were corrected to the concentration that
they would have shown at such flow rate,

UNCLASSFIED
ORML- LR-DWG 2926
35— : —— e —

 

3.0

STANDARD FLOW RATE = 210 ml/min

800°
25 A FeF, +Hy 0% pe 4+ our

800°C
O CrF, +Hy, s G 4+ 2HF

20

 

RELATIVE EF CONCENTRATION

 

 

 

o —mm——MMM ; |
0 100 200 300 400
FLOW RATE (mi/min)
Fig. 5.6. Eifect of Gas Flow Rate on Relative

Saturation of Effluent Gas with HF When Possed
Through Melten NaF-ZrF, (53-47 mole %) Con-
tominated with Known Quantities of FeF2 or Cer,

The experiments were performed by adding known
quantities of FeF, to anNaF-ZrF, (53-47 mole %)
melt in a clean nickel reactor and measuring the HF
concentration of the effluent hydrogen passed at
carefully measured flow rates. The cumulative
total of HF removed was also determined by col-
lection of the HF in caustic seclution and titration
of the excess caustic, The Fel, concentration at
any specified time could be calculated from the
Fef, found by analysis on completion of the ex-
periment and the HF yield after the specified time,
This procedure, which is simple, rapid, and
capable of high internal precision, yielded the
values plotted as the lower curve in Fig, 3.7. The
logarithm &f the HF concentration is plotted against
the logarithm of the FeF, concentration at various
flow rates of hydrogen. When these values are
corrected to the ‘‘zero flow rate' (9 ml/min) by

~2

PERIOD ENDING SEPTEMBER 10, 1954

reference to Fig. 5.6, they vyield a consistent
straight line after on initial peak. This initial peak
is probably due to the reaction of HF at its initial
high concentrations with the clean (hydrogen-fired)
nickel apparatus and its subsequent re-equilibration
with the more dilute HF-H, mixture. The straight-
line portion of the curve shows the expected slope

UNCLASSIFIED
ORNL - LR-DWG 2927

VALUES CORRECTED TG
GAS FLOW RATE OF 9 ml/min

CONGENTRATION OF HF IN EFFLUENT GAS {moles/liter)

10 2 5

 

CALCULATED FROM FREE
ENERGY DATA (BREWER)

|

EXPERIMENTAL VALUES AT
VARYING GAS FLOW RATES
C 210 ml/min
A (AS LABELED)

3
10 2 5 10

CONCENTRATION OF Fef, (ppm Fe' ')

Fig. 5.7. Equilibrium Concentration for the Reduction of Fel”"2 by H2 in Nul"'«ZrF’4 {53-47 mole %) at

800°C {Obtained by Dynamic Method).

65
ANP QUARTERLY PROGRESS REPORT

of 0.5 and agrees closely with the struight line
calculated from Brewer's value of AF° for the
reaction,

The values obtained for K and for YEeF, on the
assumptlion that the activities of HF and H, are
proportional to their partial pressures ot 800°C and
that

AFOFer = -65.4 kcal per atom of F

and

AFOHF = ~65.8 kcal per atom of

are shown in Table 5.7, The agreement of the
values obtained over a fourfold concentration range
of FeF, is extremely good. It appears that FeF,
in molten NaF-ZrF , is in nearly ideal solution,

TABLE 5.7. EQUILIBRIUM CONSTANTS AND
ACTIVITY COEFFICIENTS AT 800°C IN MOL.TEN
NflF-Zer {53-47 mole %) FOR THE REACTION
Hy o+ Ferr_—“‘\Feo + 24F

 

 

Concentration P2HF
of Fef, - VEoF, " K /146
(pom Fe) Py 'CreF,
1598 2.15 1.47
1513 2.15 1.47
1458 2.17 1.48
1372 1.98 1.36
1318 1.91 1.31
1233 2,37 1.62
1122 2.07 1.38
1067 2.00 1.37
983 2.1 1.44
900 2.18 1.49
857 2.11 1.45
760 2.17 1.49
606 2,19 1.50
550 2,24 1.54

481 2.19 1.50
Av 2.1310.22 1.46 £ 0.16

Equilibration Method, Since it was doubtful that
the lowest flow rate found to be practicable (9
ml/min) actually
tions, an attempt was made o determine the equi-
librium conditions directly. A mixture of HF and
H, was bubbled through a melt containing Fe® and
FeF, until the influent HF concentration matched

represented equilibrium condi-

66

that of the effluent. At that time a sample was
drawn through a sintered nickel filter ond analyzed
to determine the FeF, concentration,

Since the decomposition pressures of molten
KF-HF saturated with KF depend only on tempera-
ture, 19 definite mixtures of H2 and HF could be
obtained while hydrogen gas was passed over such
mixtures held ot a constant temperature. The
mixtures of H, thus obtained were used as influent
for the reactor containing the salt mixture,

In a typical experiment in which this method was
used, a batch (~3 kg) of previously purified NaF-
ZrF4 (53-47 mole %) mixiure was loaded into g
clean nickel reactor. The mixture was heated to
800°C and further purified with hydrogen until a
very low HF concentration (3 x 10~ mole/liter)
in the effluent gas denoted high purity of the melt.
The mixture was allowed to cool to rocom tempera-
ture, and FeF , and iron wire were added to the cold,
frozen mixture under an atmosphere of argon. The
mixture was then heated, melted, and brought up to
temperature while it was sparged with helium gas.
The reactor containing the KF«HF or HF generator
was heated separately to a convenient initic!
Pure
was passed in parallel through both
reacters — the HF generator and the equilibration
reactor — in order to clean out any NiF. films on
walls and to precipitate nickel impurities. After
1 to 2 hr of parallel operation, the reactors were
connected in series, ond the inlet and effluent
gas streams were continuously sampled and ana-
lyzed.  The temperature of the furnace of the
generator was then adjusted until the analyzer

temperature and connected to the system.
hydrogen

showed H -HF mixtures of the desired composition,
The gas mixture was allowed to bubble through the
system for extended periods of time, and somples
were taken intermittently for onalysis. When it was
considered that the system had approached equi-
librium, as indicated by a convergence of the in-
fluent and effluent compositions, the HF generator
was removed from the system and helium wos sub-
stituted for the influent gas mixture. A filtering
tube was introduced into the melt under an addi-
tional argon atmosphere and o filtrate was obtained,
The filtering tube with the frozen filtrate was re-
moved ond analyzed for iron, This procedure was
repeated with appropriate variation of the tempera-
ture of the HF generator and consequent variation
of the influent gas composition,

 

19G. H. Cady, J. Am. Chem. Soc. 56, 1432 (1934).
The data obtained are presented in Table 5.8,
which gives the concentration of iron found in the
filtrate, the ratio of the MF concentrations in the
influent and effluent gases, and the equilibrium
constants ond corresponding activity coefficients
for the FeF,. As is evident, the data in Table 5.8
are less precise than those obtained by the dynamic
procedure., The lowered precision is probably due
to the more complex apparatus and technique. How-
ever, when the data are plotted (Fig. 5.8), the
spread of the data is less thon the estimated un-
certainty in the AF® values.
tween the values obtained by the two methods is
remarkably good,

TABLE 5.8. EQUILIBRIUM DATA FOR THE REACTION

FeF2 + H2"""“*"‘,‘...._,_FfeD + 2HF AT 800°C IN MOLTEN
NuF-Zer {53-47 mole %)

The agreement be-

 

 

i Ratio of HF in 2
Concentration Influent Gas P HF
of of FeF2 to HF in ; yFer
(ppm Fe) Effluent Gas HZ'CFng
912 0,79 0.60 0.41
223 0.91 1.08 0.74
915 1.02 2,09 1.43
740 0.98 2.12 1.45
360 1,00 1.71 1.17
185 0.93 1.89 1.30
160 0.91 2,14 1.47
275 0.99 3.12 2.14
Av 1.34 1.25

 

Preparation of Various Fluorides

B, J. Sturm E. E. Ketchen

Materials Chemistry Division

The simple structural metal fluorides have con-
tinued to be used extensively for numerous pur-
poses. A limited interest also has been maintained
in several of the complex fluorides formed by the
interaction of alkali fluorides with structural metal
fluorides. Some effort also has been expended on
the preparation of other simple fluorides, as well as
on the purification of small quantities of the struc-
tural metal fluorides. The identity and purity of
these fluorides have been established mainly by
chemical analysis supplemented by x-ray and petro-
graphic data in some cases.

The compound AgF, with an Ag-to-F ratio of

0.98, was prepared by heating Ag,CO, under HF

PERIOD ENDING SEPTEMBER 10, 1954

at 150 to 200°C, An attempt to prepare AgF by
reacting HF with AgCN was unsuccessful; most of
the AgCN did not react. Several batches of an-
hydrous ZnF, were prepared by heating ZnF »4H.,0
under HF at 600°C., The compound TeF, was
prepared by reacting Te and HN03 and reacting the
resulting TeO , with aqueous HF 1o give TeO,F ,-H 0,
which was thermally decomposed to vyield T'eOZ,
HF, H,0, ond the product TeF,. The TeF, wos
collected in a condenser held at 130 to 140°C,

Small quantities of Rb,CrF . and Li,RbCrF  were
prepared by heating the proper quantities of CrF
RbF, and LiF in sealed capsules at 900°C. The
Rb3CrF6 has properties approximating those of o
corrosion product found in a rubidium-base fuel.
The compound RbCrF, was prepared by heating
CrF, and RbF at 800°C in a sedled capsule.
Numerous batches of structural metal fuel- or
coolant-type fluoride complexes were prepared.
Among these (prepared by methods previously
described) were: K,CrF,, K,NiF,, and KNiF,.

The compounds NiF_, FeF,, and FeF, were pre-
pared in a high state of purity by hydrofluorination
of the corresponding chlorides,

FUNDAMEMTAL CHEMISTRY CF FUSED SALTS
EMF Measurements

L. E. Topol
Materials Chemistry Division

Additional measurements of decomposition poten-
tials of KCl and of various chlorides in molten
KCl ot 850°C were muade. Morgonite (A1,0,)
crucibles were employed with all the melts studied.
Platinum and nickel were used as cathodes, and
the anodes were of nickel, carbon (graphite),
chromium, and zirconium.

The solutions of various chlorides in KCl were
prepared by melting 2 moles of KCI with 1 mole of
the anhydrous chloride in a sealed, evacuated
quartz tube, More dilute solutions were prepared by
adding appropriate quantities of KCl to the 33.3%
KCl mixtures so obtained. The data obtained
recently are given in Table 5.9.

The value of 1.5 v found upon electrolysis of KCI
with on anode of electrolytic ‘chromium agrees with
the value previously reported for chromium-piated
gold wire in an atmosphere of helium,2? The more
recent value of 1.27 v for electrolytic chromium in

 

20y | E, Topol, ANP Quar. Prog. Rep. Mar 10, 1954,
ORNL-1692, o 43,

67
ANP QUARTERLY PROGRESS REPORT

1072

UNCLASSIFIED
ORNL—-LR--DWG 2928

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

COMCENTRATION OF HF IN EFFLUENT GAS {moles /liter)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. _ CALCULATED FROM FREE—ENERGY DATA (BREWER) |
e — ]
2 T AF® FOR FeF,= ~65.4 % 1 keal |
|
|
EXPERIMENTAL DATA
_____ CALCULATED FROM FREE—ENERGY DATA (GLASSNER)
AFC® FOR FeF, = -~68.0 kcal |
107 ‘ | . [ , 1 N
102 2 5 10° 2 5 104

CONCENTRATION FeF, (ppm Fe*™*)

Fig. 5.8. Equilibrium Concentration for the Reduction of FeF, by H, in NaF-ZiF , (53-47 mole %) ot

800°C (Cbtained by Equilibration Methaod).

a hydrogen atmosphere may indicate that a hydrogen
electrode is formed.

The reaction of UCI; with zirconium metal may
be ascribed to the following equations:

2UC|3 + 3Zr = BZrClz + 20, AF° = 10 kcal,
and

4UCL, + 3Zr = 3ZrCl, + 4U, AF° = +130 kcal.

68

Current-voltage measurements indicate the first
reaction to be predominani. The E° of 0.4 v for the
reaction of UCI3 with a chromium anode, that is,

UCH; + 3Cr = 3CCH, + 2U
E® (theoretical) = 0.90 v,

seems to indicate strong complexing of the CrCl,.
There seems, from studies by other groups, to be
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 59. DECOMPOSITION POTENTIALS OF VARIOUS CHLORIDES IN KCI AT 850°C

 

Concentration of

 

 

. Electrodes Blanketing Potential Observed
Salt Salt in Solute .
Cathade Anode Atmo sphere {v)
{mole %)
KCi 100.0 Ni Cr He 1.50
UC|3 33.3 Pt Zr He Spontoneous
33.2 Pt Cr He 0.40
UCI4 33.3 Pt C H2 0.95, 0.83, (0.75)
33.3 Pt C He 0.83, 0.95
33.3 P Ni He 0.60 to 0.80
3.0 Pt C He 1,84 to 1.87
3.0 Pt C H2 (1.15)
3.0 Pt Ni He 0.34
FeCl2 33.3 Ni C Me 0.80 10 0.83
33.3 Ni Ni He 0.35 to (.45
2.0 Ni C He 1.45
i.0 Ni C He 1.50
10.0 Mi C He 1.Q0
Ni(:l2 33.3 Ni C He 0,93, 0.82
3.0 Ni C He 1.13 to 1.17
1.0 Ni C He 1.20

 

little evidence for this behavior in fluoride melts,

Dilute solutions of UCI, decompose into U and
Cl, at inert electrodes upon passage of current,
Since the measured E’s are somewhat less than
the theoretical estimate of 1.99 v, it is believed
that either some depolarization occurs or thot the
estimate is too high, or that both ore responsible,
Concentrated ‘UCl, solutions may undergo the fol-
lowing electrochemical transformations at platinum-
graphite electrodes:

()  2UCl, - 2UCl, + Ci,,
(2) ucl, -
(3 sucl,

E°= 118 v,

!

U+ 201, E° - 1.9y,

1t

4UCl; + U, E° = Lév,

Uranium metal is deposited on the cathode; so the
first reaction does not occur alone, If this process
takes place in conjunction with the transformation

(4 4ucl, - 3ucl, + U, E° - 0.85y,

a continuous depletion in the UCI, concentration,
which is low initially, would result (more UCI,
would be oxidized to UCI, than would be replaced
at the cathode). With low concentrations of UCl,

the reaction
(5) 2UC13 = 2U + 3CI2 , E°® = 2.2 v,

has been found to predominate. The high value of
E® for reaction 2 eliminates its consideration at
these voltages. The decomposition potential of
reaction 3 is high also but may be cccounted for
by an activity ratio of (UC-iS)/(UCld)S/“ that is
approximately equal to 1073 if the estimated AF®'s
of —191 and -183 kcal for UCI, and UCI, re-
spectively, are correct. The reaction of dilute and
concentrated UCI, solutions ond nickel anodes may
be attributed fo

2UCl, + Ni = 2UCly + NiCl,, E°= 0.36v.

Dilute UCl, solutions seem to show signs of re-
duction in a hydrogen atmosphere.

Concentrated ond dilute FeCl, solutions also
undergo different reactions at inert electrodes, In
concentrated solutions FeCl, is oxidized to FeCl,
at the anode (E° = 0.35 v), whereas in dilute solu-
tions decomposition into the elements occurs
(E° = L11 v). With nickel anodes, both the fol-

fowing reactions probably occur simultaneously:

FeCl, + Ni = NiCi, + Fe, E° = 0.30 v,

69
ANP QUARTERLY PROGRESS REPORT

and

3FeC|2 = 2FeCl,; + Fe, % = 0.35 v.

The NiCl, solutions, as far as is known, undergo
the same electrochemical reactions regardless of
concentration. Thus the changes in decomposition
potential with concentration may be assigned to
differences in NiCl, activity alone, With the aid
of the Nernst equation, the NiCl, activity at 33.3
mole % NiCl, may be computed to be 100 times that
in the 3.3 mole % solution and 300 times that in
the 1 mole % solution.

Solubility of Xeaon in Molten Salts

R. F. Newton, Research Director’'s Department
D. G. Hill, Consultant

The use of molten salts as reactor fuels offers
the possibility of removing the gaseous fission-
product poisons, of which zenon is the most im-
portant. {f, for exomiple, the xenon concentration of
the 60-Mw CFRE can be held at 10% of its equi-
librium value (1.1 x 1078 mole of xencn per milli-
liter of fuel), important perturbations in reactivity
can be avoided. In on circraft reactor, however,
it will be necessary to effect this removal of xenon
with an absclute minimum of auxiliary eguipment.

The available literature appears to contfain no
measurements of the solubility of gases in molten
salts.  Accordingly, it haos seemed desirable to
evaluate the solubility of xenon in molten fluorides
over the temperature range of interest, Preliminary
tests indicate that the following experimental tech-
nique is satisfactory.

The apparatus, which is constiucted of nicke!
where it will be in contact with the fluoride melt
and of glass elsewhere, consists of two sections
that can be isolated by freezing a plug of fluoride
in the U-tube connecting them,
the apparatus the melt is allowed to saturate with

In one section of

xenon under predetermined conditions of tempera-
ture and pressure. Melting of the fluoride plug
permits the saturated sali to flow into the other
section, while the U-tube seal prevents entrance of
goseous xencn into this section; refreezing of the
plug isolates the sections again.
section, xenon is siripped from the salt by repeated
circulation of the hydrogen contained in the appa-
ratus through the solt and past a liquid-nitrogen-
cooled irap. After the stripping process is com-
plete, the hydrogen is removed, the xenon is allowed

In the second

to come to room ftemperature, and the amount col-

70

lected is ascertained by measurements in a modi-
fied Mcleod gage.

Preliminary experiments at 600°C ond essentially
1 atm of xenon indicate that the solubility in the
NaF-KF-LiF eutectic is not greater than 107
mole of xenon per milliliter of salt, The method and
technique proved to be feasible, but minor experi-
mental  difficulties necessitated modification of
the nickel apparatus. More accurate data under
these and other conditions will be available in the
neor future,

Meanwhile, in order to gain familiarity with the
technique, the solubility of xenon in the KNO,-
NGNOB eutectic (66 mole % NQNO3) has been
measured in an apparaius consiructed entirely of
glass. The solubility of xenon in this melt is
about 8.5 x 1078 mole/ml at 280°C and 10-7
mole/ml at 360°C. This liquid apparently shows
the positive temperature coefficient of solubility
expected for gases in molten salts,

X-Ray Diffroction Studies in Salt Systems
P. A. Agron M. A, Bredig

Chemistry Division

Cesium Halides. [t has been suggested?1:2? that
the increase in volume on melting of simple binary
salts is indicative of structural changes in the melt,
that is, of a decrease to a lower ionic coordination.
Volume increases up to 30% have been indicated for
the atkali halides. The x-ray data?® oa the thermal
expansion of solid CsBr and Csl gave volume
changes of approximaiely 26%, but on extrapolation
of about 50°C to their melting points was required
for the computation of these changes, It was there-
fore cof interest to ascertain whether o solid-phase
transition analogous to the one known to exist in
CsCi occurs in the unexplored temperature interval.
This would replace the assumed structural change
in the melting process by one in the solid state,

High-purity CsBr and Cs! were intimately mixed
with fine nickel powder to provide an intemal
standard whose thermal expansion is known accu-
rately, 24
perature

l.attice parameters as a function of tem-
were obtained on the high-temperature

 

21y, w, Johnsan, M. A, Bredig, and W. J. Smith, Chem.
Div. Quar. Prog. Rep. Dec. 31, 1952, ORNL-1482, p 32,

22[:"‘. A. Agron and M, A. Bredig, Chem, Semiann. Prog.
Rep. June 20, 1954, QRML.-1755 (in press).

23, Wagner and L. Lippert, Z. physik. Chem. B3],
2563 (1934).

241 | Jordan and W. H. Swunger, J. Research Nat. Burn
Standards 5, 1291 (19230).
x-ray diffraction unit to within 5°C of the melting
points, No transformation was observed in either
salt to within 2 or 3°C of the fusion temperature,
The volume changes for these solids from 25°C to
the fusion point were 12% for CsBr and 11.2% for
Csl. Comparison of the molar volumes of the solids
with volumes extrapolated for these molten salts?3
down to their respective melting points shows that
an increase of approximately 27% takes place on
melting. These data indicate that a structural
chonge, similar to the lowering of the ionic coordi-
nation from 8 to 6 in solid CsCl, occurs on fusion.

NMaF-ZeF,. It wos desirable to determine the
magnitude of the volume change found in complex
salts on melting. Nickel powder was admixed to
o somple of a homogeneous phase having the com-
position of NaF-ZrF4 (53-47 mole %). The lattice
expansion was measured up to 491°C. The data
are listed in Table 5.10. The volume expansion
from 25°C to the melting point is 3.8%. The volume
is on almost linear function of temperature., The
thermal coefficient of volume expansion, «, is
7.75 x 1075 deg= !, Thus the molar volume of the
solid at the melting point of 520°C gives a density
of 4.00 g/cm?3,
molecular weight of 1287.5 g, corresponding to the
presence in the unit cell of 6.77 NaF 4+ 6 ZrF,
molecules?® for this composition, On extrapolation

The latter value is based on a

 

25t ierational Critical Tables, Yol. 3, p 24.

26p, A, Agron and M, A. Bredig, ANP Quar. Prog. Rep.
June 10, 1954, ORNL-1729, p 47.

PERIOD ENDING SEPTEMBER 10, 1954

of the liquid density?7 of NaF-ZrF , (50-50 mole %)
meosured ot above 600°C 1o 520°C, a value of
3.32 g/em® for the liquid is obtained, An adjust-
ment of this to the composition of NaF-ZrF , (53-47
mole %), by assuming an almost constant molar
volume, lowers the density to approximately 3.2
g/em?. Thus a relatively large density and volume
change of 26% is indicated for this complex salt
on melting. The volume increase may be considered
as indicative of a decrease in ionic coordination
in the fused state similor to that observed in the
CsBr and Csl salts,

The ionic coordination in the solid is assumed to
be similar to that in NoUF ., for which Zachariasen
proposed a ftluorite type of packing with equal
coordinations for both the Na¥ and Zrit of the
order of 8 or 9. It seems possible that the volume
increase on melting results from a lowering in the
coordination of the fluorines around Zr?*, which
might also be considered as radical ion formation
around zirconium; this point requires further study.

Well-formed single crystals which originated from
the central portion of a large melt of NaF-ZrF
(53-47 wmole %) have become available, Single-
crystal x-ray analysis has been initigted in an
attempt to make a complete structure determination,
especially of the positions of the fluoride ions,

273, 1. Cohen and T. N. Jones, A Summary of Density
Measurements on Molten Fluoride Mixtures und a Corre-
lation Useful in Predictin% Densities of Fluoride Mixtures

of Known Compositions, ORNML-1702 (Moy 14, 1954),

TABLE 5.10. THERMAL EXPANSION OF NuF-Zer (53-47 mole %}

 

Hexagonal Cell
O

 

 

Temperature Dimensions (A) Molar Yolume Density™

Run No. o 3 3
Q) (cn™) {g/em”)

a c

-1 25 13,79 9.407 309.690 4.158

-4 190 13.84 9.453 313.38 4.108

-2 250 13.87 9.475 315.62 4.079

-7 278 13.89 9.482 316.64 4.066

-3 378 13.90 9.497 317.90 4,050

-6 395 13.90 3.504 318.14 4.047

-8 491 13.95 ®.525 321.18 4.009

Extrapolation 520 321.50 4.005

 

cell at this composition.

*The density calculation is based on a unit cell containing 8.77 NaF and 6 ZrF , molecules per unit rhombohedral

71
ANP QUARTERLY PROGRESS REPORT

Physical Chemistry28

E. R, Yan Artsdalen
Chemistry Division

The heat of combustion and the heat of formation
have been determined for boron carbide, The com-
bustion process in a high-pressure bomb is both
incomplete and nonstoichiometric; it yields some
free carbon, but the amount depends upon certain
However, it is possible to
obtain a fairly reliable value for the heat of the

reaction

extraneous conditions.,

B,C +40,—> 2B,0, + CO,

by allowing for incomplete combustion and non-
stoichiometry as determined by direct analysis of
products. The heat of combustion is s‘\Hgga.] -
~683.5 + 2.2 kcal/mole., From this value and other
established thermal data, it is found that the heat
of formation is AHS. |, = —14.1 + 2.7 keal/mole
and that the free energy of formation is AFJog 4. =
-13.9 kcal/mole.

The low-temperature heat capacity of molybdic
oxide, MoOg, has been measured in the range from
15 to 300°K, and the results are in fair agreement
with the work of Seltz, Dunkerley, and DeWitt, 29
who measured the heat capacity above 60°K, How-
ever, low-temperature extrapolations by these
authors according to the Debye T3 jaw are in error.
Molybdic oxide has a layer structure, and its heat
capacity between 15 and 60°K varies as T2 in the
manner found for certain other crystals with layer
|attice structure, such as boron nitride.3%  The
following values were obtained for MO, at 25°C:
C% 29g.16 = 17.93 cal/molesdeg and S$9¢g 14 =
18.58 eu. Similar studies are in progress wit
molybdenum disulfide, MoS,, another layer com-
pound,

Electrical conductivity and density measurements
are nearly complete for the molten salt system,
potassium chloride—-potassium iedide, which, in a
number of respects, has been observed to resemble

systems discussed previously, 3!

 

28D)atails of this work will be published in separate
reparts and articles from the Chemistry Division. See
also Chem. Semiann. Prog. Rep. June 20, 1954, QORNL.-
1755 (in press).

29y, Seltz, F. H. Dunkerley, and &. J. DeWitt, J. Am.
Chem. Soc. 65, 600 (1943),

30A, S. Dworkin, D. J. Sasmor, and E. R, VYan Asrtsdalen,
J. Chem. Phys. 21, 954 (1953), 22, 837 (1954), and Chem.
Semiann, Prog. Rep. June 20, 1953, ORNL-1587, p 19.

72

PRODUCTION OF PURIFIED
MOLTEN FLUORIDES

F. F. Blankenship
G. J. Nessle G, M. Watson

Materials Chemistry Division

Use of Zirconivm Metal as a Scavenging Agent

C. M. Blood H. A, Friedman
F. P, Boody F. W, Miles
G. M. Watson
Materials Chemisiry Division

The most important contaminants of fluoride
melts (HF, Fef,, and NiF,) can be removed by
treatment of the melt with hydrogen, but this proc-
ess requires long periods for nearly complete re-
moval of the FeF,. Attempts have therefore been
made to demonstrate the effectiveness of zirconium
metal in removing these impurities in a short time,

In NaF.Z¢F, Mixtures. Known concentrations of
contaminants were added to 3.5-kg batches of pre-
viously purified NaF-ZrF, (53-47 mole %) in nickel
containers and allowed to remain overnight at
800°C in contact with a considerable excess of
metallic zirconium (30 g); some stirring was assurad
by continuous sparging with helium, No experi-
mental difficulties were observed either in equili-
bration or in filtration of the product,

In one experiment, the 10 g of Fel", added gave
an initial contaminant concentration of 1700 ppm
Fe. After treatment, the contaminants present in
the filtrate were 70 ppm Fe, 15 ppm Ni, and 15 ppm
Cr. In a second experiment, 9 g of CrF2 was added
to give an initial contaminant concentration of 1650
ppm Cr. The contaminants found in the filtrate in
this case were 90 ppm Fe, 25 ppm Ni, and 15 ppm
Cr. Thus these experiments indicate that the zir-
conium metal addition is on effective means of
preparing pure NaF-ZrF , mixtures,

In NoF-KF-iLiF Mixtures. The purity of several
2-kg batches of NaF-KF-LiF eutectic after treat-
ment with hydrogen and with zirconium metal ot
800°C is given in Table 5.11. It is evident thai
scavenging with metallic zirconium can be sub-
stituted for most, if not all, of the time-consuming
hydrogen stripping process. Since the impurities in
fluoride melts are initially present in low concen-
trations, it is anficipated that the small amount of

31g. R, Van Astsdalen, ANP Quar. Prog. Rep. Dec, 10,
1953, ORNL.-1649, p 58; ANP Quar. Prog. Rep. June 10,
1954, ORNL -1729, p 57.
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 5.11. COMPARISON OF HYDROGEN AND ZIRCONIUM METAL FOR REMOVING
REDUCIBLE IMPURITIES FROM NaF-KF-LiF MIXTURES

 

Zirconium

Hydrogen Used

Contaminants Found in Product

 

 

(livers) Metal Added —— : _
tx) r (wt %) Fe (ppm) Ni (ppm) Cr (ppm)-
500 None 0.0 2100 7 25
600 None 0.0 725 6 30
300 57 1.3 70 8 20
None 30 1.4 65 9 20

 

zirconium introduced will have o negligible effect
on the physical properties of the melt,
in these experiments at 800°C some alkali metal

However,

was detected in cold regions of the equipment and
some attack on the gaskets and flanges was ob-
served. If large-scale purifications were attempted,
it would probably be necessary fo use a lower tem-
perature for this treatment, '

Purification of NaF-ZrF , Mixtures by Electrolysis

C. M. Blood H., A, Friedman
F. P. Boody F. W Miles
G. M. Watson
Materials Chemistry Division

In the effort to obtain improved fuel purification
procedures it was shown that electrolysis in NakF-
ZrF, melts ot 800°C with graphite anodes and
nickel cathodes reduces dissolved iron and nickel
to relatively low concentration in a much shorter
time than is required for the customary reduction
with hydrogen. Furthermore, the reduced impurities
were collected on a removable cathode that could
easily be withdrawn from pots with welded lids.
This feature promises to be of marked advantage in
processing large batches in equipment of a current
design which ollows the reduced metal from each
batch to recontaminate subsequent batches and
thereby multiply the amount of reduction required.

Careful attention was given to current-voltage
curves at all stages during the electrolysis in the
hope that “‘brecks'’ and apparent decomposition
potentials would permit the progress of the reduc-
tion to be followed. Filtered samples were removed
during various stages and anq]yzed for iron and
nickel to determine the actual amount present, To
facilitate analysis of the current-voltage curves, the
electrolysis pot waos fitted with “floating”’ anode
and cathode probes, which were used to determine

the potential between the melt and the ‘‘working’’
anode or cathode. The probes were made of I/8-in.
nickel rod; their potentials with respect to each
other and to both electrodes were followed through
many cycles of increasing and decreasing applied
voltage, but the results have not been completely
interpreted, ' :

J. A, Mcl.aren of the Chemical Technol‘ogy
Division is of the opinion that measurements with
the reference electrodes have shown that the depo-
sition of iron takes place with very little over-
voltage, At a current of 1 amp, iron was deposited
with a difference of potential between the cathode
and the nickel reference electrode of only 15 mv.
After depletion of the iron, the cathode potential
increased to 75 mv,

The anode potential was found to have a marked
discontinuity in the graph of current density and
voltage, At some critical current density the anode
voltage would increase suddenly. At the beginning
of the electrolysis the jump was found only at high
current densities, but as electrolysis continued,
it was found at lower current densities. The anode
jump was on the order of 0.2 to 2 v, an order of
magnitude greater than the increase found on the
cathode.

All the electrolyses were carried out in NaF-
ZrF, (47-53 mole %) mixtures in nickel containers,
The electrodes were suspended from the lid of a
3-kg capacity purification readtor (4 x 17 in.) with
modified spark plugs as insulating connectors; the
electrodes were immersed fo a depth of about 4 in.
The anode was o I/2-in.-dim C-18 graphite bar
threaded to a short length of nickel rod which was
welded to the spork plug.
supplied from a Dresser Electric Co. 24-v 25-amp
selenivm rectifier; measurements were made with
the use of suitable potfentiometers, voltmeters,

Direct current was

73
ANP QUARTERLY PROGRESS REPORT

ammeters, and a recording ampere-hour meter.

In the first experiments the nickel container was
used as the cathode, and helium gas was continu-
ously bubbled through the melt in order to stir and
to remove gaseous electrode products., The melt
was first thoroughly purified by a conventional HF
treatment followed by reduction with hydrogen.
Then current-voltage curves were token before and
after the addition of 1980 ppm Fe*" as FeF,. The
effect of the added impurity was readily noticeable.
The purified melt gave a curve like plot D of
Fig. 5.9, while the FeF ,-contaminated melt yielded
a curve similar to plot A, Extrapolation of the
curve obtained from the FeF ,-bearing melt gave «
decomposition potential of 0.9 v.

An elecirolyzing current of 5 amp was employed,
but numerous current-voltage curves were also
obtained so that at the end of 5 hr, 20.9 amp-hr had
passed. This was approximately twice the theo-
retical ampere-hours required to reduce the Fe'?;
a sample removed at this time analyzed 50 ppm Fe
and demonstrated that a fairly complete reduction
had been achieved with at least 50% current effi-
ciency. No evolution of fluorine was noted, nor
was there an appreciable attack on the nickel con-
tainer. The wall which had been in contact with
the vapor appeared to be covered with a fine dust,
presumably Nif .

FFor the next series of experiments o new batch
of freshly purified melt was used, hydrogen was
substituted for helium, and a current of 10 omp was
passed. Three purifications were carried out in
succession with addition of FeF, and NiF, to
simulate impurities, as shown in the first part of
Table 5.12. Apparent current efficiencies based on

data from Table 5.12 are deceptive because some
of the reduction is accomplished by hydrogen, which
gives high values, and also becouse the current
efficiency for purification becomes very low when
the amount of impurity to be reduced is very small,
When the container was used as a cathode, 6 v was
the maximum applied potential required to obtain a
current of 10 amp. An HF concentration in the
effluent gas of about 1073 mole of HF per liter of
hydrogen was noted at 5 amp.

In another series of experiments a removable
cathode, insulated from the pot, was used. The
cathode was a %8 x 4 in, cylinder of nickel gauze
welded to a .I/B-in, nickel rod; the gavze was made
of 15-mil wire, 25 wires to the inch. The dimen-
sions were chosen to be in the correct proportion
for a small-scale model of a 250-1b capacity pro-
duction reactor. To obtain a relatively high level of
initial impurity, approximately 20 g of FeF, and
20 g of NiF , were mixed with the starting materials.
After the usual HF treatment before electrolysis,
a sample was removed to determine the iron con-
centration, The results of the trial are summarized
in the second part of Table 5.12.

At the end of 12.5 amp-hr the cathode was re-
moved; it had gained 10.3 g of metallic iron (as
computed from chemical analysis), representing a
97% yield, A new cathode was then used until a
total of 37 amp-hr had passed; this cathode gained
0.2 g of iron, and thus the percentage recovery
had been increased to 99%. A current of 10 amp
gave current densities of 117 amp/cm? of anode
surface. By doubling the nominal area of gauze
used as the cathode to obtain the total surface
area both inside and out, the current density at the

TABLE 512, ELECTROLYTIC PURIFICATION OF NaF-ZrF ; MELTS WITH HYDROGEN STIRRING

 

bimpurities Before

 

Impurities After

 

 

Cathode Electrolysis {ppm) Ampere-Hours Electrolysis (ppm)
Arrangement Passed
Fe Ni Fe Ni
Nickel container 1980 2020 15.2 95 55
as cathode 1890 2000 23.6 Not sampled
1980 Necne 11.8 100 40
Removable nickel 4660 * 12.5 150
gauze cathode * % 37.0 8qQ

 

*Unfortunately, a reliable analysis of the nicke! concentration was not obtained.

**Continued without additional impurity,

74
 

 

0.0 | :

 

PLOT A

8.0 | AMP-HR PASSED =
PN :

i :
! o DOWN |

  
  
 
  

i ! ' } Li * |
PLOT 8 ; |
|

. : PLOT G | ‘
4!‘AMP—HR PASSED:

 
 
 
 

AMP -HR PASSED: |

o 1.7 (DOWN}) o 200 (BOWN)

 

 

 

 

GURRENT (amp)

  

 

   
 
  

UNCLASSIFIED
QRNL-LR—DWG 2929

 

 

I I
PLOT D

  
   
 

 

AMP-HR PASSED:

P 36.5 (DOWN)

 

 

 

 

 

    

  

 

 

 

 

 

8O — L a 122 {UP) L i az0R2MUP b LaFTO WP
- & up N 5 : ' ‘ 5 ? : § ; ‘ § | : :
; ] | é 5 | 5 |
L } : - — <~——y—— R
i i i i : T \
| ‘ : P : :
' | L E
‘ L SN | E
| \\*BREAK {UP): 3.5 amp |
P \ | ‘ § i
: e : : ——

 

 

 

 

 

- | :
| | i
| —BREAK [UP): 1.4 amp

  

 

 

|

 

 

 

 
 

 

74

     

 

 

 

‘ ; i : i |
o BReAK {DOWN): 0.6 amp T BREAK (DOWNI: 0.6 amp ‘

4

20 1D 20 30 40 5C 80 70 20 3.0 40 50 50
APPLIED POTENTIAL {voits)

 

 

 

10
APPUIED POTENTIAL fvolts

    

20

i [
~~BREAK {DOWN): G.4 amp

 

 

= |

30 40 50 80

Fig. 5.9. Plots of Current vs Yoltage in Purification by Electrolysis of NaF-ZrF, Mixtures.

pGCéL O d39WILdIS DONIONT GOIYId
ANP QUARTERLY PROGRESS REPORT

cathode was found to be about 8 amp/cm? of
nominal gauze area,

In all runs under hydrogen, the effluent gas was
bubbled through KOM solution to remove HF. A
sample of the gas after removal of HF was analyzed
by mass spectrometry and found to contain 3.17%
CH,, 1.25% H,0 (from the KOH solution), 0.73%
N,, and only about 0.02% CF,. Evidently the
predominant anode product is HF, The anodes
disoppeared at o rate corresponding to a life ex-
pectancy longer than 50 amp-hr. The cathode de-
posits were expected to contain zirconium metal,
and this was qualitafively confirmed by the high
zirconium analysis (53.9 wt % for the cathode that
gained 0.2 g of iron) in the scrapings of mixed
metal and salt from the cathode. However, the
zirconium metal could not be distinguished by x-ray
diffraction,

 

   
      
         
 

 

 

 

 

|
’x ‘
i ;
| |
6.0 — v T t
|
INITIAL CONTAMINANT
S0 LEVELS: o - T
& Fe:= 4660 ppm /-FILTRATE
=
3 Ni= &5ppmn
2 i
X | |
< 40 o
o .
o . L
m II ;
L ‘:
O
T
3 !
- 3.0 —
<I
w
Lo
o
!
a
= i
<20 |
1.0 -
0O
O 40

Fig. 5.10, Current-Yoltage Breaks vs Ampere-Hours Used in Purification by Electrolysis

76

The conclusion was reached that electrolysis
with removable cathodes was a practical and effi-
cient means of purifying fluoride melts containing
structural metal impurities, Experiments to investi-
gate the removal of sulfates and oxides by elec-
trolysis and an attempt to electrolytically reduce
UF, to UF 5 are in progress.

A typical set of current-voltage curves is shown
in Fig. 5.9. These curves opply to the experiment
with removable cathodes and illustrate the chorac-
teristic
curves as the voltage is increased or decreased.
The “‘knees’’ are upparently due to a polarization

“knees’’ or brecks in the current-voltage

phenomenon and show considerable hysteresis,
A correlation between the current at which breoks
occur and the aompere-hours passed is presented in

Fig. 5.10.

UNCLASSIFIED
ORNL—LR—DWG 2930

NQO.1 CONTAMINANT LEVELS:
Fe=150ppm
Ni= SCppm

 

‘ TOTAlL AMP—HR PASSED = 37.0
BREAKS FOUND WITH INCREASING VOLTAGE —
BREAKS FOUND WATH DECREASING VOLTAGE

|
|
o]

FILTRATE NO.2 CONTAMINANT LEVELSI\
Fe=80ppm
Ni=35ppm

 

 

 

400

AMP-HR PASSED

of NuF-ZrFA.
Preparation of UF ,-Bearing Fuels

C. M. Blood H. A. Friedman
F. P. Boody F. W. Miles
G. M. Watson
Materials Chemistry Division

Since preliminary results have shown that UF -
bearing fuels are less corrosive to Inconel than
UF ,-bearing fuels, considerable effort has been
devoted to studying methods for preparing these
materials, The experiments . performed thus far
have given results that ore not as yet well under-
stood. :

The tabulated values for standard free energy of
formation indicate that at 800°C the reaction

%U° + JUF ==UF,, AF° = ~16 kcdl,

should proceed to essential completion if the com-
pounds are in their standard states. There is ample
evidence that UF, dissolved in NaZrF is more
stable to re_ductlon by chromium than would be
expected,” but the difference between the expected
and the actual stability is hardly sufficient to
suggest that reduction of UF, by uranium meiul
should be detectably mcomple’re.

The reduction of UF, to UF; with uranium or
zirconium metal has been attemp’red in a lorge
number of préparations by the following technique.
About 500 g of material containing the predetermined
quantity of UF:4 is treated with HF and H, in nickel
equipment in the usual fashion to render it essen-
tially free from contaminants. This purified mixture
is cooled below the melting point, ond the desired
quantity of uranium or zirconium metal turnings éis
added, The reaction is allowed to proceed while
the mixture in the reactor is stirred by a stream of
hydrogen bubbling through it. Samples of the melt
can be obtained by drawing a sample into a filter
stick of nickel containing a sintered-nickei filter
medium. The specimens removed from the filter
stick are available for petrographic, x-ray, and wet
chemical examination.

The data obtained from seveml preparations are
presented in Table 5.13. It is obvious that com-
plete reduction was not obtained in any preparation
and that reduction in the NaF-KF-LiF system was
much less complete than in the other solvents,
While the number of experiments performed to date
may be too small to establish reliable frends, the
following pertinent observations can be made on
the basis of examination of these experiments and
results, '

PERIOD ENDING SEPTEMBER 10, 1954

UF3 in ZrF, Systems. About 90% of the ontici-
pated reduction was obtained when large quantities
of reducing agent were used to obtain UF3 in
In. view of the extreme
precuuhons which must be observed to prevent
oxidation of the material during preparation and
handling, it appears safe to state that about 90% of

ZrF ;-bearing mixtures.

the UF ; can be reduced under optimum conditions.
No alkali metal was observed under the operating
conditions used. The extent of reduction in these
systems could be qualitatively dlshngUished by
petrographic examination,

UF, in Molten LiF. More than 90% of the UF in
mol’ren LiF wos reduced in each case, but complete
reduction was not accomplished. It is not possible
to say whether the difference between 0.96 at
825°C and 0.91 at 850°C is real. No alkali metal
vapor was detected. Petrographic examination
detected the presence of small quantities of UF4.

UF, in Molten NaF-KF-LiF. In five experiments
at 780 to 800°C the maximum reduchon obtained in
NaF-KF-LiF was 48% of that expected. The extent
of reduction was only slightly affected by an in-
crease in the reducing agent from 92 to 121% and
by an increase in the reaction time from 3 to 52 hr,
At lower equilibration temperatures the extenf of
reduction was increased. The highest value was
obtained at the shortest equilibration time and w:fh
the largest excess of reducing agent.

In all the experiments ot 780 to 80C°C, aikah
metal vapor was evident; however, at 600°C no
alkali vapor was observed. "It is possible that a
part of the difficulty in obtaining complete reduction
is due to reactions of the type

U + 3KF-F“———“"UF3 + 3K®, AF°® = +26.1 !(ccl,
and
UF + KF=—=

where

SUF, + KO, AF® = +30.3 keal,

the metallic potassium being relatively
volatile is removed from the reaction by the hydro-
gen stream. It is also possible that the dlspropor-
tionation reaction

UFje==YUF, + JU,  AF° =416 eal,

is responsible. However, it is especially difficult
to see why this reaction should be more important
in on NaF-KF-LiF solution than in an LiF solution.

Under petrographic exomination the products of
the eight experiments with NaF-KF-LiF solution
appeared to be identical even though by chemical
analyses they showed widely different UI::3-1c>--:UF:4

77
ANP QUARTERLY PROGRESS REPORT

T

TABLE 5.13. COMPLETENESS O

Initial Melt Compasition {mole %)

 

 

ZF, UF, NeF LiF KF Time
4 4 (hr)
63.7 363 4
47.0 53.0 5
47.0 53.0 5
45.9 40 501 3
459 40 507 3
39.9 6.5 53.6 4
39.9 6.5 53.6 4
4.5 95.5 50
4.5 95.5 26
4.5 10.6 43.8  40.9 3
A5  10.8 438  40.9 28
45 10,8  43.8  40.7 52
A5 10,8 43.8  40.9 28
A5 10.8 438 AD.9 2
5.3 0.7 435 405 28
.8 1.1 445 426 24
1.1 44.5 424 24

1.8

Equilibration

#F REDUCTION OF JF 4 IN VARIOUS FLUORIDE MEL TS

Eeducing Agent Ratio of U3+ to

 

 

*Per cent of stoichiemetric quaniity -~ amount raquirsd to saiisfy:

hue s Jur, = UFyor Y 2® 1 UF, = Uk, + )
ratios.  All were vich in red crystals (no green ot

all} dencted os “KF-UF3 complex compound.” No
tetrovalent possibla
quanfitiss) could be distinguiskad in this system,

although tetravalent uranivim was found in both the

| e

&
RELIT]

uranivim {(othar frace

Lit and the Zrf ,-bearing systems where the re-
duction was vastly more complete, These results
are guite surprising, since pelregraphic examination
has generally been guite sensitive in oscertaining
small amounts of UF, or U, in other systems.
Additional studies in these and similar systems

will be mads,

(%*) Total U
Temffg’;me W Ze°  Expected  Found
800 0 128 1.0 n.88
800 100 ¢ 1.0 0.89
800 100 0 1.0 0.88
800 0 25 (.25 0.14
800 0 25 0.25 .17
800 0 15 0.15 0.19
800 0 15 0.15 0.12
845 125 0 1.0 0.96
850 200 0 1.0 0.91
780 106 15 1.0 0.41
780 100 15 1.0 0.48
780 100 15 1.0 0.44
600 100 0 1.0 0.59
600 200 0 1.0 0.76
600 50 0 0.5 6.33
800 91.8 0 0.93& 0.30
860 921.8 0 0.9:2 0.42
ZrF
Al cliMets! Fluoride Processing Fucility
C. R. Croft 1. k. Eorgan
J. P. Bickely J Truist
Maoterials Chamistry Division

The alkali-matal fluoride processing facility com-
pletad May 15, 1954 is used chicfly for the develop-
ment of suitable processes for the production of

alkali-metal  fluoride mixivies containing UF,,

There have beon severa! preparstions, however, of
special batches of other compositions to fulfill

specific needa.  Single batches of approximately
2.5 kg caon be prepared in this equipment, During
the quarter, 27 preparations that yielded about 70 kg
of material were maode.

Repeated attempts to prepare NaF-KF-LiF mix-
tures containing 14 wt % UF,.ond 1 wt % UF, by
reduction of UF, with uranium metal yielded mix-
tures containing 3 fo 7 wt % UF, by chemical
analysis. These results, along with those reported
elsewhere in this document, seem to show that in
systems containing KF, reduction of UF, by urc-
nium metal is markedly incomplete. Pending clari-
fication of this point, production of material con-
taining 3 wt % UF ; and 10 wt % UF is under way,

Production Facility
J. P, Blakely F. .. Daley

Materials Chemistry Division

During the past quarter, 1760 kg of processed
fluorides was produced in the 250-Ib facility for
external and internal distribution, A breckdown of
the production according to composition is given

below:
Amount
Processed
{kg)
NuF-Zer-U F4 (50-46-4 mole %) 721
NoF-ZrF , (50-50 mole %) 565
NaF-ZrF ,-UF , (50-43.5-6.5 mole %) 404

Pratt & Whitney Aircraft Division received 27 kg
of NaF-ZrF, (50-50 mole %) and 113 kg of NaF-
LrF ,-UF , (50-46-4 mole %), Battelle Memorial
Institute received 46 kg of NaF-ZrF ,-UF, (50-46-4
mole %) and 68 kg of NaF-ZrF, (50-50 mole %).
The remaining processed material was distributed
to various requesters in the ANP Program, :

The difficulties associated with the long stripping
times ond large hydrogen volumes required for pro-
duction of rigorously pure fluoride melts from the
raw materials available have not yet been over-
come.3? Attempts to shorten the time required by
increasing the hydrogen flow rate from 4 to 15
liters/min ond thus decrease labor and maintenance
costs were not successful, Under these conditions

 

320 R Blankenship and G. J. Nessle, ANP Quar.
Prog, Rep. June 10, 1954, ORNL-1729, p 61.

PERIOD ENDING SEPTEMBER 10, 1954

the gas inlet tubes plugged quite frequently and
the HF concentration in the exit hydrogen feli
much below the value obtained at lower flow rates.
The over-all purification time was not appreciably
shortened by the fourfold ‘increase in hydrogen
passed.

Sublimation of crude ZrF , at a temperature lower
than that previously used by Y-12 personnel did
not appreciably improve the purity of the ZrF,
product. Accordingly, the NaF-ZrF , melts produced
future will be prepared from the
(NaF) .ZcF, that is available in moderately pure
form from o commercial source; hafnium-free ZeF,
will be used to adjust the composition as required,
since this material is available in a high state of

in the near

purity.

Electrolytic purification of NeF-ZrF, mixtures
has been shown to be complete, and the process
requires much less time than does the hydrogen
stripping process. Accordingly, one of the units
in the production facility is being modified slightly
to test this method on a large scale. It is antici-
pated that the use of the purer raw materials and
the electrolytic purification process will consider-
ably decrease the cost of fuel preparation.

in-Pile Loop Loading
J. E. Eorgan

Materials Chemistry Division

The first in-pile loop was loaded on June 11 with
a fuel concenirate, NaF-ZrF ,-UF, (62.5-12.5-25.0
mole %), prepared from enriched uranium by Y-12
personnel.  The looding apparatus and controls
performed satisfactorily, and the transfer of ma-
terial to the loop went smoothly. However, the
electrical contact which should have indicated
when the loop was filled to operating level was
not octivated even after 120% of the caiculated
charge had been added. _ |

Subsequent examination of the loop showed that
a weld had given way where the loop was connected
to the pump bowl, and it seems certain that this
leak was present during the filling operation. The
U233 j5 being salvaged from the loop. Examination
of the filling apparatus revealed that all but 68 g
of a total of 5837 g of the material was transferred
to the loop,

79
ANP QUARTERLY PROGRESS REPORT

CHEMISTRY OF ALKALI-METAL HYDROXIDES
L. G. Overholser . Kertesz

Materials Chemistry Division

Purificotion

E. E. Ketchen

Materials Chemisiry Division

Eighteen batches of NoOH were purified by filter-
ing the 50 wt % aqueous solution through a fine
sintered-glass filter to remove the Na,CO; and by
dehydration at 400°C under vacuum. Approximately
one-half the runs yielded products containing less
thon 0.1 wt % H,O and 0.1 wt % Na,CO;.  The
remaining runs yielded less than 0.1 wt % H,0, but
the values for Na,CO; ranged from0.11 to 0.2 wt %,
Four 0.5-1b portions of potassium were reacted
with water, and the resulting solutions of KOH were
dehydrated ot 400°C under vacuum. The resulting
material contained less than 0.1 wt % K,CO,, 0.1
wt % H,0, and approximately 0.1 wt % Na,

Reaction of Sedium Hydroxide with Metals

H. J. Buttram . A, Knox
F. Kertesz
Materials Chemistry Division

Work has been continued on the reaction of copper
and sodium hydroxide. Data were previously re-
ported33 on copper concentrations found in NoOH
up to 800°C, A recent determination at 900°C
shows 2500 ppm Cu in the melt after exposure.
This value is higher than would be expected from
the previous data taken from 600 to 800°C, A plot
of log Cu concentration vs 1/T yields a siraight
line from 600 to BO0O°C and thus suggests that true
equilibrium values were obtained., Extension of
this line would give a value of about 1500 ppm

rather than the 2500 ppm Cu obtained,

3B, A Knox, H. J. Buitram, F. Kertesz, ANP Quar.
Prog. Rep. June 10, 1954, ORNL-1729, p 60,

80

Reactions of copper and nickel with sodium hy-
droxide were previously investigated by determining
the equilibrium hydrogen pressures developed.
Monel, an alloy composed primarily of copper and
nickel, has now been investigated by this method.
The Mone!l reaction tube containing the melt was
surrounded by a quartz envelope, to which was
attached
hydrogen pressure.

a manometer for determination of the
Establishment of equilibrium
required long periods of time, and the values ob-
tained were not completely reproducible. Therefore,
only a range of equilibrium pressures is presented
in the following for the sodium hydroxide-—-Monel
reaction:

Tempearature (°C) Pressure (mm Hg)

600 35 to 42

700 112 to 130
800 160 to 193
900 232 to 280
1000 280 to 360

In another series of tests nickel oxide (reagent-
grade commercial heated to 700°C in
vacuum) was added to sodium hydroxide in a jacketed
nickel tube, At 350°C a pressure of 22 mm Hg was
found, Mass spectrographic analysis of the gas
revealed 5.6% hydrogen and the remainder to be
essentially water vapor, The temperature of the
test was increased to 500°C and then, in 100°C
increments, to 1000°C, The pressure in no case
exceeded 31 mm Hg, which is approximately the
saturated ambient water~vapor pressure, Analyses
of the gas showed it to be 97 to 98% water vapor,

material

with a little hydrogen. Under otherwise identical
test conditions, no water vapor was found when
nickel oxide was omitted. The net effect of the
nicke! oxide is apparently that of oxidizing the
hydrogen formed from the Ni-NaOH reaction to water

vapor.
PERIOD ENDING SEPTEMBER 10, 1954

6. CORRGSION RESEARCH

W. D. Manly

G. M. Adamson

Metallurgy Division

W. R. Grimes

F. Kertesz

Materials Chemistry Division

The static and seesaw corrosion testing facilities
were used for further studies of brazing olloys,
special Stellite heats, Hastelloy R, Inconel, graph-
ite, and various ceramics in sodium, fluoride fuel
mixtures, and other mediums. For the fabrication
of many reactor components it may be well to have
a brazing alloy that is compatible both with sodium
and with the fluoride fuels. Therefore several cor-
rosion tests were performed on Inconel joints brazed
with high-temperature brazing alloys. From this
work it was found that the brazing alloy 67% Ni~13%
Ge-11% Cr—-6% Si-2% Fe—1% Mn has good cor-
rosion resistance in fluoride fuels and fair corrosion
resistance in sodium. Two special heats of Stellite
were tested that had a lower carbon content than
those tested previously, and it was found that the
fower carbon content increased the corrosion
resistance in the fluoride fuels. '

High-density graphite was tested in sodium and
in fluoride mixtures; it has very poor resistance to
sodium attack but seems to have fair resistance to
the fluoride fuels. Various hard-facing materials,
ceramics and cermets, have been investigated from
the corrosion standpoint for use as bearings, seals,
and valve seats; Al,0,, MgAl,O,, ZrO,, and SiC
were fested in sodium, a fluoride fuel, lithiom, and
lead. In these tests, SiC showed the best corrosion
resistance.

in the thermal-convection loop corrosion studies,
emphasis has shifted to fuels with the uranium
present os UF3 and to loops of Hastelloy B, Hot-
leg attack is not found in Inconel loops in which
ZrF ,-base fluoride mixtures with the uranium as
UF, are circulated. A deposit is, however, found
on the hot-leg surface. Only preliminary information
is available, but it appears that neither attack nor
a hot-leg layer is found with alkali-metal-base fluo-
ride mixtures containing UF,. Mixtures of UF, and
UF, will result in a reduction in attack from that
found with only UF ,, but some attack is present
and in high-uranium-content systems it may be
significant,

The addition of zirconium hydride to the ZrF -
base fuel mixtures containing UF4 was found to

result in the formation of UF, and in a reduction
in corrosion. However, to obtain a complete ab-
sence of corrosion, sufficient hydride must be
added fo cause a loss of uranium,

Several Hastelloy B loops have now been suc-
cessfully operated in both the as-received and the
over-aged conditions. inboth cases, a considerable
increase in hardness occurs during operation, With
ZcF ,-bose ‘mixtures containing UF,, very little
attack is found, even after 1000 hr.

The mass transfer characteristics of type 316
stainless steel in molten lithium were studied by
using thermal-convection loops. After 1000 hr of
loop operation, there was no sign of plug formation,
and only a small amount of mass transfer was found
in one of the loops. The investigation of the cor-
rosion and mass transfer characteristics of materials
in contact with liquid iead has shown that certain
alloys possess a much greater resistonce to mass
transfer than the pure components of those alloys.
The probable reason for this resistance to mass
transfer con be related to the formation of inter-
metallic compounds in these alloys, This has been
proved in the case of alloys of 45% Cr-55% Co,
Ni-Mo alloys, and the Fe-Cr-base stainless steels,

The flammability of liquid sodium glloys has
been studied, and it has been found that only bis-
muth and mercury have an effect on the flammability
of sodium. It was found that the degree of reaction
of Na-Bi and Na-Hg alloys with air was not signifi-
cantly changed when the pressure was varied from
0.25 to 1 atm. An investigation has been started to
determine the amount of hydrogen liberated when
NaOH is heated to 900°C in an inert environment,
The container problem has been solved by using a
large single crystal of magnesium oxide. A thermo-
dynamic study of the alkali-metal hydroxides has
shown that hydrogen is important in stabilizing
NaQH at temperatures above 800°C, This is the
temperature range at which moss transfer becomes
significant in Ni-NaOH systems. These computa-
tions link the formation of hydrogen with the forma-
tion of unsaturated oxygen ions, which may be
necessary for moss transfer to occur.

81
ANP QUARTERLY PROGRESS REPORT

A plausible explanation has been developed for
the evidence of a maximum in the corrosion of In-
conel by molten fluorides in static tests at 800 to
900°C and for both void formation and chromium
concentration in the melt being less at 1000°C than
at B00°C. Static tests of type 316 stainless steel
in NaZrF ¢ and in NaF-ZrF ,-UF , substantiated the

phenomena observed with Inconel.

Additional corrosion experiments with fuel mix-
tures containing simulated fission products have
shown that the experimental procedure being fol-
lowed is entirely unsatisfactory. A controlled-
velocity corrosion testing apparatus has been con-
structed for the rapid transfer of mclten fluoride
mixtures through both heated and cooled test sec-
tions. This apparatus was designed to simulate
conditions in reactor components.

STATIC AND SEESAW CORROSION TESTS

E. E. Hoffman W. H. Cook
C. R. Brooks C. F. Leitten
Metallurgy Division

Brazing Alloys in NaF-ZrF ,~UF ard in Sodium

Specimens of the three brazing alloys described
in Table 6.1 were tested in NoF-ZrF -UF, and in
sodium. The specimens were cut from 3-in. T-joints
of Inconel which hod been brazed in a dry hydrogen
The purpose of these tests was to
find a brazing alloy with good corrosion resistance
to both NaF-ZrF ,-UF, and sodium, since such «
brazing alloy could be used to back up tube-to-
header welds, In use, the brazing alloy would be
exposed to the fluoride mixture, but resistance to
attack by sodium would also be required if o
fracture occurred in a tube-to-header weld.

atmosphere.

TABLE 6.1. RESULTS OF STATIC TESTS OF BRAZING ALLOYS IN NaF-ZrF -UF
(53.5-40-6,5 mole %) AND IN SODIUM AT 1500°F FOR 100 hr

 

 

Brazing Alloy Weight Weight
Composition Bath Change Change Metallographic Notes
(wt %) (g) (%)
Alloy A
67 Ni, 13 Ge, 11 Cr, 6 Si, 2 Fe, NaF-ZrF -UF, +0.0003 +0,024 Very little attack; slight surface
1 Mn on Inconel (brazed at roughening to a depth of 0.25
1120°C for 10 min) mil
Sodium ~0.0003 -0.017 Attack in the form of subsurface
voids to a maximum depth of 3
to 4 mils
Alloy B
65 Ni, 25 Ge, 10 Cr on Incone! NoF—ZrF4-UF4 ~0.0010 -0.077 No attack; slight uniform solu-
{(brazed at 1140°C for 10 min) tion would account for weight
loss
Sodium ~0.0013 -0.116 Braze fillet attacked tc a maxi-
mum depth of 8 mils
Alloy C
72 Ni, 18 Ge, 7 Cr, 3 P on NGF-ZrF4-UF4 ~0.0005 -0.041 Braze fillet attacked to a depth
Inconel of 2 tc 3 mils in the form of
subsurface voids
Sodium -0.0008 -0.058 Braze fillet ottacked to a depth

82

of 3 to 6 mils in the form of

subsurface voids
Three brazed Inconel specimens were exposed
for 100 hr at 1500°F in %-in.-OD, 0.035-in.-wall
Inconel container tubes to static NaF-ZrF ,-UF,
(53.5-40-6.5 mole %), and three were exposed to
static sodium. The brazed specimens were cleaned
carefully before and after testing so that valid
weight-change data could be obtained. The speci-
mens were nickel plated after exposure to prevent
rounding of the edges during metallographic polish-
ing. Presented in Table 6.1 are the data obtained
by mefullogmbhic examination of the specimens,

Alloys A and B showed very good resistance to
attack by the fluoride mixture, as shown in Figs.
6.1 and 6.2. Alloy A is also shown in Figs. 6.3
and 6.4 at low magnification in the as-brazed and
as-tested conditions. Alloy C was attacked slightly
by the fluoride mixture. All the alloys were attacked
slightly by sodium, with alloy A showing the least
attack. The maximum depth of attack of 3 to 4 mils

Ni PLATE

 

PERIOD ENDING SEPTEMBER 10, 1954

on alloy A is not serious, since sodium would be
in contact with this alloy enly in the event of a
tube-to-header weld fracture. It may be concluded
from the results of these tests that further study is
warranted on alloys A and B; both showed good re-
sistance to attack by the fluoride mixture. Special
emphasis will be placed on alloy A, since it
exhibited superior resistance to attack by sodium,

- Special Stellite Heats in
NoF-ZsF ,-UF, and in Sodium

Two special heats of Stellite, heats B and C,
were requested from the Union Carbide and Carbon
Metals Research Laboeratories for corrosion testing.
The specimens were sand cast and had Rockwell
C hardnesses as follows: heat B, 39 to 42; heat C,
42 to 43. The as-received microstructure consisted
of Cr,C, corbide particles in a cobalt-rich matrix.
The approximate compositions of these heats were

 

 

 

 

Cov-ie

0.004
>
o
o
2

0.002

0.003
T
L.
=

 

 

Fig, 6.1. Brazing Alloy A (Ni-Ge-Cr-Si-Fe-Mn) Fillet Surface After Exposure for 100 hr ot 1500°F in
NaF-ZrF -UF ; (53.5-40-6.5 meie %). Etched with oxalic acid.

83
ANP QUARTERLY PROGRESS REPORT

 

Ni PLATE

0,004

 

1000X

o
o

02

 

INCH

0.003

 

 

 

Fig. 6.2. Brozing Alloy B (Ni-Ge-Cr) Fillet Surface After Exposure for 100 hr ot 1500°F in NoF-ZrF -
UF, (53.5-40-6.5 mole %). Etched with oxalic acid.

 

 

UNCLASSIFIED |

0.005

 

0.040

 

150X

e
2
o

 

 

0.020

 

INCH

©
Q
n
o

 

 

 

 

Fig. 6.3, Brazing Alloy A on Incona! Telsint in As-Brazed Condition. Etched with oxalic acid.

 
 

Fig. 6.4.

Etched with oxalic acid.

as follows:
Composition (wt %)

Heat B Heat C
Carbon 1 0.30
Chromium 35 28
Tungsten S
Mo lybdenum 12
Manganese 0.50 0.50
Silicon 0.50 0.50
Nickel 3
Cobalt 58 55.7

The specimens were tested in static mediums for
100 hr at 1500°F.
used and duplicate tests were run.
are presented in Table 6.2.

Both Stellite heats were heawly attacked along
the carbide network in NaF-ZrF - UF, (50-46-4 mole
%), as shown in Figs. 6.5 and 6.6. The attack
varied from a minimum of 7 mils to a maximum of
20 mils. The depth of attack varied with the orien-
tation of the carbide network with relation to the

Inconel container tubes were
The test results

Brazing Alloy A on Inconel T-Joint After Testing.

PERIOD ENDING SEPTEMBER 10, 1954

 

 

 

 

0.03

CENCH:

 

 

0.04

 

Dark areas are not areas of attack,

specimen surface, being deepest when the nefwork
was perpendicular to the surface.

None of the specimens were attacked in Sodxum
however, in heat B there were no large equiaxed
carbide pamcles in a 1- to 2-mil area adjacent to
the surface. The carbide particles in this area ap-
peared to have transformed to a very fine structure.
There was also a large amount of fine precipitate
in the cobalt-rich matrix near the surface. 4

Heat C had slightly better resistance than heat B
to attack by sodium, and both heats had very poor
resistance to attack by the fluoride mixture. Several
of the specimens had shrinkage veoids near the
surface, one that was 7 mils in diameter, and the
attack was deeper in these areas. |

Hastelloy R in Various Media

Hastelloy R centerless ground and annealed rods,
?,’Q in. in diameter and 8 in. in length, were obtained
from the Haynes Stellite Company. Each rod was
machined info a specimen (h X ]/4 X ?j‘ in.), a con-
tainer, and o plug. The specimen and the medium
to which it was to be exposed were loaded into the

85
ANP QUARTERLY PROGRESS REPORT

TABLE 6.2 RESULTS OF STATIC TESTS OF SPECIAL STELLITE HEATS IN SODIUM
AND IN NunZer-UF4 (50-46-4 mole %) AT 1500°F FOR 100 hr

 

Weight Change

 

 

Metcllegraphic Notes
Bath (Q/i"-vz) e
Heut B Heat C Heat B Heat C
As received Structure consisted of large, fairly Carbide particles which sur-
equiaxed particles of complex rounded the dendrites of the
carbide (Cr7C3) surrounded by a cobalt-rich solid solution had a
eutectic structure in a cobalt- very fine structure; only a few of
rich matrix the eguiaxed carbids porticles
seen throughout Heat B could be
found
Sodium +0.008 +0.002 No attack; hawever, no globular No attack; very little difference in
particles of tha carbide phase in as-received and tested structures
a 1- to 2-mil area adjacent to the
surface; carbide particles in this
area completely transformed to a
very fine precipitate; precipite-
tion also occurred in matrix, es-
pecially near the exposed surface
+0.016 +0.003 Same as above Same gs above
Nc1l"'-Zrl:4-UF4 -0,052 —0.045 Attack on corbide phase varied Carbide network attacked to a
from 8 to 19 mils; minimum at- depth of 7 to 20 mils, depending
tack of 8 mils occurred on speci- on orientation of the network
men where dendrites of cobalt-
rich solid solution surrounded by
carbide particles were oriented
parallel to the specimen surface;
maximum attack of 19 mils oc-
curred where dendrites were
oriented perpendicular to the
specimen surface; large amount
of fine precipitation in matrix
during test due to aging
~0.060 -0.057 Same as above Same as above

 

container, and the plug was welded in place in a
dry box under a purified helivm atmosphere. The
nominal compositien of Hastelloy R is 65% Ni, 14%
Cr, 10% Fe, 6% Mo, 3% Ti, 2% Al. The specimens
were exposed to the various media for 100 hr at
1500°F, The test results are presented in Table
5.3.

The specimens showed good resistance to attack
by I‘«lc:l':-ZrF“-UF‘fii (50-46-4 mole %), Fig. 6.7, and
by lithium. Surprisingly, this alloy had better re-
sistance to attack by lithium than by sodium; the
attack by lithivm on high-nickel-content alloys (for
example, Inconel) at 1500°F is usually heavy and
is greater thon the attack by sodium. The sample
tested in sodium was attacked slightly, while the
specimen tested in lead was heavily attacked. The
specimen tested in sodium hydroxide was attacked
throughout its entire thickness.

86

Inzone! in Melten Rubidium

Tests were conducted on rubidium metal in con-
nection with an ORSORT project. Rubidium has a
low melting point, 92°F, and a low boiling point,
1270°F, and therefore is being considered as a
possible reactor coolant. The rubidium metal used
in the experiments was prepared by calcium re-
duction of the rubidium fluoride salt by the Stable
Isotope Research and Production Division,

An initial attempt to determine the vapor pressure
vs temperature relationship up to 1650°C was un-
successful. However, the test equipment has been
redesigned and another attempt will be made in the
near future, lnconel has been tested in contact
with static rubidium for 100 hr ot 1650°F, The at-
tack varied from 0.5 mil in the vapor zone to a
maximum of 1.5 mils at the liquid-gas interface, as
shown in Fig. 6.8. The rubidium used was not dis-
tilled and is known to have contained several per
PERIOD ENDING SEPTEMBER 10, f1954

Q.00

 

1000%

| o.002]|

 

0.003

 

ING &—{ -

0.004

 

 

 

 

 

Fig. 6.5. Special Stellite Heat B Before {(a) and After (b) Exposure to Static NnF-ZrF +UF, (53.5-40-
6.5 mole %) for 100 hr ot 1500°F. The area where the attack terminated is shown in (b) Efched with
KOH-K ,F e(CN) .

&

 

o.00t Y-12593

 

 

0.004

 

 

 

 

 

Fig. 6.6. Special Stellite Heat C Before (a) ond After (b) Exposure to Static NaF-ZrF +UF, (53.5-
40-6.5 mole %) for 100 hr at 1500°F. The area where the attack terminated is shown in (b). Efched with
KOM-K FE(CN) "

87
ANP QUARTERLY PROGRESS REPORT

TABLE 6.3. RESULTS OF STATIC TESTS OF HASTELLOY R IN VARIOUS MEDIA AT 1500°F FOR 100 hr

 

 

Weight Weight
Bath Change Change Metallographic Notes
(%) (g/in.?)

NuF-ZrF4-UF4 (50-46-4 mole %) ~0.05 -0.0037 Specimen unattacked except in one small area where
voids extended to a depth of 1 mil; no attack on
vapor zone of tube

NaOH +7.9 +0.6088 Specimen lfd-ina square attacked throughout entire
thickness

Sodium ~0.06 -0.0021 Specimen and bath zone oitacked to a depth of
0.25 to 0.5 mil in the form of voids; exposed
surface of bath zone of container partly de-
carburized to a depth of 1 mil

Lithium -0.07 ~0.0047 Small voids to a depth of 0.25 mil in specimen and
bath zone of tube; surface partly decarburized by
lithium

Lead

 

 

2 e Y +

¥
&
L
g

........

Specimen heavily attacked to a maximum depth of
3 mils {(average 2 mils); bath zone attacked uni-
formly to o depth of 1.5 mils; phese change prob-
ably occurred in attacked areas; vapor zone
unattacked

 

E Y.12619 :

0.004

 

4G00X

 

0.002

 

o T e 0.003

 

 

Fig. 6.7. Hastelloy R Exposed to Static NaF-ZrF ,UF, (50-46-4 mole %) for 100 hr at 1500°F. Etched

with aqua regia.

88

 
Ni

 

Fig. 6 8

PERIOD ENDING SEPTEMBER 10, 1954

 

LUNCLASSIEIED |
Y1293

10,001

 

0.6002

 

0.003

 

INCH

0.004

 

 

 

lnconel Exposed to Stuhc Rub:dlum for 100 hr at 1650°F. Note decarburization in cflacked

ared. Specnmen nickel piufed after test to pro’rec’f edge. Etched w:th glyceria reg(cl.

cent sodium and some oxygen contamination. Ad-
ditional static tests are under way with triple-
distilled rubidium, and an Inconel thermal convection
loop is being operated with boiling rubidium, Fig.
6.9. This loop has now operated for several hundred
hours without difficulty. '

Carhunzahon of Inconel by Sodium

it is well known that sodium, in addition to de-
carburizing metals, can, in some cases, carburize
them if the carbon concentration in the sodium is
sufficiently high.. Therefore, an attempt is being
made to determine whether small additions of carbon
- would prevent decarburization of Inconel specimens
during long-time creep tests in contact with sodium
at elevated temperatures. A-nickel containers are
being used for static tests ‘so that the ratio of
inconel surface area to sodium volume will be
small. The maximum solubility of carbon in nickel

ot 1500°F is approximately ;0.]%_, and therefore the
carburization of the nickel containers in these
tests is very slight. The A-nickel used for the
containers was found by analysis to contain only

0.05% carbon.

The ratio of the Inconel surface to the sodium
volume in. the tests performed to date was 0.76.
The Inconel specimens used were 0.049-in. sheet
reduced to 0.015 in. by cold rolling, and they were
annealed for 2 hr at 1650°F, The carbon additions
(1, 5, and 10 wt %) were made to the sodium in the
form of small lumps of reactor-grade graphite. :

The nickel containers were loaded with the In-
conel specimens, the grophite, ond the sodium in a
dry box in a purified helium atmosphere ond sealed
under vacuum. As shown in Table 6.4 and Fig.
6.10, all the specimens were very heavily carburized
and extremely brittle after exposure for 100 hr at
1500°F. . Also, they were partly covered v{rith a

89
ANP QUARTERLY PROGRESS REPORT

     
  

g ——— |57 OF

~g— BOILER SECTION

HEATER

| J=e— 1085°F

1549F

¥-12930

CONDENSER SECTION\ _.fi',"‘

833°F

—t— | [QUID LEVEL ——-

e 203 OF

Fig. 6.9. Inconel Thermal-Convection Loop for Circulating Boiling Rubidium.

green surface film that was identified by x-ray
analysis to be Cr,0,. The oxygen source for forme-
tion of this film was, at first, thought to be the
graphite which had not been degassed prior to the
tests; however, a Cr,C; film was later found in
other tests with degassed grophite and standard
tests with no graphite addition. Therefore prepa-
rations are being made for obtaining oxygen-free
sodium for use in these tests.

TE. E. Hoffman et al., ANP Quar, Prog. Rep. june 10,
1954, ORNL-1729, p 70.

 

90

Special Tar-Impregnated and Fired Graphite

A special tar-impregnated and fired graphite,
known commercially as Graph-i-tite, has been tested
for corrosion resistance to sodium and NaF-ZrF -
UF, (53.5-40-6.5 mole %). Also, a comparison was
made of the carburization of an austenitic stainless
steel by contact with Graph-j-tite and with C-18
graphite (reactor grade). The Graph-i-tite was
fabricated, as described previously,! by tar-impreg-
nating and firing graphite 16 times. The repeated
tar impregnations and firings produce a high density
 

PERIOD ENDING SEPTEMBER 10, 1954

 UNCLASSIFIED ;

$20 UNCLASSIFIED

y-12121

%TUnCLASSIFIED
- Y-12120

Fig. 6.10. Carburization of Inconel Exposed to Sodium with Carbon Additions for 100 hr ot 1500°F.
{a) inconel specimen before test. (b) Specimen tested in sodium plus 1 wt % carbon. (<) Specimen
tested in sodium plus 10 wt % carbon. Specimens nickel plated to protect edges. Etched with glyceria

regia. 250X.

and a “‘tough skin,”" which, it was hoped, would
reduce penetration of various liquid mediums. |

An open Graph-i-tite crucible containing sodium
and a sealed Graph-i-tite crucible containing NaF-
ZrF ,-UF were tested at 1500°F for 100 hr in
vacusm. The sodium completely penetrated the
0.38-in.-thick wall of the crucible, as in the pre-
vious tesf,] and caused it to crack end crumble.

The only macroscopically visible sign of attack on
the crucibie that contained the fluoride mixture was
that the inner surface had a higher gloss (Fig. 6.11)
after exposure. Metallographic examination of
sections of the crucibie indicated irregular pene-
tration from 0 to 5 mils deep. No attempt has been
made to determine whether reaction products ac-
companied the penetration. This attack may be

?1
ANP QUARTERLY PROGRESS REPORT

TABLE 6.4, CARBURIZATION OF INCONEL BY
EXPOSURE TO SODIUM CONTAINING CARBON
ADDITIONS FOR 100 hr AT 1500°F

Carbon content of Inconel specimen before test, 0.056 wt %

 

Carbon Added

Carbon Content™

 

. ] Specimen
to Sodium of Specimen Weight Change**
Before Test After Test (%)
(wt %) (wt %) o
1 0.882 1.09
5 1.09 1.34
10 1.24 1.63

 

*Average of three analyses.

**On all specimens part of wzight change was due to

tormation of Cr203 film on surface.

compared with the 0 to 2 mils attack by NaF-ZrF -
UF4 (50-46-4 mole %) in the previous tests.! The
carbon content of the fluoride mixture did not change
significantly in either test,

A comparison of the carburization of austenitic
stainless stee! by tar-impregnated and fired graphite
and by C-18 graphite (reactor grade) was obtained
through static tests of cylinders of the same size
of each type of graphite in type 304 stainless steel
tubes containing equal quantities of sodium in
vacuum. Tests for 100- and 500-hr periods ai
1500°C were made on each type of graphite. Type
304 stainless steel was chosen because it carbur-
izes more readily than other austenitic stainless
steels,

The results of the tests are shown in Figs. 6.12
and 6.13. In the 100-hr test, the tar-impregnated
and fired graphite carburized the type 304 stainless
steel to a slightly greater extent than the C-18
graphite did, and in the 300-hr test, it clearly car-
burized the steel to a greater extent. After the
500-hr tests of both types of graphite, steel in con-
tact with the sodium showed small subsurface
voids; the voids were pronounced in the steel tested
with the tar-impregnated and fired graphite. The
depth of carburization of the type 304 stainless
steel above the level of the molten sodium covering
the graphite cylinders was 1 mil in all the tests,
except the 500-hr test with a tar-impregnated and
fired graphite cylinder in which there was carbu-
rization of the steel to a depth of 2 mils.

92

 

¥-12414

    

CONFIDENTIAL
Y-12453

 

(b}

Fig. 6.11.
After (B) Exposure to NuF-ZrF4-UF4 (53.5-40-6.5
mole %) for 100 hr ot 1500°F. At the end of the
test the crucible was inverted so that the fluoride
mixture would flow to the top. The frozen fluoride
mixture con be seen above the shiny inner surface
of the crucible.

Graph-i-tite Crucible Before (a) and
PERIOD ENDING SEPTEMBER 10, 1954

 

  
    
     

20
S
O
>

10°0

 

 

 

 

 

 

 

UNCLASSIFIED
Y-12110

| UNCLASSIFIED
Tyaiz2i08

Fig. 6.12. Carburization of Type 304 Stainless $teel Exposed to Graph-i-tite (a) and to C-18 Graphite
(b) in Sodium for 100 hr at 1500°F in Vacuum. (a2) Depth of carburization was 18 mils (concentrated) to
32 mils (traceable). (b) Depth of carburization was 16 mils (concentrated) to 32 mils (traceable). Etched
with glyceria regia,

2 o 5

UNCLASSIFIED |

Y.12429 .
o

 

 

$0°0
z
<3
T
£6°0
200
o
<
=
+0°

 

 

 

 

 

 

 

Fig. 6.13. Carburization of Type 304 Stainless Steel Exposed to Graph-i-tite {a) and to C-18 Graphite
(b) in Sodium for 500 hr at 1500°F in Yacvum. (@) Depth of carburization was 40 mils (concentrated)
to 49+ mils (traceable through the tube wall). (b) Depth of carburization was 33 mils (concentrated)
to 49+ mils (traceable through the tube wall). Etched with glyceria regia.

93
ANP QUARTERLY PROGRESS REPORT

Ceromics in Various Mediums

Static corrosion screening tests of three ceramic
oxides and one carbide (Al,05, MgAl,0,, calcium
stabilized Zroz, and SiC) were made in a vacuum
atmosphere at 1500°F for 100 hr in each of four
mediums: sodium, NaF-ZrF ,-UF (53.5-40-6.5 mole
%}, lithium, and lead. A fourth ceramic oxide was
tested under the same conditions in sodium and in
NaF-Z¢F ,-UF, (53.5-40-6.5 mole %). The Al 0,
and MgAl,O, test pieces were cut from single,
synthesized crystals; the MgO test pieces were
cleaved from a single, synthesized crystal. The
nominal dimensions of the test pieces and the re-

sults of the tests are given in Table 6.5, Powder
x-ray exdaminations were made of the specimens
that were the least corrosion resistant in each
medium, and the results are summarized in Table

6.6.

The SiC specimens were, in general, the most
resistant to corrosion, and the Al,C; specimens
were the next most resistant. These single-run
screening tests ore qualitative, and the results
should be used with caution. Larger specimens
ore to be tested so that more accurate data can be
obtained oand metallographic examinations and

chemica!l onolyses can be made.

TABLE 6.5. SUMMARY OF STATIC TESTS OF CERAMICS IN YARIOUS MEDIUMS AT 1500°F FOR 100 hr

 

 

Remarks

 

Tested in Sodium

Thickness Weight
Material Tested Change Change
(%) (%)

a - -
A|203 0.5 0.9
sich -0.9

a
MgAl,0, 0.0
ZrD2 (CoOb stabilized) +8.6
MgO® -0.5 ~0.1

Specimen changed from a colorless transparent state to a

white semitransparent state; a small crack oppeared on

an edge

Specimen essentially unaltered; brighter and cleaner;

small part of disk edge broken off

Originally colorless transparent specimen almost com-

pletely changed to a black state; chipped slightly on

one edge

Specimen changed from light buff to blue-black; =dge

chipped

Specimen unaltered except for color change from color-

less to light blue-gray

Tested in Nc'!F-ZrF"“-UF4 (53.5-40-6.5 mole %)

A|203 -31
SiC - 1.8

Specimen broken and covered with fluoride mixture;

brown-red interface present between the fluoride mix-

ture and the specimen

Specimen broken; recovered pieces appeared to be un-

attacked

Recovered portion of specimen {more than one-half)

covered with fluoride mixture and cpparently com-

pletely altered

ZrCB2 {CaQ stabilized)
MgO

A1203

94

No visible trace of specimen found

No visible trace of specimen found

Tested in Lithium

No visible trace of specimen found
- PERIOD END!NG? SEPTEMBER 10, 1954

TABLE 6.5, (continued)

 

Thickness Weight

Remarks

 

Material Tested Change Change -
4 (%) (%)
SiC
MgA|204

ZrC’2 (CaO sfubilized)

Some small particles found
Mo visible trace of specimen found

Specimen in many small pieces; buff color changed to

charcoal-black

Tested in Lead

A|203 50,0 +0.3
5iC | +2.9
MgA1204 : (0,0
Zr02 {Ca0 s}obi!ized) . 0.0 +0.5

No visible attack; colorless transparent specimen changed
to gray; small fspot found on specimen; specimen ap-
parently covered with thin lead film

No visible attack; specimen broken  in loading

No visible attac:ik; specimen changed from colorless to
gray; broken in foading :

Mo visible utraék; specimen changed from a light buff f;:)

gray

 

?Disk cut from a single crystal; specimen 0.75 in. in diameter and 0. 021 i, thick,

bDISl‘( 0.75 in. in diameter and 0.021 in. thick.

“Cleaved from a single crystal; specimen 0.02 x 0.34 x 0.36 in.

TABLE 6,6. POWDER X-RAY IDENTIFICATION OF THE PHASES OF SOME
OF THE LEAST CORROSION-RESISTANT SPECIMENS '

Powder X-Ray Data on Specimen

 

Speci men : Te st Med iUm "‘"-u"mw""""‘-; ---------------------------------------------------------------------------------------------------------------- -
. Befcre Test ' After Test
MgAl, O, ~ Sodium : Face-centered cubic, a =7.970 Face-centered cubic, a = 7.981
NaF-ZrF,~UF, Face-cenrered cubie, a =7. 970 NaF-ZrF ,-U F4 pottern
(53.5-40-6.5 mole %) :
Zr02 (CaQ ' Sodium ' Primarily face-centered cubic ZrOz; Seme as before test
stabilized) : secondarily monoclinic ZrO2 ‘
Lithium : Primarily face-centered cubic 2502; Unidenfified;. face-centered cubicfi:,
secondarily monoclinic ZrQ, : a=4.674

 

FLUORIDE CORROSION OF INCONEL
IN THERMAL-CONVECTION LOOPS

G. M. Adamson
Metallurgy Division
Effect of UF, in ZrF“-{Base Fuels

Preliminary results of Inconel thermal- convection
loop tests of the corrosive properties of UF 4 bacrmg
fluoride mixtures were presen’red prevnousiy and

compared with the corrosive properties of UF .-
bearing mixtures. Additional tests with UF ;-bearing
ZrF ,-bose mixtures have confirmed the reduchon
or ehmmct:on of hot-leg attack, along with the
formation of a hot-leg layer. The lack of attack
and the hot-leg layer are illustrated in Fig. 6.14.

 

20, M. Adaomsan, ANP Uuar Prog. Rep. June 10, 19)4
ORNL-1729, p 72

95
ANP QUARTERLY PROGRESS REPORT

The portion of a hot leg shown in Fig. 6.14 was
taken from an Incone! thermal-convection loop which
had circulated UF3 (2.1 wt % UF, in 2.4 wt %
total U) in NaF-ZrF4 (53-47 wmole %) for 500 hr at
o hot-leg temperature of 1500°F, The fluoride mix-
ture was drained from the loop af the operating
temperature to ascertain whether the layers pre-
viously noted had formed during cooling of the loop.
Since the layer wos again present, it is now con-
sidered to have formed during operation. This
conclusion was further strengthened when the
metal lographic examination revealed o diffusion
zone hetween the loyer and the base metal. An
examination of the layer by a microspark spectro-
grophic technique® showed it to be predominantly
zirconium. Since this technique is not sensitive
to uranium, the layer could also contain uranium.

There is some indication that some of the UF
dissociates fo produce uranium metal, which re-
duces some of the ZrF4 to produce zirconium metal,
The accuracy of the analytical determination for
zirconium is not sufficient o revea! the reaction
products postulated, since some zirconium is also
lost by sublimation.

In another Incone! loop in which UF., in NaF-
ZrF, was circulated for 2000 hr at 1500°F, some

voids were found to a depth of 4 mils upon metalle-

 

3Examinufion made by C. Feldman, Chemistry Division.

 

Fig. 6.14. Hot-Lzg Surfoce of lncone! Loop
After Circuluting UF3 in MaF-ZrF , (2.1 v % UF,
in 2.4 wt % total U) for 500 hr at 1500°F. Loop
drained while hot. Unetched. 500X. Reduced
27%.

96

graphic examination. The voids did not appear to
be the same as those normally found, and it is
thought that they were formed when brittle inter-
metallic compounds were pulled out of the base
metal during polishing. No hot-leg attack could be
found, as shown in Fig. 6.15, but there was a layer
0.5 mil thick on the hot-leg surface. This layer
was also reported to be predominantly zirconium,

It has been definitely established that the use of
UF; rather than UF, in ZrF -base mixtures will
lower the corrosive attack and mass transfer in
Inconel systems. However, it has been found to be
impossible to dissolve sufficient UF, in ZrF ,-base
mixtures to obtain fuels of interest for high-temper-
oture reactors. To obtein a fuel with sufficient
vranium, it would be necessary to use a mixture
of UF, and UF,. Daia obtained from loops oper-
ated with such mixtures and with standard UF .-
bearing mixtures are presented in Table 6.7. The
data show clearly that a mixture of UF; and UF,
produces less attack than UF , alone, but the mix-
ture does not eliminate the attack, as is the case
with UF 5 alene,

Effect of UF3 in Alkali-Meta! Base Fuels

It has been found that more UF, can be dissolved
in the alkali-metal-base fluoride mixtures than in
the ZrF ,-base mixtures, and therefore additional
tests were made with NaF-KF-LiF (11.5.42.0.44.5
mole %) containing UF,, UF,, and mixtures of the
two,

 

r
.

.
¢
MILS

 

 

Fig. 6.15,
After Circulating UF3 in MuF-ZrF4 for 20060 hr
at 1500°F. Unetched. 250X. Reduced 36%.

Hot-l.eg Surface of Inconel Loop
PERIOD ENDING SEPTEMBER 10, 1954

TABLE 6.7. EFFECT OF MIXTURES OF UF; AND UF, IN NoF-ZrF, ON CORROSION OF INCONEL
THERMAL-CONVECTION LOOPS OPERATED AT A HOT-LEG TEMPERATURE OF 1500°F

 

Operating

 

 

Loop Total U Uas UF, T Metcllographic Notes
ime

MNo. (wt %) (wt %) (hr) Hot-Leg Appearance Cold-l.eg Appearance

473 8.7 1.2 500 Moderate intergranular subsurface  Occasional metal crystols
voids 1o a depth of 4 mils

491* 12.8 2.5 500 Moderate to heavy intergranular Deposit to 0.3 mil thick
subsurface voids to a depth of

_ 7 mils

492* 14.2 1.8 2000 Moderate to heavy intergranular Deposit to 1 mil thick
subsurface voids to a depth of
15 mils

469 B.S 0 . 500 Heavy general attack and inter- No deposit
granular subsurface voids to a
depth of 8 mils

462 13.9 0 500 Moderate to heavy intergranular Metallic deposit to 0.3 mil thick

subsurface voids to a depth of

10 mils

*These loops were filled from the same boteh of fluoride mixture, and the differences in uranium analysis cannot yet

be explained,

Two Inconel loops were operated for 500 hr with
NoF-KF-UF to which UF, ond UF, had been
added. Because of sampling and analytical dif-
ficulties, the relative amounts of UF, and UF, are
not known, but it appears that about one-third of
the uranium was present as U’F3. One loop con-
tained a total uranium content of 10.4 wt %, and
after the test the hot leg showed light, widely
scattered attack to a depth of 1 mil (Fig. 6.16);
the hot-leg surface was quite rough. The other
loop contained 7.4 wt % uranium, and the hot leg
showed a void type of attack that varied from light
to heavy, with a maximum penetration of 2 mils;
again, the surface was rough, In both loops, a
layer was found in the coid leg but not in the hot
leg. These mixtures appear to be superior to the
ZrF ;-base mixtures with UF, with respect to depth
of the subsurface void type of attack, and the hot-
leg deposit found with the UF -bearing ZrF ,-base
mixtures did not form. More work is needed to de-
termine whether the rough surfaces represent a
different type of attack and whether the cold-leg
deposits formed are NaK,CrF,, which would
interfere with heat transfer.

 

 

Fig. 6.16.
After Circulating o Mixture of UF3 and UF, (4.3 wt
% UF, and 6.1 wt % UF ) in NaF-KF.LiF (11.5-
42.0-46.5 mole %) for 500 hr ot 1500°F, Etched
with modified aqua regia. 250X. Reduced 36%.

Hot-Leg Surface of Inconel Loop

97
ANP QUARTERLY PROGRESS REPORT

Effect of Zirconium Hydride Additions to Fuel

Various amounts of zirconium hydride were added
to NaF-ZrF ,-UF, (50-46-4 mole %) as a means of
reducing the UF, to UF;. The hydride was added
to small portions of fluorides taken from the same
original batch, and filters were used when the
small batches were transferred to Inconel thermal-
convection loops. The data from loops operated
with these batches of fluoride mixture are given in
Table 6.8. lLayers were found in the cold legs of
all loops to which the ZrH, additions had been
made. The data show that to obtain sufficient
reducing power by the addition of ZrH, to eliminate
corrosion it may be impossible to prevent the loss
of some uranium both in the treatment pot and in
the loop.

Effect of Uranium Concentration

Two loops were operated with a high-purity NaF-
ZrF ,-UF, (53.5-40-6.5 mole %) mixture. This mix-
ture is comparable to the one to be used in the
ARE and has a higher uranium content than the
mixture normally used in thermal-convection loop
tests. The heavy hot-leg attack in both loops was
of the usual subsurface-void type with @ maximum
penetration of 10 mils. This is slightly deeper
than the 6 to 8 mils found with the lower uranium
content mixtures. Thin metallic-appearing layers
were found in the cold legs of both loops. The
results obtained with these loops confirm those
found previously with similar, but impure, mixtures.

Effect of Inconel Groin Size

Incone! pipe was annealed at two temperatures to
provide specimens with different grain sizes. A

series of loops fabricoted from the annealed pipe
was filled from the same batch of NaF-ZrF -UF,
(50-46-4 mole %) and operated for 500 hr ot 1500°F.
Two loops were made from pipe annealed at 2100°F
that had a grain size of 1 to ]]/2 gr/in.? at 100 X,
while the loop fabricated from as-received lnconel
pipe and the one fabricated from pipe annealed at
1600°F contained about 6 gr/in.2. Very little dif-
ference in hot-leg attack was found in these loops.
Those with the larger grains may have had slightly
deeper attack, but the attack was heavier and more
general and the deep penetrations were concentrated
into fewer boundaries.

FLUORIDE CORROSION OF HASTELLOY B
IN THERMAL.CONVECTION LOOPS

G. M. Adamson
Metallurgy Division

Loops fabricated from both as-received and over-
aged Hastelloy B were operated satisfactorily.
The operating mortality rate has been reduced from
90% to 0% in the last group of four loops. The
increase in hardness during operation is not so
great in the loops constructed with over-oged
material as in the loops constructed with as-
received material, but, with proper care, the loops
of as-received material can be operated.

Very little attack was found in a loop which
circulated NaF-Zrf ,-UF, (50-46-4 mole %) for
1000 hr at 1500°F. The attack appeared as a few
voids to a maximum depth of 1 mil, with possibly
some increase in surface roughness. Most of the
surface roughness was present in the as-received
tubing, as shown in Fig. 6.17.

TABLE 6.8. EFFECT OF ZrH, ADDITIONS TO NaF.ZrF o UF, (50-46.4 mole %)
CIRCULATED IN INCONEL THERMAL-CONVECTION LOOPS AT 1500°F FOR 500 hr

 

 

 

 

Loop ZrH, Uranium Content (%)
Added Hot-Leg Attack
Mo. (%) Before Test After Test
469 Heavy general atiack and intergranular voids to a depth of 8 mils 8.5 8.8
459 0.2 Moderate to heavy attack to o depth of 6 mils 8.6 8.5
470 0.5 Light to moderate attack to o depth of 3 mils 7.5 7.1
460 0.9 Thin hot-leg deposit; no attack 5.4 5.1
471 2.0 Hot-leg layer to 1 mil thick; no attack 4.0 4.0

 

 

98
MiLS

 

 

MILS

 

 

 

 

Fig. 6.17. As-Received Hastelloy B (@) and Hot-
Leg Surface of Hastelloy B Loop (b) After Circu-
lating NciF-Zrl""‘-l.lF4 (50-46-4 mole %) for 1000 hr
at 1500°F. Etched with H,GrO, + HCL.  250X.
Reduced 36%.

Chemical analysis results now available for the
Hastelloy B loop previously operated? for 500 hr
confirm the low attack rate found metallographically,
since neither the nickel nor the molybdenum content
in the fluoride mixture increased.

LITHIUM IN TYPE 316 STAINLESS STEEL

E. E. Hoffman W, H. Cook
C. R. Brooks C. F. Leitten
Metallurgy Division

Tests have recently been compieted on three
type 316 stainless steel thermal-convection loops

“AG. M. Adomson, ANP Quar. Prog. Rep. june 10, 1954,
ORNL-1729, p 77. |

PERIOD ENDING SEPTEMBER 10, 1954

in which lithium was circulated. These loops were
constructed of 0.840-in.-0D, 0.147-in.-wall pipe.
The hot and cold legs were 15 in. in length, and
the 15-in. connecting legs were inclined af an
angle of 20 deg. The welding and feading opera-
tions on these loops were performed in a dry box
in a purified helium atmosphere.? &
these tests was there any indication of plug forma-
tion. The operating conditions are given in Table
6.9, Macroscopic examination revealed no dif-
ferences between hot- and cold-leg surfaces in
loops 1 and 2. Only loop 1 has been examined
completely; loops 2 and 3 have been sectioned
and have been examined macroscopically. Loop 2
was very similar in appearance to loop 1, with no

At no time dufing

crystal deposition. Loop 3, however, revealed
mass-transfer crystals attached to the cold-zone
walls.  These crystals did not plug the loop or
noticeably affect the circulation. The crystal
deposition was heaviest on the maojor radius of the
exposed loop-bend wall in the cold zone. This
loop has not yet been examined metallographically.

TABLE 6.9. OPERATING CONDITIONS FOR TYPE 314
STAINLESS STEEL THERMAL.CONVECTION LOOPS
WHICH CIRCULATED LITHIUM FOR 1000 hr

 

 

 

Hot-Zone Cold-Zone Temperawre
L.oop
N Temperature Temperature Ditferential
o.
(°F) (°F) (°F)
1 1490 1220 270
2 1472 1355 117

3 1301 1094 207

 

Loop 1, which was operated at the highest temper-
ature and with the highest temperature differential,
had no mass-transfer crystals in the cold zone, and
the moaximum attack in the hot zone was 1 to 2
mils (Fig. 6.18). Chemical analyses of the lithium
and the amounts of crystals recovered are presented
in Tcble 6.10. At present it is not understood why
so little moss transfer occurred in loops 1 and 2,
since in all three loops the same procedures and
testing techniques were used and all were filled
from the same batch of lithium. Some as yet un-
discovered factor seems to have an effect on the
rate of mass transfer.

 

5E. E. Hoffman et al., Met. Div. Semiann. Prog. Rep.
Apr. 10, 1954, ORNL-1727, p 37.

?9
ANP QUARTERLY PROGRESS REPORT

PLATE

 

 

0.005

 

200X

0.01i0

 

0.015

 

INCH

 

Fig. 6.18. Hot-L2g Surface of Type 316 Thermal Convection Loop After Circulating Lithium for 1000 he
at 1490°F. Specimen nicke! plated after test to protect edge. Etched with glyceria regia.

FUNDAMENTAL CORROGSION RESEARCH

G. P. Smith
Metallurgy Division

Mass Transfer in Liquid Lead

J. V. Cathcart
Metallurgy Division

As previously reported,® the investigation of
corrosion and mass transfer in liquid lead has indi-
cated that certain alloys possess much greater
resistance to mass transfer than their pure com-
porients. For example, the time required for small
thermal-convection loops containing types 410 and
446 stoinless steel to plug was from two to five
times longer than that required for comparable loops
confaining pure iron or pure chromium.

It has been suggested® that the increased re-
sistance to mass fransfer of materials such as the
400 series stainless steels might be related to a
tendency toward the formation of intermetallic
compounds in these alloys. To test this hypothesis

 

3. V. Cothcart, ANP Quar. Prog. Rep. June 10, 1954,
ORNL-1729, p 79.

100

TABLE 6,70. ANALYSES OF LITHIUM CIRCULATED

IN TYPE 316 STAINLESS STEEL THERMAL-
CONVECTION L.OGCPS

 

 

Components
Other Than Weight of Crystals
Material Lithium Removed from
Analyzed (ppm) Lithium
(g)
Fe Ni Cr
Lithium, as-cast, 40 10 <10
before test
Lithium fromloop 160 150 <10 0.0190
1, after test
Lithium fromlcop 120 40 <10 0.001

2, after test

further, a loop was operated which contained speci-
mens of a specially prepared 45% Cr—55% Co alloy.
This alloy composition corresponded to a two-phase
region in the cobalt-chromium phase diagram, one
phase being the intermetallic compound CoCr, The

 
loop plugged after 768 hr of operation with hot-
and cold-leg temperotures of 820 and 510°C, re-
spectively. A transverse section of the hot-leg
specimen is shown in Fig. 6.19. '

A survey of the alloys which have been tested
revealed that the plugging time for all those in
which intermetallic compound formation is a possi-
bility [types 410 and 446 stainless steel, 2%
Si~14% Cr~84% Fe, Hastelloy B (5% Fe-28% Mo~
67% Ni), 25% Mo—75% Ni, 45% Cr—55% Co, and 16%
Ni-37% Cr—47% Fe (austenite plus sigma}] was in
excess of 500 hr, the only exception being the loop
containing the 16% Ni-37% Cr~47% Fe specimens
which plugged in 456 hr. The plugging times for
loops containing the pure coristituents of these
alloys were: Ni, 2 hr; Co, 80 hr; Cr, 100 hr; and
Fe, 250 hr. No mass transfer was observed in a
loop containing molybdenum specimens after 500

PERIOD ENDING SEPTEMBER 10, 1954

hr of operation. On the other hand, loops with
specimens of alloys for which there is virtually no
tendency toward compound formation (types 304
and 347 stainless steel, Inconel, and nichrome)
all plugged in less than 150 hr. On the basis of
these data, it was concluded that resistance to
mass transfer in liquid lead is considerably greater
in alloys in which intermetallic compound formation
is possible.

Future work will include tests of two alloys: one
with the composition of 50% Cr—50% Fe and the
other with approximately 50% Fe~50% Mo. These
compositions correspond closely to the composi-
tions of intermetallic compounds in types 410 and
446 stainless steel, and if the ideas presented
above are correct, both should show relatively
high resistance to mass transfer in liquid lead.

e UNCLASSIFIED
e e T Y gy

 

Fig. 6.19. Tronsverse Sect‘ion of 45% Cr~35% Co Specimen Exposed to Liquid Lead at 820°C in the
Hot-Leg Section of a Quartz Thermal Convection Loop. '

101
ANP QUARTERLY PROGRESS REPORT

Flammability of Sedium Alleys

M. E. Steidlitz L. L. Hall
G. P. Smith
Metallurgy Division

The studies of the flammability of liquid sodium
alloys, which were reported previously,” were ex-
tended.
action of jets of sodium-bismuth and sodium-mercury
alloys is not significantly changed when the pres-
sure of air is varied from 0.25 to 1.0 atm,

Additional data on the reactivity of sodium-
bismuth solutions in dry air were obtained, and
the results are summarized in Figs. 6.20 and 6.21.
Figure 6.20 shows the limits of the region of no
reaction on o temperature-composition diagram. A

tt has been found that the degree of re-

careful calibration of the flammability apparatus
has shown that there wos a substantial error in the
temperatures previously reported in these tests.
[n this respect the earlier data have been corrected.
The circles represent the temperature-composition
values at which tests were conducted. The numeral
by each circle gives the number of tests. The un-
shaded circles represent tests which showed re-

 

7G. P.Smith and M. E. Steidlitz, ANP Quar. Prog. Rep.
Dec, 10, 1953, ORNL-164%9, p 26,

UNCLASSIFIED
ORNI~LR-DWG 3030

 

 

 

 

 

 

Co 3
800 | -~ O OO SE— -
| | [REACTION
‘ \ !
2 2 | | |
5 o 0 5
a . i e ////4 _
MR R
= 1 , ! v
" ] %
NO REACTION | / |
. / |
| ‘ ‘ 1‘ 2 2}/// 2 2
‘ | ’ ! c
o 0z 04 T e
MOLE FRACTION OF SCOIUM
Fig. 6.20. Temperature-Composition Diagram

Showing Regions Within Which lets of Sodium-
Bismuth Seolutions Did end Did Net React with Dry
Ajr.

102

action, while the rest showed slight reaction. The
cross-hatched band demarks the approximate limits
within which the line separating the region of
reaction from the region of no reaction lies. Figure
6.21 shows how the rate of reaction was found to
vary with composition at constant temperature. The
rate scale is arbitrary, with no reaction given the
value zero and that of pure sodium given the value
4. Further work is being done to determine the
flammability temperature for pure bismuth,

UNCLASSIFIED
ORNL-LR-DWG 3031

 

  

4 ] o —*“'7’—@{7*0
a 3 :
o i
= ; |
o ;
E o2
<7
Ll
[0l
i
>
!,_,__,
<I ‘
lIJj i
& © |
1
0.3 04 0.5 0.6 o
MOLE FRACTION OF SODIUM
Fig. 6.21. Relotive Reaction Rate of Sodium-

Bismuth Solutions ot 700°C aos o Function of
Temperature.

Thermodynamics of Alkali-Metal Hydroxides

G. P. Smith C. R. Boston
Metallurgy Division

Free Energy of Formation Data. Numerical data
for the stondard free energies of formation of the
alkali-meta! hydroxides and related compounds are
an essential guide in studies of hydroxide chemistry.
Therefore the free energy data have been computed
as a function of temperature for the hydroxide, the
peroxide, and the superoxide of sodium. The data
are presented in Fig. 6.22, together with the data
for the oxide and water which are used in subsequent
computations in this report,

Values of the stondord free energy of formotion
at 25°C are available® for sodium oxide, sodium
peroxide, and sodium superoxide. The value for
sodium hydroxide may be computed readily from
the values for the standard heat of formation of

8

 
UNCLASSIFIED
ORNl—LR—DWG 3032

S

 

     

 

 

 

 

 

300 500 700 900 {100
TEMPERATURE {°K)

1300

Fig. 6.22. Free Energy of Formation Data as a
Function of Temperature for the Hydroxide, the
Peroxide, and the Superoxide of Sodium.

sodium hydroxide® and the entropies of sodium
hydrcoxi«:ir:a,10 hydrogen, oxygen, and sodium in their
standard reference states. The computation shows
that AF;%(NQOH) = ~91.0 kcal/mole. At higher
temperatures, values of the standard free energy of
formation have been given in the literature for only
sodium oxide.!! However, values have been esti-
mated for sodium peroxide, sodium superoxide, and
sodium hydroxide in the following way. The curves
of the standard free energy of formation vs temper-
ature for the class of compounds dealt with here
are known to a fairly good degree of approximation
to consist of segmented straight lines with dis-
continuous changes in slope at the transition
points of the compound and of its component ele-
ments. One point on each of these curves can be
fixed from the values for the standard free energy
of formation data at 25°C, mentioned above. The

 

TUnless otherwise specified, literature values used
herein are token from F, D, Rossini, Selected Values of
Chemical Themodynamic Properties, National Bureau of

Standards, Washington, D. C., 1952,

10J. C. R. Kelly and P. E. Snyder, J. Am. Chem. Soc.
73, 4114 {1951).

”F. D. Richardson and J. H. E, Jeffes, J. [ron Steel
Inst. (London) 160, 261 (1948).

PERIOD ENDING SEPTEMBER 10, 1954

slope at this point is determined from values of
the standard entropy of formation by using the

relation
(8.’.\1«";)
e = =AS% .
ar /

P

These entropy data are obtained from the standard-
state entropy values. The change in slope ot each
transition point is determined from the transition
entropy, which, of course, is readily computed from
the heat of transition.

Recently cbtained data'? on the heat capacity of
liquid sodium hydroxide wili permit a more precise
evaluation to be made of the slope of the free
energy curve from the melting point up to 1000°C.
The heat of fusion of sodium peroxide (TM =~ 980°K)
was not available, but it should introduce only a
small change in slope, which may be considered to
be negligible for the computations which follow.

Decomposition of Hydroxides. In considering any
solvent as a reaction medium it is important to de-
termine whether the solvent tends to decompose to
an appreciable extent, particuiarly if any of the
products of the decomposition take part in the re-
action. The classic example is, of course, water.
Here the covalent water molecules decompose
slightly into ions. The resulting equilibrium fre-
quently exerts a controlling influence on the course
of aqueous reactions. The following is an outline
of the results obtained to date in a theoretical
search for the significant decomposition equilibria
in fused alkali-metal hydroxides. It should be
emphasized that the work presented here is in no
sense definitive or complete.

The most important result of the treatment given
here is the evidence for the possible occurrence of
appreciable quantities of peroxide as the result of
decomposition equilibria. The way in which this
peroxide may play a decisive role in mass transfer
and corrosion will be shown in o later report.

In a fused alkali-metal hydroxide in an inert
environment such that the volume of any gas phase
is small compared with that of the liquid phase,
the fused hydroxide may be regarded thermodynami-
cally as being in equilibrium with every chemical
species composed of nothihg more than oxygen,
hydrogen, and the alkali metal. In practice, many
of these chemical species can be ignored for either

 

Y2y D. Powers and G. C. Blalock, Enthalpies and
Specific Heats of Alkali and Alkaline Earth Hydroxides
at High Temperatures, ORNL-1653 (Jan. 7, 1954).

103
ANP QUARTERLY PROGRESS REPORT

First, the equilibrium concentra-
tion of the species in question maybe insignificant.
Second, the rate of formation of the species in
question may be so slow that it does not approach
equilibrium in any time period which may be con-
sidered,

of two reasons.

Since in the following discussion the
thermodynamic approach is used, only the first
reason will be treated.

There are three particularly useful theoretical
measures of the significance of a decomposition
They are the equilibrium constant, the
standard free energy change, and the concentration
of the products of decomposition. The third measure
is the most directly related to experiment and hence,
at the present stage, the most useful. The eonly
concentration in o fused hydroxide which can be

reaction.

meosvred at temperature by present methods is the
partial pressure of a gaseous product. Hence,
considerable emphasis is placed on possible de-
composition reactions which produce gaseous
products and on an expression of their equilibrium
concentrations in terms of the decomposition
pressure,

The decomposition pressure of a pure substance
with regard to a gaseous product is referred to here
as that partial pressure of the gaseous product
which, when in equilibrium with the condensed
phase, is required to maintain the over-all composi-
tion of the condensed phase equal to the composition
of the pure substance. According to the phase
rule, the pressure so defined is unique for a given
temperature. If there are several gaseous products,
there is o decomposition pressure for each such
product.

The standard states used here ore those con-
ventionally chosen in defining the standord free
energy of formation. Although this is not the usual
choice of standard states in the case of solutes,
it is definitely the most convenient choice for the
computations which follow.

The four possible decomposition reactions which
give a gaseous product at the temperatures at which
the hydroxides are liquid are considered here,
They are as follows:

K
(1) 280H =' M,0 + H,0, AFS
K2 o
(2) 2MOH =2 M,0, + H, , AFS
K3 o
(3) 2MOH = MO, + M + H, , AFS -,

104

K
(4) 2MOH =* MO, + MH + LH, ,  AFY,

where all substances are partitioned among all
phases present and where K. and AF? are the
equilibrium constants and standard free energies
of reaction respectively.

The sparsity of reliable information on the alkali
metal-oxygen compounds raises serious questions
about the application of all the above squilibric to
all the alkali metals. This subject is treated by
Brewer® and will not be repeated here, However,
it is the essential point of this treatment to de-
termine whether unsaturated oxygen ions such as
the peroxide and superoxide ions may occur as
decomposition products, in addition to the saturated
oxide ion. The existence of some sort of unsatu-
rated oxygen ion is certain for all the alkali metals,
and it is known that the stability of the unsaturated
species increases in going from [ithium to cesium.
Therefore, the basic competition referred to below
between hydrogen and the mono-oxide for water
should be a characteristic of all hydroxide de-
composition equilibria and should lead to a water-
hydrogen equilibrium in the gas phase of the type
discussed. The data for sodium compounds ore
reasonobly reliable. Hence, sodium hydroxide will
be treated separately.

By definition of the equilibrium constant,

a a N N
M,0 TH,O VM20 M,0 YH20 H,0
K = ™ ~
2
AMOH

 

7
2 2
YuoH VmoH

where o, is the activity of substance X in the
hydroxide phase and N, and y, are the corre-
sponding mole fraction and activity coefficient,
respectively. Let the male fractions be those
which are in equilibrium with the decomposition
ressure. lThen N = N , and it is possible
F:o write #2© 120

 

2
“H,0 yM2O
Ky = [y = , [y o
“MOH YH,0
Likewise, for Eqs. 2 through 4,
2
a.
Ky = FZ a ! I2 = '
MOH sz
 

“/Moz YMH
K, =1, =, Uy = 4——

2
S HOH Yi

 

The above equations are exact. However, a more
useful form can be obtained by assuming that the
extent of decomposition is very small. As will be
shown later, this assumption involves a negligible
error for all four reactions. From this assumption
two conclusions can be drawn: (1) the activity of
the hydroxide as used in the above equations may
be taken to be unity because of the choice of
standard states; (2) the partial pressures of water
and hydrogen in the gas phase may be taken to be
equal to the fugacities in the gas phase and hence
equal to the fugacities in the liquid phase. Finally,
by definition, the fugacity of the gases in the
standard reference state is unity. Substituting
these conclusions into the above squations, selving
for the partial pressures, and applying the reaction
isotherm give

 

 

 

 

\1/2 o
(50) i k o
| — = e S K —a e z
’ szo [ F 1/2 P 2RT
(56) 2 : o
7 R = exp {——— ¥
(pHZ)z L, I,"? P 2/t ]
. 3 o
(5¢) ( G 7 APy
C =1 - = e EXD A — ,
(PH2)3 1y ) e 3RT
K4 2/5 ] ZAFE‘
(5d) (1t y = | = 2 e @XP § e o
( 2)4 I, F42/s SRTJ

where (sz) is to be read as the partial pressure
\ .

of hydragenl provided that no other hydrogen-
producing reaction is operative save the /th,

From the original choice of mole fractions if
moy be noted that the above partial pressures are
the decomposition pressures and that there are

PERIOD ENDING SEPTEMBER 10, 1954

three values of the decomposition pressure with
respect to hydrogen. Except over unique temper-
ature ranges, one of these three decomposition
pressures will be much larger than the other two
and hence, by the law of mass action, will stabilize
the others. Thus, only one of the three pressures
will be the decomposition pressure of MOH with
regard to hydrogen.

Because of the lack of thermodynamic dota on
all the alkali-metal hydroxides except sodium, it
is necessary, in most opplications, to use ratios
of decomposition pressures.. Thus, it is assumed
that reactions 1 and 2 are simultaneocusly at equi-
fibrivm and that they determine the gaseous water
and hydrogen pressures, respectively. By dividing
Eq. 2% inte 5a, the decomposition pressure ratic
becomes '

 

 

 

 

 

 

: 1/2
pHZO Ky 4
(6“) e = ‘it;*‘-"*-‘
sz 12 1,2
] AFO
= exp —_ .
1 i 21/2 2RT
where
AF;’ = A - AFY
K Kl
1,2 7 T
Kq
. I, f’f»izo VH2
I 1,2 T = T

2 VA1202 }”Hzo

it will be noted that AFFY 20 unlike AF] and AFY,
does not contain the Free energy of formahon of
the hydroxide.

The concentrations of the exide in reaction 1 and
of the peroxide in reaction £ are not independent
but are related through the equilibrium

>

K
7y M0 + Ko, =7

Ao
, , = M0, AF3

Furthermore, the concentrations of oxygen in this
equilibrium ond of hydregen in reoction 2 are
related through the equilibriom

K

1 - 8 o
8 H, + O, =8 H0 AF

105
ANP QUARTERLY PROGRESS REPORT

The existence of these equilibria is compatible
with Eq. 6a, as may be seen from the fact that
AFS 5 = AFg — AF3, Ky 5 = Kg/K;. Hence it
would have been possible to derive Eq. 6a by using
equilibria 7 and 8. This is because the two sets
of equilibria 1 and 2 and equilibria 7 and 8 are
both equivalent to the one equilibrium
9) M0, + H, =12 M0 + H,0 AF?S

Maba + My = Mgl + Hal 1,2 °
From Eq. 9 it will be noted that the pressure ratio
(pHZO/sz)] , is determined by the competition

between the oxide and hydrogen for oxygen, with
the one trying to form the peroxide and the other
trying to form water.

In like manner, an expression analogous to Egq.
6a can be derived under the assumption that equi-
libria 1 and 3 determine the water and hydrogen
pressures, respectively. This expression is

 

 

 

P
H,0 13
(65) ot
Py 1,3 L
1 AF?,:%
= eXp { - —- ),
s 6RT
where
AFS o = 3AFS - 2AFS
1/2
K KI
1,3 T T e
' 1/3
Ks
1/2
r 't
1,3 ~ oo
’ N 1/3
I’y

By using the free energy data computed above
for sodium hydroxide, the representative values

shown in Table 6.11 were obtained. The pressure-
activity coefficient products in this table give the
pressures which would be obtained if the solutions
were ideal. Figure (.23 shows the relative im-
portance of equilibria 1 and 2 in terms of the logo-
rithm of the pressure ratio as a function of the
reciprocal of the absolute temperature. Figure
6.24 gives the free energy differences (AF‘I’ , and
AF‘]"3) as functions of absolute fempemfure.'

Examination of these data shows the following:
(1) Apart from unexpectedly large deviations from
ideality, equilibrium 1 will predominate at all
temperatures.  (2) Assuming ideality, hydrogen
evolution from fused sodium hydroxide in an inert
container will be small at all temperatures, but it
should be measurable at 900°C. Likewise, the
peroxide concentration at all temperatures may be
small, but it should be appreciable at 900°C. (3)
Deviations from ideality of a moderate amount
could make hydrogen and peroxide formation either
insignificant or greatly important at high temper-
atures,

UNCLASSIFIED
ORNL—LR—DWG 3033

   

 

 

30 Lo — L

60 80 100 120 140 160
{1/ TEMPERATURE ] x40* k)™
Fig. 6.23. Decompesition Pressure Functions
for NaOH,

TABLE 6.11. DATA ON THE DECOMPQOSITION OF SODIUM HYDROXIDE

 

 

 

.o -0 - T
MMk R
25 35.3 76.0 . 10726 10736 s
600 21.4 64.4 4.7 x 1078 10718 10727 1.7 7.4 x 1078
900 13.5 59.9 3 x1073 8 x 10712 16718 42.0 2.1x107°

 

106

 
UNGLASSIFIED
ORNL-LR-DWG 3034

 

 

 

Fe (keal/mote)

I
|

100

 

 

 

 

  

 

10 1 ‘
200 HOO 1300

TEMPERATURE (°K)

Fig. 6.24, Free Energy Differences as Functions
of Temperature.

There are no satisfactory heat of formation
dota for any of the alkali-metal hydroxides other
than those for sodium. Therefore, only the rela-
tive importance of the decomposition equilibria
can be studied. Some of the entropy values for
the higher oxides have been estimated and are
in the literature. However, the heats
rubidium and cesium oxides

reported
of formation for
are quite uncertain, and it is. presently believed
that the accuracy of the AF‘{ 5 and [\I*1 5 values
given in Table 6.12 would not be improved by
use of these entropy estimates. Therefore, only
enthalpy values will be used, but it will be shown
that this restriction is not so serious as it might
at first seem to be.
Comparing equilibria 1 and 2 shows

AF°
1,2

= AHA‘;20 + AF;fizo - AHM 0, T (Z,\ij

All the data required in the above equation are
available, except those for the entropy term. Since
this term represents o difference in entropy for two
similar substances, it may be supposed that its
omission will not be serious. A similar situation
obtains for AFI 3» Hence, the following approxi-
mations will be used.

PERIOD ENDING SEPTEMBER 10, 1954

AFS , X AHS . + AF? ~ — AHS ,
—o ~ A » .

The degree of approximation can be checked by
using the dato for sodium compounds. In this case
AFS g s -35.5 kcal, when using the approximation,
as compcred with fhe correct value of -40.7 keal.
Likewise, AFS ~94.2 kcal compared with the
correct value of -94 9 keal. The approximate
values of f\[‘{ and AFS 5 at 25°C are tabulated
in Table 6.12 for all the ‘alkali-metal hydroxides.
Table 6.12 shows that hydrogen and the peroxide
should be produced most readily in potassium’ hy-
droxide and least readily in lithium hydroxide.
Nevertheless, equilibrium 1 is predominant, and
equilibrium 3 and hence equilibrium 4 are insigni-
ficant.

TABLE 46.12. APPROXIMATE /_\Ft]’ 2
AND AF‘; 5 AT 25°C

 

 

 

O 20

{kcal) {kcal}
Li ~47
MNa -36 —94
K =25 ~76
Rb —-34 ~73
Cs ~35 =70

Various investigations of corrosion and mass

transfer have shown the fundomental role played
by hydrogen. The above computations show that
hydrogen is important in stabilizing sodium hy-
droxide at temperatures above 600°C, as shown in

 

- ASMzoz)

 

Table 6.11, This is the temperature range at which
mass transfer becomes significant. These compu-
tations also link the formation of hydrogen with
the formation of unsaturated oxygen ions. As will
be shown in a subsequent report, the presence of
these unsoturated ions may be necessary for mass

107
ANP QUARTERLY PROGRESS REPORT

transfer to occur. As may be seen in Table 6.12,
hydrogen and the unsaturated oxygen ions are most
important for the hydroxides of potassium, rubidium,
and cesium.

Furthermore, these computations provide the first
estimate of the thermal stability of the alkali-metal
hydroxides (other than LiOH). These estimates
confirm the frequently made assumption that the
alkali-metal hydroxides are by far the most thermally
stable hydrogen-containing liquids known.

It must be emphasized that these computations
are given as o progress report in a continuing in-
vestigation. Thus, many of the valuesobtained,
such as the standard free energies of formation, do
not in themselves provide direct information about
hydroxide corrosion, but they are essential values
for any theoretical investigations of corrosion,

Experimental Studies (with M. E. Steidlitz, Metal-
lurgy Division). An investigation is under way to
determine experimentally whether a significant
amount of hydrogen is liberated when NaOH is
heated to 900°C in an inert environment. The pre-
liminary method being tested is tc use a pair of
automatic Toepler pumps to concentrate any gaseous
products liberated from the hydroxide. The col-
lected gas is analyzed with a mass spectrometer,
The most difficult aspect of this research has been
the endeavor to provide an inert container for the
NaOH, which reacts with all metals studied at
900°C (including the most noble metals) to pro-
duce hydrogen. It reacts strongly with nearly all
ceramics except magnesium oxide. Unfortunately,
nonporous crucibles of pure MgO are not available
because of the difficulty in sintering this material.
Ordinary MgQ crucibles, which usually contain a
small amount of binder such as silica, breck down
due to attack on the binder. The problem was
salved by machining out large single crystals of
pure magnesium oxide prepared by the Ceramics
Group.

In the experiments which have been performed
to date, appreciable quontities of hydrogen have
been found in addition to much larger quantities of
water. However, there was evidence of contamina-
tion by organic material and the hydrogen may have
come from this source. These studiesare continuing.

 

134. J. Butiram et al., ANP Quar. Prog. Rep. june 10,
1954, ORNL-1729, o 63,

108

CHEMICAL STUDIES OF CORROSION

F. Kertesz
Materials Chemistry Division

Effect of Temperature on Corrosion of
lncone!l and Type 316 Stainless Steel

H. J. Buttram R. E. Meadows
N. V. Smith

Materials Chemistry Division

The effect of temperature on the corrosion of
Inconel by molten fluorides in static tests, as
indicated by the extent of void formation observed
and by chromium concentration of the melt, was
discussed in a previous repart.'? In those studies
there was soime evidence that a maximum in the
corrosion could be observed at 800 to 900°C and
that both void formation and chromium concentration
in the melt were less at 1000°C than at 800°C.
While it is possible to rationalize the decreased
void formation on the basis that a high rate of dif-
fusion for chromium in the metal may minimize the
formation of voids, the decreased chromium con-
centration in the melt is, however, more difficult
to explain.

During the past quarter NaZrF ; and NaF-ZrF ,-UF
(53.5-40-6.5 mole %) were tested in type 316 stain-
less steel under static and nearly isothermal con-
ditions for 100 hr at 100°C intervals over the range
600 to 1000°C, as a check of the findings with
Inconel. Metallographic examination of the cap-
sules exposed at 600°C revealed light intergranular
penetration up to 1 mil in depth. At 1000°C the
attack was still light, but one isolated case of
intergranular penetration to a depth of 18 mils was
noted. When the chromium concentration of the
melt was plotted against test temperature for the
NaZrF ., a slight maximum in the curve appeared
between 800 and 900°C; a similar plot for the UF ,-
bearing mixture shows a plateau between 700 to
1000°C. However, more chromium appeared in
solution in the NeZrF ¢ samples than in the UF,-
bearing mixtures. The tendency toward less cor-
rosion at 1000°C than ot 800°C never appears to
be pronounced, but it has been sufficiently persistent
in these tests to require explanation.

A plausible explanation can be evolved in the
following manner. The NaZrF, melt which has
been used in these experiments gove rise to ab-
normally high chromium concentrotions (1000 ppm)
compared with equilibrium values (200 ppm) ob-
tained in measurements carried out with carefully
handled melts and pure chromium metal. It appears,
therefore, that impurities were predominantly re-
sponsible for the amounts of corrosion measured.
The most likely impurity is HF, which con result
from hydrolysis of adsorbed water or hydrated water
in salts containing Na Zr,F,, or from inadequate
purification.

If HF were responsible, the ascending portion of
the curve could be due to an increasing rate of
reaction with increasing temperature. (There is
little reason to believe that equilibrium conditions
are reached in static capsules in 100 hr.) The
descending portion of the curve may be ascribed to
a marked decrease in corrosiveness of HF at
increasing temperatures.

Standard free energy estimates'* show that AF®
for the reaction of HF with Ni to form NiF, becomes
positive near 300°C, with Fe to form FeF, ot
800°C, and with Cr to form Crf, at 1400°C. In
the experiments under discussion here, the melt at
the end of a 100-hr test contains mostly Cr*¥, with
very little Fe*" and Ni*™,

There is some evidence that at the lower temper-
ature ali three elements are attacked rather slowly
and indiscriminately by HF and that the resulting
Fe'' and Ni*' are replaced by Cr*™¥ in o fast
secondary reaction. The effect of increasing
temperature is to increase the degree of approach
to equilibrium in 100 hr in a static capsule; how-
ever, at higher temperatures the effect of a more

unfavorable equilibrium constant becomes manifest.

Iron and nickel are relatively unreactive toward HF
at 1000°C, and the amount of Cr™" pickup in 100 hr
could well be less than that noted at 800°C for
three reasons: the free energy change for the
reaction of chromium with HF is smaller, fewer
Fe** and Nit" ions are present for reaction with
Cr**, and there is mechanical interference by the
unreacted nickel and iron. In other words, chromium
is most readily oxidized from an alloy by HF if the
accompanying alloy constituents are alsc attocked.

Corrosion by Fission Products

H. J. Buttram R, E. Meadows

Materials Chemistry Division

in a previously reported experiment,'® a fuel
mixture containing simulated fission products ot
1000 times the concentration expected in the ARE

My | Brewer et al., The Thermodynamic Propierties and
Equilibria at High Temperatures of Uranium Halides,
Oxides, Nitrides, and Carbides, MDDC-1543 {Sept, 20,
1945, rev. Apr. 1, 1947),

PERIOD ENDING SEPTEMBER 10, 1954

was corrosion tested. Very heavy attack was
observed; this was expected, not becouse normal
amounts of fission products are particularly cor-
rosive but because oxidizing agents such as RuF ,
MoBr3, and elemental tellurium were used in suffi-
cient concentration fo cause excessive corrosion
in any case,

Additional experiments were performed with the
use of the same additives, individually, in amounts
as small as possible in an attempt to measure the
relative activities as corrosive agents. Preliminary
results, such as those previously reported which
showed YF, to be extremely corrosive, and oddi-
tional trials during the past quarter, which showed
that such compounds as CsF were also very cor-
rosive, made it obvious that the techniques em-
ployed were unsuitable and that very misleading
indications were being obtained. Free energy
considerations, as well as general experience with
fluoride systems in closed capsules, make it ap-
parent that impurities such as HF and water were
responsible for the effects noted. Hence it must
be concluded that none of the experiments per-
formed to date on corrosion by simulated fission
products have heen sotisfactory for the intended
purpose.

Controlled-Yelocity Corrosion Testing Apparatys

N. V. Smith F. A, Knox
Materials Chemistry Division

In order to more nearly simulate corrosion con-
ditions of molten fluorides circulating through
reactor components, a controlled-velocity corrosion
testing apparatus has been constructed. This ap-
paratus, similar to one previously described, '® al-
lows the rapid transfer of molten fluorides through
both heated and cooled test sections. The apparatus
consists of two 4-in.~1D, 24-in.-high Inconel cylin-
drical pots connected by means of three sections of
1/‘(in. inconel tubing with ¢ wall thickness of 0.035
in. The two outside sections are 5 t long and the
center section, which was initially 1 ft long, was
later increased to 3]/2 ft. The center section is
cooled by either oir or water. Suitable furnaces
and heaters permit holding the pots and transfer
lines at desired temperatures. FEach Inconel pot
has a gos inlet which allows opplication of helium

 

Y5, 3. Buttram et al., ANP Quar. Prog. Rep. June 10,
1954, ORNL-1729, ¢ 65,

1614, W, Hoffman ond J. Lones, ANP (Quar. Prog, ‘Reb.
Mar. 10, 1954, QRNL-1692, p 98.

109
ANP QUARTERLY PROGRESS REPORT

at predetermined pressures to transfer the melt, A
probe indicates when the desired liquid level is
reached in the second pot and activates a relay
which closes the helium inlet to the first pot and
opens a corresponding one in the second pot to
reverse the cycle.

Velocities obtained with the apparatus have
reached 5 fps, which corresponds to a Reynolds
number of only about 1500. It is hoped that an
improved design now being assembled will make
possible turbulent flow. A 200°C gradient across
the cooling section was desired; however, it has
not yet been possible to achieve a differential of
more than &0°C. The transfer cycle consists in
having the melt contained in either pot at 700°C,
being heated to 800°C in the first section of the
transfer line, cooled in the middle section, brought
back to 700°C in the third section, momentarily
stored in the opposite pot at 700°C, and then the
cycle is reversed, A complete cycle, that is, from
one pot to the other and back, required approxi-
mately 5 min. It is believed that this apparatus will
yield corrosion data under conditions not easily
obtainable by present means.

Reaction Between Grophite cad Flueride Melt
F. A. Kinox

Materials Chemisiry Division

Graphite has been used extensively as a con-
tainer material during preparations of fuel and

110

coolant materials; therefore, an investigation was
made of possible reaction between it and a fluoride
melt. In the procedure used an lnconel reaction
tube was loaded with either NaF-ZrF -UF, (53.5-
40-6.5 mole %)} alone or with the fluoride mixture
plus graphite which had been treated under vacuum
for degassing. The reaction tube was connected
to @ manometer and then held at 1000°C until the
gas pressure became constant.

It was found that the rather large gas pressures
obtained were due chiefly to SiF,. Analysis of
the graphite indicated the presence of nearly 1%
Si. The use of spectrographically pure graphite in
this apparatus has yielded pressures very nearly
the same as the pressure of the pure fluoride. No
HF or CF, was detected during these tests.

Lithium Fivoride Castings
N. V. Smith F. Kertesz

Materials Chemistry Division

At the request of the Instrumentation and Controls
Division, high-density lithium fluoride cylinders
were cast from powdered material for use in an
x-ray spectrometer. These cylinders were pre-
pared by charging the powder to grophite crucibles
and heaiing in an inert atmosphere to 900°C. The
graphite liner containing the melt was then re-
moved from the furnace and a strong air jet was
applied to its periphery. The resulting cylinders
were uniform and acceptable.
PERIOD ENDING SEPTEMBER 10, 1954

7. METALLURGY

W. D. Maniy
Metallurgy Division

Creep and stress-rupture testing have, in the
past, been performed primarily on uniaxially stressed
specimens. lests of this type are continuing, but
the interest is now centfered on tube-burst tests in
which a tube that is closed at one end is stressed
with on internal gas pressure. The stress pattern
introduced into the specimen in this test more
nearly approaches the stress pattern that will be
found in ANP-type reactors. The ratio of the tan-
gential stress to the longitudinal stress is quite
important in the ductility of the metal being tested,
and therefore equipment in which the lengitudinal
and the tangential stresses can be varied has been
constructed. A theoretical analysis of a stressed
cylindrical pressure vessel has been made with
which a check on the experimental results con be
obtained. Also, to determine the effect of tan-
gential and compressive stresses on the corrosion
rate of metal in contact with fused salts, an ap-
paratus for loading a sheet specimen in bending
was designed and constructed; tests are now being
made.

In the investigation of high-thermal-conductivity
materials for fins for sodium-to-air radiators, stresse
rupture and creep tests were made on copper fins
with various types of cladding at stress levels be-
tween 500 and 2000 psi at 1500°F. The tests have
shown that for a 1000-hr exposure in air, stresses
greater than 500 psi and less than 1000 psi are
tolerable; that is, in this stress range there is no
indication of brittleness in the core or oxidation of
the core due to cladding failure, Thermal conduc-
tivity measurements of a 6% Al-94% Cu aluminum
bronze were made in the temperature range 212 to
1562°F, '

Tests of brazing alloys have shown that in braz-
ing high-conductivity fin materials to Inconel tubing
there are four alloys that can be used: low-melting-
point Nicrobraz (LMNB)}, Coast Metals alloy 52,
electroless nickel, and an Ni-P-Cr alloy. From the
over-all considerations of melting point, oxidation
resistance, dilution of fin and tube wall, formation
of low-melting eutectics, and flowability, it was
found that Coast Metals alloy 52 was the best
alloy for the construction of radiators with high-
conductivity fins.

A sodium-to-air radiator with 6 in. of type-430
stainless-steel-clad copper high-conductivity fins
was fabricated by using a combination heliarc
welding and brazing procedure. The tube-to-fin
sections were assembled and brozed with Coast
Metals alloy 52. The tube-to-header joints were
made by using the semiautomatic welding equipment
and then back-brazing. The header sections were
closed by manual heliarc welding.

In the construction of a 100-kw gas-fired liquid-
metal-heater system, packed-rod nozzle assemblies
were needed as the inlets for gir and gas. These
devices were made by brazing ]/B-in.-dicu stainless
steel rods in o tight ossembly with a ductile,
oxidation-resistant alloy of 82% Au-18% Ni.

Work has started on the formation of duplex tubing
in an attempt to prepare alloy composites that have
good corrosijon resistance on the inner surface and
oxidation resistance on the outer surface. Com-
posites of copper and type 310 stainiess steel,
Inconel and type 310 stainless steel, Inconel and
Hastelloy B, and Hastelloy B and type 310 stainless
steel have been prepared. Tubing produced from
stainless-steel-clad molybdenum and columbium and
from several special Inconel-type alloys was fabri-
cated into thermal convection loops for fluoride
corrosion testing. Several new alloys were produced
for the corrosion tests with liquid lead.

Attempts are being made to find new olloys in
the nickel-molybdenum system that will have better
high-temperature strength and fluoride corrosion
resistance than Inconel has. Hastelloy B satisfies
these requirements, but it has poor fabrication
properties and oxidation resistance, and it loses
its ductility in the temperature range of interest for
application in high-temperature circulating-fuel re-
actors. Investigations are under way in an effort
to find a suitable melting and heating treatment that
will increase the ductility of Hastelloy B in the
temperature range of interest,

Investigations of methods for producing boron
carbide shield pieces of the required density and
shape indicate that the pieces should be molded
with nonmetallic bonding material by cold pressing
followed by sintering. The bonding materials being

m
ANP QUARTERLY PROGRESS REPORT

studied include sodium silicate, silica, silicon
nitride, and boric oxide.

The fluoride-to-sodium intermediate heat ex-
changer, which failed in a life test after 1680 hr of
cyclic service in the temperature range 1080 to
1500°F, was exomined, |t is probable that the
failures in the tube-to-header welds were caused
by unequal thermal expansion, which caused stress
concentration at the roots of the tube-to-header
welds. These stresses, combined with expansion
and contraction of the tubes, would tend to propa-
gate cracks through the walls. These failures
emphasize the extreme desirability of using back-
brazing as o means of minimizing the notch effect
in tube-to-header welds.

STRESS-RUPTURE TESTS OF INCONEL

R. B. Oliver D. A. Douglas
J. H. DeVan J. W. Woods
Metallurgy Division

The tube-burst test for obtaining information on
the stress-rupture properties of Inconel has been
studied intensively. The siress pattern introduced
intfo the specimen in this test simulates to some
extent the stress pattern that will be present in
circulating-fuel reactors. The test consists of
stressing a closed-end tube with internal gos pres-
sure. In tests of this type reported previously, | it
was observed that Inconel specimens stressed in
this manner showed less ductility and much shorter
rupture life than the uniaxially stressed specimens,
and therefore an intensive study of the multiaxial
stress system was initiated. W. Jordan of the
Mechanics Deportment of the University of Alobama
is investigating this problem.

Part of the investigation consists of a study of
the theory of stresses in cylindrical pressure ves-
sels, with particular attention to the variations in
siresses calculated by the thin-wal! formula vs the
stresses determined by the more exact Lamé theory.
The stresses under discussion are those in the
walls only, with no consideration given to the end
closure shape, except that the ends are assumed to
be completely closed. The three principal stresses
at a given point are the radial stress (o), the tan-
gential (hoop) stress (o), and the longitudinal
(axial) stress (o ). The theory of elasticity yields

 

2. B. Oliver, D. A. Douglas, and J. W. Woods, ANP
Quar, Prog. Rep. June 10, 1954, ORNL-1729, p 89.

112

the following equations:

o pir? - por?J ( 1 ) r?ri (PO Pz-)
o = e L |
ri — r? T2, Tg - T
Piff"‘POTi | rro(po~{71)
o, MHE (P
rz -7 \f2 r — T
pir? - porg
(3) O‘a = _‘2’—;“ ’
reo— T

where r is the radial distance to the point at which
the stress is desired, 7 and r are the internal and
external radii, and P and p, ore the infernal and
external pressures, For the special case of the
cylinder subjected to internal pressure only (po =0},
the equations reduce to:

(4) Opi = Pi s Yo T 0,
r2 + r2 2r2
5) _ Q 2 _
( oti - pzl 2 2 ! g[o o pz —-2— ----------- '
r  — 7. r r.
o i o i
2
7i
(6) Oaz’ = Gao = pl 2 !
r°- - r

where the additional subscripts 7 and o on the
stress terms indicate stresses ot the inner and
outer surfaces. To simplify these equations for
application to cylinders with thin walls, it is fre-
quently assumed that the tangential stress does
not vary across the wall of the vessel. This simpli-
fication results in the following equations for in-
ternal pressures only;

(7) S,. =5

(8) s . =395 = 1"5; ’

where the letter S is used to denote stress and ¢
is the thickness of the wall (¢ = r- ri)' A simpli-
fied expression for radial stress in thin-walled
cylinders is not commonly used.
In order to determine the error involved in using
the formulas for thin-walled vessels, ratios of the
stresses computed by the exact and the approximate
theories are given below.

 

 

¢
U,. ' r.
(9 e
S, r t
te z 2 4 —
Ti
1o 2
(10) 2 ,
Sto 2 +_i_
7
o
(11) a2
Sa 2 +L

Equations 9 through 11 are plotted in Fig. 7.1.

The so-called ‘“‘exact’’ formulas (often called the
Lamé formulas) are valid only while the stresses
are elastic. The approximate formulas are based
on equilibrium considerations and the assumption
that the stresses are distributed uniformly across
the wall. Under inelastic (including creep) con-
ditions, it is believed that the tangential (hoop)
stress distribution lies somewhere between those
assumed in the Lamé and in the approximate elastic
theories. With increasing stress, increasing temper-
ature, and increasing time, the actual stress dis-
tribution will depart from the Lamé assumption and
approach the thin-wall theory assumption. Thus
the two theories give the upper and lower bounds
for the actual stress distribution.

For the longitudinal (axial) stresses, the Lame
and the thin-wall theories both use the assumption
that the stresses are uniformly distributed over the
cross section, Since the Lame theory uses the
correct value for the cross-sectional area, while
the thin-wall theory uses an approximate area, the
Lamé theory is believed to be more correct in both
the elastic and the inelastic ranges of stress. The
majority of the tube-burst tests made to date have
been of 0.010- and 0.020-in.-wall tubing, and thus
the approximate formulao applies. Specimens are
now being tested which have 0.060-in. walls, Since
it is predicted that under the test conditions the

PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
ORNL~-LR-DWG 2848

 

   

 

  

 

 

 

 

 

 

1.28 :
I ] | | { i
/"2
i ’.
- " e !
1,24 — faa"‘“f/'a (\Sf'—PT
A
= N
1.20 B |
2.2
T T THIN-wALL
118 L. r,e - ,72 THEORY ]
LAME THEORY
Ve LI
1.08 |
a
=
108 Lo
$.00
096
032 |-
0881 o STRESS COMPUTED 8Y LAME THEGRY - T~ |
S STRESS COMPUTED BY THIN —WALL THECRY
084 | | J |
o) 005 010 015 0.20 0.25 0.30
.
I‘}'
Fig. 7.1, Comparison of Elastic Stress in

Cylinders as Computed by Lamé Theory and by
Thin-Walled Pressure Vessel Theory.

stress will be uniform across the wall thickness
the approximate formula will be used for calculating
the stresses, and the slight error thus introduced
will be neglected,

It is well known that in the case of a thin-walled
closed-cylinder pressure vessel the ratio of the
tangential (hoop) stress to the longitudinal (axial)
stress is 2:1. Some investigators have shown that
this stress distribution results in the minimum
ductility for a given material, Therefore, in order
to better understond the criteria for predicting
failure of tubular specimens and in order to study
the effects of anisotropy of the tubular material, it
would be desirable to be able to test a series of
specimens with varying stress ratios. Jordan has
devised the apparatus described below with which
it is possible to stress a tubular specimen purely
tangentially, purely axially, or in ony desired ratio
of these stresses. '

113
ANP QUARTERLY PROGRESS REPORT

The most obvious method of changing the stress
ratic is by loading a tubular specimen axially with
dead weaights ond producing an axial stress without
the use of pressure, so that o /0 =0, By using
movable pistons for the ends of the cylinder, as
shown in Fig. 7.2, all the axial load caused by the
pressure acting on the pistons is borne by the rod
connecting the pistons. This will reduce the axial
stress in the cylinder wall to zero (assuming no
friction between piston and cylinder) without af-
fecting the tangential stress. In actual practice,
one piston can be reploced by a fixed end. By
using @ piston of larger net area than the opposite
cylinder end, compressive axial stresses can be
introduced into the cylinder walls without altering
the tangential stresses. This arrongement is shown
in Fig. 7.3. By using thin-walled cylinder approxi-
mations, the siress ratio is given by the equation

vy 24 {d + 1)

Ya a’2 -~ D2

UNCLASSIFIED
ORNL-LR-DWG 2849

 

| ——— ——PISTON

 

SNONCNY
/

 

 

___—SPECIMEN

 

 

o= PISTON ROD

 

 

 

 

 

NN

 

e PISTON

—

 

 

 

_Z\

Fig. 7.2. Apparatus for Producing Tangentia!
Stress in o Tubular Specimen,

 

114

(Note that the diameter a of the piston rod does not
enter into the calculation of the stress ratio.)

In order to obtain a positive stress ratio of 0 /o,
an arrangement similar to Fig. 7.4a can be used,
The stress ratio is given by the equation

In the event that it is necessary to extend the piston
rod through the top end of the cylinder {to provide o
thermocouple well, for example), the arrangement
shown in Fig. 7.45 may be considered. The stress
ratio is given by the equation

Uz 24 (d + t)

Ua D2 __.bz

UNCLASSIFIED
ORNL-LR-DWG 2850

 

   
 
   

 

 

 

 

/A - \\B e PISTON

/ AN = ) ———PISTON ROD
A I

/ ———SPEGIMEN

 

 

  

 

YN\ \\R

8

 

 

 

 

 

 

 

 

Fig. 7.3. Apparetus for Producing Compressive
Axial Stresses in a Tubular Specimen Without
Altering the Tangential Stresses.
PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
ORNL-LR-DWG 285t

DIMENSION O MAY BE GREATER
OR LESS THAN ¢

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

s
X,
7 7
//. 7 /)
i e
7
Y
| e f
A
a b

Fig. 7.4. Apparatus for Obtaining a Positive Stress Rafio of a,/oru in o Tubular Specimen. (o) With-
out thermocouple well. (&) With thermocouple well.

115
ANP QUARTERLY PROGRESS REPORT

As before, the stress ratio is independent of the
diameter 2 of the piston rod. Since the piston rod
is in compression, however, it should be made
large enough to prevent buckling. The bleed lines
shown in Figs. 7.4a and & should be kept open
during testing to assure that no pressure leaks
from the large-diameter chamber to the smaller one.

A test rig has been built that imposes the con-
dition 7 /o Strain measurements made with a
strain-gage bridge fastened to the gage area showed
that the friction in the system was negligible and
that the only active stress was tangential. A speci-
men is now being tested at 1500°F under this stress
condition,

= oo,

There has been considerable speculation as to
the possibility of o difference in the rate of cor-
rosive attack of fused salts in contact with Inconel
under tensile siresses as compared with lnconel in
compression, In order to observe any difference, it
is desirable that the compressive and tensile
stresses be imposed on the same spscimen and be
of the same magnitude. One approach to this prob-
lem is to test a specimen under pure bending con-
ditions. This would ensure that the magnitude of
the maximum tensile stress would always be equal
to that of the maximum compressive stress during a
given test. An apparatus which would produce this
effect was also designed by Jordan, and a specimen
is now being tested in a fused salt medium. A
drawing of the specimen and the loading opparatus
is presented in Fig. 7.5, If it is assumed that no
friction occurs between pins C and D and the
specimen, all the force P will be transmitted across
section |-l by the link L and no axial force will
exist in the specimen. Likewise, since link L is
loaded axially, it cannot transmit any bending
moment and therefore the specimen must resist any
bending moment existing across section |-l. The
force in link L will be equal to P to satisfy the
equilibrium of forces. The bending moment in the
specimen will then be equal to the product of P and
the horizontal distance between the centers of pins
A and B. In order to minimize variations inthe
bending moment (and therefore stresses) in the
specimen during a test, the horizontal distance
between A and B should be held as nearly constant
as possible. Since the specimen deformsunder load,
the ends will rotate and some change in the distance
AB must occur, In order to minimize the change,
the pins A and B are placed with their centers
slightly above and below, respectively, ahorizontal

116

UNCLASSIFIED
ORNL—LR—DWG 28582

    
   

VTR

T

    

 

 

 

 

SPECIMEN

 

Fig. 7.5. Apparatus for Testing a Sheet Speci-
men Under Pure Bending Conditions.

diameter, This will allow some rotation to occur
with very little change in the horizontal distance,

The apparatus shown in Fig. 7.5 has been con-
structed. Strain gages were placed on the tension
and compression sides of a specimen, and load was
applied. In the range of loads contemplated in a
creep test, the tensile and compressive strains
were essentially of the same magnitude; thus very
little friction existed between the pins and the
specimen,

HIGH-COMDUCTIVITY-.FIN
SODIUM-TO-AIR RADPIATOR

Developmental work hos continued on o sodium-
to-air radiator with fins of a high-thermal-conductivity
material, The developmental effort includes in-
vestigations of materials with high therma! con-
ductivity, the development and testing of brazing
alloys for use in fabricating the radiator, and the
fabrication of radiators for service testing.

Investigations of Fin Materials

J. H. Coobs
Metallurgy Division

H. Inouye

Stress-rupture and creep tests of type-310 stain-
less-steel-clad copper fins that were 8 mils thick
were made at 1500°F at stresses between 500 and
2000 psi. The material was obtained from the
General Plate Div. of Metals & Controls Corp. and
had an average cladding thickness of 1.87 mils per
side. The maximum cladding thickness was 2.3
mils per side. The results of the several tests
made show that some alteration of the copper oc-
curs even though the specimen does not fracture.
This was indicated by the brittleness of the normally
ductile composite. For long-time exposures (1000
hours}, stresses greater than 500 psi and less than
1000 psi are tolerable and there is no indication of
brittleness in the core or oxidafion due to cladding
failure. In all the specimens which ruptured, oxi-
dation of the copper at the point of fracture was
complete, and the copper which did not appear to
be oxidized was brittle. It was also noted that
when the strain in 2.5 in. was 10%, the copper re-
mained ductile, while strains above 20% (regardless
of exposure time} produced brittleness. The data
obtained are given in Table 7.1, -

The aluminum bronzes are among the materials
being considered as cladding for copper fins. Al-
though extensive diffusion occurs between these
alloys and copper, increases in the thermal con-
ductivity of the fin can be realized by cladding the
copper with these alloys. The composition gradient
existing in a diffused composite results in an in-
crease in the conductivity of the cladding material
(because of the outward diffusion of aluminum) and
a decrease in the conductivity of the copper core.

Thermal conductivity measurements of two alumi-
num bronzes were made by the Solid State Division
and are reported in Table 7.2.  The values given
for measurements at temperatures up to 392°F com-
pare favorably with the data of Smith and Polmer;z
but for temperatures above 392°F they exceed the
values extrapolated from their data. Also, the

 

2¢, S, Smith and E. W, Palmer, Am, Inst. Mining Met,
fing%rg), Metals Technol. Tech. Pub. No. 648, 1-21

PERIOD ENDING SEPTEMBER 10, 1954

TABLE 7.1. 5TRESS.RUPTURE PROPERTIES OF
TYPE-310 STAINLESS-STEEL-CLAD
COPPER FINS AT 1500°F

2-mil cladding on each side of 4-mil copper sheet

 

Rupture Elongation

 

Stress

(p5|) Tlme in 2.5 in. Remqus
(hr) (%)

2000 96 39 Copper embrittled

1800 336 50
1650 528 40

Copper embrittlied

Copper embrittled

1500 887 Copper embrittied

1400 1220 In progress

1000 27.5 Test terminated at 1000
hr; specimen oxidized
throughout

1000 .20 Test terminated at 500 hr;
oxide stringers on
surface of specimen

500 10 Test terminated at IBOO
hi; no indicatjon of
failure

500 5 Test terminated at 500: hr;

no indication of failure

 

TABLE 7.2, THERMAL CONDUCTIVITY OF
ALUMINUM BRONZES AT VARIOUS TEMPERATURES

 

Therma! Conductivity
Temperoture [C"'/SBC‘sz (DC/Cm)]

(°F) For For
6.2% Al.-93.8% Cu 8.4% A1-91.6% Cu

 

212 0.178

0.176

302 0.240 0.210

392 0.280 0.250
1156 0.55
1562 0.77

 

estimated value of six times the thermal conductivity
of stainless steel ot 1500°F was exceaded by the
experimental value for the 6.2% Al-93.8% Cu alloy.
The values given in Table 7.2 were not corrected
for the volume expansion of the alloy with temper-
ature and are therefore accurate only to within

+5%.

117
ANP QUARTERLY PROGRESS REPORT

Development of Brazing Alloys for
Use in Fabricating Rediators

P. Patriarca
K. W. Reber G. M. Slaughter
Metallurgy Division

J. M, Cisar

Aircraft Reactor Engineering Division

Several alloys were investigated for brazing
copper fins clad with Inconel, type 310 stainless
steel, or type 430 stainless steel in order to es-
tablish an optimum combination for fabrication of a
sodium-to-air radiator, The evaluation of these
alloys was bosed on metallographic examination of
tube-to-fin specimens before and after exposure to
static air for periods up to 600 hr, The alloys
tested are listed in Table 7.3.

TABLE 7.3, BRAZING ALLOYS FOR

 

 

 

. o Brazing
Alloy Name ~-omposition Temperature
(et %) (°F)

Low-melting-point 80 Ni, 6 Fe, 5 Cr, 1920

Nicrobraz (LMNB) 558i,3B,1C
Coast Metals alloy 89 Ni, 55i, 4 B, 1840

52 ?2 Fe
Electroless nickel 88 Ni, 12 P 1800
Ni-P-Cr

80 Ni, 10 P, 10 Cr 1800

A microsectionis shown in Fig. 7.6 of an Inconel-
clad copper fin (2 mils of Incone! on each side of
6-mil-thick copper) joined to gflé-in.-OD, 0.025-in.-
wall Inconel! tubing with 88% Ni~12% P alloy and
then exposed to staftic air at 1500°F for 600 hr.
Solution of the tube wall during the brazing cycle
was negligible, but rather severe solution of the
fin lip occurred. It is evident, however, that the
oxidation resistonce of the joint was not affected.
Examinations of specimens brazed with the other
alloys listed in Table 7.3 revealed similar results,
that is, minimum solution of tube walls and ex-
cellent oxidation resistance of the braze joints.

The effects of diffusion, however, were in evi-
dence after each test, Color changes in the copper
core, along with the formation of voids, indicate
that the use of Inconel-clad copper in a radiotor
would not be advisable, since the void formation

118

ond the change in core chemistry from copper to a
copper-nickel alloy would seriously reduce the heat
transfer efficiency of the fin. Other claddings, such
as types 310 and 430 stainless steel, have been
found to be relatively immune to diffusion effects,
however, and a number of brazed joints were evalu-
ated to determine the optimum alloy for brazing each
of these materials to Inconel tubing.

A joint of type-310 stainless-steel-clad copper
brazed to Inconel tubing with the 88% Ni-12% P
alloy is shown in Fig. 7.7. It may be seen that fin
dilution and intergranular penetration of the resulting
brazing alloy into the tube wall occurred. This
penetration, undetected in the case of Inconel-clad
copper, appears to be due to the formation of o
complex eutectic resulting from the presence of an
iron-base cladding material rather than from «a
nickel-base material. Similar intergranular pene-
tration was found when the 88% Ni-12% P ailoy

A
, UNCLASSIFIED
Y-12604

 

 

0.01

 

 

 

|0.04

 

0.05

 

 

 

0.c8

 

 

 

Fig. 7.6,
Incone! Tubing with 88% Ni.12% P Alloy ond

Incone!-Clod Copper Fins Brazed to

Exposed te Stotic Air at 1500°F for 600 hr. Note
dilution of fin, hole formation in copper due to
diffusion, and the excellent oxidation resistance
of the bronze joint. As polished. 50X. Reduced
15%.
was used with type-430 stainless-steel-clad copper,
os shown in Fig. 7.8, In this case, the intergranular
penetration was completely through the tube wall.
The etfect of the intergranular penetration on the
physical properties of the tubing has not yet been
determined, but it is believed that it would be ad-
visable to use a brazing alloy that does not react
with the iron-base cladding material. |

The tube-to-fin joints shown in Figs. 7.9 and
7.10 are, respectively, type-310 stainless-steel-clad
and type-430 stainless-steel-clad copper fins that
were brozed with Coast Metals alloy 52 to Inconel
and then exposed to static air for 400 hr at 1500°F,
A negligible amount of dilution and o relatively
minor amount of boron diffusion into the Inconel
tubing occurred. "

Although low-melting-point Nicrobmz (LMNB) has
been used successfully to braze all the proposed
cladding materials to Inconel, its relatively high

. UNCLASSIFIED ~
: S D.02

 

 

0.04

0.05

 

0.06

 

 

 

0.09

 

 

 

 

 

Fig. 7.7. Type-310 Stainless-Steel-Clad Copper
Fins Brozed to Inconel Tubing with 88% Ni~12% P
Alloy. Note dilution of fin and intergranular pene-
tration into tube wall. As polished. 50X. Re-
duced 15%. :

PERIOD ENDING SEPTEMBER 10, 1954

flow point of 1925°F is so close to the melting
point of copper that it cannot be used with ease,
Since the oxidation resistance of the Coast Metals
alloy 52 is comparable to that of LMNB and the
effect of boron diffusion is the same for both alloys,
the Coast Metals alloy was selected for the fabri-
cation of the first high-conductivity-fin sodium-to-air
radiator. ' ‘

The problem of edge protection of the clad copper
fins has been resolved. Aluminum is opplied to the
edges of fins by dip coating or spraying, or it is
painted on as aluminum powder in a Nicrobraz
cement slurry. The fin is then heated ot 800°C for
approximately 3/2 hr. The diffusion of the aluminum
into the copper core results in an aluminum bronze
which exhibits excellent resistance to oxidation
when exposed to static air for 500 hr at 1500°F,
It appears that the depth of bronzing, which con-
tinues during oxidation testing, is about 3/32 to 1/8
in. in 500 hr. The feasibility of simultaneously
bronzing large numbers of fins is to be determined.

 

NCLASSIFIED &
¥ 12602

T

L 0.01

 

 

 

0.04

 

 

kL 0.05

 

f (o

 

g.08

 

 

 

 

Fig. 7.8. Type-430 Stainless.Steel-Clad Copper
Fins Brazed to inconel Tubing with 88% Ni-12% P
Alloy. Note dilution of the fin and severe inter-
granular penetration into the tube wall. As pol-

ished. 50X. Reduced 19%.

119
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIED
Y-12601

 

 

 

0.04

 

0.05

 

0.06

 

 

 

 

 

 

Fig. 7.9. Type-310 Stainless-Steel-Clod Copper
Brazed to Inzone! Tubing with Coast Metals Alloy
52 and Exposed to Static Air at 1500°F for 400 hr.
As polished. 50X. Reduced 16%.

Radiator Fabrication

P. Patriarca
K. W. Reber G. M. Slaughter
Metallurgy Division

J. M. Cisar

Aircraft Reactor Engineering Division

A sodium-to-air radiator with 6 in. of type-430
stainless-steel-clad copper fins wos fobricated by
using a combination heliarc welding ond brazing
procedure. The tube-to-fin section was assembled
and brazed with Coast Metals alloy 52 at 1020°C,
The split headers were then assembled and heliarc
welded by wusing the semiautomatic equipment
available in the Welding Labeoratory. A photograph
of the radiator and welding torch after the welding
of the 108 tube-to-header joints is shown in Fig.
7.11.

The remaining header sections were then monually
heliorc welded by utilizing the complete penetration
technique. Although o pressure test indicated that
six tube-to-header joints were not leak-tight, a

120

 

0.

 

.02

 

)4

0.03

 

©.04

 

£ 10.05

 

 

 

0.08

 

 

 

 

Fig. 7.10. Type-430 Stainlass-Steel-Clad Copper
Brazed with Coast Metals Alloy 52 and Exposed
to Static Air ot 1500°F for 400 hr. As polished.
o0X. Reduced 17.5%.

bock-brazing operation satisfactorily sealed the
leaks ondremoved the effects of notches as sources
of stress concentrations. The completed unit, as
shown in Fig. 7.12, was helium leak-tight ofter the
rebrazing operotion. This unit is to be used in the
100-kw gos-fired liquid-metal-heating system.

NOZZLES FOR THE GAS-FIRED
LIQUID-METAL-HEATER SYSTEM

P. Patriarca
K. W. Reber G. M. Slaughter
Metallurgy Division

J. M. Cisar
Aircraft Reactor Engineering Division

One of the design features of the 100-kw gas-fired,
liquid-metal-heater system now heing constructed is
the packed-rod nozzle assembly to be used for inlet
cantrol of the gir and the gos flow. Test data and
theoretical considerations indicated that ]/B-in.-dia
stainless stee!l triangulorly packed rods would be
satisfactory for use in this nozzle, |t was expected
that these rods could be joined in o rigid bundle
UNCLABSIFIED
Y-12916

Fig. 7.11. Sodium-to-Air Radiator Showing Semi-
automatic Heliore Tube-to-Header Welds.

with @ minimum of plugged interstices by furnace
brazing with a ductile oxidation-resistant high-
temperature brazing alloy such as 82% Au-18% Ni.

Experiments revealed that proper jigging was
extremely important in obtaining the completed air-
nozzle blank (Fig. 7.13). Each layer of rods was
tack-welded for rigidity, ond the layers were then
tack-welded to form the assembly. Since the de-
termination of the exact amount of brazing alloy
required was difficult, if not impossible, sufficient
material was used to ensure good bonding at all
capillary joints. The excess material was then
removed from the nozzle interstices by placing a
flat cross section against a metal sheet on which

 

PERIOD ENDING SEPTEMBER 10, 1954

B UNCLASSIFIED
¥-13019

Fig. 7.12. Completed Sodium-to-Air Radiator
Showing Tube-to-Fin Manifold Construction.

many fine lines had been scribed. Upon rebrazing,
these fine scratches acted as capillaries and the
excess material was removed.

The design of the natural-gas nozzle specified
that a regular pattern of interstices should be
selectively closed. The closures were obtained by
inserfing wires into the designated holes ond brazing
them to the rods.

SPECIAL MATERIALS
FABRICATION RESEARCH

J. H. Coobs H. lnouye
Metallurgy Division

Stainiess-Steel-Clad Molybdenum ond Columbium

Two thermal-convection loops constructed of
stainless-steel-clad molybdenum were assembled by

121
ANP QUARTERLY PROGRESS REPORT

UNCLASS|FIED
Y-12934
o 1

 
 
     

Brazad Air-Mozzle Bilank.
bonding at the points of tangency of the stainless
steel rods.

Fig. 7.13. Note

using threadad molybdenum joints, which were then
protected from oxidation by the use of welded cover
plotes. Both loops failed early in corrosion tests
because of fraciure of the cover plates and leakage

frem the |oints.

It is believed that overheating of the cover plates
occurred while the loops were being initially heated
by alectrical resistance., Since no metallurgical
bond existed between the molybdenum and the
cladding, separation became more pronounced with
increased temperatyre, until finally all the current
was being carried by the rather thin cover plates,

Future molybdenum thermal-convection loops will
be made from arc-melted molybdenum, which is
weldoble, and the heavy nickel electrodes required
for resistance heating will be welded directly to
the molybdenum tube instead of to the cladding.

Columbium clad with type 310 stainless stee! has
been fabricated into a thermal-convection loop by
heliarc welding in air. This loop will be tested as
soon as the accessory parts are available.,

122

Inconel-Type Allcys

Eight thermal-convection loops of Inconel-type
alloys were fabricated. The compositions and room
temperature properties of the alloys are given in
Table 7.4, The alloys are being studied as a part
of the effort to obtain alloys with better high-
temperature corrosion resistance than that of Inconel.

TABLE 7.4, PROPERTIES OF HIGH-PURITY
INCONEL-TYPE ALLOYS”®

 

 

¥ Nominal Tensile Yield - _
Neof Composition Strength Point or(u(c;c)mon

e (wt %) {psi) (psi) °

14 18 Cr, 7 Fe, 75 Ni 85,000 41,500 49

15 10 Cr, 15 Fe, 75 Ni 77,700 38,800 44

16 5 Cr, 20 Fe, 75 Ni 72,700 34,200 47

17 10 Cr, 7 Fe, 83 Ni 76,000 36,000 44

18 5 Cr, 10 Fe, 75 Nj,
10 Mo

*Data obtained from the Superior Tube Company.

Nickel-Molybdenum-Base Alloys

The nickel-molybdenum-base alloys are also being
studied as part of the effort to obtain alloys with
better high-temperature strength and corrosion re-
sistance than are offered by Inconel. These re-
guirements are met by Hastelloy B; however, some
difficulties have been encountered in attempts to
fabricate this material. In castings and in fusion
welds, o whole range of composition results because
of severe coring during freezing, and thus the
properties of the fobricated specimen vary from one
area to another. Even long-time high-temperature
aging treatments of such material will not produce
an equilibrium structure. The aging treatment also
results in o great increase in hardness with a
cortesponding decrease inductility. At temperatures
between 1200 and 1800°F, the wrought material
exhibits its poorest ductility because of hot short-
ness, which can arise from trace impurities or the
precipitation of an age-hardening constituent, which
is probably the case for Hastelloy B. Thus the
temperature range of poorest ductility of Hastelloy
B coincides with the temperature range of con-
templated use in high-temperature circulating-fuel
Hastelloy B is not recommended for use
in air at temperatures above 1400°F because of

reqactors,
excessive oxidation. A potential source of trouble
exists because of the molybdenum content; molybde-
num oxide has been reported to cause catastrophic
oxidation of surrounding structures, especially the
steels.

The solutions to these problems are being ap-
proached in two major ways. First, attempts are
being made to find suitable melting and heating
treatments of Hastelloy B that will reduce the
degree of age hardening so that both the cast and
wrought materials will possess acceptable properties
in the temperature range of interest. Second,
Hastelloy-type alloys are being siudied. Both
phases of the investigation will involve attempts to
extrude seamless tubing, since, at present, Hastelloy
B tubing must be made by welding strips. Aging
studies of the following materials are now in prog-
ress at temperatures between 1200 and 1700°F:
commercial, wrought, Haostelloy B plate; vacuum-
melted Hastelloy B, as cast; cast and wrought 20%
Mo—80% Ni alloy; wrought 24% Mo—76% Ni alloy.

Some of the properties of the 20% Mo--80% Ni
alloy already obtoined indicate that this alloy will
be easier to fabricate than Hastelloy B from an
extrusion standpoint; sound tubing could be made,
whereas none was obtained from Hastelloy B,

The recrystallization temperature of a 60% cold-
worked sheet of the 20% Mo-80% Ni alioy was ap-
proximately 1900°F for an annealing time of 30
min.  However, this alloy oxidizes in air at 1500°F
at a faster rate than does Hastelloy B, although
under a temperature cycle between room temperature
and 1500°F the oxidation rates are similar. The
scale formed spalis near rocom temperature. A
sheet, 0.065 in. thick, could be heliarc welded
without producing porosity, but an argon arc pro-
duced porosity. In strength tests at 1500°F, the
20% Mo~—80% Ni alloy had a rupture life of 90 hr ot
8000 psi; when tested at 5000 psi for 600 hr, it had
a 3% elongation; during a short-time test the alloy
had atensile strength of 40,000 psi and an elongation
of 10% in a 2-in. gage length.

Oxidation and oxidation-protection studies are in
nrogress to find suitable methods of protecting the
nickel-molybdenum-base alloys during tests in ther-
mal- convection loops and creep-testing apparatus.
This work will be the ground work for further studies
of the protection of reactor components, The coat-
ings being studied are those that can be applied by
methods readily available, that is, chromium electro-
plate, chromium plus nickel electroplate, nickel

PERIOD ENDING SEPTEMBER 10, 1954

plus aluminum spray, and aluminum spray. In
constant-temperature tests, g chromium electroplate
was found to be beneficial. In oxidation tests in-
volving «cycling between room temperature and
1500°F, these coatings eliminate the spalling
characteristics of Hastelloy B. The chromium elec-
troplate and the chromium plus nickel electroplate
afford the best protection.

Duplex Tubing

Attempts have been made to fobricate duplex
seamless tubing that will have good corrosion
resistance on the inner surface ond oxidation
resistance on the outer surface. In the first ex-
periments, attempts were made to deep draw duplex
blanks 4 in. in diometer. The blanks were made by
hot pressing and also by hot rolling. The following
materials were combined by hot pressing 0.025-in,-
thick sheets under an argon atmosphere with Al,0,-
coated graphite dies at 1500 psi.

Pressing
Temperature

(°Q

Copper plus type 310 stainless 1000
steel

Inconel plus type 310 stainless - 1200
steel

Inconel plus Hastelloy B 1200

Hastelloy B plus type 310 stajnless 1200
steel

Copper was also combined with type 310 stainless
steel by hot rolling at 1000°C, For this, a three-ply
composite with copper in the center and cover plates
of the stainless steel was used. One of the cover
plates was oxidized ot 1100°C for 2 hr to prevent
bonding during rolling. |

included in these experiments were some tests
involving the fabrication of composites of molybde-
num and Inconel. The purposes of these tests were
to determine the conditions necessary to obtain a
metallurgical bond by hot rolling, to determine the
extent of directional properties due to the molybde-
num, and to determine whether such a combination
could be deep drawn. The first rolling experiments
were made on an evacuated copsule at 1225°C with
reductions of 10% per pass. The total reduction was
79% in thickness. No bonding was obtained by
using nickel os an intermediate layer between the

123
ANP QUARTERLY PROGRESS REPORT

Inconel and the molybdenum. In o second series of
experimenis the rolling temperature was reduced to
1000°C, with reductions of 40% per pass and a
total reduction of 85% in thickness., The lower
rolling temperature reduced the tendency for inter-
metallic compounds to form af the interface. Bonding
was achieved by using nickel as the intermediate
layer, and evidences of intermetallic compounds
were found only by metallographic examination af
high magnifications. The final thickness of the
composite was 0.010 in. of molybdenum on 0.040 in.
of Inconel.
bination, the maximum ratio of molybdenum to Incone!
without the directional properties of wrought
molybdenum will be determined.

All the composites mentioned above were suc-
cessfully reduced from 4-in. blanks to 2‘/2~in. cups
by the ‘“Guerin’’ forming process (male parts of die
are hard rubber) at room temperature, Attempts to
deep draw the composites are to be made.

In future experiments with this com-

Sigma-Phase Alloys

For the continuing study of sigma-phase alloy
corrosion by liquid lead (cf. Sec. 6, ‘“Corrosion
Research’’), some additional tubing was made., Two
ingots of a 48.2% Cr—51.8% Fe alloy and two ingots
of o 45% Fe~40% r-15% Ni dlloy were vacuum
cast into 1]/2-in.-did ingots 6 in. long. These ingots
were then hot rolled in an air atmosphere at 1250°C
to rod 5/B in. in diameter. After onnealing at 1100°C
for 2 hr and quenching, the rod was readily machin-
able info tubing.

The transformation from the ferrite to the sigma
phose will be accomplished by aging at 1425°F,
Two of the tubes (Fe-Cr) were cold swaged 12% so
that the effect of cold swaging on the distribution
of the sigma phase could be determined.

Boron Carbide Shielding

The experimental fabrication of !fs-in.-fhick shield
pieces by warm pressing boron carbide bonded with
copper or silver has been discontinued in favor of
the fabrication of pieces molded with nonmetallic
bonding material by cold pressing followed by
sintering to develop the desired properties. Non-
metallic bonding materials that would be suitable
from nuclear considerations and would be stoble af
the expected service temperature include sodium
silicate, silica, silicon nitride, and boric oxide.
Arrangements are being made to secure samples of
these materials for compatibility tests with Inconel
under operating conditions.

124

The density of the proposed shield pieces is to
be about 1.8 to 1.9 g/cm? of B ,C, equivalent to 1.4
to 1.5 g/cm?® of boron. They would be molded in
the form of equilateral triangles or diamond shapes,
2 to 3 in. on a side, so as to cover a sphere with
an integral number of pieces,

Additional tests of the compatibility of Incone!
with pure boron carbide are being made. Previous
tests indicated the formation of a diffusion layer
5 mils thick in a 100-hr test ot the operating temper-
ature of 1500°F,

Tubular Fuel Elements

Twelve tubular fuel elements are being prepared
for further drawing experiments at Superior Tube
Company. They are being assembled with high-
fired UQ, in.the 30- to 55-p particle size range, as
well as with very fine UO,, and with prealloyed
and elemental stainless steel and iron core matrixes.

Plans are also being made to try hot swaging of
the tubes in an attempt to reduce stringer formation
in the cores by reducing the amount of cold working.
Hot swaging on a mandrel moy prove to be suc-
cessful if bonding to the mandre! can be prevented
by an oxide scole or coating.

METALLOGRAFPHIC EXAMINATION OF A
FLUORIDE-TO-SODIUM HEAT EXCHAMGER

R. J. Gray P. Patriareca

G. M, Slaughter
Metallurgy Division

The fluoride-to-sodium intermediate heat ex-
changer, which in a life test failed after 1680 hr in
cyclic service in the temperature range 1080 to
1500°F, was examined metallographically. All the
tube-to-header joints were manually heliarc welded,
and tests showed the heat exchanger to be helium
leak-tight before assembly into the test rig.

Visual examination indicated the presence of 19
fissures in the tube-to-header welds in two of the
three adjacent headers at the hot sodium inlet end.
The probable progressive propagation of these
fissures is shown in Figs. 7.14, 7.15, and 7.16,
The fissure shown in Fig. 7.16 probably extends to
the surface of the weld, If so, there was a leak in
the system at this point. It seems likely that dif-
ferential thermal exponsion between the tubes and
the casing caused stress concentrations at the
roots of the tube-to-header welds.
concentrations

These stress
would tend to propagate cracks
through the welds in the course of thermal cycling,
PERIOD ENDING SEPTEMBER 10, 1954

UHGLASSIFIED
E  y.raoe

 

 

 

 

 

 

 

 

Fig. 7.15. Téube-io-Hender Joint witha $light Crack Extending into;the Weld, Etched. 100X, Reduced 16%.

125
ANP QUARTERLY PROGRESS REPORT

 

UNCLASSIFIED
Y-13004

Fig. 7.16. Tube-to-Header Joint Showing Severe Cracking. Etched. 100X.

particularly since the columnar dendrites, which
are typical of a weld structure, are aligned in such
a way as to aid parallel fractures,

This investigation emphasizes the extreme de-
sirability of using back-brazing as a means for
minimizing the notch effect and for reducing the
possibilities of leaks developing in the system,

126

Corrosion tests in static sodium and in fluoride
mixtures indicate that a 67% Ni-13% Ge-11% Cr—-6%
Si~2% Fe-1% Mn alloy may be useful for this ap-
plication. This alloy, which flows well at 2050°F,
was found to be virtually unattacked in the fluoride
fuel and corroded o a maximum of 4 mils in 100 hr

at 1500°F in sodium.
PERIOD ENDING SEPTEMBER 10, 1954

8. HEAT TRANSFER AND PHYSICAL PROPERTIES

H. F. Poppendiek

Reactor Experimental Engineering Division

The enthalpies and heat capacities of NoF-ZrF -
UF, (65-15-20 mole %) were determined; the heat
capacity in the solid state over the temperature
range 90 to 614°C was found to be 0.17 cal/g-°C,
and the heat capacity in the liquid state over the
temperature range 653 to 924°C was found to be
0.20 cal/g-°C. The enthalpies and heat capacitiés
of LiF-NqF-UF4 (57.6-38.4-4.0 mole %) were also
obtained; the heat capacity in the solid state over
the temperature range 97 to 594°C was found fo
be 0.23 cal/g-°C, and in the liquid state over the
range 655 to 916°C it was found to be (.53
cal/g°C. The thermal conductivities of three
fluoride mixtures in the solid state at normal
temperatures were determined. The conductivity
of NaF~KF—UF4 (46.5-26.0-27.5 mole %) was 0.7
Btu/hr-ft? (°F/#), that for KF-LiF-NaF-UF, (43.5-
44.5-10,9-1.1 mole %) was 2.0 Btu/hr-ft?2 (°F/ft),
and that for LEF-KF-UF4 (48.0-48.0-4.0 mole %)
was 1.4 Btu/hr-f12 (°F/ft). A new electrical con-
ductivity device has been constructed and has
been successfully checked with molten salts of
known conductivity.

Additional forced-convection heat transfer meas-
urements of molten NaF-KF-LiF (11.5-42.0-46.5
mole %) have been made. Measured thermal con-
ductivities and thicknesses of insoluble deposits
on the inner walls of Inconel heat transfer tubes
made it possible to calculate the thermal re-
sistance of the deposit. The calculated value was
in good agreement with values previously deduced
from the heat transfer measurements. A device
for studying the rates of growth of tube-wall de-
posits has been successfully tested with a simple
heat transfer salt. A hydrodynamic flow system
to be used for studying the reflector-moderated
reactor flow structure has been tested. A mathe-
matical study of the temperature structure in can-
verging and diverging channel systems that duct
fluids with volume heat sources has been made.
A study of wall-cooling requirements in circu-
lating-fuel reactors was made.

 

'W. D. Powers and G. C. Blalock, Heat Capuacities
of Compositions No. 3% and 101, ORMNL CF.54-8-135

(te be issued).

PHYSICAL PROPERTIES MEASUREMENTS
Heat Capacity
W. D. Powers

Reactor Experimental Engineering Division

The enthalpies and heat capacities of twe fluo-
ride mixtures have been determined with Bunsen
ice calorimeters:!

NaF-ZrF4—UF4 (65-15-20 mole %)
Solid (90 to 614°C)
fHy - Hpoe = =3 + 07T
C, = 0.17 + 0.01
Liquid (653 to 924°C)
Hy - Hypo = 22 + 0,207
Cp = (.20 * 0.02
LiF-NaF-UF  (57.6-38.4-4.0 mole %)

Solid (97 to 594°C)

Hy — Hyor = 0.22(7)T + 0.0001772

C, = 0.22(7) + 0.000337

b
Liguid (655 1o 916°C)
Hp = Hpo = ~68 + 0.53T
CP = 0.53 + 0.04

In these expressions H is the enthalpy in cal/g,
CP is the heat capacity in cal/g-°C, and T is the
temperature in °C. The experimental enthalpy data
for LiF-NaF-UF, (57.6-38.4-4.0 mole %) are plotted
in Fig. 8.1.

Automatic Simplytrol units have been added to
the calorimeter systems to more accurately control
the furnace temperatures. This modification has
increased the precision of the experimental data.
Currently, enthalpies of a fluoride mixture are
being measured with the copper block calorimeter,

Density ond VYiscosity
S. 1. Cohen

Reactor Experimental Engineering Division

Preliminary density measurements were made on
rubidium metal, and the density was found to be
represented by the equation

plg/em®) = 1.52 — 0.00054 (T — 39°C) ,

127
ANP QUARTERLY PROGRESS REPORT

ORNL - LR-0OW6G 2747

 

450 [
400 |- o : Co
350

|
|
300 | = e e e e s '
|
|

i
Jo——— b
i |
i :

|

 

 

 

 

 

250 | -
200 {
|

ENTHALPY (cai/g)

 

 

 

{OO S e e

 

 

 

 

 

 

 

 

 

 

 

 

400

600 80C

TEMPERATURE ({°G)

Fig. 8.1. Enthalpy vs Temperature for LiF-NaF-UF , (57.6-38.4-4.0 mole %).

where T is the liquid temperature in °C. Measure-
ments were made by the buoyancy principle. A
calculated valve for the liquid density at the
melting point, given in the Liguid Metals Hand-
book,? falls within 3% of the value yielded by this
study.

This experiment was carried out at the request
of the ORSORT group currently studying the use
of rubidium vapor in an aircraft reactor cycle.
Measurements were made on a sample of metal

 

2R. N. Lyon (ed), Ligquid Metals Handbook, p 52,
AEC-Department of the Navy, NAVEXOS P-733 (rev),
June 1952,

128

produced by the Stable Isotope Research and Pro-
duction Division by reduction of a high quality
fluoride which was prepared by the ORSORT group
from RbF furnished by the Materials Chemistry
Division.

In order to obtain more accurate viscosity data
it was decided to initiate a program of refining
viscometry techniques. In particular, the influence
of fluid density on calibrations of both the ro-
tational and the capillary devices is being studied.
Easily hondled fused salts are being used as
calibrating fluids. The present studies indicate
that some of the old calibration curves, which were
based on low-density calibrating liquids, were
shifted by about 30%. This means that some of
the previously reported viscosity measurements are
about 30% high. :

Becaouse of significant improvements made in
fluoride preporation ond handling, the mixtures
which are now being obtained from the Materials
Chemistry Division are exiremely pure, and they
no longer cause the difficulties encountered in the
early stages of the project, such as fouling of
equipment and variations in physical character-
istics. No doubt some of the earlier viscosity
measurements were high because of the impurities
in the materials.

Thermal Conductivity
W. D. Powers

Reactor Experimental Engineering Division
The thermal conductivities of three solid fluoride
mixtures ot normal temperotures were determined
by the transient cooling technique:

Thermal Conductivity
{Btu/hr-#2 (OF /i)

NaF-KE-UF , (46.5-26.0-27.5 0.7
mole %)

KF-LiF NaF-UF , (43.5-44.5- 2.0
10.9-1.1 mole %)

LiF-KF-UF , (48.0-48.0-4.0 1.4
mole %)

The thermal conductivity of solid LiF-KF-UF,
(48.0-48.0-4.0 mole %) was measured by the steady-
state flat-plate technique; a value of 1.5 Btu/hr.ft?
(°F/ft) was obtained. The ratios of corresponding
liquid to solid thermal conductivity measurements
for the fluoride mixtures studied to date are near
unity.

Electrical Conductivity
N. D. Greene

Reactor Experimental Engineering Division

The experimental current-potential conductivity
cell has been successfully tested and stand-
ardized. The electrical conductivity of a KNO,
melt was determined to within a few per cent of
the values reported in the literature.

As predicted, the effects of polarization within
this cell were greatly reduced by the use of the
alternate method.® This reduction in the effect

PERIOD ENDING SEPTEMBER 10, 1954

of polarization was further substantiated by the
small observable increase of conductivity with
frequency in contrast to the usual large depend-
ence of conductivity upon frequency, as observed
in the low-resistance conductivity cell. The
measurements on KNO. were obtained during a
single run without the necessity of having to
replatinize the electrodes after each determination.
A refinement of this current-potential cell which
will permit measurements of greater accuracy is
now under way.

Preliminary measurements of several fluorides,
the conductivities of which were determined previ-
ously by an alternate method, were found to be in
good agreement within the inherent limitations of
accuracy. A summary of the preliminary measure-
ments of the conductivities of KF-LiF-NaF-UF,
(43.5-44.5-10.9-1.1 mole %), NaF-ZrF  -UF, (50.0-
46.0-4.0 mole %), and NaF—'ZrF4-UF4 (53.5-40.0-
6.5 mole %), as well as the medsurements of
several other salts, has been compiled in o sepa-
rofe reporh""

Vapor Pressure

R. E. Moore
Materials Chemistry Division

Yapor pressure measurements on the mixture
N0F¢ZrF4 (25-75 mole %), which were begun last
quarter,” have now been completed. The method
and apparatus, originally described by Rodebush
and Dixon,® were discussed in previous reports.” 8
The data, given in Table 8.1, can be represented
by the equation '

9368
T(°K)
from which the calculoted pressures and the heat

of vaporization {43 kecal/mole) were obtained. The
approximate liquidus temperature, obtained from

log,q P(mm Hg) = - + 10.57 ,

 

3N. D. Greene, ANP Quar. Prog. Rep. fune 10, 1954,
ORNL-1729, p 100.

N, D. Greene, Measurements of the Electrical Con-
ductivity of Molien Fluorides, ORMNL CF-54-8-64 {to be

issued).

R, E. Moore and C. J. Barton, ANP Ouar. Prog. Rep.
June 10, 1954, ORNL.1729, p 101,

SW. H. Rodebush and A. L. Dixon, Phys., Kev., 26, 851
(1925).

7R. E. Moore and C. J. Barton, ANFP Quar. Prog. Rep.
Sept. 10, 1951, ORNL-1154, p 136,

BR. E. Moore, ANP Quar. Prog. Rep. Dec. 10, 1931,
ORNL-1170, p 126.

129
ANP QUARTERLY PROGRESS REPORT

TABL.E 8.1. YAPOR PRESSURE IN NaF-ZrF4
(25-75 mole %) SYSTEM

 

Temperature Observed Calculated

©C) Pressure Pressure
(mm Hag) (mm Hg)

754 29 28
769 37 38
776 44 44
786 54 54
796 66 65
BDS 74 76
311 83 83
828 115 115
844 152 155

 

the intersection of the vapor pressure curve with
that of pure ZrF ,, is about 760°C.

FUSED-SALT HEAT TRANSFER
H. W. Hoffman

Reactor Experimental Engineering Division

Additional data for NaF-KF-LiF (11.5-42.0-46.5
mole %) flowing for short-time periods in an elec-
trically heated type 316 stainless steel tube have
been analyzed. The results are in agreement with
the general turbulent flow correlations for ordinary
tluids.

The thermal conductivity of K,CrF, was found
to be 0.13 Btu/hr-ft? CF/f)° in the temperature
range 100 to 200°F, Measurements from a photo-
micrograph, Fig. 8.2, of a section of the tube used
in the Incone! experiment show o film thickness
of about 0.4 mil. The thermal resistance calcu-
lated from these results, 0.00025 hr-ft2.°F /Btu, is
in agreement with the measured thermal resistance
obtained from the heat transfer experiments in the
Inconel system.

The forced-convection apparatus,'® which is to
be used for studying the rate of growth of wall
deposits, as well as for heat transfer measure-
ments of fused fluoride mixtures, has been operated
successfully for more than 100 hr with the heat

 

4. W. Hoffman and J. Lones, ANP Quar. Prog. Rep.
June 10, 1954, ORNL-1729, p 101,

1044, W, Hoffman and J. Lones, ANP Quar. Prog. Rep.
Mar, 10, 1954, ORNL.-1692, p 98.

130

 

FILM; 0.4 mil ‘~—~—I

Fig. 8.2.
K3CrF6.

incone! Tube with Film Deposit of

transfer salt NaNOz-NqNfia-KNO3 (40-7-53 wt %)
as the circulated fluid. The system is now being
modified so that resistance-hected, as well as
externally heated, test sections can be inserted
in the system. With resistance-heated sections,
larger axial fluid temperature rises can be at-
tained. It is planned to study next a zirconium-
uranium fluoride salt mixture in lnconel.

TRANSIENT BOILING RESEARCH
M. W. Rosenthal

Reactor Experimental Engineering Division

A study of the heat transfer phenomena which
occur when a water-cocled and -moderated solid-
fuel-element reactor suddenly becomes super-
critical has been undertaken. The time required
for generation of steam in water which is initially
below the boiling point will be determined. The
effects of water temperature, rate of power in-
crease, air concentration, and surface condition
will be investigated.

Construction of the equipment for generation of
power surges by electrical means in aluminum fila-
ments has been completed, and the entire experi-
mental apparatus has been assembled and tested.
Satisfactory operation has been achieved. The
transient-power generator consists of 100 thy-
ratrons connected in parallel. A different bias
can be put on the grid of each thyratron, and the
system can be set so that a signal of increasing
voltage applied simultaneously to all tubes fires
them in a sequence that approximates the desired
A peak current of about 500 amp can be
produced, with the duration of the power surge
ranging from 30 to 500 msec.

function.

The instantanecus values of voltage drop across
the test filament and current through it are re-
corded by a fast-response Hathaway oscillograph.
A TFastax high-speed motion picture camera is
used to photograph the event, and illumination is
supplied by an Edgerton stroboscopic flash unit
synchronized with the camera. A repetition rate
of 6000 frames per second is obtainable with a
flash duration of 2 usec. Closing a single switch
initiates the operation of all the equipment; a time
delay is provided in which the camera can ac-
celerate to full speed before current is passed
through the filament.

FLUID FLOW STUDIES FOR CIRCULATING-
FUEL REACTORS

J. O. Bradfute L. D. Palmer
F. E. Lynch

Reactor Experimental Engineering Divisieon

The flow system for the hydrodynamic studies of
the reflector-moderated reactor core, shown sche-
matically in Fig. 8.3, has been fabricated, as-
sembled, and tested. A detailed sketch of the test
section is also shown in Fig. 8.3. Since it will
be necessary to insert, photograph, and then re-
move a coordinate grid from time to time, the
system was designed so that it could be readily
assembled and disassembled. The cores are po-
sitioned and supported by a brass sleeve, which
was attached to the test section support along its
center line before it was machined. The sleeve
was carefully bored and lopped so that the shafts
protruding downward from the cores would fit with
minimum clearance and yet loosely enough so that
they could be removed by hand. The upper surface
of the test section support flange was machined
perpendicular to the center line of the brass
sleeve. This arrangement permits accurate align-
ment of the core within the Plexiglas test section
without supporting members in or ahead of the test
region. Attempts are currently being made to
obtain some preliminary fluid flow information for
the reflector-moderated reactor core for the straight
entrance condition shown in Fig. 8.3 by photo-
graphing particle flow through the Plexiglas test
section.

The phosphorescent flow-visualization system,
which was described previously,!! has been further
modified. The ZnCd sulfide phosphor has been
replaced by a ZnS (Cu-activated) phosphor whose
luminosity was found to be greater after excitation.
The duct work in the flow circuit has been im-

 

11 . D. Palmer and G. M. Winn, A Feasibility Study
of Flow Visualization Using a Phosphorescent Particle

Method, ORNL CF-54-4-205 (Aprit 30, 1954).

PERIOD ENDING SEPTEMBER 10, 1954

proved so that higher Reynolds numbers are now
attainable, and a diverging lucite channel has
been constructed. Some of the flow features, such
as asymmetry, flow separation, and the effect of
entrance lengths, are to be observed for different
entrance configurations in such channel systems.
The phosphorescent flow-visualization method is
also to be used in the reflector-moderated reactor
core experiment.

HEAT TRANSFER STUDIES FOR CIRCULATING-
FUEL REACTORS

H. F. Poppendiek L. D. Palmer

Reactor Experimental Engineering Division

The classical hydrodynamics study, conducted
by Nikuradse,'? of converging and diverging
channels was reviewed. It is felt that some of
the fundamental flow features observed in con-
verging and diverging channel systems will alsc
exist in the reflector-moderated reactor. The
results obtained by Nikuradse indicate that for
the case of turbulent flow in convergent channels
the flow becomes more stable or less turbulent in
nature and that the eddy ditfusivities are signifi-
cantly lower than in a parallel-plate channel
system. For the case of turbulent, symmetrical
flow in diverging channels the tlow is found to
be more turbulent than in the parallel-plate system.
When the total channel angle of a divergent system
is increased beyond 8 deg, asymmetrical flow re-
sults. For o 10-deg angle, a thin, low-velocity
layer of fluid flowing in a reverse direction to the
main stream is noted near one of the channel
walls. This separation phenomenon becomes more
pronounced at larger channel angles.

Some radial temperature distributions were calcu-
lated in the established-tlow region of a few con-
verging and diverging flow channel systems con-
taining volume heat sources within the ducted
fluide The experimentel velocity and eddy dif-
fusivity data obtoined by Nikuradse were used in
the analyses. The results indicated that wall-fluid
temperature differences can be significantly higher
in converging flow channels than in parallel-plate
systems; the converse is frue for symmetrical
diverging flow channels. Asymmetrical divergent
flow passages are difficult to analyze at this time
because the diffusivity structure is not known.

 

12,, Nikuradse, Forschungsarbeiten YDl 289, 1-49
(1929).

131
el

UNCLASSIFIED
ORNL—LR ~DWG 2749

+
L

MANOMETER

UPPER PIPL FROM WATER MAIN ——=

%

BUILDING WATER MAIN

COLLAR —

 

PARTICLE SLURRY STORAGE

  

ROTAMETER

  

ENTRANCE SECTION——=

 

 

 

 

 

  
  
  
 

 

 

 

 

 

 

~
@é / T T
i A
o : / - N
TEST SECTION — / /\ \
CORE : \ : z-ni i }
!
S
\ 1 /
TEST SECTION \\‘ 7
BRASS SLEEVE y

 

 

TEST SECTION SUPPORT—— =

 

 

 

LOWER PIPE TO DRAIN—==

 

 

TO DRAIN

Fig. 8.3. Flow System for Hydrodynamic Study of Reflector-Moderated Reactor Core.

14Oy SSFYO0Ud ATIFLIVNO NV
However, it appears reasonable to conclude that
high temperatures probably would exist in the thin,
low-velocity layer of fluid in the 10-deg divergent
channel because the flow there is probobly laminar
in nature.

A report is being prepared which will be useful
to designers for predicting wall-cooling require-
ments in circulating-fuel reactors which have pipe,
parallel-plate, or annular fuel-duct geometries.
The temperature solutions, which were presented
previously,”"” have been tabulated in detail so

PERIOD ENDING SEPTEMBER 10, 1954

that the superposition process outlined in those
reports can be quickly effected. Cooling require-
ments and resulting fluid temperature distributions
have been determined for typical circulating-fuel
reactor systems.

 

By F, Poppendiek and L. D. Palmer, Forced Con-
vection Heat Transfer in Pifies with Volume Heat
Sources Within the Fluids, ORNL-1395 (Nov. 5, 1953).

4y F Poppendiek and L. D. Palmer, Forced Con-
vection Heat Transfer Between Parallel Plates and in
Annuli with Volume Heat Sources Within the Fluids,
ORNL-1701 (May 11, 1954).

133
ANP QUARTERLY PROGRESS REPORT

9. RADIATION DAMAGE

J. B. Trice
Solid State Division
A, J. Miller
ANP Project

Additional irradiations of lnconel capsules cone
taining fluoride fuels were carried out in the MTR,
Only one capsule containing a UF j-bearing fuel
has been examined, ond it shows no corrosion,
This is in contrast to the capsules containing
UF ,-bearing fuel in which minor increases in
penetration have at times been observed in the
irradiated specimens as compared with the out-ofs
However, it should be emphasized
that even in the UF ,-bearing capsules, the ap-
pearance of small deviations from the controls
could be due to experimental difficulties.

Inspection of the previously described LITR in-
pile circulating-fuel loop, which failed prior to the
startup, disclosed that the failure was caused by a
bregk in the weld connecting the pump discharge
nipple to the fuel line, Design revisions and
refabrication of some components are in progress.

pile controls.

Developmental work continued on a smaller loop
for operation in an LITR A-piece. Work on a loop
for insertion in the MTR is described in Sec. 3,
‘‘Experimental Reactor Engineering.”

Development of equipment continued for in-pile
stress-corrosion tests on Inconel in contact with
fluoride mixtures, and testing of the MTR tensile-
creep rig neared completion. Detailed examination
was completed of an Inconel loop in which sodium
was circulated at high temperature in the ORNL
Graphite Reactor, and no evidence of radiation-
induced corrosion was found in the Inconel,

Metallographic examinations were carried out in
the hot cells on irradiated fuel plates for the
Pratt & Whitney Aircraft Division, and studies were
made on annealing.out of fission-fragment damage.
Work continued on examination of wire and multiple-
plate type units for GE-ANP,

MTR STATIC CORROSION TESTS

W. E. Browning G. W. Keilholtz
Solid State Division

A new series of capsules was irradiated in the
MTR to compare NaF-ZirF, fuels containing UF,
with those containing UF,. The property to be
compared is the effect of fission on static cor-

134

rosion of Incone! by the fuels at 1500°F. The
series includes five capsules containing 1,79
mole % UF,, 49.3 mole % ZrF,, and 48.9 mole %
NaF and five capsules containing 1.74 mole %
UF,, 48.2 mole % ZrF,, and 50.1 mole % NaF.
Analyses are being run to confirm the purity of the
UF; with respect to UF, contamination. For
every capsule, there will be an out-of-pile control
capsule. Each fuel generates 1100 w/cm® in the
A-38 position in the MTR,

Six capsules have been irradiated for the speci-
fied two-week period. Five have been received
at ORNL after irradiation; three of them have been
opened and two have been examined - one con-
taining UF ; and one containing UF ,. Bothcapsules
had been irradiated with the metal-liquid interface
at 1500 + 50°F., The UF;-bearing capsule had
experienced one temperature excursion to 1660 +
50°F for less than 30 sec, but the temperature of
the UF ,-bearing capsule never exceeded 1550 +
50°F. The fuel in each capsule was frozen two or
three times during its history. No evidence of
corrosion or film could be seen in the UF j-bearing
Inconel There was subsurface void
formation in the UF ;-bearing Inconel capsule to a
depth of about 2 mils. No intergranular corrosion
or concentration of voids at the grain boundaries
was evident,

Analytical samples of the irradiated fuel were
taken with two sizes of drills. The uranium concen-
trations
identical, within experimental error (+3%), and were
of the wvalue expected after burnup correction.
Analyses for iron, chromium, and nickel gave the
following results: Fe, 0.03 to 0.14 wt %; Ni, 0.01
to 0.05 wt %; Cr, less than 1074 wt %. There were
no significant differences in the final iron, nickel,

capsule,

in the ceniral and outer samples were

and chromium compositions between the two cap-
sules. Analyses of the unirradiated fuels have
not been completed.

It would be unwise to draw conclusions on the
basis of only one pair of samples; however, if
these first results are confirmed in future capsules,
UF, fuels may be considered quite desirable as
far as radiation-induced corrosion is concerned.
Two UF;-bearing capsules and two UF ;-bearing
capsules in this series will be irradiated for six
weeks each to amplify the apparent differences in
the effects of UF; and UF, on Inconel. It is inter-
esting to note that the UF ; capsule in this series is
the first one run in the MTR and the first one run at
800 w/cm? or greater that has not shown a tendency
toward intergranular corrosion. - Whether the differ-
ence between eorlier UF -bearing capsules and
the one in this series is due to improved temper-
ature control or to uranium concentration effects
will be determined by additional irradiations.

FISSION PRODUCT CORROSION STUDY

C. C. Webster G. W. Keilholtz
Solid State Division

Steps are being taken to perform corrosion tests
out-of-pile with irradiated fuel in order to separately
study the effects on Inconel of fission-product
concentration and of Inconel irradiation.

A water-cooled facility for irradiation of solid
fuel was constructed and has been installed in
position C-46 of the LITR., The unperturbed
thermal flux is on the order of 4 x 1073, An Inconel
tube has been constructed for casting solid bars
of fuel 1 in, long and 0.1 in. in diameter. The
solid bars of fuel will be transferred to Inconel
capsules for irradiation. The capsule has been
constructed so that it can be welded shut without
the bars being melted and can be reopened in the
hot cell for transfer of the fuel to the corrosion
test capsule.

Upon completion of an out-of-pile heating cycle,
the test capsule will be opened in the hot cell.
The fuel will be drilled out and divided into three
samples: a portion for petrographic study to detect
any oxidation or reduction, a portion for mass
spectrographic  study to determine burnup, and
a portion for chemical analysis of the fuel. The
capsule will be slit on the remote slitting machine
described below for a metallographic study. At
least two concentric samples can be taken from a
capsule of this size by remote means.

FACILITIES FOR HANDLING IRRADIATED
CAPSULES

C.C. Wébster G. W. Keilholtz
Solid State Division

it was observed metallographically, in some
that when transverse sections

were taken from a single fluoride fuel corrosion

cases, several

PERIOD ENDING SEPTEMBER 10, 1954

capsule the corrosion pattern was not consistent
over the length of the capsule but appeared to be a
function of the longitudinal temperature gradient.
One method of obtaining a longitudinal section
would be to grind away half of the capsule on the
milling machine. However, by such a methed it
would be difficult to ascertain that a section
through the diameter of the capsule had been made.
Also, the milling operation would create a serious
contamination problem,

The method being used consists in passing the
capsule longitudinally by a 10-mil-thick silicon
carbide fine-grit wheel and thus slitting the cap-
sule down the center. The capsule is clamped
into a vise and then fed into the wheel by means of
a mechanical linkage which is to be replaced by a
variable-speed motor and o contact cutoff switch,
When the capsule is put into the gripping adapter
in the vise, it is automatically aligned so that the
wheel will cut in the plane of o diameter through
the capsule, By making different gripping adapters,
various sizes of tube can be slit lengthwise. Thus
far the apparatus has been used on only MTR-type
fluoride fuel capsules (0.100-in.-ID and 0.200-in.~
0OD); it has worked very satisfactorily. :

Since the abrasive wheel has a rubber base, it
must be kept cool. The darea around the cut must
also be kept cool to prevent any high thermal
stresses from affecting the corrosion results. The
wheel and specimen are over a tray and are covered
with a splash shield so that a liquid coclant can
be used. The coolant level is always kept above
the bottom edge of the abrasive wheel so that by
cutting upward into the specimen the coolant is
carried into the cut; therefore the specimen need
not be in the coolant, Carbon tetrachloride is
used as the coolant because it does not react with
the fluoride fuel at high temperatures. It is forced
into the tray from outside ‘the hot cell through
Tygon tubing by air pressure. The tray and splash
shield are attached to the mount for the vise and
they move with the specimen.

Once installed within the hot cell, the remotely
operated tube-slitting machine can be used with
ease in cutting longitudinal sections from various
sizes of tubing. Thicker abrasive wheels can be
used for heavier walled tubing. The advantages of
this method of cutting are that it permits visual
observation of the amount of fuel left on the tube
walls, it permits observation of constrictions
without the necessity of removing or disturbing the

135
ANP QUARTERLY PROGRESS REPORY

material causing the trouble, it gives o well-exposed
surface for chemically removing the last traces of
fuel, it permits a cut to he made through the bottom
plug of the capsule so that an examination can be
made for crevice corrosion, and it leaves nearly
one half of the capsule for further and more com-
plete study, if desired. After the fuel is removed
chemically without attacking the metal wall, the
specimen can be electroplated so that the fuel-
metal
metallographic polishing operation.

The stability of the fivoride fuel under irradiation
and the extent of the corrosion of the container by

interface will not be rounded during the

the fuel are determined, in part, by analyzing
samples of the irradiated fuel for the concentration
of the original components ond for the major con-
stituents of the container material.  The fuel
samples are put into solution within the masier-
slave hot cells and are then transferred to the
Analytical Chemistry Division for analysis. Work
on the necessary transfer shields has been coordi-
nated with the design and construction of the lead
barrier being provided in the analytical chemistry
hot cells to improve transfer conditions and to
reduce personnel exposure.

Two transfer carriers weighing 150 Ib each and
giving 2 in, of lead shielding have been constructed
and are now in use. FEach carrier holds four 30-ml
bottles that are held in place by spring clamps.
A motor-driven decapper for removing the bottle
cap and a remotely controlled pipet thai can be
lowered into the bottle are provided. The bottles
may be inserted into the carrier within the cells
and can be left in the caorrier during all sample
removal operations. The whole unit moy then be
returned to the hot cells for reuse.

Two lead storage units weighing 450 |b each and
giving 3 in. of lead shielding were also built and
are available for use when all the transfer carriers
are in use. The lead storage units have a capacity
of four 50-ml volumetric flasks and can be loaded
within the master-slave cells.

ANALYSIS OF iRRADIATED FLUORIDE
FUELS FOR URANIUM

M. T. Robinson G. W. Keilholtz
Solid State Division

One of the regular steps in in-pile static cor-
rosion testing of fluoride fuels has been the anal-
ysis of the irradiated fuel for uranium. However,
there are several major problems connected with

136

sampling operations, described previously,'s?

which have never been fully resolved. In the
cutting and drilling operations it is possible for
small chips of metal, either from the capsule or
the drill, to get into the fuel sample. To minimize
this possibility, drills used recently have been
A search is also
In the
it
was possible for dirt from the hot cells to inad-
vertently enter the fuel samples during cutting
and drilling. This possibility has been minimized
in the newer equipment.? The transfer operations,
in several of which the salt is exposed to con-
tamination from hot cell dirt, pieces of rubber
from manipulator grips, and the like, are all possible

carefully honed to remove burrs.
being conducted for new drill materials.
early use of the present sampling technique,

sources of contamination. The observation of
opaque material, in some cases in large amounts,
during petrographic examination of irradiated fuel
samples® indicates that at one or the other of the
stages discussed above foreign material may have
been introduced info some of the samples. Such
contaminants make the resulting uranium analyses

u! 1
QV{G

The samples obtained for chemical analysis
have, at times, been very small, often being as
small as 10 mg. They are weighed in tared weighing
bottles on a conventional chain-type analytical
balance, and, since they are so small, the precision
of weighing is about 4% instead of the usual
negligible amount. This lack of precision results
in an obnormally high uncertainty in the uranium
analyses. After being weighed, the samples are
dissolved and the solutions are made up to volume
in the hot cells. The resulting solutions often
contain as little as 500 ppm of uranium, Such
high dilution is undesirable, since the titer may
change appreciobly on fong standing. It seems
reasocnable to conclude that the uncertainty in
uraitium analyses may be as high as +5% or more,
even for the usually precise (+2%) potentiometric
titration, An increase of ten times in sample size
(easily attainable) would restore this technique
to its normal precision.

]J. G. Morgan et al,, Solid State Div. Semiann. Prog,
Rep. Aug. 3, 1953, ORNL.-1606, p 40,

2C. C. Webster and J. G. Morgan, Solid State Div.
Semiann. Prog. Rep. Feb, 28, 1954, ORNL-1677, p 27.

3(3. D. White and M. T. Robinsen, Sclid State Div.
Semiann. Prog, Rep. Feb. 28, 1954, ORNL-1677, p 28.
Similar remarks apply, in part, to the mass spec.
trographic analysis. The samples taken are weighed
by the analysts on a microbalance to alleviate the
effect of small sample size. However, the handling
of the very small crucibles in the hot cells is riot
so delicate an operation as might be desired. The
possibilities of contamination entering the sample
are high.

No highly useful conclusions hove emerged from
much of the analytical work done in the past,
probably because of the difficulties described
above. Recently, however, marked improvement
has been made as a result of continved efforts to
improve sampling techniques, sample sizes, and
handling procedures, ond uranium analyses have
been carried out in a satisfactory manner. Since
the analytical results obtained recently show thot
no measurable segregation of uranium takes place,
the procedure of taking more than one sample from
each capsule will be abandoned. This will allow
very much larger samples to be obtained, will
make the analyses more sure, and will minimize the
effects of foreign additions to the salt.

HIGH-TEMPERATURE CHECK VALVE TESTS

W. R. Willis G. W. Keilholtz
Solid State Division

As part of the development program for a small
in-pile pump loop for operation in a berylium
A-piece in the LITR, an attempt was made to
develop a high-temperature check-valve pump,
Such a pump has the advantage of being o compact,
completely sealed unit.

A test rig was run with Inconel, stainless steel,
and Stellite 25 check balls. None of these materials
gave successful tests. Time of operation before
failure varied from 30 min for the Inconel to 2 hr
for the stainless steel. Except with the stainless
steel balls, operation of the valves was not con-
tinuous; they would occasionally stick in both
open and closed positions. When failure occurred,
it was abrupt, and therefore sticking of the valves
was indicated rather than o gradual stoppage of the
system,

MINIATURE IN-PILE LOOP -~ BENCH TEST

J. G. Morgan G. W. Keilholiz
Solid State Division

An exploded view of the in-pile pump described
previously? is shown in Fig, 9.1, The RPM meter
coil measures the rotation of the shaft on which

PERIOD ENDING SEPTEMBER 10, 1954

is mounted the motor armature. The Graphitar
bearing is mounted in the bearing casing above the
impeller housing. The assembled loop is shown
in Fig. 9.2. The pump contains an extra reservoir
to enable metered volumes of salt to be pumped
into the surge tank for calibrating the venturi. The
furnace for heating the pump and reservoir is also
From performance tests with this setup,
pump speed vs head and pump speed vs flow have
been determined. At 5000 rpm the flow was 4 fps,
corresponding to a Reynolds number of 3000 in o
200-mil tube. At the 540 w/cm” generated in-pile,
the temperature differential is expected to be about

100°F.

shown,

LIFE TESTS OM AN RPM METER, BEARINGS,
AND A SMALL ELECTRIC MOTOR UNDER
IRRADIATION

J. G. Mergan H. E. Robertson
G. W, Keilholtz
Solid State Division

Tests were conducted on an RPM meter, bearings,
and o small electric motor as o part of the de-
velopment of o loop for insertion in an LITR A-
piece. In order to drive the pump, it is desirable
to use a variable-spesed electric motor and to align
the shaft with lubricated bearings. Also, o record
of shaft speed would be desivable, The entire
motor assembly will be just above the upper grid
guide in the LITR and in an estimoted thermal
flux of 5 x 1077 nevtrons/cm?.sec. Hole 53 of the
ORNL Grophite Reactor provided a comparable flux
intensity and was used for these tests,

The in-pile apparatus, with a single lower bearing
as in the second test, is shown in Fig. 9.3. A
motor was mounted in g horizontal position with
upper ond lower armature shaft bearings, In the
first test two lower bearings were used. The motor
and RPM meter were canned and inserted into
the center line of the reactor. The variables
measured were resistance of the motor winding,
resistance of the RPM coil, the voltage applied
to the motor, the current drawn by the motor, the
temperature of the lower Fafnir bearing, the temper-
ature of the motor field, and the speed of the motor
shaft,

The Delco ac-dec motor used is rated at 115 v,
0.9 amp, and }’15 hp and has insulated wires. Some
substitutions of materials were made. In the first

 

4.!. . Morgan, Solid State Div. Semiosnn. Prog. Rep.
Feb., 28, 1954, ORMNL-~1677, p 30,

137
ANP QUARTERLY PROGRESS REPORT

UHMCLASSIFIED
PHOTO 1273%

 

  

g

0

RPM METER COIL

 

@

BEARING CASING
IMPELLER HOUSING

 

 

Fig. 9.1. In-Pile Pump.

Py UHTL ASSIFIED
1 PUOTG 12738

SURGE TANK

   

(2
FURNACE

     

 

 

B W T O W™ W

‘a PUMP AND EXTRA RESERVOGIR

VENTURI

rig., 2.2. Bench Test Loop.

138
ARMATURE -

 

THERMOCOUPLt\\

e

 

T SO PR O N,

oA
N \;\;\\\

\‘\ o \"\\

2 NN

o
DN

  

EARING, FAFNIR OW-8

MOTOR: DELCO, AC-DC, "5 hp, SERVICE NO. 504743t
STATOR TAPE: GLASS

BRUSH HOLDER PLATE: MICA

RPM COIL: NO.'36 AW.G. CEROC, 200-5500 TURNS
BEARING GREASE: RRG-2 (NLGI GRADE 2)

~STATOR

      
 
  

    
   

\\THERMOCOUPLE

PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
ORNL-LR~DWG 20814

BEARING, NEW DEPARTURE 77R6

 

   

)

TLREM COIL INSTRUMENT LEADS

Fig. 9.3. Motor and RPM Meter Test Assembly.

test a brush holder plate of mica backed with a
stainless steel plate was used. In the second test
fired Lavite asbestos covered wire was used from
the brush holder to the field coil. Glass tape was
used around the field coils.

The rest of the construction features were un-
altered; that is, fish-paper-coil supports, plastic
(paper-backed  Bakelite) segment
spacers, and cardboard armature shaft spacers
were used.

The upper bearing (New Departure 77R6) and
lower bearing (Fafnir DW-6) were packed with
California Research Corporation grease, identified
as RRG-2 [NLG1 Grade 21.

The RPM meter consists of a magnet fastened to
the shaft which will rotate inside the gas-tight
motor shell. A coil wound around a C-shaped iron
core is mounted outside the shell. As the shaft
rotates, pulses are generated in the coil and counted
remotely by a frequency meter calibrated in rpm.
The coil was wound with 36-g Ceroc 200 magnet
wire covered with glass tape impregnated with
Glyptol. Summary data on the test conditions are
given in Table 9.1.

The totors were examined, after irradiation, in
the, hot cell, Assistance was given in hot cell
examination by J. Noaks and J. Crudele, Pratt &
Whitney Aircraft Division. A ‘withdrawal cask and
its facilities were loaned by L. E. Stanford, General
Electric Company.

commutator

TABLE 9.1. CONDITIONS OF TEST MOTOR AND
RPM METER ASSEMBLY

 

 

Test No, 1 Test No, 2
Motor load
d-c voltage 60 33
Current, amp 0.45 0.20
Shaft speed, rpm 4500 4500
Fafir bearing 45 45
temperature, OC
Motor field plate 65 40
temperature, °C
Time of operation
At 4500 rpm, hr 100 648
In flux,* hr 84 613

 

*Thermal flux, 5 % 10” neutrons/cmz-sec; ratio of fast

to thermal neutrons, 1; gamma field, 4.4 r/hr.

in the first test, motor load was abnormally great
from the beginning and the field plate temperature
When the motor failed after 100 hr,
the stator temperature rose to 182°C while the
power was kept on. The suspicion that there was
a cocked upper bearing plate was confirmed upon
postirradiation examination (Fig. 9.4). The sensing
coil for the RPM meter functioned well; there was
no resistance change or visual deterioration.

was high.

139
ANP QUARTERLY PROGRESS REPORT

BHCLASSIFED
PHOTC 42903

 

Fig. 9.4. Motor lrradiated in the First Test.

In the second test the motor operated as long as
it would be expected to under normal conditions,
However, postirradiation inspection shewed that
the commutator segments had been loosened as a
result of the deterioration of the plastic commutator
segment spocers under irradiation (Fig. 9.5). The
bearing grease proved to be adequate and was
essentiolly unaffected, as evidenced by the constant
power demand by the motor at constant rpm. The
RPM coil also showed no resistance change or
visible breakdown (Fig. 9.6). The brush wear
appeared to be normal, A high-tempercture motor
is being constructed for testing that will have
glass-insulated-wire, mica, and high-altitude
brushes.

REMOVAL OF XENON FROM FUSED FLUORIDES

M. T. Robinson . W. Keilholtz
Solid State Division
D. E. Guss
United States Air Force
W. A, Brooksbank
Analytical Chemistry Division

A program has been initiated to study removal of
Xe'35 from fluoride fuels by means of helium

140

B uneLassIfIED |
e PHOTO 12302

   
 

Fig. 9.5,

Commutator Segments from Motor Ir-
radiated in the Second Test.

purging. A stainless steel capsule will be con-
structed in a manner thot will allow bubbles of
helium to pass through a filter plate and through
the fissioning fuel in hole HB-3 of the LITR.
Plastic mockups of the capsule are being used to
study the characteristics of the helium flow. A
four-channe! spectrometer has been
constructed by the Instrument Department to monitor
the Xe!3® concentration in the flowing stream of

gamma-ray

helium.

CyF 6 UF, IRRADIATION

W. E. Browning G. W. Keilholiz
Solid State Division

An experiment is being conducted to determine
in a very rough fashion the effect of pile irradiation
on a 20% solution of UF in C;F g This material
is of interest tothe ShieldingSection of the Physics
Division for simulating liquid fuels in dynamic
shielding experiments, and the experiment is being

conducted with the aid of personnel from that
 

Fig. 9.6. Motor lrradiated in Second Test.

division. Three nickel capsules were charged with
the two components to make a 7-g solution in each.
Normal uranium was used. Two capsules were
irradiated in appropriate secondary containers in
a water-cooled hole in the ORNL Graphite Reactor
at a thermal flux of 6 x 10'7 neutrons/cm?.sec for
63 hr. This exposure corresponds to 25 times the
fission density expected in the shielding experi-
ment with enriched uranium, The capsules were
opened in a special gos manifold, and the gas
pressures were determined. The pressure inside
the irradiated capsules was shown to have been
greater than 50 and less than 100 psia. The pres-
sure inthe unirradiated capsule was not perceptibly
different from the vapor pressures of the two
components (about 4 psi). The gas in the capsules
was analyzed by infrared absorption spectrometry
at K-25 and was found to contain CF,, C,F,,
C,F ¢ and other fluorocarbon compounds not
identified. No UF, could be detected in the gas
sample. Pressure measurements at ~80 and —190°C
were consistent with these analyses. The ir-
radiated capsules were full of o nonvolatile (at
25°C) greenish powder. The powder was insoluble
in C,F

160 It will be analyzed for uranium. The

PERIOD ENDING SEPTEMBER 10, 1954

unirradiated capsule contained no residue except
for a faint chalkiness on its inner wall. Additionai
capsules are being irradiated for 1/100 of the
accumulated exposure,

LITR HORIZONTAL-BEAM-HOLE
FL.UORIDE-FUEL LOOP

W. E. Brundage C. Ellis

C. D. Baumann M. T. Morgan
F. M. Blacksher A. S. Olson

R. M. Carroll W. W. Parkinson
J. R. Duckworth 0. Sisman

Solid State Division

A loop for circulating fluoride fuel was inserted
in hole HB-2 of the LITR and filled with fluoride
fuel NaF-ZrF -UF, (62.5-12,5-25 mole %) during a
protracted reactor shutdown. The pump was started
as soon as sufficient fuel had been added, but
circulation of the fluoride mixture became erratic
after 5 to 10 min and finally stopped completely.
Fluoride fumes in the off-gas line from the jacket
around the pump indicated a leak,® and therefore
the loop was removed before the reactor was started.

The loop was carefully disassembled to aveid
loss of uranium, since over 5 kg of enriched fuel
mixture had been used to charge the loop. Recovery
of the fuel from the ‘‘nosepiece’’ at the in-pile
end of the loop had to be carried out in a hot cell
because of neutron activation acquired when the
unfilled loop was exposed to reactor flux before
the fuel was added. The leak was found in-the
weld joining the loop tubing to the discharge nipple
ot the bottom of the pump bowl. Complete breaking
of the weld occurred probably during disassembly
of the loop or during cutting of the pump jacket,
rather than while the pump was in operation,

Before insertion in the reactor the jacket for
enclosing the loop and the pump in an inert atmos-
phere was tested for gas tightness., The loop,
with the pump superstructure replaced by a blank
flange, was vocuum tested at 800 to 1000°F with
a helium leak detector. After the loop was inserted
in the reactor, it was pressure tested at 1200 to
1500°F ot 16 psig, and there was no perceptible
loss in pressure over a l-hr period, even though
the pump superstructure with its rotating shaft seal
was in place.

 

5W. E. Brundage et al,, ANP Quar, Prog. Rep. June
16, 1954, ORNL-1729, p 107.

141
ANP QUARTERLY PROGRESS REPORT

The connections of the loop tubing at the pump
have been redesigned for the second loop to provide
a filling and drain line. This will permit charging
the loop with a non-uranium-bearing flucride mixture,
circulating it at operating temperature, and draining
it before the loop is inserted into the reactor. The
danger of tube rupture becouse of expansion of the
fuel on melting necessitates the draining and
refilling feature, if operating tests are to be carried
out. The necessary changes in the pump shield
to accommodate the drain and fill line are being
made.

The Metallurgy Division will supervise all critical
welding and has provided specifications for an
improved weld joint developed for Incone! tubing.
The components fabricated previously for the
second loop are being changed to conform to the
specifications for the new weld joints. A new
flange for closing the rear of the jacket has been
made in order to give more space for making pipe
connections to the large air lines required for the
heat exchanger.

During installation of the loop in the reactor an
unexpectedly long time was required to connect
heater leads, thermocouple leads, and air, water,
and helium piping in the confined space available
between the shielding and the instrument pane!,

The piping at the reacter face is being revised
to facilitate connection when the second loop is
ready for insertion so that installation may be
performed more ropidly. Modifications are also
being made to remove lag in the flow alarms for
the jacket cooling water, The oil systemfor cooling
the pumps is being rebuilt to incorporate more
reliable pumps ond improved flow and alarm instrue
ments.

A hydraulically operated door to match the with-
drawal shield has been installed in o wall of a hot
cell in preparation for handling the loop after
A horizontal band saw has been modi-
fied for almost complete remote operation and is
being installed in the hot cell. Other remote han-
dling tools for disassembling the loop have been
made and are being tested.

irradiation.

ORNL GRAPHITE REACTOR SODIUM-INCOKNEL
LLOOP

W. E. Brundage W. W. Parkinson
Solid State Division

Metallographic examination was completed on an
Inconel loop which had circulated sodium in hole

142

58-H of the ORNL Graphite Reactor for 218 hr with
the irradiated section at 1500°F. The unperiurbed
thermal flux in this hole is about 7 x 1011, and
the flux above 1 ev is a factor of 10 lower. There
was no evidence of radiation induced corrosion of
the Inconel. The temperature of the pump cell
section was held between 1025 and 1100°F, and
that of the line entering the reactor was somewhat
higher; the temperature of the line emerging from
the reactor was about 1500°F. Before insertion
in the reactar the loop had been operated at 800
to 1000°F for 80 hr while sedium was circulated
through the loop and through a bypass filter circuit
connected to the loop.

CREEP AHMD STRESS-CORROSION TESTS

W. W. Davis J. C. Wilson
N. E. Hinkle J. C. Zukas
Solid State Division

Development of a suitable apparatus for in-pile
stress-corrosion experiments hos been slow be-
cause of difficulty in securing constant hegt-
transfer rotes from the fuel to a heat sink where
the fission heat can be removed. In the design
shown in Fig. 9.7, fission heat is conducted from
the fuel through the stressed specimen wall to a
sodium-filled annulus. The sodium acts as an
efficient heat conductor that exerts no constraint
upon the deformation of the specimen tube as it
creeps under applied stress. Part of the fission
heat is removed by gaseous convection at the outer
surface of the sodium-containing chamber and part
by conduction through o metal ring to a water
jacket. Simpler designs in which the sodium was
contained in o constant-diameter annuius did not
permit good temperature control. The temperature
gradient that resulted from simulated fission heat
input at the bottom and heat abstraction at the
top caused heated sodium to rise discontinuously
in the annulus, and temperature excursions of
100°F in 1 sec were not uncommon. A substantial
temperature gradient (500 to 800°F per in.) is
required to meet headroom requirements in the
experimental hole and to prevent sodium distil-
lation out of the annulus.

Several designs of the sodium annulus were
tried, such as baffles and spirals, ond ratios of
outside-to-inside sodium-annulus diameters were
varied but did not result in a constant temperature
gradient, The apporatus shown in Fig, 9.7 appears
to operate satisfactorily, but more operating time
PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
ORNL - LR-DWG 3109

 

FURNACE -

   
      

L1
"
\¢
G -
INCONEL SPECIMEN — 3 ¥4

P

l,u'_!
BAFFLE ¥ ill

 

X

LIRS

22 |

TR

"

“q i

FIXED BASE~ \ FUEL CHAMBER

.~ COCLANT CHAMBER

== THERMOCOUPLE WELL

Fig. 9.7. Stress-Corrosion Apparatus,

must be accumulated before it will be certain that
sodium distillation is not taking place in sufficient
quantities to cause trouble.

In the apparatus a tubular Inconel specimen con-
taining 0.2 g of NaF-Zer-UFd {53.5-40.0-6.5 mole
%) in the fuel chamber attached to the fixed base
is eccenirically loaded through a lever arm by a
weight. A stress pattern is produced across the
horizontal cross section which varies from a
maximum compressive stress.  Surrounding the
specimen is a sodium coolant chamber with baffies
which reduce the magnitude of the thermal-con-
vection currents and allow closer control of the
specimen temperature. A furnace helps to maintain
the temperature at 1500°F throughout the test.
Two thermocouple wells project into the sodium

chomber to a position neor the walls of the speci-
men at the neutral axis. A double-walled, con-
centric, sodium fill line, with only the outer tube
welded to the chamber, and a vacuum exhaust tube
complete the apparatus. All surfaces in confact
with the fuel and sodium are made of Inconel.

The assembly is hydrogen fired for 1 he at
1750°F, filled with fuel, and welded in a dry
box wnder purified helium. The other parts of
the container are welded into subassemblies, hy-
drogen fired, and then welded together, with helium
backing up the welds. The lever arm and weight
are adjusted and spot welded onto the specimen,
and the chomber covers are welded into place.
Betore the rig is filled with sodium, it is filled
with NoK and heated to 1000°F to remove  any

143
ANP QUARTERLY PROGRESS REPORT

The rig
is then washed with butyl alcohol, which reacts
with the NaK ond leaves the surfaces clean ond
ready to be filled with sodium.

The sodium is metered out inte a tube that has
first been cleaned with NaK, sealed with Swagelok
fittings, and provided with a helium valve. A
micrometallic filter sealed into the line removes

oxide remaining on the metal surfaces.

any oxides of sodium that may be formed when the
connection to the test rig is made. Once the con-
nection is made, the apparatus is evacuated and
the sodium is melted and forced into the chamber
by helium pressure through the valve; a combi-
nation of high surface tension and back pressure
made filling the chaember somewhat more difficult
without this procedure. For the bench tests, in
sticks of sodium were used, the filled
chamber was heated while a vacuum was pulled
on the system, and a rise in pressure was noted

which

at about 525°F, presumably from dissociation of.

sodium hydride. It remains to be seen whether
such a procedure is necessary to ensure the purity
of the sodium now available. The dry-box work
is completed by cutting the outer fill tube, re-
moving the inner (contaminated) tube, and crimping
and welding both the fill and exhaust tubes.

The entire assembly is welded into a stainless
steel water jacket furnished with Kovar seals and
a capillary tube through which helium flows
throughout the test. A probe beneath the weight
cuts off the current to the furnace at a prede-
termined deflection of the lever arm and indicates
the time required for a given degree of deformation.
A gravity ‘‘tilt"" indicator aids in plumbing the
specimen in the exposure can. A base plate with
a conical cup furnishes a well for the sodium in
case of rupture during operation, A suitable safety
system is being developed.

Approval by the LITR Experiment Review Com-
mittee is being withheld until additional tests are
made on the compatibility of sodium with the
various component materials it might contact if
a rupture occurred at the test temperaiure. Several
tests aimed at a qualitative determination of the
speed of reaction and the heat generated by such
reactions should be completed within a short time.

After an irradiation period of from two to six
weeks in hole HB-3 of the LITR and a suitable
decay period, the specimen is to be sectioned by
the remote metallography group and examined for
stress-carrosion effects.

144

Bench tests on the specimen tube alone are
under way with the outside surface in air, The
inside of the tubes contains either air, helium, or
NaF-ZrF ,-UF, (50-46-4 mole %). An atmosphere
chamber has been built for testing four specimens
simultaneously in vacuum or any desired atmos-
phere. One 200-hr test has been completed on
barren NcF-ZrF4-UF4 (50-46-4 mole %) at 1500°F
and about 1500 psi. No stress dependence of
corrosion was found; another specimen was tested
for 400 hr but has not yet been examined. A
melting and overturning technique for removing
fluoride fuel from the section that is to be ex-
amined metallographically has been successful;
this will greatly simplify handling of irradiated
capsules.

The MTR tensile creep rig is undergoing final
testing and is scheduled for irradiation tests on
Inconel in Qctober.

REMOTE METALLOGRAPHY

M. J. Feldman W. Parsley
R. N. Ramsey A. E. Richt
Solid State Division

Work was continued on the examination of the
solid-type fuel sandwiches of interest to the Pratt
& Whitney Aircraft Division. In addition to the
two experiments mentioned previously,® six other
specimens have been exaomined. The results on
three of them (capsules 1-4, 1-6, 1-7) have been
published.” A report covering the other three
capsules (1-5, 1-8, 1-9) is being prepared. Ex-
amination of the eight capsules processed to date
has indicated greater resistance to cracking and
core-clad separation upon bending of the stainless-
steel-matrix samples than upon bending of the
iron-matrix samples, better resistance to cracking
upon bending of the large UQ, particle size core
than upon bending of the smaller particle size
core, and a greafer increase in hardness upon
irradiation of the small particle size core than
upon irradiation of the larger particle size core.

Work on solid-type fuel elements for the GE-ANP
group has continued. Examinations of two ni-

 

6A. E. Richt, E. Schwartz, and M. J. Feldman, Solid
State Div. Semiann. Prog. Rep., Feb. 28, 1954, ORNL-
1677, p 9.

7M. J. Feldman et al., Metallographic Analysis of
Pratt and Whitney Capsules 14, 1-6, and -7, ORNL
CF-54.5-41 (May 5, 1954).
chrome V fuel elements have been completed.?
Examinafions of two single-plate fuel elements®
ond one two-stage multiple fuel plate test as-
sembly!? were also completed. Work will continue
on both wire and multiple-plate elements.
Expansion of remote metallographic facilities is
proceeding. '

FISSION-FRAGMENT ANNEALING STUDIES

M. J. Feldmon W. Porsley
Solid State Division

Because of interest in the possibility of the
removal of the radiation damage to fuel plates by
annealing, a preliminary study was wndertaken.
The eight samples in Pratt & Whitney capsule
1-9 (four were small particle size, <3-u UG, and
tour were large particle size, 15- to 44 UO.Q)
were selected for the study. Cne sample from each
group (large and small particle size) was examined
after the following treatment: (1) as irradiated,
(2) 900°F anneal for 24 hr, (3) 1100°F anneal for
24 hr, and (4) 1400°F anneal for 24 hr,
showing the results of this preliminary study is
shown in Fig. 9.8, The conclusion drawn fron:
this initial study is that a portion of the neutron

A graph

damage to the samples has been annealed out.
Since the full annealing temperature for type 347
stainless steel is about 1800°F (for complete
removal of the effects of cold work), a complete
anneal for neutron damage at 1400°F was not
expected. For the core, where the major portion
of the damage is by fission fragments, it is felt
that the curve shows a reduction in hardness be-
cause of a partial anneal of the neutron damage
with no reduction in the fragment damage. An
attempt will be made to determine the anneal nec-
essary to remove all the neutron damage from this
type of material and, if possible, the fission-
fragment annealing temperatures.

HIGH-TEMPERATURE, SHORT-TIME, GRAIN-
GROWTH CHARACTERISTICS OF INCONEL
M. J. Feldman

A. E. Richt
Solid State Division

W. Parsley

As an aid in the analysis of the static corrosion
capsules, a study of the short-time, high-tempera-
ture, grain-growth characteristics of the Inconel

 

8M. J. Feldman et al, Metallographical Analysis of
?.E%bsfll;;’re Fuel Element No. 1, ORNL CF-54-4-8 {April

PERIOD ENDING SEPTEMBER 10, 1954

stock used for the tests was made. Information
is availoble in the literature on the usual times
and temperatures for Inconel grain growth, but
because of the nature of the experiment, with its
possibilities of local short-time hot spots in the
fuel, this study was initiated. Figure 9.9 is a
graph of the data obtained. Of major interest were
the definite temperature dependence of the carbide
solubility and the extremely short times (relative
to the corrosion test times) at which large grains
could be produced. Details of the experiment have
been published.!!

BENL NEUTRON SPECTRUM - RADIATION
DAMAGE 5TUDY

J. B. Trice P. M. Uthe
Solid State Division

R. Bolt J. C. Carroll
N. Shiells

California Research Corporation

V. Walsh

Brookhaven National Laboratory

A series of measurements were made in hole
E-25 of the Brookhaven reactor to determine neu-
tron flux energy distributions, The project was
a cooperative effort with California Research Cor-
poratien and Brookhaven National Laboratory. In-
formation obtained from these measurements is to
be used to correlate neutron flux intensity and
energy distribution with radiation damage to cer-
tain lubricants and organic compounds that have
been irradiated in E-25 and examined by CRC.
Data for the irradiated samples have been coded
for the ORACLE by the ORNL Mathematics Panel
so that the lorge number of arithmetical compu-
totions usuolly required by o threshold detector
experiment can be performed in o shorter period
of time than is usual.

 

?A. E. Richt and R. N. Ramsey, Metallographical
Analysis of Single Plate Fuel Elements GE-ANP 3B
and 3C, ORNL (CF-54-3-42 (March 9, 1954),

XOM. Jo Feldman et al.,, Metallographic Analysis of
TwosStage MTR Test Specimen GE-ANP-1B, ORNL
CF+54-7-77 (July 9, 1954),

”M. J. Feldman et al., Short Time—High-Temperature

Grain Growth Characteristic of Inconel, ORNL CF-54-
6-70 (June 8, 1954).

145
ANP QUARTERLY PROGRESS REPORT

$36° DIAMOND)

£ESS (DPH

HARDN

600

550

500

450

400

350

300

250

200

150

100

UNCLASSIFIED

CORNL-LR-DWG 268TA

 

|

SAMPLE 17,

SAMPLE 19,

15 TO 44 UO,

3 UO,

 

 

 

 

 
 

 

 

 

 

  
  
    

 

 

 

 

 

 

 

.fu44fizy

 

 

 

 

 

AT HTOOF

ANMNEALED FOR 24 hr

 

|
Oy ~ 1 1 o Jkiifi__
B R
< | & | < |
a | el i
x | oo x|
eS8 L E |
W W, 0
ql gk
| zl o
| %I ul
< | Z
| ]
I I
\\\\\ \\\\\\\\\\‘Lil\\\\\\\\\
//I: A
rllfi/ ;
Yy

1/1[///

AT 1400°F

 

 

 

 

 

 

bbbk P T T T 71T

 

 

 

IRRADIATED
633 hr TO~12.49, U%3d

BURNUP

 

et e e e— e ey e e — e E— —

Fig. 9.8. Preliminary Annealing Study of Stainless Steel-U0, Fuel Plates in Pratt and Whitney Capsule
1.9,

146
UNCLASSIFIED
ORNL—LR—-DWG 1552A

 

  
     
  

&

 

TiME {min)
[es]

 

 

-

e " e "
Py - -
AN s
LN
AL
AN

o5 LA
1800 1900 2000 2100 2200 2300 2400
TEMPERATURE (°F)

 

 

 

 

Fig. 9.9. Grain Growth Characteristics of In-
conel. '

PERIOD ENDING SEPTEMBER 10, 1954

MTR NEUTRON-FLUX SPECTRA - RADIATION
DAMAGE STUDY

T. L. Trent J. B. Trice
P. M. Uthe
Solid State Division

J. Moteff
General Electric Company

Fission probes with different fissionable ma-
terials were used to maoke neutron flux traverses
in MB-3 of the MTR. The fissionable materials
used included U238, Y236 Np237  gnd Th2%2,
whose fission threshold energies span the energy
region between Y% and 2 Mev. The data from these
traverses will be used with the results from acti-
vations in HB-3 of other, nonfissionable, threshold
detectors that are in the process of being irradi-
ated.

After measurements have been completed in HB-3
with the hole empty, several inches of beryllium
will be introduced into the hole in such a way as
to change the energy distribution of the trans-
mitted neutrons. The new energy distribution will
then be measured, and materials will be irradiated
in both moderated and unmoderated neutron spectra.
An attempt will then be made to correlate spectra
and radiation damage.'?

 

12.1. B, Trice and P. M. Uthe, Solid State Semiann.
Prog. Rep. Feb, 28, 1954, ORNL-1677, p. 14.

147
ANP QUARTERLY PROGRESS REPORT

10. ANALYTICAL STUDIES OF REACTOR MATERIALS

C. D, Susano
Analytical Chemistry Division

J. M. Warde
Metallurgy Division

The primary analytical problem continues to be
the separation and determingtion of trivelent ond
tetravalent uranium in both NaZers- and NaF-Kf-
LiF-base fuels. A successful potentiometric titra-
tion of UF, in molten NaZrF . with metallic zir-
conium was accomplished by meons of polarized
platinum electrodes. The observed end point
occurred ot a weight of titrant which corresponded
to one-half an equivalent of zirconium metal per
mole of UF,, This stoichiometry indicates reduc-
tion of UF, to a mixed oxidation state. Polaro-
grophic studies of uranium fluorides in reactor fuels
dissolved in molten ammonium formate ot 125°C
were initiated. Preliminary experiments revealed
that tetravalent uranium was not reduced at the
dropping-mercury electrode,
formote appeors to be o useful solvent for fluoride-
base fuels containing various oxidation states of
uranium. The conversicn of UF, to UCIH in fluoride
fuels was shown to be quantitative when the fuel
was heated with boron trichloride at 400°C for
90 min. Trivalent uranium fluoride is converted to
UCl; under these conditions. The solubility of
UF, in NaAICI, was found to be 18 mg/g at 200°C
as contrasted to a solubility of UF, of less than
1 mg/g. Molten NaAlCl, is expected to dissolve
tetravalent uranium selectively from the fuels.

Calibration measurements have been completed

Molten ammonium

on the apparatus for the determination of oxygen as
metallic oxides in reactor fuels, The reaction in-
the hydrofluorination of the oxide ond
measurement of the increase in conductivity of
liquid HF as a function of the water formed.
Investigations were also made on the oxidation

valves

of UF, with oxygen at elevated temperatures. At
300°C, UF; is converted quantitatively to UF,
and UO,. The mechanism of the oxidation of UF
with oxygen is being studied. An oxide residue of
about 11% of the original material was found when
UF, was heated in oxygen at 300°C for 2 hr,

An improved method for the determination of
lithium in NaF-KF-LiF-base fuels woas developead.
The chloride salt is exiracted with 2-ethylhexanol,
and the chloride is titrated in the nonogueous
mediyums,

148

Studies were also made on the solubilities of
potassium, rubidium, and cesium tetraphenylborates
in various organic solvents to ascertain differential
solubilities,  The feasibility of a spectrophoto-
metric titration of fluoride in NaF-KF(RbF)-LiF.
base materials by decolorization of zirconium or
aluminum complexes was investigated, Determina-
tion of the concentration of oil in helium was made
by absorbing the hydrocarbons in petroleum ether
and weighing the residue obtained after evaporation.
It was found that prior bleeding of the line effec-
tively eliminated contamination by oil.

ANALYTICAL CHEMISTRY OF
REACTOR MATERIALS

1. C, White
Analytical Chemistry Division

Reseorch and development were continued on the
problem of determining UF,; and UF, in NaZrFs-
and NaF-KF-liF-base reactor fuels. A number of
new opproaches to this problem were investigated,
including redox titrations of the fluoride mixtures in
the molten state, for which polarized platinum
electrodes were used, and pelarographic studies of
fluoride mixtures in molten ammonium formate,
Work was continued on the determination of oxygen
in fluoride fuels and on the determination of alkali
metals and fluoride ion in NaF-KF-LiF-base mao-
teriols,

Determination of Oxygen in Fluoride Fuels

A, S, Meyer, Jr, J. M, Peele
Analytical Chemistry Division

The apparatus for the determination of oxygen as
oxides in fluoride fuels, which was described
briefly in a previous report, 'has been constructed.
A schematic drawing of this apparatus is shown in
Fig. 10.1,

The hydrogen fluoride purification still, which
was fabricated from mild steel, is equipped with
a 24 x 1 in, packed column with on estimated

efficiency of 25 theoretical plotes. Conductivity

 

]A. 3. Meyer, Jr, and J. M. Peele, ANP Quar. Prog.
Rep. June 10, 1954, ORNL-1729, p 110.
 
 

THERMOMETER WELL -——— 3| | &= “‘LQ»—V—COPPER ColL

N

 

    
   

(\ PRESSURE GAGE
' >

PERIOD ENDING SEPTEMBER 10, 1954

UNCLASSIFIED
ORNL-LR~DW3 2724

MONEL NEEDLE VALVES

()  MONEL BELLOWS VALVES

 

 

 

. Tfi\\ ___________________ e
-
<
8
\ :@::_
COPPER TURNINGS ——n __ %93 s
:::{(. il
Y J\fo W'
\\ , |
3 |
— rm."L__
THERMOMETER WELL ~
5
REACTOR

 

 

 

 

ANHYDROUS
HYDROGEN FLUORIDE

  
 
    

TO VACUUM PUMP

s -

L

   

 

 

e

20 % KOH SODA~LIME TRAP

Fig. 10.1. Apparatus for Determination of Oxygen in Fluoride Salts.

measurements are carried cut ina 25-mi fluorothene
cell fitted with platinized platinum electrodes,
The reaction vessel, a modified Parr bomb, is
silver plated on the inner surfaces, as are the
lines and the fittings in the reactor-conductivity
cell section, :

In the procedure now being tested the somple is
placed in the reaction vessel and the system is
evacuated to valve 5 (see Fig. 10.1}. After the
cell has been filled with a measured volume of
the distillate, the conductivity is measured and
the acid is transferred to the reaction vessel by
distillation. When the oxides have been converted
to water, the solution of water in hydrofluoric

acid is distilled to the cell where the measurement
of conductivity is made. Transfers of the acid
are repeated until constant readings are obtained.
The distillation apporatus hgs been tested and
has been found to deliver anhydrous acid at a rate
sufficient to fill the cell in less than 1 min. When
the still was charged with 99% HF, the conduc-
tivity of the distillate corresponded to a maximum
concentration of water of approximately 1 ppm..
Weighed samples of anhydrous LiOH are being
taken through the procedure outlined above as a
Calibration
measurements have been completed over the range

0.03 to 0.25% H,0 in HF,

means of calibrating the opparatus,

149
ANP QUARTERLY PROGRESS REPORT

Oxidation-Reduction Titrations in Fused Salts

A, S. Meyer, Jr.
Analytical Chemistry Division

Experiments were carried out to determine whether
the equivalence point of a titration in which UF
dissolved in a molten fluoride fuel solvent is re-
duced by the addition of small increments of an
electrodeposition metal could be observed by one
of the conventional electrometric techniques, The
first system studied involved the reduction of UF,
in NaZrF . by zirconium metal according to the
proposed reaction

AUF, + Zr° > 4UF, + ZrF, .

Since no insulating material which is compatible
with molten NaZrF. was available, it was not
possible to devise o reference elecirode for meas-
urement of the potentials of the solutions, Accord-
ingly, polarized platinum electrodes were selected
as indicating clectrodes. In order to obtain more
information about the nature of the polarization
phenomena, the area of one of the electrodes was
made approximately 80 times that of the other.
Thus, since the current density at the smaller
electrode waos 80 times as great as thot ot the
larger, the potential difference between the elec-
was approximately equal to that of the
smaller electrode. By reversing the direction of
the polarizing current, it was possible to determine
the potential at the cathode and at the anode,

The solution of UF, in NaZrF g was contained in
a 25-ml platinum crucible which was placed in the
helium-filled quartz liner of o pot furnace, Agi-
tation was accomplished by means of a mechani-
cally driven platinum stirrer which was constructed
of 0.05in.-dia platinum wire, Additions o the
solution were made through o quartz tube which
extended from the top of the liner to a point about
1 in. above the melt.

In pure NaZrF. the onode was more strongly
polarized than the cothode. At polarizing currents
of 80 pa the potential of both electrodes exceeded

trodes

1 v. Approximately ]/2 hr was required for the
electrodes to reach equilibrium. Platinum elec-
trodes could not be oppreciably polarized in NaF-
KF-LiF-base fuels,

When UF , wos added to the molten NaZrF, the

potential decreased rapidly. For a solution which

. L. Colichman, Polaro§rap!3y in Molten Ammonium
Formate, LRILL.-117 (April 1954)

150

contained 1.25 mole % UF the potential of both
electrodes was approximately 300 mv, with no
significant difference between the potentials of the
cathode and ancde.

During the early part of the titration of the molten
fluoride mixture with zirconium metal, the potential
of the cathode decreased to values less thon 10 mv
and remained constant throughout the titration. At
the same time the potential of the anode increased
rapidly ond attained potentials in excess of 1 v
before the end point of the titration was reached,
When an equivalence point was reached, the poten-
tial of the anode decreased rapidly to about 50 mv
and then decreased gradually on further addition of
Only one breck in the potential vs
The curve is shown

reductant,
titrant curve was observed.

 

 

 

 

 

 

 

Fig. 10.2,
N
“w ORNL—LR—DWG 2725
TP . —— - Y —mseey
s ' ; |
‘ -2 |
2 -~
< | r’/ i
o 08 r T e i -
LeJ S :
™ ¢
s |
< 0.6 ‘ II C r - ’
— / I |
O / i |
0 ‘I |
o) ;
&j i
o N |
= ® SO L
O o - S . i 1 A AT
a-
0 01 0.2 03 04 s 06 c9 10
EQUIVALENTS OF ZIRCONIUM PER MOLE OF UF,
Fig. 10.2. Titratien of UF, in Molten NaZrF,

with Zirconium Metal,

The cbserved end points of two titrations occurred
at a weight of titrant which corresponded to one-
half an equivalent of zirconium metal per mole of
UF,. This stoichiometry indicates the reduction of
UF, to a mixed oxidation state that corresponds
to the postulated formula U, F,.

Polarographic Studies in Fused Ammonium Formate

A. S, Meyer, Jr, D. L. Manning
Analytical Chemistry Division

Colichman? reported that when solutions of ucl,
in molten ammeonium formate were reduced at the
dropping-mercury electrode, three well-defined re-
duction waves with halfewave potentials of 0.1 to

~0.2, =0.75 to —-0.80, and ~0.95 v (vs pool) were
obtained, The following reactions were postulated
to explain these waves:
UM-EP—’# U+3£~> ) +2
and : :
Ut e > U+3———%1 U+:Z %Ew} ue .

He also reported that anhydrous hexavalent uranium
gave no reduction wave until after an aging period
during which the compounds were reduced by the
ammonium formate to a lower valence state, probably
guadrivalent, |

Since it appeared that this techmque could be
appiied to the determination of tetravalent uranium
in the presence of large quantities of trivalent
uranium, expérimmfs were carried out to obtain
polarograms of solutions of UF ,
experiments it was found that UF,
ond NaF-KF- L:F base fuels dassoived readily i
molten ammonium formate ot 125°C to form green
solutions containing as much as 1 mg of uranium
per milliliter. Although UF, alone was found to be
much less soluble, UF; which had been fused with
NaZrF . was readily soluble, Similor relationships
were found when molten ammoniom acetate was
used as a solvent, :

No U(lV) reduction waves were obtained when
polarograms of solutions of UF, were recorded
immediately after dissolution, and no reduction
waves were  observed in the polarograms of a
sample of NaF-KF-LiF-base fuel that contained
both UF, and UF,. When these solutions were
gerated, a single reduction wave with a halk
wave potential of approximately ~0.2 v was ob-
served. The potential and the diffusion current of
this wave corresponded closely to those recorded
immediately after the dissolution of an equivalent
quantity of K;UO,F, in molten formate. Since the
U(IV) waves moy have been shifted to potentials
outside the range of the solvent (0 to ~0.9 v) by
the complexing action of the fluoride ions, oddi-
tional polarograms were token of solutions con-
taining UCi,.  Except for a small reduction wave
ot about —~0.2'v, no evidence of reduction was found,
The diffusion current of this wave was increased
by aeration to values corresponding to those ob-
tained for hexavalent uranium, and therefore the
increase was atiributed to the oxidation of UC!M
It appears that in molten ammonium formate no
oxidation state of uranium other thon hexavalent is
reduced ot the dropping-mercury electrode.

In preliminary

and NaZrF -

PERIOD ENDING SEPTEMBER 10, 1954

Conversion of UF,; and UF to the Respective
Chiorides with BCI3

A, S, Meyer, Jr. D. L. Manning
Anquhcal Chemlstry Division

It was reported previously? that, on the bas:s of
free-energy ' calevlations, UF,, UF, ZrF,, ‘and
NaF are quantitatively converted to the chlorides
when at equilibrium with boron trichloride. Since
the exchange reaction in organic solvents at moder-
ate femperdtures was found to be too slow for
analytical application, tests were carried out with
gaseous BCl, ot elevated temperatures. This tech-
nique offered an addifional advantage in that the
resulting chlorides couid be sepumfed by subllmu-
tion.

When o somple of NaZrF, UF: fuel was heated
in a stream of BCl, and hehum a‘t 350°C, the bulk
of the zirconium was sublimed as ZrCl, within 90
min, and all but fraces were removed ofier heating
for an additional hour at 450°C, It has been re-
ported? that uranium can be sublimed as Uucl, by
treating UF:‘i with BCl, af 600 to 650°C, but in
these tests only o small fraction of the UCl, was
volatilized - when the UC! -NaCl residue, which
fused at about 450°C, was heated to temperatures
as high as 750°C. ‘

Analyses of the residues showed that when the
conversion was carried out at 400°C for 90 min the
vranium in - the NoZrF UF samples was quanti-
tatively converted to UCl, ond that all the uranium
remained in the residue. When the conversion was
carried out at 500°C, 2%  of the uranium ‘was
volatilized, while at 750°C, 4% was volatilized.

When UF, samples were converted to UCI, by
this method, significant froctions of the uranium
were oxidized to UC|,, It is believed that this
oxidation will be eliminated when all traces of
oxidizing contaminants are removed from the gase-
ous reactants, &

Salubiiiéi‘y of Tri- and Tetravalent Uranium
Flyorides in Fused NaAIC],

A, 5. Meyer, Jr. W, J. Ross
Analytical Chemistry Division

in the continuation of experiments to defermine
the optimum conditions for the conversion of UF,

 

3A. S, Maver, Jr., D. L. Manning, and W. J. Ross, ANP
Quar, Prog, Rep, fune 10, 1954, ORNL-1729, p 110,

4y, P. Caolkin 5, Divect Conwversion of TFfi to TCIA,
CEW.TEC C.0.350.4 (Sept. 23, 1945).

 

151
ANP QUARTERLY PROGRESS REPORT

and UF, to the corresponding chlorides,® it was
noted that UF , was dissolved much more readily
thon UF; by fused NoAICl,. Measurements of the
solubilities were carried out by equilibrating sam-
ples of the uranium fluorides with molten NaAICH,
for periods up to 5 hr, under an inert atmosphere,
followed by the separation of the undissolved
solids by filtration ond calculation of the solubility
from the concentration of uranium in the filtrate,

Solubilities of 18 and 16 mg of UF , per gram of
NaAICl, were found at 200 and 175°C, respectively.
However, the solubility of UF3 decreased from
> mg/g of NaAICl, at 225°C fo less than 1 mg/g at
200°C. Since UCI; was also found to be sparingly
soluble at the lower temperatures, the above values
are indicative of the true solubility of trivalent uro-
nium in this medium rather than of incomplete con-
version of the fluoride to chloride. At temperatures
lower than 200°C, molten NaAlCi‘i is expected to
dissolve tetravalent uranium selectively from UF,
which has been complexed by potassium fluoride in
the fuel.

Since all available UF; samples are contaminated
with tetravalent uranium, it has not been possible
to carry out accurate solubility measurements at
the lower temperatures by relying on the determina-
tion of total dissolved uranium. An apparatus has
been constructed for the determination of microgram
quantities of frivalent uronium in the filtrates by
hydrogen evelution, This apparatus, which is based
on the measurement of evolved gases at reduced
pressures, is now being calibrated against cou-
lometrically generated hydrogen. In order to carry
out dissolutions at lower temperatures, LiCI-AlCI
fluxes which melt at approximately 120°C have
been prepared.

Oxidation of UF; with Oxygen

A, S, Meyer, Jr. W. J. Ross
Analytical Chemistry Division

Preliminary investigations have been initiated to
determine the extent of the oxidation of UF3 by
oxygen at elevated temperatures. The reaction is

assumed to be
In the tests conducted to date the reaction products

are refluxed with 2.5 w/v % ammonium oxalate
solution, ond the soluble portion is analyzed for

SW. J. Rodd and J. L. Mattern, ANP Quar. Prog. Rep.
Mar. 10, 1954, ORNL-1692, p 111,

152

its uronium content, which is assumed to be UF ,.
In these tests the reaction time was held constant
at 2 hr and the temperafure was varied from 200 to
600°C, with the results shown in the following:

Temperoture Reactivity
(°C) (%)
200 30
300 100
200 91
600 25

At 300°C the reoction proceeds quantitatively as
written, At higher temperatures, however, the con-
centration of soluble uranium (UF4 in ammonium
oxalate) decreased rapidly from the theoretical
content with increasing temperature, probably as o
consequence of a further reaction of oxygen with
UF .

The experiment was then repeated with UF, and
oxygen. At 200°C the entire sample was soluble
in ammonium oxalate, as expected, which indicated
no oxide formation. At higher temperatures, a black
insoluble residue was formed upon dissolution of
The amount of
residue increased at higher temperatures from 11%
at 300°C to 27% at 600°C., The major portion of
the sample dissolved rapidly, and a yellow filtrate
was formed. It is to be noted that Kraus® reported
that the following reaction occurs ot 800°C:

UF, + 0,—> UF, + UO,F, .

the somple in oxalate solution,

The yellow color of the filtrate substantiates this
reaction in part, but there is no way of accounting
for the black residue, obviously an oxide of ura-
nium. 1t is known,” however, that UF, is converted
quantitatively to U,0g when ignited ot 900°C in
air, Petrographic and x-ray siudies of the material
will be made. Further tests are planned at various
reaction times and temperatures.

Determination of Lithium, Potassium, Rubidium,

and Fluoride lon in NaF-KF{RbF)-LiF-Base Fuels

J. C. White G. Goldberg
B, L. MeDowell
Analytical Chemistry Division

A critical review of the analytical methods for
the determination of lithium, potassium, rubidium,

6C. A, Kraus, Technical Report for the Months of April,
May and June, 1944: Chemical Research — General,
CC-1717, July 29, 1944,

7). J. Kotz and E. Rabinowitch, The Chemistry of
Uranium, p 374, McGraw-Hill, New York, 1951.
and fluoride ion is being made to ascertain those
applicable to the ondlysis of NaF-KF(RbF)-LiF-
base reactor fuels.

Lithium. An indirect volumetric method for the
determination of lithium in the presence of the
large concentrations of sodium and potassium and/or
rubidium which are found in NaF-KF(RbF)-L.iF-base
fuels was developed. The method was designed
specifically to provide a rapid, and yet precise,
means of determining lithium. The alkali metals
are first separated from uranium and then converted
to chlorides. Lithium is extracted as the soluble
chloride salt by heating ot about 135°C with
2-ethylhexanol® until the slightly soluble salts
of sodium and potassium become *‘free-flowing”’
and no longer cling to the walls of the vessel,
These salts are removed by filtration, and then the
tiltrate is diluted with ethanol. The chloride in the
filtrate is titrated by the Volhard method,? and the
lithium concentration is calculated, The coefficient
of varigtion is 0.5% for 1 to 50 mg of lithium. A
topical report of this work is being written,

Potassium and Rubidium. Since the concentration
of either potassium or rubidium in NaF-KF(RbF}-
lLiF-base fuels is of the order of 30 wt %, the de-
termination of potassium or rubidium by means of
the flome photometer has been neither sufficiently
accurate nor precise, The conventional method for
this determination is the perchlorate separation !0
of potassiumor rubidium perchlorate from the sodium
and lithium salts in ethyl acetate. The methed is
satisfactory but requires considerable
operator time. |herefore the application of possible

entirely

substitute procedures which are specific for potas-
sium and/or rubidium was investigated.
cedure!! involving the precipitation of potassium
bitartrate in ethanol-water and titration of the
solution of the salt with standard base was tested.
The method is extremely simple and rapid, and the
bitartrate salt is readily filterable. Rubidium
behaves similarly to potassium, but cesium does

A pro-

not precipitate as the bitartrate salt under these
conditions. Further work is necessary to determine
the effect of acid and salt concenirations on the

 

8E. R, Caley and H. D. Axilrod, Ind, Eng. Chem. Anal.
Ed, 14, 242 (1942).

1. M. Kolthoff and E. B. Sandell, Testbook of Quanti-
tlatéve Inorganic Analysis, p 477, MacMillan, New York,

Y010:d,, pp 41415,

YA, F. levinsh and Ya. K. Ozol, J. Anal. Chem. USSR
8, 57 (1953).

PERIOD ENDING SEPTEMBER 10, 1954

solubility of potassium bitartrate before a decision
can bereached as to the applicability of this method
to NaF-KF-LiF-base fuels.

A second method under consideration is the use of
sodium ftetraphenyl boron'? as a reagent for the
direct determination of potassium, rubidium, and
cesium.  The potassium salt is precipitated in
aqueous acidic solution, filtered, dried at 110°C,
and weighed as potassium tetraphenylborate. The
method has the advantage of being simple and rapid,
and it has a favorable weight factor for potassium,
Themain interference in the use of tetraphenylboron
is the ammonium ion, The reagent is expensive,
but it is readily available at the present time,
Preliminary work on the solubilities of potassium,
rubidium, and cesium tetraphenylborates in non-
aqueous solvents has shown some interesting dif-
ferences in solubilities. The possibility of ex-
ploiting differential solubility as a means of
separating the salts is being explored, Acetyl-
acetone, diethylketone, and methy! isopropylketone
are three possible solvents for this separation,
The order of solubility in organic solvents has been
established as: KTPB > RbTPB > CsTPB (TPB =
tetraphenylboron).

Fluoride lon, The determinatian of fluoride ion
in fluoride fuels was previously conducted by the
pyrohydrolysis'® procedure which was particularly
successtul for NaZrF -base fuels. Alkali metal
fluorides, however, are resistant to pyrohydrolysis,
and thus considerable time, of the order of 3 to 4 hr,
is required to pyrohydrolyze completely 100 mg of
NoF-KF-LiF-base material.

Two alternative methods have been studied,
therefore, for opplication to the determination of
fluoride ion in NaF-KF-LiF-base materials. A pro-
cedure, which was formulated by Chilton, ' was
tested. In this procedure the fluoride ion is titrated
with a standard solution of aluminum (potassium
alum), and the resulting change in acidity is re-
corded. Considerable difficulties have been en-
countered, the foremost of which is the extremely
small change in pH at the equivalence point. The
method in its present form does not appear to be
applicable to this work.

The feasibility of a spectrophotometric titration
of fluoride ion based on the decolorization of o

 

12, Geilmann and W. Gebaukr, Z. anal. Chem. 139,
161 (1953).

13), C. Warf, W. D. Cline, and R. D. Tevebaugh, Anal.
Chem. 26, 342 (1954).

145, A Chilton, personul communication.

153
zirconium or aluminum complex, or lake, is also
being investigated, In this method, the solution of
fluoride is titrated with o colored zirconium com-
plex, and the change in absorbance of the solution
is plotted as a function of the concentration of
in the complex. Zirconium alizarin
sulfonate and zirconium eriochrome cyanine both
react rapidly with fluoride ion and will receive

zirconium

further study. Preliminary results indicate that the
reactions of these compounds with the fluoride ion
are reproducible, A simple titration cell is being
designed.

Qil Contamination in ARE Helium

G. Goldberg
Anolytical Chemistry Division

The contamination of the helium for the ARE with
oil from valves and fittings was checked before the
gas at the site was used, The method used for
these tests was to pass the gas through dual
scrubbing towers containing petroleum ether and
then to evaporate the ether to dryness and weigh
the residue. In the initial test on a tank car of
helium, 360 pg/ft3 was found. A tower containing
70 in.3 of charcoal was placed in the gas line, but
it had little effect in eliminating the oil. Further
tests on the same tank car showed that bleeding the
line before samples were taken markedly lowered
For example, in three
successive samples of 15.5 fi3 of gas each, the
contamination decreased from 100 to 10 ug/ft3.
These tests indicate that bieeding the line with
50 to 100 ft2 of gas before the equipment is coupled

the contamingtion level,

into the system will effectively remove oil from
fittings and valves,

PETROGRAPHIC INVESTIGATIONS OF
FLUORIDE FUEL

G, D. White, Metallurgy Division
T. N. McVay, Consultant

Petrographic studies of UF , in alkali-metal fluo-
ride systems have shown that two phases exist in
the KF-UF,; system. One is red and is neor
3KF-UF ;, and the other is blue and is near KF.UF .
Only one binary phase is present in the NaF-UF,
system, and it is probably 3NaF.2UF,, Lithium
fluoride has not been found to combine with UF,
in either the binary system LiF-UF,, in the ternary

154

systems LiF-KF-UF, or LiF-NaF-UF,, or in the
quarternary  system LiF-NaF-KF-UF,  Cesium
fluoride was found to form a complex with UF,,
with a composition near 3CsF-UF,, Some work was
done on the systems LiF-ZnF, and NaF-CsF, and
no binary compounds were found,

Part of a loop in which NaF-ZrF -UF, (50-46-4
mole %) had been circulated was received from
Battelle Memorial Institute, Samples token from the
wall contained about 50% ZrO,, and the balance
was an NaF-ZrF, (53-47 mole %) compound. In
addition, several loops containing NaF and ZrF,
with reduced UF, have been examined, and no
appreciable oxidation of the UF; could be found.

SUMMARY OF SERVICE AMALYSES

J. C. White W, F. Yaughan
C. R. Williams
Anglytical Chemistry Division

Most of the somples received during the past
quarter were from NaZrF .- and NaF-KF-LiF-base
matericls.  Other types of samples that were
analyzed included silver fluoride, chromium fluo-
rides, iron fluorides, beryllium, sodium, and
rubidium metals, Inconel, and alkali-metal caor-
bonates., A revised procedure was established fo
determine UF; and UF, in NaF-KF-LiF-base fuels,
Also, the method of White, Ross, and Rowan !5 was
adapted successfully to the determination of oxygen
and metallic impurities in rubidium metal.

A total of 1,165 samples, involving 10,702 de-
terminations, was analyzed., A breckdown of the
work is given in Table 10.1.

 

15), C. White, W. J. Rass, and R, Rowan, Jr., Anal
Chem. 26, 210 (1954).

TABLE 10.1, SUMMARY OF SERVICE
ANALYSES REPORTED

 

 

Number of Number of
Samples Determinations
Reactor Chemistry 679 6,694
Fuel Production 38 395
E xperimental Engineering 379 3,406
Miscellaneous 59 207
1,165 10,702

 
Part Il

SHIELDING RESEARCH
11, SHIEL.DING ANALYSIS

E. P. Blizard

J. E. Faulkner

H. E. Hungerford

F. H. Murray
C. D. Zerby

Physics Division

H. E. Stern
Consolidated Yultee Aircraft Corporation

Theoretical analyses have been made so that an
understanding of the air and ground scattering
measurements: at the Tower Shielding Facility could
be obtained and the attenuation in the side wall of
an aircraft crew shield could be determined. In
connection with the cir and ground scattering analy-
ses, special attention was given to the calculation
of the reduction of air scattering due to ground
interruption and to the effects of multiple scattering.
The gamma-ray slant penetration has been calcu-
lated for the side wall of an aircraft crew shield,
and the variation of radiation intensity throughout
the crew volume has been briefly explored.

SLANT PENETRATION OZF COMPOSITE
SLAB SHIELDS BY GAMMA RAYS

C. D. Zerby

The total radiation dose in the crew compartment
of an airplane is partially dependent on the gammo-
ray flux penetrating the compartment shield. To
obtain o fundamental, but practical, knowledge of
the gamma-ray penetrations, the problem has been
treated theoretically by using stochastie, or Monte
Carlo methods.

The problem was programed for solution on the
ORACLE ond was arranged for investigating the
effects of the variations of all the parameters in-
volved. The shield was taken as o composite slab
of two materials comprising a thick layer of Compton
scattering material ' followed by a thin layer of
fead. In each of many cases the initial incident
photons were taken as monoenergetic and incident
on the slab at a particular angle with the normal'to
the slab. The stochastic process used the exact
physical analog of the probability laws known to
govern the life of a photon.

 

]In the range of energy considered, the Compiocn
scatterer has opproximately the same physical chorac-
teristics relative to a photon passing through it as does
concrete or polyethylene (CH,).

The parameters and their variations that have
been investigated to date are listed below:

Thickness of Compton SCOQ’ferEr,z in. 3,9, 15
Thickness of lead, in. ]/]0, :}/10’ }'2
Initial photon energy,3 me? units 2,6

Initial angle of incidence, deg 0, 30, 60

All combinations of these parameters were investi-
gated. &

The following solution of a typical problem indi-
cates part of the information obtained:
Initial Conditions

Thickness of Compton scatterer 3in.
Thickness of lead ]/]0 in,
Initial photon energy 6 mc2
Initial angle of incidence 60 deg
Results
Fraction of initial energy
1. reflected 0.0234
2. obsorbed in the Compton scatterer owing 0.318
to scattering collisions
3. absorbed in lead owing to scattering 0.0568
collisions
4, obsorbed in lead owing to absorption 0.0515
collisions
5. penetrating shield 0.550
6. penetrating shield without being 0.435
degraded in energy
Energy build-up factor 1.265

The energy spectrum for the reflected photons
was also obtained in this typical problem. The
spectra for the photons penetrating the shield were
obtained in the angle intervals 0 to 15, 15 to 30,
30 to 45, 45 to 60, and 40 to 90 deg. These spectra

will be available in a forthcoming report.

2 .
The electron density of the Compton scatterer was
token as the some as that for polyethylene,

3Mu|tiply mcz units by 0,5108 to obtain Mev.

157
ANP QUARTERLY PROGRESS REPORT

AIR SCATTERING OF NEUTRONS

In an effort to understand the air-scattered neutron
intensities which have been measured af the Tower
Shielding Facility, a number of calculations have
been carried out with a variety of conditions and
assumptions. The predominant difference between
the cuirent work and that which was done pre-
viousiy? is that the interference of the ground has

The ground scattering is
In an attempt to obtain a

been taken info account.
calculated separately.
better fit to the data, such aspects as onisotropy of
scattering and slant penetration at the detector tank
(crew shield) have been taken into account in some
of the work. Furthermore, measurements have been
made of the angular distribution of the radiation
leaving the reactor shield, and these data have
been used in some of the calculations.

Because the importance of the multiply scottered
neutrons has appeared to be greater than was antici-
pated, considerable effort has been mads to extend
the calculations to beyond the singly scattered
case. Some results ore reported for two scatterings,
and o general method has been developed for all
orders of scattering.

Single Anisofropic Air Scattering in the Presence

of the Ground (Shielded Detector)
F. H. Murray

For a cenvenient machine calculation, the at-
tenuation in air was neglected, The detector was
placed with 10 cm of water between it and the face
of the crew shield which, when extended, contained
the reactor source. The beam from the source was
about an oxis through the detector, in theory, and
contained terms cos a and cos® @ a calculation
was made separately for the terms 1, cos o, and
cos? ain ithe expansion of the source flux. I D is
the reactor-detector distance and b the height of
both above the ground, the flux at the detector is
expressed as the sum of the terms:

g w /2

and

 

12 _ o i h/~W
16a2p B=0 ¢=7/2

158

where

Il

gla + B) 1 +0.3cos(a+ ),
2

Aa) = a, 1; b, cos a; ¢, cos” a .

The angle ¢ is the angle between the vertical
plane taken as the face of the crew compartment
and the plane friangle formed by the source, the
scattering volume, and the detector. The results
of this calculation® will be published later.

Single !sotropic Air Scattering in the Presence

of the Ground (Unshielded Detector)
J. E. Favulkner

A colculation was made of the reading of an iso-
tropic detector caused by first-scattered neutrons
in air as the distance above the ground was varied.
The following assumptions were made:

1. The source and detector have the same alti-
tude,

2. The ground is an infinite plone.

3. The source is isoiropic.

4. The scattering in air is isotropic.

5. There is no energy degradation on air scatter-
ing.

6. The attenuation is pure inverse square.

With these assumptions the reading on the detector
may be shown to be proportional to

»/(5)

b \p)
where D is the separation of source and detector,
b is the altitude, and / may be expressed in closed

 

e~1:25/sin Bsin ¢ gg g4 f

e 1-25/sin Bsin ¢ df de¢ fflomfif(a) gla + B) do

4See, for example, H. Goldstein, Chap. 2.9, p 831, in
Reactor Handbook, ed by J, F. Hogerton and R. C.
Grass, Technica! lnformation Service, AEC, 1953,

F. He Murray, Single Anisotropic Air Scattering of
Neutrons in the Presence of the Ground (Shielded

Detector), ORNL. CF.54-8-104 {to be issved).

 

arc cot {D/h cos ¢ —~ cot )

f(a) gla + B) da,

a=0

for0§<;c‘5§n/2;

forn/2<h<m,
form in terms of the Spence functions. A plot of
f(h/D) has been made (Fig. 11.1) by using the re-
sults of ORACLE computations. A report that
gives details of the calculation is being prepared. 8

Formulas for Multiple Scettering in Uniform Medium
F. H. Murray

Method for Anisotropic Source and Scattering.
The Fourier analysis can be applied to multiple
scattering in such a manner as to lead to explicit

UNGL ASSIFIED
ORNL —{.R—DWG 2645

 

"

 

 

 

e B I O

INTENSITY NORMALIZED TO 1 AT INFINITY

 

 

 

INTENSITY, Fib/0)

 

 

0.3 : e -

 

 

0.2 s

04 fore —

 

 

 

 

 

 

 

 

 

0 1.0 2.0 3.0 4.0 5.0
24/0

Fig. 11.1. Variation of Radiation intensity with
Altitude for o Constant Source-Detector Separation
Distance.

PERIOD ENDING SEPTEMBER 10, 1954

formulas, similar to some derived by Wigner? but
which do not require a knowledge of group theory
for their application, Here the scattering law is
represented by a general formulag, (»0, ) dwy do,
for the probability that a particle arriving wnfhm a
unit volume at P and moving in a direction a)o
within the solid angle dw, will be scattered into
the solid angle dw in the direction D,

Let O and P be the points (x4,y,,2,) and (x,¥,2),
Fig. 11.2, and let f (Q,mo) dwy dV be the number
of particles which have suffered their mth collision
in the volume element dV , = dx, dy, dz, and
emerged in the solid angle dw, with a direction 50.
If the next collision occurs in a volume element
dv and if o is the total cross section, the number

 

6J. E. Faulkner, Single Isotropic Air Scattering of
f the Ground (Unshielded

Neutrons in the Presence o

Detector), ORNL CF.-54-8-96 (1o be issued).
7E. P. Wigner, Phys, Rev. 94, 17 (1954).

UNCLASSIFIED
ORNL—LR-—-DWG 2646

Plx, ¥, 2

 

 

Fig. 11.2. Geometry for Multiple Scattering Cal-
culation.

159
ANP QUARTERLY PROGRESS REPORT

of collisions in dV, for which the particle emerges in the solid angle dw with direction & is equal to

fpsy P @V pde = do [ff S(&,8) e~ 12 dr duy [, (0,&,) Awy

ni
where
Y

roe= x—-x02+y—y0 + (2~ z4 .
|G~ xg)? 4 (= yp)? + (e = 2)? |2

Here A, is the solid angle subtended at Q by dV  or by dV /72; by putting 72 dr doy gy = Vo then

M [P jjjf (@) S(6ya) ST vy

et Y“(fi) be the normalized spherical surface harmonics. Forconvenience, these functions canbe ordered
in a single series Y :
C

 

Y, = Y ;Y

0 0,0’ p = Y

Y Y

1,17/ Y2 = fq,01 '3 1,04 f4 = Yo, 24 75 2,-1¢ "

Let
A0 @) =F 7, 0y () .

fnsr (P 2 55“‘ fm C @) S(B3y,6) e":f v,

7

From Eq. 1,

 

Introducing a Fourier transformation of both sides and using
k1x + koy + kyz = kr cos (k,r) ,

with
1

7

2 2 2 |2

e [ K2 k34 k22,

gives

—ilk xtk,ytk.,z)
ije PR XTRyYTR4Z fmfl(p,gj) dVP
-ik7 ( ,k) . - L7 -_—
_ 2 J*fj- . ikrg cos (7 fm,.;(Q) dVQ J‘J‘J‘e—zk,r cos (r,k) YC(CSO) 5(630,63) e 2 ~ dVP ,
C

r

 

where
AV, = dlx ~ xp) dly - vo) dlz ~ z4) .
The direction & is the direction from O to P, so that, as indicated, the transform of the right side of Eq. 1

breaks up into a sum of products of Fourier transforms. The last integrals on the right can evidently be
evaluated independently of the position O. By expanding S in the form

o © = LAY (6) Y (@),

160
these integrals become

PERIOD ENDING SEPTEMBER 10, 1954

 

fff o—ikr cos (r,k) y ((—3 ) 5(5013) _iiai dii;p
T2
: @ :
- ~ikr cos {k,r)~ar: -2 - :
- Z qu Yq(cu ff'C::O e Yp(mo) Y (&g) dag dr
= L Agp Yq((“) Toe v
where
di?P = 12 dr d&g)o ; and
> . v v mm B §
Y (@,) Y _(0,) doy L Y s Fob = [ e - 1Py} .
p c
be a + ik cos(k,r)
The source function may be anisotropic; however,
Let if an isotropic source is present or if the scottering
, ' law possesses axial symmetry, some simplificafions
= f ffe"zk"o C°s('0'k)f c(Q) C{VQ ’ occur.
: e ? When the scaftering density has been obfmned by
and then a Fourier inversion, one further infegration gives

Lr o, Y& =Ere LA 1, v().

Equating coefficients of id

m+l E qu pc

Cy P

where

entical Y's gives

e = L Pac P

quc = Z Aqr: [pc *

 

 

 

 

 

the flux inte an arbitrary unit volume from: any
direction; if the detector sensitivity depends on
direction, the total scattering to be counted at any
point P “is represented by a convolution and can be
calculated by another Fourier transformation and
inversion. . The method is to be applied te air
scattering measurements taken at the TSF and will
make possible subsequent calculations for unusual
reactor shield shapes,

 

 

v
Thus, if lec» is represented as a vector (leo,F?fifl’,,Fm’z, ...) for all m,
Fm+1,0 Fm,O
m+1,1 o, 1
N s, 5wl - al-lrl
Fm+l,q ‘m,c
Then, formally, | |
en, tormd y Method for Isotropic Source and Scattering. |f
(Pl = B 3;:0; the scattering is isotropic after each collision, the
formula to be used for calculating the total flux at
N [a ¥ :
WLt = BEF I = B2 Fo the point P after the nth coliision is a convolution:
F § = B™{F,} FPY = o [ f [F, {0 v, 0P

161
ANP QUARTERLY PROGRESS REPORT

with

WQRPQ

SOP) =,

2
471R PO

Since the Fourier tronsform of a convolution is the
product of the transforms of the individual func-
tions, for an isotropic source with no energy loss
ofter each collision the transform TF becomes

(Tqfi)”” O_Sn

with
1 A+ po
T =——1| et -
b gt ()= s
where
]/2
p =1 Cr)§+é)§+m§] .

In taking the inverse of TF_,the path of integration
is deformed inte the closed curve about the negative
rea! axis to the left of —a. Calculations of these
scattering functions for the first- and second-
scattered flux give

F

| = 1.635,

F, = 0,19167 .

2

These figures are to be multiplied by 1/4772 AD,
for A = 130 meters, to obtain the number of fast
neutrons at a distance D feet from the source.

GROUND SCATTERING OF NEUTRONS
A. Simon H. E. Stern

An investigation is being carried out to derive
formulas for the ground-scattered neutron flux to
be obtained from a general source distribution. The
model used is that of a source whose axis, or
direction of maximum intensity, forms an arbitrary
angle 0 with the source-receiver axis but lies in
the same horizontal plane as the latter. The
differential source strength is assumed to be a
function of only the angle between the emergent
ray and the source axis. The receiver is isotropic
under the assumptions of no air attenuation and a
constant albedo for the ground. The following
expressions for the flux at the receiver are ob-

tained:

162

For isotropic re-emission from the ground,

N OA
F .

ffl (e} sin a da

a=0
X f
P

8722
7w/ 2

dep

 

 

mi(m]x+m2y+w3z)

e Ax,y,2) dx dy dz

 

For cosine re-emission from the ground,

NoA
T _
F = —— fla) sin a da
4?72192 a=0
/.2 d
X f ___qs__,
D= mtt/ 2 [g(a,q'))]?’/z
where
N, = total source strength {(neutrons/sec),
A = reflection coefficient of the ground
(albedo),
fla) = relative source strength per steradian at
angle a between emergent ray and source
axis,
b = height of source and receiver above the
ground,
]
gla,g) - ——
sin? a cos? b
D
-2 o cos ( eot a sec ¢
2 D 9 / 2y
- — | sin G tan & + | —
b T\
) = angle between source axis and source-
receiver axis,
D = separation distance between source and

receiver.
PERIOD ENDING SEPTEMBER 10, 1954

Alternatively, if a single-scatter approach with attenuation by a *‘removal’’ mechanism is used, the fol-

fowing formula is obtained:

 

 

F o= f fla) sinadaf - .
167252 Zr a=0 b —1/2 gla,d) 1 + sina cos & gla,d)
where
% = scattering cross section,
Y = removal cross section, ORNL-LR-DWG 2547

A detailed report on this calculation is being
prepared. 8

FOCUSING OF RADIATION IN A
CYLINDRICAL CREW COMPARTMENT

J. E. Faulkner

A study of the focusing of radiation in o cylindri-
cal crew compartment has been made, with particular
attention being given to the case with the following
conditions: |

1. The cylinder is infinite.

2. The
along the inner eylinder walls.

3. The distribution of the radiation
depends only on the angle between the direction of
emergence and the normal to the cylinder.

4. The attenuation inside the crew space is pure

surface radiation density is constant

angulor

inverse square.

5. The detector is isotropic.

Under these conditions the angular distribution
may be represented as an isotropic component plus
a series in odd powers of the cosine. The contri-
bution of each term at a given point in the com-
ponent in the series may be expressed in an even
polynomial in x (see Fig. 11.3), where x is the
distance of the given point from the axis-of the
cylinder divided by the radius of the cylinder.
For pure cosine distribution the reading of an
isotropic detector is independent of position in the
crew compartment,

 

8 A. Simon and H. E. Stern, Some Calculational Methods
for Air) and Ground Scattering, ORNL CF-54-8-103 (to be
issued).

 

T

CROSS SECTION OF
CREW COMPARTMENT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

x=h/R i
cos 8
L
ol bl L
\
P = RADIATION PER cm® OF INSIDE SURFACE
PER STERADIAN IN THE NORMAL DiREGTION
O.B e L ,...!.,UA AAAAAAAA deinaienn ...____L .......
0.6 [ofeoeereerfressseeferen G 2
E': “'“L*--fi_ \\\<v co8
> —
= T~ \
Z 0.4 B T~
z TN N
2 \< cos 8 \
\\
0.2 ‘ -------- T~
o L L4 b

 

 

 

 

 

Fig. 11.3. Variation of Radiation Intensity with
Distance from Center of Crew Compartment.

163
ANP QUARTERLY PROGRESS REPORT

12. LID TANK SHIELDING FACILITY
G. T. Chapman

J. M. Miller

W, Steyert

D. K. Trubey
Physics Division

J. B, Dee
Pratt ond Whitney Aircraft Division

Preliminary work for further circulating-fuel
reflector-moderated reactor shield experiments at
the Lid Tank Shielding Facility (LTSF) has con-
tinved with in the ORNL Graphite

Reactor of a somple of the UF -C, 7 ¢ mixture

irradiation

that may be used to simulate the reactor fuel in
the shield mockup. Also, a new effective removal
cross section for carbon has been obtained for the

case in which the carbon is distributed uniformly
throughout the shield.
Thermal- and fost-neutron measurements have

been made around an array of three of the GE-ANP
helical air ducts, and a 35-duct array is being
Difficulties
in fabricating the enriched uranium plate for the
new LTSF source have delayed the completion of
this project; however, it appears at present that
these difficulties have been surmounted. It is
anticipated that the installation will he completed
within a month.

assembled for further measurements.

REFLECTOR-MODERATED REACTOR
AND SHIELD MOCKUP TESTS

J. B. Dee

D. K. Trubey W. Steyert

A second series of mockup tests for the circu-
fating-fuel reflector-moderated reactor (RMR) and
shield is being initiated at the LTSF.! For this
series o larger tank (approximately a 6-ft cube) has
been constructed to hold all the dry components of
the configurations, as well as an exponsible plastic
bag for containing boroted water. The beryllium
blocks to be used in the mockups will also be
placed inplastic containers for additional protection.

The dry tank face adjacent to the source plate
has a ]/B-in.-’rhick Incone! window, which corre-
sponds to the RMR core shell. In order to determine
the effect of the Inconel on the gamma-ray dose,

 

TFor first series see C, L. Storrs et al., ANP Quar,
Prog. Rep. Sept. 10, 1953, ORNL-1609, p 128.

164

gamma measurements were taken in pure water in
the tank. The dose was higher by a foctor of 2
than that normally observed in the LTSF, and the
increase agrees closely with the calculated 9-Mev
capture gomma-ray dose from the Inconel.

I addition to the presently planned static fission
source tests for RMR designs, a dynamic fission
source test is being considered for measuring the
sodium activation from delayed neutrons released
in the heat exchanger and the attenuation of gamma
rays from shori-lived fission products. A liquid
being considered is C;F,  containing 20 wi % UF .
[n cooperation with the Radiation Damage group of
the Solid State Division, a sample containing natural
irradiated in a fluorinated nicke!
capsule in the ORNL Graphite Reactor for an inte-
grated flux of 10'7 avt, which compares with an
integrated flux of 10'3 nvt for the enriched fuel.
After the volatile products had been removed by
vacuum distillation, a precipitate containing most

uranium was

of the radicactivity and a large part of the uranium
was found in the irradiated capsules; thus, this fuel
mixture could not be used for long exposures. A
similar capsule test is being prepared for a shorter
exposure, and alternate solutions are being explored.

EFFECTIVE REMOVAL
CROSS SECTION OF CARBON

D. K. Trubey

Megsurements of the removal cross section of
carbon have been made in a continuous carbon
medium obtained by dissolving sugar (C,,H,,0,,)
The solution (density = 1.312 + 0.001
g/cm3) contained 64.2 wt % sugar, which gave
0.354 g/ecm® of carbon and o hydrogen and oxygen

in water,

density that was 96% of that of plain water. It was
contained in a large tank that had a I/B-in.-thick
Inconel window on the source side.

The thermal-neutron flux, the fast-neutron dose,
and the gomma-ray dose are shown as functions of
distance from the source in Figs. 12.1, 12.2, and
 

 

 

 

 

 

 

 

 

 

 

Va CigHaply a0

‘ | ‘

. /{"“_ J.::f;;; T CTTTIIT TV L
CHRHOT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THERMAL-NEUTRON FLUX {nv,, )

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 26 40 60 80 100 120 140 60
Z, DISTANCE FROM SOURCE (em)

Fig. 12.1.
tion.

Thermal-Neutron Flux in Sugar Solu-

12.3, respectively. Plain water curves are also
shown for comparison. The gamma-roy water curve
is higher than the normal LTSF water curve because
of the high-energy capture gamma ray from the
Inconel window of the tank (see preceding discus-
sion).

Since the medium contained almost as much hy-
drogen and oxygen os does plain water and con-
tained them inthe same ratio, no geometric correction
was made in calculating the effective removal cross
section, o . The average o_ for the range of 90 to
140 cm from the source was 0.750 barn. This com-
pares with a value of 0.81 £ 0.05 barn froman LTSF
measurement behind a slab of graphite that con-
tained 51.3 g/cm?® of carbon, which corresponds to

PERIOD ENDING SEPTEMBER 10, 1954

ORNL~LR-DWG 2649

 

 

 

 

 

 

& 2 -

~

1.

¥

E

0 1™

Q

0

=

0

-

B

=

1

—

192

& N o
2 - — b fil I

CyoHpz 0y M0 \\

    

 

 

 

 

 

 

 

 

5l et \ e
7”......““”,.‘ ............................................. ‘,i.,vr...fi‘_‘}.. —_—
2 e e e il —_— —
13 !
70 80 20 100 110 120 130

Zz, DISTANCE FROM SOURCE {(cm)

Fig. 12.2. Fast-Neutron Dose in Sugar Solution.

the sugar-water solution at 145 ¢m,

fn order to observe the neutron attenuation in a
medium removed from the source, the sugar-water
solution is being placed in a 36-in.-long aluminum
tank located 48.2 cm from the source. Thermal-
neutron, fast-neutron, and gamma-ray measurements
will be made in the tank. The measurements at
the interface will indicate the difference in neutron
age between water and the sugar solution for the
more penetrating of the fission neutrons.

165
ANP QUARTERLY PROGRESS REPORT

GE.ANP HELICAL helical air ducts at the LTSF. The ducts (Fig.
AIR DUCT EXPERIMENTATION 12.4) were fabricated from flexible steel conduit
J. M. Miller shaped around a 9-in. core. After removal of the

core, the ducts were stiffened by Fiberglas wrapping.
Thermal-neutron measurements have been made

around 3-in.-dia by 4-ft (developed length) GE-ANP

ORNL—LR-DWG 2651

 

— 10
4 ORNL—LR—DWG 2650 A
0 - —_ - e e e e JP i

 

 

 

 

 

 

 

 

 

 

 

 

GAMMA-RAY DOSE {mr/hr}

 

 

 

 

 

 

 

THERMAL-NEUTRON FLUX (nvih)

60 70 80 90 100 10
7, DISTANCE FROM SOURCE (cm)

 

 

 

 

Fig. 12.3. Gamma-Ray Dose in Sugar Solution,

 

 

YNCZLASSIFIED
GT - GNP PHGTO U-32438

 

 

 

 

 

 

 

 

 

 

 

120 130 140 150 160
z, DISTANCE FROM SOURCE {cm)

 

 

G e e
Fig, 12.5. Thermal-Neutron Flux Measurements
Beyond One and Three GE-ANP Helical Air Ducts —
Fig. 12.4, GE-ANP Helical Air Duct. Horizontal Traverses.

166
PERIOD ENDING SEPTEMBER 10, 1954

Two configurations, o single duct and a friangular  neutron traverses parcllel to the source plate. An
array of three ducts (5 in. from center to center), array of 35 ducts is to be placed in a medium of
were placed in the LTSF with one end adjacent to  steel Raschig rings (35 vol %) and borated water
the source plate. A comparison of the thermal-  for a study of the effect of the addition of ducts to
neutron flux beyond the two configurations is given  gamma shielding,

in Fig. 12.5, and Figs. 12.6 and 12.7 are thermal-

  
 

 

 

 

 

 

     
    

 

 

 

 

 

 

 

 

 

 

     

 

    

 

      

 
 

 

 

 

 

 

 

 

 

g

S ORNL--LR--DWG 2653

ORNL-LR—DWG 2852 ' B T

I JnllJln{ ,,,,, T[ T I b ]

,,,,,,,,,,,,,, S TN SR ST PR SR e i
i e
S I O T N
T
‘= o:c £ ( E
£ .| HOCURVE AT £
£ =
x 3
2 i@
L z
- g
E [ DA = R0 em T NGIN >
wl . - =
z H ‘e ‘ e e 4
3 e e N g
5 ‘ _ N &
5 5 e — 0 g N e L:"-:|
u ~H,0 CURVE AT z =130 om | =
= 9 /A 1 T,,_f_f [ER —_1 N e ]
: _ ™
{/' —H,0 ATl z =140 cm \‘
i - ._..._._,‘WL_..-.._._..__ _...__.[....“L..n..______.“............_.t.._--___‘.
-70 ~50 -30 -0 10 30 50 70

 

-70 =30 -3 -0 t0 30 50 70 ¥, VERTICAL DISTANCE FROM SOURCE AXIS (cm)
¥, VERTICAL DISTANCE FROM SOURCE AXIS (cm)

Fig. 12.7. Thermal-Neutron Flux Measurements
Fig. 12.6. Thermal-Neutron Flux Measurements  Beyond Three GE-ANP Helical Air Ducts ~ Vertical
Beyond One GE-ANP Helical Air Duct — Vertical  Traverses.

Traverses.

167
ANP QUARTERLY PROGRESS REPORT

13, BULK SHIELDING FACILITY

R. G. Cochran

F. C. Maienschein

G. M. Estabrook K. M. Henry
J. D. Flynn E. B, Johnson
M. P. Haydon T. A Love

R. W. Peelle

Physics Division

At the Bulk Shielding Facility (B35F) measure-
ments were made of reactor radiations through
thick slabs of graphite, In this experiment the
fast-neutron spectrum through graphite and a re-
moval cross section of carbon were also determined,
The light to be given off from a nuclear-powered
airplane has been further investigated, and some
quantitative measurements are reported. In ad-
dition, o method of determining the power of the
ARE by means of fuel activation technigues is

described.

REACTOR RADIATIONS THROUGH
SLABS OF GRAPHITE

R. G. Cochran J. D, Flynn
G. M. Estabrook K. M. Henry

Measurements been completed for de-
termining the attenuation of large thicknesses of
graphite next fo a reactor, !

have

These measurements
are of interest for evaluating a graphite retlector as
a shield component, and they also provide a direct
comparison with LTSF determinations of the carbon
removal cross section. Graphite thicknesses of
1, 2, and 3 ft were used, and the usual gamma-ray,
thermal-neutron, and fast-nsutron dosc measure-
ments were made behind each slab thickness., In
addition, the fast-neutron spectrum (above 1.3
Mev) through 1 ft of graphite was measured,

A graphite slab 1 ft thick and one 2 ft thick were
constructed, and the 3-ft thickness was obtained by
strapping thess two slabs together. When strapped
together, there was no possibility of water getting
between the slabs. Fach slab was constructed
fargely of graphite blocks, 4 in. x 12 in, x 5 ft,
stacked
there could be no streaming of radiation through
cracks extending through on entire slab thickness.

in aluminum tanks in such @ way that

 

Vihe details of this experiment will be reported in a
memorandum by R. G. Cochran et al., Reactor Radi-
ations Through Slabs of Graphite, ORNL CF-54-7-105
(to be issuzd).

168

Thin graphite shims were added to fill the tanks,
which were sealed by heliarc welding and were
pressure tested for leaks., The total aluminum wall
thickness for each tank was 1.3 cm,

Fulk Shielding Reactor loading No, 26 (Fig. 13.1)
was used. This configuration is very similar to
loading 22, which has been studied in considerable
detail for other experiments,? but loading 26 uses
less fuel to compensate for the presence of the
graphite, which is a better reflector than the water
it replaces. Since the presence of the graphite
perturbed the neutron flux in the reactor core
rather severely, the neutron flux distributions and
thus the reactor power were determined in the
usual way by means of cobalt foils,

 

2R, G. Cochran et al., Reactivity Measurements with
the Bulk Shielding Reactor, ORNL-1682 (to be issued).

OSNL-—LR 7056 2654

CREACTOR GR.O FLATE
STANDARD FUEL ASSENSLY

,SAFETY RODS

   

 

e
S N o

 

N M ELEMENT

WAaTER REFLECTOR'

¥, CLEMENT :
CONTROL ROD

fig. 13.1.
Reactor.

L.oading 26 of the Bulk Shielding
The thermal-neutron flux {(Fig. 13.2) was measured
behind each of the three graphite slab thicknesses
to a distance of 300 cm from the reactor face.
Measurements were made with a 3-in. U?3> fission
chamber, an 8-in. BF; counter, and a ]23’2-in. BF,
counter, and the data were normalized to indium
foil data.

Fast-neutron measurements (Fig. 13.3) were
made with o three-section neutron dosimeter,
Gomma-ray measurements (Fig. 13.4) were taken

 

ORNL-LR-DWG 265
S

 

 

 

 

3-in, FISBION CHAMBER,
" B-in. BF, COUNTER, AND
= 2% -in. By COUNTER;
DATA NORMALIZED

 

 

B T F

—
T

THERMAIL-NEUTRON FLUX {n,, /watt)
o

g,
W

 

 

1
21 OF

........

 

 

 

G AC’HITE”

“,,u,.,,,,,...,.*,,,,

 

 

 

 

 

.....

 

 

T TO INDILM FOIL DATA

 

 

 

A ft OF FRAPH\TF’ TUNARON T T

 

 

400
DISTANCE FROM REAGTOR {cm)

 

 

 

 

 

 

 

 

 

   
 

 

 

 

Fig. 13.2. Thermal-Neutron Flux Measurements
Behind AGHT Graphite.

PERIOD ENDING SEPTEMBER 10, 1954

behind each of the three graphite slab thicknesses,
both to determine the magnitude and relaxation
lengths of the gamma-ray dose through graphite and
to ensure that none of the neutron detectors was
being used in a too high gamma-ray field, which
would have caused their readings to be high.

The fast-neutron spectrum from 1.3 to 10 Mev was
measured with the BSF fast-neutron spectrometer,®
A spectrum with reasonable statistics could be
obtained only through the 1-ft graphite slab; the
neutron intensities through the 2- and 3-ft sections

were too loew. The spectrum measured with the

 

3R G. Cochran and K. Henry, A Proton Recoil Type
Fast-Neutron Spectrometer, ORNL-1479 (April 2, 1953).

 

 

 

 

 

 

 

FAST-REUTRON DOSE {mrep/hr/watt}

 

 

 

 

   
 
 
 

GRAPHITE SLABQ_", 7
I8 ALUMINUM TANKS

  
 

  

A0 &0 30 GO
DISTANCE FROM

a0

20
REACTOR lam)

0 20

Fig. 13.3. Fuast-Neutron Dose Measurements Be-
hind AGHT Graphite.

169
ANP QUARTERLY PROGRESS REPORT

ORNL-LR-DWG 2657

10
.. 1t OF GRAPHITE
. .
BN S
\_\( - 2 H OF GRAPHITE
\ ST
= 1 N - 3 ft OF GRAPHITE |
et N ]
< : : ,_,’% - o !
L e , |
L A A |
; R \
a . . . P & .. .. .- i
> Ve
& ) /\‘\[/ efi
I . . . . &
5 ‘ N
2 : 50-cc ION CHAMBER ; \i\ i
", \
h \
\‘(-\‘ |
- - -fSLAE Ry
- 2-fislAg | | SR
v BOTH SLABS ’ | S . l
. l ;
i ‘ ‘
‘ GRAPHITE SLABS
| INALUMINUM TANKS
LO-4 I l . ] - . - . oo
0 40 80 120 160 200 240

DISTANCE FROM REACTOR {cm)

Fig. 13.4. Gamma-Roy Dose Measurements Be-
hind AGHT Graphite.

1-ft graphite thickness between the reactor face
and the end of the spectrometer collimator is shown
in Fig. 13.5. The spectrum measured with the end
of the collimator against the reactor face is also
shown.

Removal cross sections have been calculated for
each thicknass of graphite by use of a method
suggested by E. P. Blizard.* The resulting cross
sections are 0,82 barn/atom for the 1-ft slab,
0.84 barn/atom for the 2-ft slab, and 0.80 barn/atom
for the 3-ft slab, These removal cross sections are
in good agreement with measurements made on
graphite at the L. TSF,

REACTOR AIR GLOW

R. G, Cochran T. A. Love
K. M. Henry F. C. Maienschein
R. W, Peelle

Attempts to theoretically determine the amount of
visible light which may surround a nuclear-powered
airplane in flight have resulted in widely differing

170

values,®s® Therefore an experiment was performed
at the BSF to provide an experimental basis for
future estimates.

The end of an air-filled aluminum periscope tube
was placed at the reactor face, and the amount of
light produced in the tube was measured by a
photomultiplier with spectral response
to that of the average human eye. Other relative
measurements were taken with a photomultiplier
which was sensitive chiefly in the blue and near-
ultraviolet range., The latter measurements are
plotted in Fig. 13.6 as functions of the air pressure
in the tube, Measurements with argon in the tube
demonstrate that neither the approximate amount of
light produced nor the exact spectrum emitted is
strongly dependent on the atomic number of the
gas. It is also interesting to note that the light
production in a given volume of air appears to have
a maximum at a pressure corresponding to an
altitude of about 30,000 fi.

It is demonstrated in Figs. 13.7 and 13.8 that
the air glow is largely caused by gomma radiation
rather than by neutrons. Figure 13.7 shows the
decay of the light plotted along with the decay of
the reactor gamma ion chamber current just after
reactor shutdown. The attenuation by water of
the radiation which produces the air glow is shown
in Fig. 13.8. This attenuation rather closely
follows that of gamma rays.

similar

The photomultiplier was used to compare the
quantity of light given off in the air-glow tube with
that from a small tungsten lamp mounted at the
reactor end of the tube, This comparison showed
that 7.2 x 107° lumen was given off by the glow
for a reactor power of 100 kw and atfmospheric
pressure, Presumably, the amount of light should
be proportional to the integral of the gamma-ray
dose rate over the volume of the air inthe measuring
tube. This integral wos estimated to be 1.1 x 1010
(r-em)/hr. Therefore the effective light production
per unit volume of air is

L = 6.5 %1071 (lumen/cm3)/(r/hr) .
If it is assumed thot all the light is given off at

4E, P. Blizard, Procedure for Obtaining Effective
Removal Cross Sections from Lid Tank Data, ORNL
CF-54-6-164 (June 22, 1954).

5T, A. Welton as quoted by C. E. Moore, Visual
Detecrability of Aircrajt at Night, LAC-15, p 24 (Aug.
14, 1953).

6). E. Faulkner, Visible Light Produced in Air Around
Reactors, ORNL CF-54-8-99 (tc be issved).
PERIOD ENDING SEPTEMBER 10, 1954

ORNL-—@ 2658

FLUX (neutrons/cmZ/sec/watt/ Mev)

4-ft COLLIMATOR AGAINST
i{-ft OF GRAPHITE

 

0 2 4 6 8 10 12 14
ENERGY (Mev)

Fig. 13.5. Fast-Neutron Spectrum of the Bulk Shielding Reactor Observed Through 1§t of Graphite.

171
ANP QUARTERLY PROGRESS REPORT

ORNL -LR-DWG 2859

EQUIVALENT ALTITUDE ()
60,000 40,000 20,000

 

9
7 x10 T i T T
| ;
‘ i
\
\

 

 

 

 

—3 0O

 

 

 

 

 

 

 

 

 

 

|
E

 

 

LIGHT CURRENT {amps)

 

 

 

 

S A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PRESSURE (cm Hg)

Fig. 13.6. Variation of Glow with Pressure,

4600 /i, which is in the blue region,

7.2 x 10=3 lumen = 1,9 x 10~ 7 watt of visible light.

The amount of air glow can then be characterized
as the fraction of energy transformed into visible
light:

watts of visible light in tube

watts of energy dissipoted in tube by gomma radiation

1.9 x 10~¢

T e = 6 x 107 for atmospheric pressure.
3.2 x 102

This experimental value of light yield is in disa-
greement with the theoretical estimates of Welton®
(0.05) and of Faulkner® (2 x 107%), In neither of
these calculations was the effect of pressure
considered,

172

The value of L may be used to calculate the
light emission from a typical nuclear-powered
aircraft, For a reactor power of 200 Mw and a
reactor shield of 147 em of water (similar to the
shield design of the 1950 ANP Shielding Soard?),
the BSF measurements indicate that the gamma-ray
dose just outside the edge of the shield would be
1.36 » 10% r/hr with a relaxation length of 22 cm.
The dose integrated over space outside the shield
is 1.2 x 10'5 (recm®)/hr. Multiplying this by L,
the luminosity would be 75 lumens, which would
have roughly the equivalent brightness to the eye
of a 10-w (+7 w) incandescent light if the aircraft
were far enough away to appear as a point source

 

7Report of the Shielding Board for the Aircraft Nuclear
Propulsion Program, ANP-53 (Oct. 16, 1950).
ORNL~LR--DWG 2860

 

 

  
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

—~8
107 L . ;
Lo e ‘E ' ]
I Iffifiifiifififi:[fifiiff’J'f'","fif",fffffifffififi.{fffifififfifi
............. o ]
5 | e Jo AIR GLOW — ]
» e GAMMA RADIATION
1’?
T B T ]
5 ;
]
= ;': PRESSURE = 30cm Hg
2
- g \
= |
T 5 '
r:_l:) o
& {5
- M
I =
Qo
53
<
o
=
=
=
=
19
w0 I
0 5 10 15 zZ0 25 30 35
TIME (min}

Fig. 13.7. Decay of Air Glow and Reactor Gumma
Rays.

of light., This large uncertainty is primarily at-
tributable to the low sensitivity of the eye-equiva-
lent photomultiplier. A more detailed discussion
of these preliminary velues will be given in a
forthcoming report,®

FUEL ACTIVATION METHOD FOR POWER
DETERMINATION OF THE ARE

E. B. Johnson

When the ARE goes into operation, many measure-
ments of its performance will be made. Among
them will be the power level at which it operates,
which will be measured in more than one way. A
method suggested by J. L. Meem is based on the
measurement of the relative activity of fuel samples
exposed in the ARE and in a known flux in another
reactor (BSR}). An experiment has been initiated
at the BSF to implement this suggestion.

 

8F. C. Maienschein et al., Measurements of a Reactor-

Induced Air Glow, ORNL CF-54-9-1 (to be issued).

PERIOD ENDING SEPTEMBER 10, 1954

ORNL-LR—DWG 2661

 

 

 

   
 
 

 

1
oo AR GLOW . . § -
¢ (ATMOSPHERIS PRESSURE )

 

 

 

LIGHT CURRENT {amp)

GANMMA~RAY DOSE [{r/hr) /watt]

FAST NEUT
ol (arbitrary units) N 1.

o 10 20 20
DISTANGE FROM REACTCR {cm)

 

 

 

 

Fig. 13.8. Attenuation in Water of Radiation
Causing Air Glow,

The activity resulting from the irradiation of
fissionable material is not easy to predict because
of the large number of isotopes for which calcu-
lations would be required. However, for identical
flux, exposure time, and waiting period after
exposure, the specific activity of two samples
should be the same. Furthermore, at low neutron
fluxes the activity should be proportional to the
flux.

[t is contemplated that the ARE will be operated
at a nominal power level of gpproximately 1 w for
1 hr, that it wiil then be shut down, and that a
sample of the irradiated fuel will be withdrawn for
counting and analyzing, The ARE experimental
program will then proceed as scheduled. A decay
curve will be run on this fuel sample from the ARE
and compared with the decay curve obtained on o
similor fuel capsule irradiated for the same length
of time in the BSR in a known neutron flux of about
the same magnitude. From the relative activity of
the two samples, the operating power of the ARE

173
ANP QUARTERLY PROGRESS REPORT

can be determined readily, For the power com-

parison it will be necessary to know the uranium
content of the sample and of the ARE, as wel! as

the energy per fission.” A small correction will

95 L. Meem, L. B. Holland, and G. M. McCammon,
Determination of the Power of the Bulk Shielding
Reactor — Part [Il. Measurement of the Energy Released

per Fission, ORNL-1537 (Feb. 15, 1954).

174

be necessary for the local depression of flux by
the uranium sample. This factorhas been estimated
by W, K. Ergen to be 0.80. Details for the appli-
cation of this method are given in a separate
repori’.]D

10¢, 3, Johnson, Fuel Activation Methods for Power
Determination of the ARE, ORNL CF-54-7-11 (in press),
PERIOD ENDING SEPTEMBER 10, 1954

14. TOWER SHIELDING FACILITY

C. E. Clifford
T. V. Blosser J. L. Hull
L. B. Holland F. N. Watson

Physics Division

D. L. Gilliland, General Electric Company
M. F. Valerino, National Advisory Committee for Aeronautics
J. Yan Hoomissen, Boeing Airplane Company

The Tower Shielding Facility (TSF) experimental
program has, thus far, included measurements of
ground- and air-scattered fast neutrons and the
development of a new procedure for the determi-
nation of the power of the reactor. Tests on the
GE-ANP R-1 divided-shield mockup haove been

started.

FAST-NEUTRON GROUND AND AIR
SCATTERING MEASUREMENTS

T. V. Blosser J. Von Hoomissen

D. L. Giililand F. N. Watson

The performance of neutron and gamma-ray air
scottering experiments that are free from an ex-
cessive background of ground-scattered radiation
is a primary objective of the TSF, Therefore,
measurements of the scattered fast neutrons aos
a function of reactor-detector altitude were neces-
sary to determine the contribution of ground-
scattered radiation to the total flux, particularly
at the maximum altitude. These measurements will
help to indicate the magnitude of the ground-
scattered neutron background to be expected in
future differential experiments, and they will aid
in an understanding of the variation of ground and
air scattering as the ground is approached.

Measurements of the thermal-neutron distribution
were token in the detector tank, which is es-
sentially a 5-ft cube of water and which was, for
these experiments, situated 64 ft from the reactor
tank. The reactor was placed at an angle ¢ of
330 dey from the 4 axis (Fig. 14.1), and a BF,
counter was moved along a line normal to and near
the right side of the detector tank. In this region,
contributions from other faces of the tank were
negligible; thus the neutrons detected by fthe
counter were the air- or ground-scattered fast
neutrons which entered the side wall and were
thermalized in the water near the detector. The
reactor and detector tank altitudes were varied
simultaneously, in discrete steps, from 0 to 195 ft.

T R WA A ——— i

A composite plot of the measurements {(Fig. 14.2)
indicated only small differences in slope but ap-
precicble differences in magnitude between the
curves for the various altitudes. A plot of the
flux vs the daltitude (Fig. 14.3) showed o .pro-
nounced peck in the region. between the 15- and
20-ft altitudes, which indicated the importance of
ground-scattered neutrons in this region. [t should
be noted that the reading ot zero altitude was
obtained with the reactor half-immersed in the
ground pool, the upper half being shielded as
before; so essentially half of the source was
occluded, as was half of the scattering medium.
Thus the air-scattered neutrons should be no more
than half of the value at full altitude and no less
than one quarter. Becnuse the intensities at the
150- and 195-ft altitudes were the same, it was
concluded that the ground-scattered contribution
at these altitudes was small,

fn a preliminary analysis of the dota, it was
estimated that this contribution was 2 to 5% of
the total scattered neutrons. While the detailed
caleulations of ground and air scattering being
undertaken in connection with this investigation
are not yet completed, o preliminary comparison
of the dato with calculations carried out ot the
Boeing Airplane Company for a not too dissimilar
situation has been made. The qualitotive agree-
ment is very good, although the pedk in the total
flux, as measured, seems to be at o lower altitude.
A report on this experiment is being prepared.!

CALORIMETRIC REACTOR POWER
DPETERMINATION

D. L. Gilliland L. B. Holland
3L, Hull

A procedure hos been developed and tested for
a colorimetric determination of the power of the

 

YeL B Clifford e al., A Preliminary Study of Fast-

Neutren Ground and Air Scattering at the Tower Shielding
Facility, ORNL CF.54-8.95 (1o bs issued).
ANP QUARTERLY PROGRESS REPORT

  
 
   

REACTOR TANK

d — distance from the center of the reactor tank to

the 0, 0, 0 point.

b — altitude of both the 0, 0, 0 point and the center
of the reactor.

@ — horizontal angle between p and 4 axes.

a — detector orientation angle, horizontal angle

between 4 axis and a perpendicular to the
broad side of the triplet BFF, chamber,

UNCLASSIFIED
CRNL-LR-DWG 843

 

DETECTOR TANK

iLEFT SIDE

 

DETECTOR

  

RIGHT SIDE
TOP VIEW OF DETECTOR TANK
SHOWING DETECTOR ORIENTATION ANGILE @

0, 0, 0 point location — on the outside of the de-
tector tank 2.5 ft above the inside bottom of the
tank.

x, y, z — coordinates of geometric center of de-
tector. Since the center of detection of the various
counters will shift with the location of the detector
in the tank, the x, y, z coordinates indicate the
geometric center of the detector.

p — thickness of water between outer reactor face
and reactor tank wall, measured along a radius.

Fig, 14.1. Geometrical Convention Adopted for Tower Shielding Facility Experiments.

176
RELATIVE THERMAL-NEUTRON FLUX (ARBITRARY SCALE)

PERIQD ENDING SEPTEMBER 10, 1954

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g
3 ORNL~-LR-DWG 2772
10 | I [ 1
i l | |
_________ _ A = VARIABLE 8 = 330 deg CENTER OF TRIPLET BF, CHAMBER AT: __|
- d = 64 ft = —90 deg x = VARIABLE y = 79.6cm
S T p=452cm z=00cm T
> CURVE  ALTITUDE
1 15 ft
‘ , o 25 ft
3 8 fi
o2 \\\ 3 4 4t
SON 4 5 50 ft
\\\\\\ e 6 96 f1
6 7 150 ft
\i\\\\)K - 8 197 ft
3 \ \\\ 9 O ft
N\ ‘\\\\\
N\
NN
, N\
N \\
XA 1
10 X OO 3
87 A AN LT 5
N \\\\\>( -
9 N NN\
5 N\ "“\\\\\\
\\ \\\\
\ \
2 t
1
-75 ~65 -55 ~45 -35

X, HORIZONTAL DISTANCE FROM ¢ AXIS TQ DETECTOR CENTER (cm)

Fig. 14.2. Attenuation of Scattered Neutron Flux in Detector Tank ot Various Altitudes.

177
ANP QUARTERLY PROGRESS REPORT

SECRET
s—0I-056-2- 72

 

 

 

 

 

O S T T
|/ \ . . ' ‘ . |
o % ) _ \ . ¥ =-67.5 cm ,
sl -8
S 5 ./ \ . LT v—m___L,,__,, _________________ % i
w .
> ' \ . x=-65cm
[ R
= e s
& L T ;
m / \o
z Ll N
: 2 h. - . . .
> \__\. ¥=—60 cm
% “—""*‘“"J—“'-—-uo----u—v "
Lt
}__
=
_a = -390 deg . A _
=z = 54 ft o - - . . e
& L g=sd _ A S
5 p =452 cm
= . 6 = 330 deg ‘ _ ‘ ,
“ CENTER OF TRIPLET BFy CHAMBER AT:
E x = SEE ABOVE tYy-in IRCN WALL AT x=-77.5cm)
£ y=79.6 cm o . i .
W z=02¢cm ‘ '
>
< e
o
01 : - e —
0 24 18 72 96 120 444 168 {92 215

&, ALTITUDE (i1}

Variation of Scottered Neutren Flux

Fig., 14.3.
with Altitude,

TSR,
water
are made while the water is being stirred suf-
ficiently to ensure that no significant difference
in temperature exists throughout the tank. The

Thermocouple measurements of the bulk
temperature in the 12-ff-dia reactor tank

rate of temperature rise of the water when the
reactor is started is then proportional to the power
of the reactor. Small corrections must be made
for the power input due to the mixing and for the
heat loss.

This procedure was first used for a 5 by 6 fuel
element loading in the TSR. The reactor tank
was lowered into the drained handling pool, and
four mixers were placed at various positions within
the tank. The turbulence created by the mixers
was sufficient for thorough mixing of the water,
but the power added by the mixers was not enough
to give a measurable increase of the water temper-

ature. Before the reactor was started, the temper-

178

ature remained constant for more than 1 hr with
the mixers running.

Soine heat insulation was obtained by creating,
with a double layer of canvas, a dead-air space
above the water and around the reactor tank. The
temperature rise during a reactor run was measured
by 12 thermocouples placed in the tank. The
thermocouples had previously been calibrated at
temperatures of 0, 25, and 50°C and agreed to
within 0.1°C.

The first run of the experiment was started at
24°C, and the total temperature rise was approxi-
mately 14°C.
higher temperature level but with the some heat
input initial
temperature for this run was dpproximately 15°C
above the ambient temperature, and the rise was
again 14°C. The cooling rate ot the conclusion
of this run was 0.3°C/hr. In a third run the initial
temperature was brought 15°C below the ambient
by cooling the reactor tank with ice before the run.
The results of the three runs were consistent to
within 1%. A detailed report of this experiment
is being prepared.?

The experiment was repeated at a

and total temperature rise. The

GE-ANP R«1 DIVIDED-SHIELD MOCKUF TESTS

T.V. Blosser J. Van Hoomissen

D. L. Giliiland . N, Watson

Tests on the GE-ANP R-1 divided-shield mockup
(Figs. 14.4 and 14.5) will begin with traverses
around the reactor-shield section in the handling
pool for comparison with measurements made at
the BSF.3 Measuraments will then be made in
the TSI detector tank, which will be situated 64
from the reactor shield ot an altitude of 195 fi.
[t is anticipated that the detector tank will be
replaced by the crew-compartment mockup in mid-
September.

 

2c. e, Clifford et ul., Calorimetric Power Determi-
ngtion of the Tower Shielding Reactor, ORNL CF-54-8-
105 (to be issued).

3. E. Hungerford, Bulk Shielding Fuacility Tests on
the GE-ANP R-1 Divided Shield Mockup, ORNL CF-54-8-
G4 (to be issued).
PERIOD ENDING SEPTEMBER 10, 1954

s
o
N
O N
T
Sl
e
a

¥ . o
oy 2L

 

-ANP R-1 Divided-Shield Mockup.

GE

Fig. 14.4. Reactor-Shield Section of the

179
 

Fig. 14.5. Crew-Compartment Section (Ends Removed) of the GE-ANP R-1 Divided-Shield Mockup.

180
Part IV

APPENDIX
REPORT NO.

CF-54-7-11

CF-54-7-143
CF-54-7-144
CF-54-8-171

CF-54-6-201
CF-54-7-1

CF.54-7-187

CF-54-8-200

CF-54-7-166

CF-54-8-199

CF-54-7-159

CF-54-8-180

CF-54-8-27

CF-54-4-188

CF-54-7-145

15. LIST OF REPORTS [SSUED DURING THE QUARTER

TITLE OGF REPORT

l. Aircroft Reactor Experiment

Fuel Activation Method for Power Determination of the ARE

ARE Operating Procedures, Part |, Pre-Nuclear Operation

ARE Operoting Procedures,

Critical Mass of the

A Reactor Design P

Part H, Nuclear Operation

ARE Reactor

1, General Design

arameter Study

The Effect of the RMR Shield Weight of Varying the Neutron
and Gamma Dose Components Token by the Crew and

Comparison of the
ized Shield

Test Results and Design Comparisons for Liquid Metal-to-Air

Radiators

RMR Shield Weight to that for an ldeal-

High Conductivity Fin Test Results

I, Experimental Engineering

Morse Silent Chain Drive Test on ARE Pump Hot Shakedown

Test Stand

Aircraft Reactor Experiment Fuel-Removal System Mockup

IV, Critical Experiments

The First Assembly of the Small Two-Region Reflector

Moderated Reactor

The Second Assembly of the Small Two-Region Reflector

Moderated Reactor

V. Metallurgy

Preliminory Metallographic Examination of K-25 Heat Ex-

changer Loop

Vi. Heat Transfer ond Physica! Properties

Physical Property Charts for Some Reacter Fuels, Coolants,

and Miscellanecus

Measurement of the Therma! Conductivity of Melten Fluoride

Mixture No. 104

Materials (4th Edition)

AUTHOR

E. B. Jehnson
W. B. Cottrell
J. L. Meem
C. B. Mills

>

. H. Fox et al.

R. M. Spencer
H. J. Stumpf

H. J. Stumpf

R. E. MacPhersen
J. G. Gallagher

J

. E. Ahern

A. G. Grindell

J. Y. Estabrook

Dunlap Scort

Dunlap Scott

R. S. Crause
G. M. Adamson

S. I. Cehen et al.

W. D. Powers
S. J. Claiborne

DATE ISSUED

7-31-54
7-20-54
7-27-54
8-28-54

6-25-54
8-26-54

7-19-54

8.27.54

7-16-54

8-26-54

7-26-54

8-26-54

8-5-54

6-21-54

8-24-54

183
REPORT NO.

CF-54-8-10

CF-54-6-4

CF-54-8-135

CF-54-6-164

CF-54-8-95

CF-54-8-97
CF-54-8-98

CF-54-6-216
CF-54-6-217
CF-54-7-35

CF-54-7-162

184

TITLE OF REPORT

Measurement of the Density of Liquid Rubidium

VYil. Radiation Damage

Release of Xenon from Fluoride Fuels: Proposal for an Ex-

perimental Program

Heot Copocities of Compositions No. 39 and No, 101

Vill. Shielding

Procedure for Obtaining Effective Removal Cross Sections

from Lid Tank Data

Preliminary Study of Fast Neutron Ground and Aijr Scattering
at the Lower Shielding Facility

Calculation of Fission Neutran Age in NeZrF¢

Age Measurements in LiF
I1X. Miscellansous

ANP Research Conference of May 19, 1954
ANP Research Conference of June 8, 1954
ANP Research Conference of June 29, 1954
ANP Research Conference of July 20, 1934

AUT HOR
S. 1. Cohen
T. N. Jones

M. T. Robinson

W. D. Powers
G. C. Blalock

E. P. Blizard

Tower Shielding
Group

J. E. Faulkner
J. E. Faulkner

A. W, Savolainen
A. W, Savolainen
A, W. Savolainen

A. W. Savelainen

DATE ISSUED

B-4-54

6-2-54

8-17-54

6-22-54

8-23-54

8-31-54
8-31-54

6-25-54
6-25-54
7-2-54

7-26-54
THE AIRCRAFT NUCLEAR PROPULSION PROJEGT
THE OAK RIDGE NATIONAL LABORATORY

SEPTCMBER 1, 1954

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

C. R BALLOCK®
A R, sITESY

8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANP PROJECT DIRECTOR W. H. JORDAN RD
CO-DIRECTOR 5. J, CROMER RD
ASSOCIATE DIRECTOR R, I, STROUGH PW
ASSISTANT DIRECTOR A ) MILLER RO
D. HILYER, SEC, RD
REPORTS
ASSISTAMT TO DIRECTOR AL W. SAVOLAINEN ARE
L. 4, COOK ARE P, HARMAN, SEC. ARE o
H. MCFATRIDGE, SEC, ARE LITERATURE SEARCHES
A, L, DAYIS ARE
MRCRAFT REACTUR ENGINEE RING DIVISION
5. 1. CROMER, DIRECTOR RD
H. MCFATRIDGE, SEC. ARE SUPPORTING RESEARCH
¥, H. JORDAN
A J.MILLER
ASSISTANT TQ DIRECTOR
L
L. M. COOK ARE
1
| i i { ! I i i i
REACTOR PHYSICS ENGINEERING DESI Gal POWER PLANT ENGINEERING EXPERIMEN TAL ENGINEERING ARE OPERATIONS STAFF ASSISTANT STAFF ASSISTANT | STAFF ASSISTANT STAFF ASUSTANT STAFF ASSISTANT STAFF ASSISTANT
W. K. ERGEN ARE H. C. GRAY P¥ A, P. FRAAS ARE H. W. SAYAGE ARE E.5. BETTIS ARE W, R. GRIMES WC W.D. MANLY " l E. P. RLIZARD P A, J. MILLER RD W. K, ERGEN ARE H. F. POPPENDIEK REE
J. L. MEEM ARE
J. BENGSTON ARE A. A. ABBATIELLO ARE C. 5. BURTNETTE USAF ¥. G, COBB ARE P. HARMAM, SEC. ARE
R. K. OSBORN ARE J. Y. ESTABROOK ARE R. W. BUSSARD ARE 3. A CONLIN ARE
R. E. HELMS ARE W. T, FURGERSON* K-25 C. P. COUGHLEN ARE G, A, CRISTY EMm
COMPUTERS G L. HOLLIS ARE M, E. LAVERNE ARE R DREISBACH P R. E. HARRIS EM CHEMISTRY METALLURGY SHIELDING RESEARCH RADIATION DAMAGE REACTOR CALCULATIONS HEAT TRANSFER AND PHYSICAL
F. L. MAGLEY ARE F. W, MCQUILKEN ARE A. G, GRINDELL ARE E. R MANN® i W, R. GRIMES re . PROPERTIES RESEARCH
A. FORBES ARE C. A MILLS ARE C B MLLS ARE /. HELTON ARE DAY SHIET T - £ D BLIIARD z T e « & A ChaRPE ® H. F. POPPENDIEK REE
M. TSAGARIS ARE J. A. GLSON P G, SAMUELS ARE . R, HUNTLEY ARE P. A, AGRON ¢ G. M, ADAMSON M L. S ABBOTT P i . -
H. R. ROCK PW W. L. 5COTT ARE J. W, KINGSLEY ARE w. B, COTTRELL, SUPV. ARE C, J. BARTON MC c R BOSTON M J.EI FAULKNER P R. R. COVEYOU ARE
CONSULTANTS W, J. STELZMAN ARE E. M. LLEES ARE J. G. GALLAGHER ALC J. P BLAKELY MC w' H’ BRIDGES M . HI C. D. BAUMANN 85 C. L. GERBERICH" MP J. 0. BRADFUTE REE
B. M. WILNER AGC R, E. MacPHERSON ARE H. E. HUNGERFDRD P F.F. BLANKENSHIP MC "o R Do M F. H. HURRAY P W. E. BROWNING 55 M, . GIVEN* P S, 1. COHEN REE
T. ANDERSON M. M, YARDSH ARE L. A, HANN ARE $. C. SHUFORD P €. M. BLOGD we O e " ne.steRn & ¥. E. BRUNDAGE 55 W, E. KINNEY ARE N. D, GREENE REE
A. NOHEL, GEQRGIA TECH, E. M. EISSEMBERG, SEC.  ARE %. B. MCDONALD ARE A, L, SOUTHERN éERE g‘ :. :g;g:fl ME v AR ARF " R, RICKMAN, SEC 5- :‘ g:sgou- gg W. C. SANGREN MP g- 5 :S?LLJ&N :EE
W.R. C. STORRS P, M / . . , SEC. W ™
CONSULTANTS o DesoRN ARE - I . BRATCHER MC R B, CLAUSING . M. J. FELDMAN 5§ B LYNCH REE
DL L H. . E. . D, PALM
A. H. FOX, UNION COLLEGE O. F. SALMON 4RE ECHMICIARS ol e . H, COOK " CONSULTANT R fifixms e CRITICAL EXPERINENTS W, 0. POWERS Ree
W, LOWEN, UNION COLLEGE P.G. SMITH ARE 4 5. ADDISON ARE - X 1. H. DE VAN M H. A, BETHE, CORNELL UNIVERSITY U IR e MW, ROSENTHAL REE
R L. MAXELL E.$TORTD ARE J. R. CROLEY ARE D. R. CUNEQ e . DO AnaRE W C.L.HIL A. . CALLIRANY P . ¥
UNIVERSITY OF TENNESSEE W T TUNNELL ARE F. J. SCHAFER ARE F.L.DALEY e o n BOUGLAS . m hE :E:E:guz g: T. K. CARLSMITH, SEC. REE
6. F. WISL ICENUS D. R, WARD ARE C. F. WEST ARE J. E. EORGAN W 2 L HLESTAND " CONTRACTORS 8. % KEWLHOL 5 . 5. CRUDELE PW TECHNICIANS
PENNSYLVANIA STATE URIVERSITY D. ALEXANDER, SEC. ARE EVENING SHIFT H. A PRIZOMAN e E.E. HOFFMAN M METAL HYDRIDES, INC. M. T. MORGAN 55 9. L. DREENSTREET ARE C. 6. BLALOCK REE
D. HARRIS, STENO. ARE F. KERTLSZ He H. INOUYE " NUCLEAR DEVELOPMENT ASSOCIATES, . OLson « V. G, HARNESS* P R W BURNETT REE
D. STOREY, REC. CL. ARE G. D. WHITHAN, SUPY. ARE E.E.KETCHEN MC C.F.LEITTEN " iNC. 3 e 1. TN e . J. CLAIB
A. MONTGOMERY, CL. TYP.  ARE K. LESLIE ARE F. A, KNOX e R. B. OLIVER M M. . OSRORME 5 . H. MARABLE* P - J. CLAIBORNE REE
E. L. MORRISON USAF S. LANGER MC N B TRIAGCA " W. W. PARKINSON 55 1 W NOAKS Pw T. N. JONES REE
TECHNICIANS F. A, FIELD UsAF R, E. MEADOWS MC o n H. E. ROBERTSON 55 E. R. ROHRER* P J. LONES REE
T. ARNWINE ARE G. C. ROBINSON ARE R, P WMETCALE Mg K. W. REBER " LID TaNk W o s D. SCOTT, JR. ARE i M
G. 5. CHIL F.W. MILES MC - - - > R. M. SPENCER USAF -
! aRE TECHNICIANS R, E. MODRE M C- . SLAUGHTER " G. 7. CHAPHAN P DSCAR SISMAN 5§ DV, P WILLIAS® P
. M. CUNNINGHAM ARE T.E. CRABTREE ARE G, J. NESSLE e Wt STEIOLITZ " L. . TEMPLETON 5 E. L. ZIMMERMAN® P
J. R. DUCKWORTH ARE R. E.DIAL ARE R. F. NEWTOH RD A TABODA " 1. B. DEE PW C.C. WEBSTER 55 M. L. RUEFF,* SEC. P
%K. R, FINNELL ARE ., . JENKINS ARE L. G. OVERHOLSE R Us W, WOODS " J. M. MILLER P R. A WEEKS 55
W, D, GHORMLEY ARE £. . PERRIN ARE 4. 0. REDMAN M "1 THOMAS, SEC M D. K. TRUBEY r 3 & wieon o
R. 1. SHEIL ue - : .C.
C. J, GREEN e <
o L :EE NWGHT SHIFT M.V, SMITH e TECHNICIANS TECHNICIANS J. C. ZUKAS S5
R. A. HAMRICK ARE R. G, AFFEL, SUPY, ARE B. L STURM M W. A. ANDERSON M E. BECKHAM P TECHNICIANS
8. L. JOHNSON ARE 4. W. ALLEN ARE 1. E. SUTHERLAND MC G. D. BRADY “ R, TAYLOR e LLis 55
W. 1. MAYNARD ARE R.L. BREWSTER ARE R, E. THOMA e 2. TLEAST M 1%, WAMPLER . B
D. E. MCCARTY ARE J. M. CUNNINGHAM ARE L. E. TOPOL MC M. GOMZALEZ M J. SELLERS Ic F. M. BLACKSHER 55
6. E. MILLS ARE 4. J. HARSTON REE 1. TRUITT MC L L HALL M
C. C. NANCE ARE - G. #h WATSON MC 5. D, HUDSON W CONSULTANT
J. 1. PARSONS ARE TECHNICIANS C. F.WEAVER MC R. W. JOHNSON M D. F. WEEKES, TEXAS A & M COLLEGE
M. A, REDDEN ARE T. L. GREGORY ARE D. E. CALDWELL, SEC. MC .G LANE " BULK SHIELDING FACILITY
g- ?;T%i':féw ARE g- :2:5 ke TECHNICIANS 5. MCOONALD ¥ 1 &6 cocHRAN P
. G AR . . MCNA *
B. C. WILLIAMS ARE R, K. BAGWELL MC B Hotnet : F. €. MAIENSCHEIN P
J. M, DIDLAKE MC .
CONSULTANTS F. A DOSS ARE ‘L: ;' ?’;‘é‘fig " G. M. ESTABROOK 3
4. F. BAILEY, UNIVERSITY OF TENNESSEE CONSULTANT 1. P, ELBANKS MC R Thomas " M. P. HATDON* P
¥, R. CHAMBERS J. H. BUCK, WELL SURVEYS, INC. B, F. HITCH MC W ALKER . 1.0, FLYNN P
J. F, HAINES W JEHNINGS MC - K. . HENRY P
W. K. STAIR, UNIVERSITY OF TENNESSEE F. G. KILPATRIC MC E. B, JOKNSON P
G, A, PAIMER MO METALLOGRAPHY C. BOUNDS, 5EC. P
REACTOR COHTROL B, C. THOMAS MO
CONTRACTOR ) ) C H. TIPTON e R. 1. GRAY M TECHNICIANS
YITRC CORPCRATION E.P.EPLER < R. A, WILEY ARE R. 5. CROUSE I H. JARVIS Ic
T. M. KEGLEY M P e
F. P, GREENT I CONSULTANTS C.L.LONG L R, th SIMMONS P
5. H. HANAUER® ic J. M. CARTER, CARTER LABORATORIES TECHNICIANS S. SMIDDIE Ic
E. R. MANN® Ic D, G. HILL, DUKE UNIYERSITY G. G. STOUT ic
W. F. MRUK! Ic H, INSLEY LR " H. WEAVER P
L. C. GAKES* Ic : . F.
LG SaKES, Ic T. N, MCVAY, UNIVERSITY OF Al ABAMA L P zeeRALD "
R. S, STONE* Ic CONTRACTORS B C. LESLIF M
C.S. WALKER® ic TOWER SHIELDING FACILITY
P. GROOVER " SEC. e AME'Sr CULE MR STITUTE CONSULTANTS C. E. CLIFFGRD P
NOTE: THIS CHART SHOWS ONLY THE LINES OF TECHMICAL COORDINATION OF THE ANP PROJECT, THE H. A, BISHOP,* DRAFTSMAN  IC g:gékfig‘ggfi%‘fl'—lgg ! -
YARIOUS (MDIYIDUALS AND GROUPS OF PEOPLE LISTED ARE ENGAGED EITHER WHOLLY OR PART TIME ON METAL HYDRIDES, INC, N. CABRERA, UNIVERSITY OF V”‘G““T'Ar T. ¥. BLOSSER P
RESEARCH AND DESIGN WHICH IS COORDINATED FOR THE BENEFIT OF THE ANP PROJECT IN THE MANNER MOUND LABORATORY e A S n TUTE F. N. WATSON P
INDICATED ON THE CHART. EACH GROUP, HOWEVER, IS ALSO RESPONSIBLE TO IT5 DIVISION DIRECTOR FUEL REPROCESSING A R NIZHOLS, SANDIEGD STATE ). L. GREGS, CORNELL UNIVERSITY L. B poLLAND k
FOR THE DETAILED PROGRESS OF ITS RESEARCH AND FOR ADMINISTRATIVE MATTERS, —] F R BRUCE LNIVERSITY OF ARKANSAS 2 LNS:Ii;ES 3. VAN HOOMISSEH BAT
. F. . M. F. VALERING NACA
M. R, BENNETT cr RENSSFLAER FOLYTECHNIC INSTITUTE B L. GILLILAND GE
THE KEY TO THE ABBREVIATIONS USED IS GIVEN BELOW. G. | CATHERS cr W. F. SAVAGE, E. MCBEE, SEC. P
D. E. FERGUSON cr ANAL YTICAL CHEMISTRY RENSSELAER POLYTECHNIC INSTITUTE
C. E. LITTLEJOHN T . A E. C. WRIGHT, UNIVERSITY OF ALAGAMA TECHNICIANS ANP STEERING COMMITTEE
AC  ANALYTICAL CHEMISTRY DIVISION ~ ORNL T. LONG T ©. D. SUSANOD < P, C. SHARRAH, 1 N, MONEY P
T R. P, MILFDRD CcT 8. £. YOUNG,* SEC AC UNIVERSITY OF ARKANSAS
AGC  AEROJET.GENERAL CORPORATION O b WATSON T E.D. CARRO:‘- ic W. H. JORDAN, CHAIRMAN
ALC  AMERICAN LOCOMOTIVE COMPANY TECHNICIAN 1. C.WHITE AC G. G. UNDERWOOD '«
ARE  AIRCRAFT REACTOR ENGINE ERING DIVISION — ORNL s CaLowect o D.E CARPENTER re CONTRACTORS E.S. BETTIS
BAC  BOEING AIRPLANE COMPANY — D, L. MANNING AC BALDWIN-LIvA-tAMILTON CORPORATION o e aove
€ CHEMISTRY DHVISION - ORNL R. 1, MECUTCHEN® AC g;a;fi;fi?fmfl%o%fi«?w 5. J. CROMER
CT  CHEMICAL TECHNOLOGY DIVISION - ORNL B. 1. MCDOWELL AC . K. ERGEN
A. 5. MEYER 4 AC TERROTHERM ¥. . EREN
CV  CONSOLIDATED VULTEE AIRCRAFT CORPORATION . PEELE AC MASSACHUSETTS INSTITUTE YR GRIVES
EM  ENGINEERING AND MAINTENANCE DIVISION - ORNL . 1. ROSS AC oF TEC”"OLSOG"C W. D, MANLY
W. . VAUGHAN AC METAL HYDRIDES, INC.
GE  GENERAL ELECTRIC COMPANY R WILLIAMS AC RENSSELAER POLYTECHNIC INSTITUTE : .Js ‘,‘Jééfé’umsx
e {NSTRUMENTATION AND CONTROLS DIVISION - ORNL E. M. ZARZECKI, SEC. AC SUPERIOR TUBE COMPANY H. W, SAY
M METALLURGY DIVISION — ORNL GLENN L, MARTIN COMPANY E.D. SHIPLEY
TECHNICIANS
MC  MATERMLS CHEMISTRY DIVISION — ORNL 7. P. BACON AC JA- ?i. -‘;:g::‘;gg;
MP  MATHEMATICS PANEL - ORNL R. G. BRYANT AC CERAMIC RESEARCH
NACA MATIONAL ADVISORY COMMITTEE FOR AERONAUTICS M. R. CHILDS AC L )
P PHYSICS DIVISION — ORNL t *;A ."5‘3& :2 J. M. WARDE M
PW  PRATT AND WHITNEY AIRCRAFT DIVISION — UAC M. A MARLER AC €. E. CURTIS™ M
RD  RESEARCH DMRECTUR'S DEPARTMENT - ORNL T.G. MILLFR AC L. M.DOREY® M
REE  REACTOR EXPERIMENTAL ENGINEERING DIVISION - ORNL ;7 3 ;FSB:ER :g > ?‘ mflii‘m" ::
s1 STABLE ISOTGPES DIVISION - ORNL i M. P. HAYDON* P
$S  SOLIG STATE DIVISION = GRNC CONSULTANT & R. fir‘:NSON“ "
USAF  UN{TED STATES AR FORCE H. H. WILLARD, UNIVERSITY OF MICHIGAN o5 witer "
SPECTROGRAPHIC ANALYSIS A. HOBBS,” SEC. "‘
.
PART TIME iR, MCNALLY® o TEGHNICIAN
.
J. A, NORRIS® H J. A, GRIFFIN "
AND OTHERS
CONSULTANTS
MASS SPECTROMETRY T. N. MCVAY, UNIVERSITY OF ALABAMA
T. 5. SHEVLIN, OHIO STATE UNIVERSITY