ORNL-2012.txt

TR i
o \
L'-r-.-'l.'d ! . '
-. .
a

“'ORNL-2012, Parf 1,11, 11’
C-B4 - Reoctors-Special Features of Aircroft Reactors

AEC RESEARCH AND DEVELOPMENT REPORT

oy. 129A

    
    

*

|
ol

T

      
   
  

=3
. \
ey i 3 445k 0349900 &
&;i 4
S
Crs c&; I
&7 3
* | d‘fi! AIRCRAFT NUCLEAR PROPULSION PROJECT
C—_i’« Tfl QUARTERLY PROGRESS REPORT
IE ‘3."3‘{ FOR PERIOD ENDING DECEMBER 10, 1955
W 3

CrassrrIcATION (1
Py Al'l'!fl‘lé
B -"flt

OAK RIDGE NATIONAL LABORATORY

OPERATED BY
UNION CARBIDE NUCLEAR COMPANY

A Division of Union Carbide Lund Carbon Corporation i b N

POST OFFICE BOX P * OAK RIDGE, TENNESSEE

 

 

P ERSE RS R N

   

-
MIHS E'E iri i §
tl its contents n,
in any L

 
 

ORNL-2012, Part I, II, Il}

C-84 ~ Reactors-Special Features of Aircraft Reactors

This document consists of 244 pages.

Copy /.,2701‘ 324 copies. Series A.

Contract No, W-7405-eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT

For Period Ending December 10, 1955

W. H. Jordan, Director
S. J. Cromer, Co-Director
A. J. Miller, Assistant Director
A. W. Savolainen, Editor

DATE ISSUED

FER 20 1958

OAK RIDGE MATIONAL LABORATORY
Operated by
UNION CARBIDE NUCLEAR COMPANY
A Division of Union Carbide and Carbon Corporation
Post Office Box P
Oak Ridge, Tennessee

     
 

mic

ts cantents
MARTIN MARIETTA ENERGY SYSTEMS LIBRARIE

N

3 4456 03494900 b

NS R e b aaa £ ol UL
- -.a—-a-u_q_.

in any mannef"

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
R. G, Affel
2 R. Baldock
3. Borton
4. D, S;Wgillington
5. F. F. BWNgkenship
6. E.P. Blizdg
7. C. J. Borkow
8. G. E. Boyd
9. M. A, Bredig
10. F. R, Bruce
11. A. D, Callihan
12.+ D, W, Cardwell
13. J. V. Cathcart
14. C. E. Center (K-25)
15. R. A. Charpie
16. G, H. Clewett
17. C. E. Clifford
18. J. H. Coobs
19. W. B, Cottrell
20-21. D. D, Cowen
22. S. Cromer
23. R. S, Crouse
24. F. L. Culler
25. J. H. DeVan
26. D. A, Douglas
27. E. R, Dytko
28. L. B. Emlet (K-25)
29. M. J. Feldman
30. D. E. Ferguson
31. A.P. Fraas
32. J. H. Frye
33. W. T, Furgerson
34. H. C, Gray
35. W, R. Grimes
36. E. E. Hoffman
37. A. Hollaende
38. A. S, Housgffolder
39. J. T.Ho
40. H. K kson
41. W. HBordan
42. G. Y. Keilholtz
43, P, Keim

44. M. T. Kelley

45, F. Kertesz

46. E. M. King
47-48. J. A. Lane

w

 

ORNL-2012, Part I, 11, 1l

C-84 - Reactors-Special Features of Aircraft Reactors

INTERNAL DISTRIBUTION

 

o 64.

- .94,

49. R, S, Livingston

50. R.N. Lyon

51, F. C, Maienschein

52. W. D, Manly

53. E. R. Mann

54. L. A, Mann

55. W. B. McDongjd

56, F.R., McQug i

57. R. V. M rebllan

58. R. P. Mfliford

59. A. Jg lller

60. RgE. Moore

o1, j _ Z, Morgan

6g¥ E. J. Murphy

3. J. P. Murray (Y-12)
G. J. Nessle

65. R. B, Oliver

66. L. G. Overholser

67. P. Patriarca

68. R. W, Peelle

69. A. M. Perry

70. J. C, Pigg

71. W. G, Piper

72. H, F. Poppendiek

73. P. M. Reyling

74. H. W, Savage
A. W, Savolainen

RS, R. D. Schultheiss

W E.D. Shipley

78. %. Simon

79. OQSisman

80, G. Wy Smith

81. A. H,%pell

82. C. D, S&ano

83. J. A. SwaRgut

84. E.H. Taylo

85. R, E. Thoma

86. D. B, Trauger %

87. E.R. VanArtsdaly

88. F. C. VonderLage ¥

89. G. M. Watson

90. A. M. Weinberg

91. J. C, White

92. G, D. Whitman

93. E. P. Wigner (consultant)
'@, C. Williams
95. J. C. Wilson
96. C. E. Winters

*1***

 

ORNL-2012, Part I, I, Il
C-84 — Reactors-Special Features of Aircraft Reactors

107-126. Laboratory Records Department
127. Laboratory Recorgs, ORNL R.C,

97-1084 ORNL —Y-12 Technical Library, 128-130. Central Researchii® ibrary

Document Reference Section

31.

]
134
135.
136.
137.
138.
139.
140.
141.
142-144.
145.
146.
147.
148-153.
154,
155.
156.
157.
158.
159.
160.
161-162.
163.
164.
165.
166.
167-169.
170-172.
173.
174-177
17§
1.
0.
81.
82.
183.
184.
185.
186.

 
 

EXTERNAL DISTRIBUTION

AF Plant Representative, Baltimore

2. AF Plant Representative, Burbank
. AF Plant Representative, Marietta

  
  
  
 

AF Plant Representative, Santa Monica
F Plant Representative, Seattle
Plant Representative, Wood-RidgedFf
Aflateriel Area ;
Air Y search and Development Copgfiiand (RDGN)
Air Rqearch and Development Cglimand (RDZPA)
Air TeMigical Intelligence Cenj
Aircraft Mlboratory Design Boich (WADC)
ANP ProjeN@ Office, Fort Wgl
Argonne Naffinal Laboratg
Armed ForcesNpecial Wellpons Project, Sandia
Assistant SecreWixy offfle Air Force, R&D
Atomic Energy Coamifsion, Washington
Battelle Memorial JMtitute
Bettis Plant ‘
Bureau of Aerongltics
Bureau of Aerqfifiutics (ede 24)
Bureau of Aeglifautics Geral Representative
Chicago Opgiiitions Office\g
Chicago PgiBnt Group
Chief of iiifval Research
Convairgieneral Dynamics Corpiation
Directglfof Laboratories (WCL) \
Dire of Requirements (AFDRQ&
Dirgfifor of Research and Developmefl (AFDRD-ANP)
Digtorate of Systems Management (A7 -15N)
e ctorate of Systems Management (R(¥ge 1SS)
quipment Laboratory (WADC)
eneral Electric Company (ANPD)
Hartford Area Office _
Headquarters, Air Force Special Weapons Cerfg
idaho Operations Office
Knolls Atomic Power Laboratory
Lockland Area Office
Los Alamos Scientific Laboratory
Materials Laboratory Plans Office (WADC)
Mound Laboratory \
National Advisory Committee for Aeronautics, Clevelan¥

      
   
       
 

  
   
      
  
     
   
    
  
   
 
 
   
   
 
   
  

     
  
 
187.
188.
189.
190.
191.
192.
193-195.
196-199.
200.
201.
202.
203,
204.
205.
206-208.
209-323,
324.

 

-

ORNL-2012, Part |, Il, 1]
C-84 = Reactors-Special Features of Aircraft Reactors

National Advisory Committee for Aeronautics, Washington
Naval Air Development Center

New York Operations Office

North American Aviation, Inc, (Aerophysics Division)
Nuclear Development Corporation

Patent Branch, Washington

Powerplant Laboratory (WADC)

Pratt & Whitney Aircraft Division (Fox Project)

San Francisco Operations Office

Sandia Corporation

School of Aviation Medicine

Sylvania Electric Products, Inc.

USAF Project RAND

University of California Radiation Laboratory, Livermore
Wright Air Development Center (WCOSI-3)

Technical Information Extension, Oak Ridge

Division of Research and Development, AEC, ORO

 
 

-

Reports previously issued in this series are as follows:

ORNL-528
ORNL-629
ORNL-768
ORNL.-858
ORNL-219
ANP-60
ANP-65
ORNL-1154
ORNL-1170
ORNL-1227
ORNL-1294
ORNL-1375
ORNL-1439
ORNL-1515
ORNL-1556
ORNL-1609
ORNL- 1649
ORNL-1692
ORNL-1729
ORNL-1771
ORNL-1816
ORNL-1864
ORNL-1896
ORNL.-1947

Period Ending November 30, 1949
Period Ending February 28, 1950
Period Ending May 31, 1950
Period Ending August 31, 1950
Period Ending December 10, 1950
Period Ending March 10, 1951
Period Ending June 10, 1951
Period Ending September 10, 1951
Period Ending December 10, 1951
Period Ending March 10, 1952
Period Ending June 10, 1952
Period Ending September 10, 1952
Period Ending December 10, 1952
Period Ending March 10, 1953
Period Ending June 10, 1953
Period Ending September 10, 1953
Period Ending December 10, 1953
Period Ending March 10, 1954
Period Ending June 10, 1954
Period Ending September 10, 1954
Period Ending December 10, 1954
Period Ending March 10, 1955
Period Ending June 10, 1955
Period Ending September 10, 1955
 

 

 
 

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 other
ANP research ot the Laboratory under its Contract W.7405-eng-26. The report is divided
into three major parts: |. Reactor Theory, Component Development, and Construction,
H. Materials Research, and lll. Shielding Research,

The ANP Project is comprised of about 530 technical and scientific personnel en-
gaged 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 national ANP effort. However, the bulk of the
ANP research at ORNL is directed toward the development of a circulating-fuel type
of reactor.

The design, construction, and operation of the Aircraft Reactor Test (ART), with
the cooperation of the Pratt & Whitney Aircraft Division, are the specific objectives of
the project, The ART is to be a power plant system that will include a 60-Mw circulating-
fuel reflector-moderated reactor and adequate means for heat disposal. 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-power, circulating-fuel, reflector-

moderated aircraft reactor system,

vii
 

 
FOREWORD
SUMMARY

1.

...............................................

...................................................

 

.........................................................................................................

.........................................................................................................

PART I. REACTOR THEORY, COMPONENT DEVELOPMENT, AND CONSTRUCTION

REFLECTOR-MODERATED REACTOR ..ottt ne e s rs s
ART Facility Design and Construction ... st s s scne et et e enens
ART DESIGN civirirriiieiirereee ittt cseste s s sa e st s bt st a e e s st e st s resaesbesaeasesesaanbastsresseneenaneaseesbantn e saentnsnenresesne s

System Flow Sheets .ocevvvevvirenens

.........................................................................................................

Fuel-to-NaK Heat Exchanger ..ot et et he s enens
Fuel Filleand-Drain System s seee ettt abesae s st e a e s e sae e s s aeneasnnsreens
ReaCtor Shield ..o e

Core Flow Studies............. et are e

---------------------------------------------------------------------------------------------------------

Engineering Test Unit oottt ettt e s e e st eb e e ae et et es e s sasasteesemesseeaeeenenee

CoNtr OIS AN IS trUMEM O i O e coiiiieeerii s eiesceeereaesssrresarsnnsteeneesnarennnsssasesssstasasasssrssresansnsrasessssssesnsssesssessessenees

Procurement of Special Reactor Materials and Components .....ccocceeverieivceiiieciecieee e,

Beryllium oo
Shell Fabrication .coveecerveeciireeeennene.
CX-900 lnconel .ccoceoveeerccerreeeeeeeeene,

.........................................................................................................

.........................................................................................................

.........................................................................................................

Main Heat Exchangers and Radiators ..ottt e

Operation of ZrF , Vapor Traps in the High-Temperature Critical Experiment ...cocovvveviviiiiinnes,
EXPERIMENTAL REACTOR ENGINEERING ...oooioioe et esver st e sve e sas e eb e

In-Pile Loop Development and Tests ..ottt et et e st ene st esene e eanese s oo
[MeP ile LLoop OPEration v eoccoeeeiieteieeveee ettt ee st eeae et et s et ee s e bt s s ab s bt ebes e eaetensenserentesessnsenes
L 00D P UrGe SYSTEM ittt ettt e bbb bbb bbbk eb bt n b b s b bt a b e b ne b

Forced-Circulation Corrosion and Mass Transfer Tests v eessieeeceeeeereerseeeereeeeteeeseeaeeeesaresnens
Fused-Salt—lncone] Sy stems ..ottt ettt en et

Liquid Metals in Multimetal Loops

........................................................................................................

P UMD DEVEIOPMENT oueitiiiiiiiriiie ettt s e et snae s et et sabs e s era e ese e ae e ses et e e e ar e seannrsnnens
Bearing=and-Seal T ests «i ittt e e e et a s s raaeeare s
Sodium-Pump Water Performance Tests ...t irnnenie e neee e sere s s sane e ses e rnesee s

High-Temperature Tests of ART MF-2 Fuel Pump with NaK.......ccooi s

High-Temperature Pump-Performan

8 TSt SHANAS tooreriiee et e et eeese e e e te e ae e e e e erraneaass

Heat Exchanger Development ...ttt e s
Intermediate Heat Exchanger Tests oottt st st ane e e
Small Heat Exchanger Tests ittt et e s e s en
Heat-Transfer and Pressure-Drop Correlations ..o

U CEUIA] T @S S ettt ettt i e ete e e eiat s e tbesesarresabeaesabbaesbe e sb b s nesnsesn e ab b e s e eR b e e e an s e rabeesesararnsnrerrnneeeabreannnrs
Outer-Core-Shell Thermal-Stability Test............ SO PP U OO U S PRTPI RO

Inconel Strain-Cycling Tests .......

..........................................................................................................

Thermal-Cycling Test of Sodium-lnconel-Beryllium System ..,

Reactor Component Development ...
Dump Valve crciiie e
Cold Trap and Plugging Indicator

..........................................................................................................

..........................................................................................................

..........................................................................................................

vii

15
15

19
19
19
20
20

23
25
25

25
25
26
26
26

26

27

27
27
29

30
30
32

33
33
35
36
41

4]
41
49
49

54
54
56
58

59

59
60
 

Zirconium FIuoride VAPOr TIAP v ste e s es e sasseesnes b ene st e snesbe e e ssassneesssssensesanns
Water Test of Aluminum Mockup of Top of ART ..ot e b s

3. CRITICAL EXPERIMENTS ...............

........................................................................................................

Room-T emperature Reflector-Moderated-Reactor Critical Experiments....c.c.ccecermveninnrcennreneninians
Power and Neutron-Flux Distributions ...t s sae e anens
Neutron Production in the Fuel-to-NaK Heat Exchanger .......ccccoccoininnccveiiieeeieeenieans
Radial Importance of Uranium in the Fuel Annulus. ..o e see e,
Importance of Beryllium at End of Reactor ....cvecveieiceiiniicrrce et see e
Axial Importance of @ Neutron SOUFCE ..ottt

High-Temperature Reflector-Moderated-Reactor Critical Experiments ...ccccoccevrnnvinviniinecinininnninee.

Compact-Core Reflector-Moderated-Reactor Critical Experiments ...

PART Hl. MATERIALS RESEARCH
4. CHEMISTRY OF REACTOR MATERIALS cc..coiici ittt et

Phase Equilibrium Studies.....c............
The System ZrF -UF
The System LiF-UF , oo,

The System NaF-LiF-UF ...

The System KF-UF4 ..........................
The System NaF-KF-Z¢F,................
The System NaF-LiF-BeF,...........
The System NaF-LiF-BeF ,-UF , ...

The System KF-BeF ..o,
The System NcF-KF-BeF2 ..............

Systems Containing Alkaline-Earth

Chemical Reactions in Molten Salts..

........................................................................................................

........................................................................................................

........................................................................................................

........................................................................................................

........................................................................................................ -

........................................................................................................

.......................................................................................................

--------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------

........................................................................................................

FlUOr IS eoeeeeeee oot e e et st st e v e e s s sa s bmns

........................................................................................................

Equilibrium Reduction of FeF, by H, in NaZrF ¢ e
Reduction of UF4 By Structural Metals (it e st et
Stability and Solubility of Chromium Fluorides in Various Molten Fluorides ..occoveeiicvececcvernnccennne.
Reaction of UF3 with Alkali Fluorides ..ciiiiiienniiiciiceeiee Eeereereeesneanr et rrresb e aasenener naaaaesaraen
Reduction of Alkali Fluorides by Uranium Metal......ccocoooiiiiiiie e e
Experimental Preparation of Pure FIuorides .....ccciceoiiiiiiiieceees e

Production of Purified Fluorides ......

........................................................................................................

Removal of CrF, from Nc:F-ZrF“-UF4 MEXEUEES ettt e eeee e e et e m e et ene e e seneeeeneeen
Laboratory-Scale Purification Operations .........ueiiririeininienc et ene e seses e e sn e as
Special Preparation of NaF-ZrF ,-UF J-UF | s
Evaluation of Raw Materials for Fuel Preparation ..o et

Pilot-Scale Purification Operations
Production-Scale Operations............

........................................................................................................

--------------------------------------------------------------------------------------------------------

Batching and Dispensing Operations. ..ot rers s ene et e st

Special Services ...oovvvceveiiceeneiiennns

........................................................................................................

Fundamental Chemistry of Fused Salts ...ttt e
Relative Viscosity-Composition Studies of Nch-LiF-ZrF4 Mixtures at 600°C...........covrniviieneee.

EMF Measurements in Fused Salts

--------------------------------------------------------------------------------------------------------

 

60
64

66

66
66
66
67
67
70

71
73
 

F o lUOPIAE S 1n TGO e oeee oot e et ee et e e e e e e e e e et v ereerere e e aa s aesaaaseasasaassaasssssasssnsesnsassaessenasnsesnsasessnenses 103

SOdTUM N TNCONEL vt et e e s sbesne e s s st ea st st e e e e naasresnnesteabeesrensteasens 106
Thermal-Convection STUdIies . oottt ettt s e s b e sas s e ebe s e nen e aesans 107
Effect of Difference Between Loop Wall Temperature and Fluid Temperature ......c.oooveeeerecennnnne. 107
Effect of Zirconium Hydride Additions to Fluoride Mixture......coooimeiiiiiiincin et 107
Loops Fabricated from Special Inconel-Type Alloys ..ccoveriiecimree e 110
Nickel-Molybdenum Alloy Loops ittt r et eb et e es bbb er e ees 111
Effect of Nitrogen Atmosphere ...ttt st e 111
General Corrosion STUAIES oo e s et s et e st bbese s s e eanesaesreesaesrneesbensesanabensessans 112
Thermal-Convection Loop Tests of Brazed Inconel T-Joints in NaF-ZrF -UF, s 112
Seesaw Tests of Brazed Inconel T-Joints in Sodium and in Fuel Mixtures.....cccoeeivvivccricniiniennnn., 113
Static Tests of Brazed Materials ..ottt e se st e abs et ene st 117
Static Tests of Brazed Hastelloy B T-Joints in Sodium and in NaF-ZrF ,-UF . 120
Seesaw Tests of Chromel-Alumel Thermocouple Joins to Inconel Thermocouple Wells ................ 120
Effect of Ruthenium on Physical Properties of Inconel oo, 122
Boiling Sodium in INConel Loop .ottt sasa e r et e s 122
Decarburization of Mild Steel by Sodium....cciiieii s e e e 123
Static Tests of Special Stellite Heats in Lithium it 124
Solubility of Lithium in NaK ... s e 125
Seesaw Tests of Titanium Carbide Cermets in NaF-ZrF -UF oo, 127
Thermal-Cycling Tests of Inconel Valve Disks and Seats Flame-Plated with a
Mixture of Tungsten Carbide and Cobalt ..o 130
Static Tests in NaF-ZrF -UF , of Kentanium Cermet Valve Parts Brazed to Inconel ..........c..c...... 131
Effect of an Air Leak Into an Inconel~Fused-Salt Test System ..cccocevvvveiiiiiceicci e 131
Fundamental Corrosion Research ..ottt sa b s 133
Seif-Decomposition of Fused Hydroxides.......ococcvueiciniiiiiciinec e e 133
Mass Transfer and Corrosion of Various Materials in Fused.Sodium Hydroxide .......ccccocevninrinenn, 133
Chemical Studies of Corrosion ...ttt st st s bbb asteaansaeeris 138
Reaction of Inconel with Sodium and NaK .. 138
Reaction of Sodium Hydroxide with Nickel ..o 139
Study of Eutectic Mixtures by Zone Melting........c.ocooii ittt e 139
METALLURGY AND CERAMICS ..ottt ettt et et bbb ettt enes s 141
Fabrication of Test Components ..ottt e 141
NaK-to-Air High-Conductivity-Fin Radiators ....ccocoeiiiiniin e 141
Twenty-Tube Fuel-to-NaK Heat Exchangers ... 143
Intermediate Heat Exchanger No. 3 ...t st ennes 143
Intermediate Heat Exchanger Job-Sample Evaluations ..o 144
Examination of NaK-to-Air Radiators That Failed in Service ..o, 145
Brazing-Alloy Development ... 149
Mechanical-Property Studies of Nickel-Molybdenum Alloys ..o 152
Influence of Aging Heat Treatments on the Creep Properties of Hastelloy B.......ccoccoeninieii, 152
Preliminary Investigation of Creep Properties of Hastelloy W ..o 153
Investigation of the Creep Properties of Some New Nickel-Molybdenum Alloys.......oc 153
Special Materials Studies...coccovrriiiiiiience e rr et e e et b et e e et b eba st e rene s 155
Extrusion of Seamless Duplex Tubing ... 155
Neutron Shielding Material for ART .o 162
D ORE - N T O UMM et eeee ettt ettt e eteseseses e atarareete s sasasaet s e sasasas sbstssbnns rsnrssnaransnsransnaransnrsrnrns 163
Gamma-Ray Shield Material for ART PuUmps . vttt e 163

  

X1
xii

 

Contr Ol R od FabriCation o ieeieieieeecceeteiessetetasnnesseensenntsnsssasnensaassrsaseertrenssssnnnnen

NoNdestructive TeSHING .ovvcveiicieiiiiieieieie ettt e ent bbb aae s b
Ceramic Research ..ot e e s sane s

Rare-Earth Cermet Fabrication............. et e re e areirreiebeeseataerseree s aennereeseeantarteearees

Boron Carbide Shield Material ..ot e cre e s se e e
HEAT TRANSFER AND PHYSICAL PROPERTIES ..o
Fused-Salt Heat Transfer.... e s ese e e e ses e eeas
ART Fuel-to-NaK Heat Exchanger ...t
ART Core Hydrodynamics ...ecciiiiiiiiiciincrite e em st e
ART Core Heat TransSfer .. et veeestiete e s r e resee e sresesa e sessestraeeas
VYolume-Heat-Source Convection Analyses ...cocccviiiiiiiniiicnceiincncincenee
Heat Removal from Fuel Dump Tank ..o esiiiee s esteee e ssrve e sssevenns
Heat Transfer in Helical Pipes. ..ottt s
Heat Capacity Measurements on Fluoride Mixtures .....ccocoocieriiiniienin e
Heats of Fusion of Fluoride Mixtures .....cccoicieieicoieiiviniee e seeresse e
Viscosity Measurements on Fluoride Mixtures ...ocooveiiiriiiiniiieiciecce e
Thermal Conductivities of Liquids..cccocciiviniieiiieicirccre e ser e e s sese e s
RADIATION DAMAGE ..ot e ten e raer e s sbe s r e b e ere s rserae s b e
Disassembly and Examination of Irradiated Equipment......ccccovinciiricicicnnnnnenne.

MTR In=Pile Lo0p ottt e b e e e b

LITR Miniature 1n-Pile Loop .ttt

ARE ComPonents ..ovviceeeiiieierieressieeesrts e e ee e ssre e serteeseseeeena g e sesnneesrensrneesrraeeasas
Thermocouple Errors in In-Pile lLoop Temperature Measurements..........c...........
Holdup of Fission Gases by Charcoal Traps .....ccceievceininccnireiere e
Creep and Stress CorroSion vt re e seveee e cee e seenas e st nssaras

In-Pile Tube-Burst Creep Tests .iieiiecciieiieieieeeeereeeeeteecee b e cveeesae st aens

Alternate In-Pile Apparatus ...t e ebe s e s snenae
ANALYTICAL CHEMISTRY OF REACTOR MATERIALS ...
Determination of Oxygen in Sodium ...

n-Butyl Bromide Method. ..ot

Distillation Method ...ttt et e ste e e s et e e s s as e sb b e e
Determination of Traces of Rare-Earth Elements in Stainless Steels................

Spectrophotometric Determination of Aluminum in Flucride Salts with Aurin

Tricarboxylie AT et se st st v s es e s ne e e ere s
Determination of Water in Hydrogen Fluoride Gas ....ccooceeeeiieiiciiinnecc e
Determination of Oxygen in Zirconium Oxide by Bromination .....cccoevvvrvrerrinnnnne.
Determination of Tantalum in Fused Fluoride Salts .....ocoooioiiiniiiincciiieane.

Direct Determination of Traces of Fe(lll) in NaF-KF-LiF-UF4 ..........................

Determination of Traces of Fe(lll) in Mixtures of Alkali-Metal Fluoride Salts
ANP Service Laboratory

..................................

----------------------------------

..................................

..................................

----------------------------------

----------------------------------

..................................

----------------------------------

..................................

----------------------------------

----------------------------------

----------------------------------

..................................

----------------------------------

----------------------------------

..................................

----------------------------------

..................................

..................................

----------------------------------

..................................

..................................

----------------------------------

----------------------------------

..................................

..................................

----------------------------------

..................................

----------------------------------

..................................

..................................

----------------------------------

----------------------------------

..................................

----------------------------------

..................................

----------------------------------

----------------------------------

 

163
164
167
167
168
170
170
170
171
174
177
177
177
178
179
179
181

182

182
182
182
183

183
183
184
184
184
186

186
186
187

189

189
190
191
192
192
192
193
10.

11,

12,

13.

 

RECOVERY AND REPROCESSING OF REACTOR FUEL ..ottt s e 194
Pilot PEAnt DeSign wocoecoiiceee ettt e ettt et b et s ie st ks e b e ke ket e bt s e b be e te st nas s s abenranens 194
Engineering Developments ..ot b s e s 194
CONTACTOF vitiiieeeeeete st st sse s saeste st s anarar e e e s be s ressetsebeese e s eset e beaaas ssssesnasaeamtaeee s snaneseneeraasbasntssat e anarassnerasas 194
Fre@ze Valves ettt cee et ra et e e s he et ae e s e e e ba e raaare s erenee 194
Resistance Heating of Transfer Pipes and Waste-Discharge Nozzle ..o 194
Process Development ..o ittt eb e a bbb eab et e bR s ae b s b be et et e bt e et aes 195
Fused-Salt FIuorination SHudies .. r e ree s ere s s e sbasr et bsebs s st 195
NaF Absorption Capacity and UF , Loss on Desorption.........occoierminieiminncincii s 196
UF6 Decontamination in NaF Absorption Step ...ttt s 197
PART Hll. SHIELDING RESEARCH
SHIELDING ANALYSIS Lot ieee s st e st ae s e b s b s e e s eaesasasseseesasssnesassassesaasensnseneaes 203
Air Scattering of Co®? Gamma Rays: Theory vs EXperiment.......coooooomeieceeioniiiesoneeneiensenceseeesnieees 203
Energy Absorption Resulting from Gamma Radiation Incident on a Multiregion Shield
With S1Ob GEOMEITY c.c.viitieini ettt s re e ettt s arn e sr e aesr e s 203
Integral Equations for the Flux Density near a Thin Foil and for Neutron Scattering in
Air in the Presence of the Ground ... et se e ae s st s b ane s 203
SHIELD DESIGN L.ttt ettt es s s tese et st st s aa st es e s b ts e sab s e e s s st se st esseresnessoanansebanns e 204
Calculation of the Sodium Activation in the Heat Exchangers of Circulating-Fuel Reactors .......... .204
Calculation of Activation by Core Neutrons.......co.oceiiiiii ettt st et ee st 204
Calculation of Activation by Delayed Neutrons.......coevevciriincniiienc e 207
Calculated Total Activations for Several Reactors......c.cocoiverirecerinierireeriesecersssss e ssrsssssnne, 209
LID TANK SHIELDING FACILITY Lottt sttt et v b ee st st st et st reean s e emns 210
Static Source Tests with Mockups of a Reflector-Moderated Reactor and Shield .......oooevveiviveinenne. 210
Effect of Varying Lead Thicknes s ...ttt nee e 210
Study of Secondary Gamma-Ray Production ...t 213
Effect of Heavy Metals in the Reflector ..o 218
Effect of Borating the Water Shield ..ot eer et e sse s s 219
Dynamic Source Tests with Mockups of a Reflector-Moderated Reactor and Shield .....ccocccceiiicis. 219
Sodium Activation in the Heat Exchanger ...t e 220
Fission-Product Gamma-Ray SPectrum ..ottt ettt et e bena s 223

- xiii
 

R
Ty

O 2
X

P

 

   
 

ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART I. REACTOR THEORY, COMPONENT
DEVELOPMENT, AND CONSTRUCTION

1. Reflector-Moderated Reactor

Construction work is now under way on the
Aircraft Reactor Test facility, Package 1 con-
struction, which consists in alterations to the
existing building, an addition to the building,
and installation of the reactor cell, was 10%
complete at the end of the quarter. A major design
change was made in order to provide more space
for NaK system piping and equipment, as well as
for a special heat-dump facility for the removal of
heat from the fuel fill-and-drain tank, Drawings
and specifications for a modified version of pack-
age 2 work were completed. This work now con-
sists in the installation of the diesel generators
and facility, the electrical control center, and the
spectrometer room electrical and air-conditioning
The piping work for nitrogen, air,
cooling water, helium, lubricating oil, and hydraulic
oil drive systems was removed from the package 2
work unit and was designated as package A,
Design work continued on package 3, which covers
the experimental instruments, controls, process
lines, and process equipment to be installed by
ORNL forces,

System flowsheets and instrumentation lists
were prepared in an initial attempt to define the
entire ART. The flowsheets show the various
components of the system; the annunciator pickup
and presentation locations; the operating con-
ditions of temperature, pressure, and flow rates;
the control stations for electrically operated
components; the normal valve positions and uniform
valve identification; and line sizes,

The fuel-to-NaK heat exchanger design was
modified to overcome interferences at the headers
and to provide additional space in the region of the
headers for the beryllium support struts., Di-
mensional and operating data for the heat ex-
changers were revised accordingly,

The problem of cooling the fuel fill-and-drain
tank was studied. Dual NaK systems are to be
utilized for both cooling and heating the tank,
Each system is to be capable of operating individu-

equipment,

ally and of removing the total maximum heat load
of 1.75 Mw from the fuel in the tank.

The basic mechanical design work on the ART
lead-and-water shield and supports was completed,

and the procedure for assembly and installation

was studied, A comparative evaluation of rubber
and aluminum containers for the borated shield
water showed aluminum to be preferable on all
counts, including weight.

Additional core flow studies were made on both
the full-scale aluminum and plastic core models,
Photographs were made of the flow potterns ob-
tained with various core inlet configurations.
Intensive studies of the various flow patterns are
under way,

An equipment layout was prepared for the Engi-
neering Test Unit (ETU), and work was started on
design of the facilities required for operation of
the unit, Design problems have delayed the
issuance of firm drawings of the reactor com-
ponents, and it is estimated that operation of the
ETU will be about three months behind the origi-
nally scheduled date of September 1, 1956.

Instrumentation |ists were prepared for each ART
system, and instrument
designated. Five permanent instrument information
centers will be used, in addition to three temporary
ones, and some instruments will be located and
read on the equipment, Layouts of control and
instrument panels and boards are being prepared,
and elementary wiring diagrams have been com-
pleted. The layout of the control rod and drive
mechanism is being prepared.

sensor IOCO‘HOI’IS were

The problems of procurement of special reactor
materials and components are being investigafed.
Equipment for machining the large beryllium re-
flector blocks is being prepared by The Brush
Beryllium Co. A newly developed Hydrospin
process shows promise of being satisfactory for
the fabrication of the required six sets of thin
Inconel shells, which vary from 18 to 54 in, in
The processing of the 100,000 Ib of
special CX-900 Inconel required for tubing is
under way. Welders qualified under the special
ORNL procedures are being trained for several

diameter,

- |
ANP PROJECT PROGRESS REPORT

vendors who are interested in fabricating the heat
exchangers and radiators.

Operation of the high-temperature critical experi-
ment revealed that the ZrF, vapor traps were
inadequate, Therefore a program for the develop-
ment of suitable equipment is under way,

2. Experimental Reactor Engineering

An in-pile loop was inserted and successfully
operated in the HB-3 beam hole of the Materials
Testing Reactor, after attempts to operate two
other loops had failed because of difficulties with
electrical heating circuits,  The fuel mixture
circulated was NaF-ZrF4-UF4 (53.5-40-6.5 mole %);
the maximum fuel temperature was 1500°F; the
maximum mixed-mean temperature differential was
150°F; and the total operating time was more than
400 hr, The loop power generation was estimated
to have been 20 kw at the position of maximum
insertion, The power density was calculated to
have been 0.6 kw per cubic centimeter of fuel
mixture., The loop is now being disassembled for
shipment to ORNL for examination. Operating
difficulties are being studied in connection with
design work on the next in-pile loop.

Studies of the electrical failures which prevented
startup of the first two loops have indicated that
the helium atmosphere used allowed electrical
breakdown and that the Savereisen cement used
as insulation tended to conduct current at high
temperatures, Since nitrogen is a better insulator
than helium, tests are being conducted to de-
termine the effectiveness of nitrogen as a purge
gas for the nose section. Grade A Lava sleeves
are to be substituted for Sauereisen cement at the
heater terminals in future loops.

Five electrically heated and three gas-heated
fused-salt—Inconel forced-circulation loops were
operated during the quarter. A study was made of
temperature measurements and heat distribution in
the gas-heated loops, and the electrical heaters of
the electrical-resistance-heated loops were modi-
fied so that fluid-to-wall temperature ratios could
be varied. Also, six forced-circulation loops were
operated with sodium and noneutectic NaK. Further
evidence was obtained that the plug indicator does
give a measure of the oxide contamination in the
sodium,

Seals of various designs are being tested for
possible use in the lower position in the ART fuel
pump., Several seals have been found to meet the

2 %@_ Cm

t.‘i Ve Ak g

fiff&

1

stringent specification of ‘'leakage of oil into
helium across a pressure differential of 0 to 5 psi
of not more than 2 cm?3 per 24 hr."’

Full-size models of the impeller and the volute
for the ART sodium pump are being tested with
water. Initial tests indicated that the impeller
would meet head and flow requirements at slightly
lower speeds than predicted. The cavitation tests
performed, to date, indicate that the pump will
operate safely at inlet pressures below the pres-
sure planned for the reactor sodium pump,

Tests of an ART fuel pump are under way with
high-temperature (1400°F) NaK as the pumped
fluid. A heat-source plug was used in the initial
tests to simulate gamma-ray heating, but equipment
failures necessitated discontinuing this phase of
the experiments., Tests with a heat-source plug
will be resumed later, if more satisfactory means
can be found to simulate reactor conditions, Two
high-temperature loops are being fabricated for
testing ART-type fuel pumps with NaK and with
fuel mixtures, and loops are being designed for
testing ART-type sodium pumps and ART-type NaK
pumps at high temperatures.

Operation of intermediate heat exchanger test
stand A was continued, and test stand B was
placed in operation. Heat exchangers were de-
signed for future operation in these tests stands
and in test stand C, which is being fabricated.
York radiator units Nos. T and 2 were tested on
stand A, and unit No. 3 is being installed, Ex-
amination of unit No, 1, which failed after 140 hr
of operation, revealed severe buckling of the side
plates of the radiator core because of differential
thermal expansion between the tube matrix and the
The side plates of unit No. 2 were
split to allow the tubes to move freely, and this
unit operated for 435 hr. The diffusion cold trap
installed in the loop was found to be ineffective,
and it was replaced with a circulating cold trap,
which, although somewhat more effective, is not
satisfactory,

Test stand B has completed a total of 670 hr of
operation, including 247 hr of bifluid operation,
in a series of radiator, heat exchanger, and circu-
lating-cold-trap tests. The system is now in
steady-state operation with a temperature differ-
ential imposed,

side plates,

A small heat exchanger test stand is also oper-
ating with a 20-tube fuel-to-NaK heat exchanger,
a 500-kw NaK-to-air radiator, and a circulating

e
R

2 . :...
5
cold trap. The NaK circulating in the cold trap
is cooled by an economizer located in the inlet
line to the cold trap and by air circulated in the
cold-trap cooling coil. With this arrangement it
has been possible to operate at cold-trap temper-
atures below 150°F, with system temperatures as
high as 1500°F. A preliminary analysis of the
data indicates that cold-trap performance is repro-
ducible and that the oxide concentration of the
NaK can be repeatedly reduced,

The test data obtained with the intermediate
and small heat exchanger test stands have been
correlated. The seven radiators tested, which
were built by three fabricators from the same
specifications, have given a wide range of per-
formance. An intensive investigation is under way
to determine the reasons for the differences, The
correlation of the heat exchanger data revealed
an apparent rise in the NaK pressure drop during
operation of the system with a temperature differ-
ential imposed, which is also being investigated,

Work has continued on the fabrication of a one-
fourth-scale outer core shell model that is to be
subjected to cyclic thermal and pressure tests.
Two anvil bending-test assemblies were put in
operation to obtain basic information on the be-
havior of Inconel under strain cycling at elevated
temperatures in both inert and corrosive atmos-
pheres.  The preliminary tests have indicated
that temperature has a major effect on the initia-
tion of cracks. The number of cracks produced for
a given number of cycles was greater at 1200°F
than at 1400°F and was greater at 1400°F than
at 1600°F.

A test assembly for use in the development and
testing of cold traps and plugging indicators is
‘being fabricated, and a test assembly for use in
the development of a ZrF, vapor trap was placed
in operation. Preliminary results indicate that
there must be very close control of the temperature
of the inlet line to the trap. The ZrF, vapor
precipitated and plugged the line when there was a
temperature differential of 25°F or more on the
0.5-in.-0D, 0.025-in.-wall inlet line,

A full-scale aluminum model of the top portion
of the ART is being fabricated, Problems of
design and operation will be investigated with
this assembly, which will be equipped with Inconel
fuel and sodium pumps and the external piping
needed to permit the pumping of water under simu-
lated reactor flow conditions.

PERIOD ENDING DECEMBER 10, 1955

3. Critical Experiments

room-temperature experiments on the
reflector-moderated-reactor critical assembly have
been completed. Power-distribution measurements
across the fuel section at the equatorial plane of
the reactor show an edge-to-center ratio of 4.5,
with only about 10% of the fissions at the center
being produced by thermal neutrons; adjacent to
the beryllium, this fraction is 70%. Neutron flux
measurements, with gold foils, show a similar
depression and energy distribution in the fuel. In
a mockup of the fuel-to-NaK heat exchanger,
shielded from the reactor by a ‘é-in. thickness of
boral, the fissioning rate was about 0.1% of that
in the reactor fuel region. The contribution to the
over-all critical assembly reactivity by a sample of

Several

vranium in the fuel region was shown to vary by a
factor of 2 as the sample was moved from the edge
to the center of the annulus. The effect of the
berytlium at the end of the assembly on the over-all
reactivity was found to be about 20 cents per
linear inch.

Experiments with the reflector-moderated critical
assembly operated with liquid fuel (NoF-ZrF4-UF4)
at 1200°F were also completed, The mid-plane
power-production distribution in the liquid fuel
showed an edge-to-center ratio of about 2, The
value was determined from the fission-product
activity induced in uranium disks which remained
in the annulus throughout the test,

A preliminary loading has been made of a mockup
of a solid-fuel modification of the refiector-
moderated reactor proposed by the Nuclear De-
velopment Corporation of America (NDA). In this
reactor design the solid fuel is cooled with sodium.
It was found that the critical mass of the mockup
had been underestimated by NDA, A 35% increase
in the uranium loading and a 23% decrease in the
steel content of the core were necessary in order

to make the assembly critical, The core now con-
tains 31 kg of U235,

PART Il. MATERIALS RESEARCH

4, Chemistry of Reactor Materials

Phase equilibrium studies of various fluoride
systems were continued in order to obtain a better
understanding of their structures and to devise
and detect improved fuels. In the ZrF ,-UF,
system, in which solid solutions are the only
crystalline phases that have been observed, it was
ANP PROJECT PROGRESS REPORT

shown that there is no major miscibility gap.
Diagrams of the LiF-UF, and KF-UF, systems,
which had been based almost entirely on thermal-
analysis data, were revised on the basis of ex-
aminations carried out on quenched specimens.

The study of mixtures in the NaF-LiF-'-UF4
system gave strong indications that no fuel mixture
with a suitably low melting point was available.
Thermal-analysis experiments and examinations of
slowly cooled melts in the NcF-KF-Zer system
were continued, and o series of quenching experi-
ments will be carried out to obtain additional
information,

Work continued on determining the phase re-
lationships in the NaF-LiF-BeF,-UF, system.
Fuel mixtures with low melting points appear to
be available, The viscosity of the mixtures was
improved by lowering the BeF, content, but it
appears to be likely that only very little additional
lowering can be achieved by reducing the BeF,
content to below about 20 mole %. Additional
clarification of the phase relationships in the
KF-BeF , and NaF-KF-BeF, systems was achieved.

As a matter of long-range interest in materials
for producing higher reactor temperatures, some
phase studies were carried out on alkali-metal-
fluoride~MgF , and —CaF, systems,

Improved equilibration methods gave values of
K, =2.5and 0.64 at 800 and 600°C for the reaction
FeF,(d) + Hy(g) & Fe(s) + 2HF(g), where the
concentrations in solution are expressed as mole
fractions and the gas pressures are in atmos-
pheres.

Studies of the stability of vanadium and nicbium
in NaF-LiF-KF-UF, at 600 and 800°C indicated
that vanadium was quite stable and that niobium
was not, Preliminary studies indicated that the
behavior of tungsten was similar to that of niobium,
Analytical data from equilibrations with tantalum
are not yet available,

Additional data were obtained on the solubility
and stability of CrF, in NaF-ZrF, and of CrF,
in NaF-LiF-KF. In each case the major portion
of the dissolved chromium retained its original
valence., The solubility of CrF2 in NaF-ZrF,
increased as the available CrF, was increased and
thus confirmed previous indications that solid
CrF,.UF , is formed.

Experiments were carried out in which UF_ was
added to molten fluorides at 900°C in copper

containers. No reduction occurred in the case of

LiF, but, with NaF and, to an even greater extent,
with KF, alkali metal was formed and partially
vaporized from the melt, A study was made of the
reactions of uranium metal with LiF and KF., In
the case of KF, metallic potassium was produced,
but LiF was stable, Since NaF-LiF is unstable
to uranium metal at 750°C, it can be concluded that
NaF was the unstable constituent.

Preliminary experiments indicated that used fuel
material can be freed from chromium by reduction
with zirconium followed by standard hydrofluori-
nation-hydrogenation treatment, if its oxide content
is kept low by careful handling.

A satisfactory commercial supply of ZrF4 has
been located for use in fuel preparations. The
hafnium content of the material limits it to use
in tests that do not involve nuclear activity.

Fuel samples for small-scale corrosion and
physical-property tests continued to be prepared
in laboratory- and pilot-scale facilities. The
stockpile of fuel from the large-scale production
operation reached an all-time high, and a temporary
shutdown will occur if usage continues to be less
than the predicted requirements, Low-carbon
nickel reactor cans are on order to replace the
A-nickel cans which have experienced frequent
failures.

The fuel mixture was removed from the high-

temperature ART critical experiment, with the
exception of the enricher system. Almost all the
fuel is still in this system, and, since it is a

prototype of the ART enricher, it will be used for
additional experiments. Two in-pile loops were
filled with enriched fuels,

Relative viscosity-composition studies were
made in the NaF-LiF-ZrF, system. Additional
emf measurements were carried out in the fused
salts, and additional optical and x-ray data were
obtained on various compounds in the fluoride

systems,

5. Corrosion Research

Examinations were completed of several Inconel
forced-circulation loops in which fluoride fuel
mixtures were circulated. Additional data were
obtained which confirmed previous observations
that, for a constant
temperature, increases in wall temperature signifi-
cantly increase the amount of corrosion, The

bulk-fluoride-fuel-mixture

attack in the loop which operated with a maximum
wall temperature of about 1710°F was 8 mils after
332 hr, in comparison with attack to a depth of
5 mils in 681 hr in a loop which had a maximum
wall temperature of about 1582°F, These data
were obtained with NaF-ZrF .UF, as the circu-
lated fuel mixture. The alkali-metal-base mixture
NaF-KF-LiF (11.5-42-46.5 mole %), with sufficient
UF, and UF, added to give 11.9 wt %, was circu-
lated in another Inconel forced-circulation loop,
and, as previously, the ottack was negligible, but
there was o continuous metal deposit in the cold
leg.

Several Inconel forced-circulation loops were
operated in a study of the effect of the oxide
content of the sodium being circulated. Mass-
transferred deposits were found in the cold legs
of loops operated with and without added oxygen,
as Na,0,, and with and without cold traps. The
addition of barium as a deoxidant did not measur-
ably decrease the mass-transferred deposit. A loop
operated with a maximum sodium temperature of
1000°F demonstrated the significant effect of
temperature on mass transfer; the loop was free
of deposited material, and there was no hot-leg
attack,

A special thermal-convection loop was operated
with thermocouples attached to the hot-leg wall,
and it was found that a temperature difference of
160°F existed between the maximum fluid temper-
ature and the hottest section of the wail, This
temperature difference is higher than that found
for the forced-convection loops and accounts,
in part, for the attack in thermal-convection loops
being deeper than that in forced-circulation loops
with similar bulk-fluid temperatures.

A series of six thermal-convection loops was
operated with NaF-ZrF -UF , (50-46-4 mole %)
to which various amounts of ZrH, had been added
to reduce the UF, present to UF,. The total
uranium content of the fuel mixture decreased
during operation of the loop in each case, and the
decrease was apparently associated with the
appearance of a metallic layer on the loop walls.
However, the corrosive attack was negligible
(0.5 mil) compared with that in a standard loop
operated without ZrH, added to the fuel mixture,

A group of thermal-convection loops was fabri-
cated from special Inconel-type alloys in which the
chromium content was the primary variable. When
the loops were operated under standard conditions
with NaF-ZrF -UF , (50-46-4 mole %) as the circu-
lated fluid, lf was found that the depth of attack

PERIOD ENDING DECEMBER 10, 1955

in the loops that contained less than 10% chromium
was appreciably lower than it was in a standard
Inconel loop; that is, 2 to 3 mils in comparison
with about 13 mils in 1000 hr. Loops fabricated
from several nickel-molybdenum alloys in which
NaF-ZrF 4+ UF, (50-46-4 mole %) was circulated
also showed little attack, 1 to 2 mils, in 1000 hr,

A special thermal-convection loop was fabri-
cated for testing brazed Inconel segments in the
hot-leg section.  This loop was operated for
1000 hr at a hot-leg temperature of 1500°F and a
cold-leg temperature of 1100°F with NcF-ZrF4-UF4
(50-46-4 mole %) as the circulated fluid. The
Inconel segments were brazed with Coast Metals
alloy No. 52 (89% Ni-5% Si-4% B-2% Fe). No
evidence of mass transfer was found, but the
brazed joints were porous and shrinkage voids
were present, In a series of seesaw tests of
Inconel T-joints in sodium and in fuel mixtures, a
75% Ni—25% Ge mixture was found to have fair
resistance to both mediums, Buttons of six differ-
ent braze materials were also tested in static
sodium and in static NGF'ZFF4-UF4 (50-46-4
mole %), and two palladium-rich buttons were
tested in NaOH and in the fluoride mixture. Coast
Metals alloy No. 52 and General Electric alloy
No. 81 showed fair corrosion resistance to sodium,
and the 60% Pd-40% Ni alloy showed good re-
sistance to both the fluoride mixture and NaOH.

Thermocouple assemblies were exposed to
sodium and to NaF-ZrF -UF, (50-46-4 mole %) i
seesaw apparatus so fhat 1119 effect of various
amounts of Chromel-Alumel in the weld nugget
could be studied, The nuggets with low Chromel-
Alumel content were unattacked after 100 hr ot
1500°F in sodium and in the fuel mixture, and
those with high Chromel-Alumel content were
attacked to a depth of about 0.5 mil in the fluoride
mixture, The Inconel tubes on which the welds
with high Chromel-Alumel content were made
were more heavily attacked in the nonweld areas
than were the tubes with the welds of low Chromel-
Alumel content,

A  second boiling-sodium—-Inconel loop was
operated, and the scheduled 1000-hr test was
completed, in contrast to the 400 hr of operation
for the first such loop. After the 400-hr test no
mass-transferred crystals were found in the cold
trap, but heavy intergranular cracking to a depth
of 50 mils was found. After the 1000-hr test,

mass-transferred crystals were visible in the cold
ANP PROJECT PROGRESS REPORT

trap, and there were cracks to a depth of 25 mils,
The temperatures in areas where cracks were found
varied from 1150 to 1325°F. A loop fabricated
from type 348 stainless steel is now in operation,

In a test of type 1043 mild steel (0.433% C) in
static sodium at 1830°F for 100 hr, significant
decarburization occurred, The carbon content of
the mild-steel decreased, and the
carbon content of the capsule walls in contact
with the increased; the two types of
capsules used were Armco iron {(0.019% C) and
type 304 ELC stainless steel (0.022% C). Two
special heats of Stellite were exposed to static
lithium at 1500°F for 100 hr. They were attacked
to a depth of 2 to 3 mils, in general, but there
were isolated areas that suffered very heavy
attack, The variations in susceptibility to attack
may be due to composition variations in the
specimens,

Titanium carbide cermets with cobalt- and nickel-
base alloys as binding material were exposed to
NGF-Zer-UF4 (53.5-40-6.5 mole %) for 200 hr
in seesaw apparatus, The hot and cold zones of
the apparatus 1500 and 1200°F, re-
spectively, Ten specimens were tested and only
two showed attack; the other specimens remained
unattacked, The compositions and methods of
fabrication of these materials, which were sub-
mitted by the Sintercast Corp. of America, are not

specimens

sodium

were at

known,

Inconel valve disks and seats were flame-plated
with a mixture of tungsten carbide and cobalt for
testing to determine the suitability of the coating
as a hard-facing material, To be suitable, the
coating must be resistant to solid-phase bonding,
galfling, and wearing. In an initial thermal-cycling
test from room temperature to 1500°F, the coating
did not crack, but, under conditions
similar to those used in brazing, the coatings
cracked. Kentanium cermet valve parts were held
in contact at a pressure of 10,000 psi and exposed
to static NQF-ZTF4-UF4 (53.5-40-6.5 mole %) for
150 hr at 1500°F. There was no indication of
bonding and the seating was uniform., The joint
between the cermet and the Inconel showed no
signs of attack; there was no distortion of the
z-in.-thick nickel shim; and the cermet pieces
did not show any signs of cracking.

Several experiments were performed to determine
the effect of a small air leak into a fused-salt—
Inconel system. The test capsules were heavily
attacked by the air-contaminated fused salt,

thermal

Measurements were made of the self-decomposition
of fused sodium hydroxide into water and sodium
oxide, Data obtained in recent measurements of
the vapor pressure of water over fused NaOH give
a value for AH® that agrees within 8% with a value
Tests of mass
transfer and corrosion of structural materials in
fused NaOH were continued, The materials studied
were Inconel, nickel, iron, Hastelloy B, and
types 310 and 405 stainless steel.

An apparatus is being constructed for studying
the solubility and the rate of solution of the
constituents of Inconel in sodium and in NaK as
a function of temperature and of oxygen content
of the liquid metal, in order to clarify the mecha-
nism of the mass-transfer reaction observed when
fiquid metals are circulated in inconel loops.
Also, an apparatus was constructed for studying
eutectic mixtures by zone melting.

known from other measurements.

6. Metallurgy and Ceramics

A third 500-kw NaK-to-qir high-conductivity-fin
radiator, two 20-tube fuel-to-NaK heat exchangers,
and two 100-tube bundles for intermediate heat
exchanger No. 3 were fabricated. Job samples
containing welded and brazed joints of the type
used in the fabrication of intermediate heat ex-
changer No. 3, submitted by outside vendors, were
examined and evaluated.

The two NaK-to-air radiators that failed in
service were examined. One of the radiators was
fabricated by ORNL and the other by the York
Corp. to the same specifications. The ORNL
radiater was found to have been so badly damaged
by fire that the cause of failure could not be
detected. However, it was found that the York
Corp. radiator failed because of the initiation of
a fracture in a braze fillet by shear forces and
the propagation of this fracture through the tube
wall by tensile forces or combinations of shear
and tensile forces, The tensile loading was
caused by differences in cooling rates between
the support members and the finned tubes when
cooling air was forced across the high-conductivity-
fin surfaces. The shear forces were caused by
a difference in cooling rate between the support
members and the bottom flanged plate and the
finned tubes. Modifications have been made to
the radiator design to relieve the tensile and shear
forces.

The results obtained to date in the brazing-alloy
development program were correlated with cor-
rosion data, and the alloys that are satisfactory
for use in the fabrication of heat exchangers and
radiators were selected. The alloys selected for
radiator fabrication on the basis of compatibility
with both liquid sodium and air were Coast Metals
alloy Nos. 50, 52, and 53; standard and low-
melting Nicrobraz; General Electric alloy No. 81;
and an 80% Ni—10% Cr-10% P alloy. The alloys
selected for heat exchanger fabrication on the
basis of compatibility with both fluoride fuel
mixtures and sodium were an 80% Ni-10% Cr—10% P
alloy; standard and low-melting Nicrobraz; Coast
Metals alloys Nos. 50, 52, 53, and NP; a 70%
Ni-13% Ge-11% Cr—6% Si alloy; and a 50%
Ni—-25% Ge-25% Mo alloy.

Alloys in the nickel-molybdenum system are
being investigated in the search for structural
materials with sufficient corrosion resistance and
high-temperature strength for use in circulating-
fuel reactors. The strength and corrosion re-
sistance of Hastelloy B have proved to be superior
to those of any previously tested material. How-
ever, this alloy has a tendency to age-harden in
the 1200 to 1500°F temperature range. The re-
sulting decrease in ductility is so severe that its
use as a structural material in a circulating-fuel
reactor would be limited.

Hastelloy W has been found to have strength
properties similar to those of Hastelloy B, and
the results of preliminary tests of the creep
properties show that Hastelloy W has less tendency
to age at the temperatures of interest.

Several new vacuum-melted nickel-melybdenum
alioys have also been creep-tested. Several of the
alloys showed poor ductility that may be attributed
to insufficient degassing in the vacuum-melting
process. Despite the poor ductility of the alloys,
their stress-rupture properties are equal, or su-
perior, to those of Inconel,

Preliminary work on the fabrication of seamless
duplex tubing was completed, in that tube blanks
of nickel-, Inconel-, and Monel-clad type 316
stainless steel were extruded, without difficulty.
The conditions for extrusion of Hastelloy B are
still not definite. Three Hastelloy B—clad type
316 stainless steel duplex billets were extruded,
but they were unsatisfactory because of cracking
and roughness that resulted from poor lubrication.
Attempts to improve lubrication of Hastelloy B by
flame-spraying with a heavy layer of type 304
stainless steel were unsuccessful. The extrusion

PERIOD ENDING DECEMBER 10, 1955

of Hastelloy W tube blanks from forged billets
was unsuccessful because of hot-shortness of the
material,

Attempts are being made to produce suitable
B,C-base tiles for the ART neutron shield and
to evaluate the properties of such tiles, Com-
patibility tests of Inconel and B,C with various
coatings and diffusion barrier materials in the
temperature range of 1500 to 2000°F are under
way. An irradiation tesf is being run to determine
whether helium and lithium will be released from
B,C at high temperatures. If these gases are
released, it may be necessary to vent the shielding
layer. Several other methods of preparing neutron
shielding are being studied, such as casting
mixtures of borides and metal and fabricating
cermet compositions that are high in B,C or boride
content,

The problems associated with the fabrication of
Inconel-clad niobium are being studied. A number
of specimens fabricated with copper—stainless
steel foil or tantalum foil diffusion barriers have
been prepared for mechanical-property tests.

Several compositions are being tested for use
as a gamma-ray shield of low thermal conductivity
for the ART pump impeller shafts. Two compo-
sitions which are of satisfactory density are
tantalum-constantan and tungsten carbide—con-
stantan,  Thermal-conductivity data are being
obtained for these materials.

It was demonstrated that control rods containing
30 vol % rare-earth oxide could be fabricated by
canning a cermet-type core in a capsule of suitable
cladding material and hot swaging the composite.
Physical-property data were obtained for the ex-
truded control-red parts containing rare-earth oxides
and for iron-zirconium alloys.

Investigation of the various methods by which
small-diameter tubing may be inspected has con-
tinved. The Cyclograph (an eddy-current instru-
ment) gives very effective, high-speed indications
of flaws as small as 10% of the tube-wall thick-
ness. For detection of minute flaws, the ultra-
sonic method is effective and can be applied to
large-scale inspection. The eddy-current-probe
instrument is being developed for use with the
ultrasonic method as a simultaneous inspection
on the same mechanical scanning operation. This
will provide close correlation of the defect indi-
cations from both methods, Tubing that is ]41 in,
in diameter and 7 ft 6 in. in length can now be
ANP PROJECT PROGRESS REPORT

inspected with this dual inspection method, and,
upon completion of a new probe coil, 3/]‘S-in.-dic:;
tubing 28 ft long can also be inspected.

7. Heat Transfer and Physical Properties

A forced-convection fused-salt heat-transfer
system that includes a mechanical pump has been
fabricated., This system is to be used to study
the heat-transfer characteristics of the fused salts
at Reynolds numbers higher than those covered
previously. The friction characteristics of small,
drawn, Inconel tubes such as those to be used in
the ART fuel-to-NaK heat exchanger were de-
termined over the Reynolds-number range 50,000
to 200,000; the friction factors of the Inconel
tubes fell only about 6% above the normal corre-
lation for flow in smooth tubes.

Additional hydrodynamic studies were made of
annuli of the type being considered for the fuel
channel of the ART reactor core. Calculations
of the pressure drop through a core fabricated
according to the present ART design were made
for straight-through
rotational-flow case. A study of flow distribution
in the annulus in which sodium will be circulated
to cool the outer core shell of the ART has been
initiated.

The conceptual design, as well as the detail

flow, as well as for a

designs of the test section, entrance sections,
exit section, and mixing chambers, of the volume-
heat-source experiment for determining the temper-
ature structure to be expected in the ART core
has been completed, and fabrication of the test
section is in progress. Several analytical heat-
transfer solutions for volume-heat-source systems
were developed; for example, turbulent flow in
curved channels was investigated, and the thermo-
couple error that results when a thermocouple is
used to measure the temperature of a flowing fluid
having a volume heat source was studied. A
general study of the problem of heat removal from
a fuel dump tank of an ART-type aircraft reactor
after shutdown was initiated. The wall-temperature
asymmetry predicted to occur in a helical pipe
that has fluids with volume heat sources flowing
through it was demonstrated experimentally in o
glass helix that contained a circulating fluid with
an electrically generated volume heat source. The
predicted results correlated well with data ob-
tained from operation of the MTR in-pile loop.

The enthalpies and heat capacities of two
rubidium-bearing fluoride mixtures were determined:

namely, RbF-ZrF,.UF, (48-48-4 mole %) and
LiF-RbF (43-57 mole %). The heats of fusion of
nearly all fluoride mixtures that have been studied
to date are presented. Viscosity measurements
were made for nine different fluoride mixtures. A
transient-cell method that has been used to
measure the thermal conductivity of liquids was
studied and used successfully for measurements
on water and molten sodium hydroxide., Pre-
liminary measurements made by the constant-gap
method gave a value of about 1.2 Btu/hr-#12(°F /1)
for the thermal conductivity of RbF-ZrF ,-UF
(48-48-4 mole %).

8. Radiation Damage

The in<pile loop in which a fluoride fuel mixture
was recently circulated in the Materials Testing
Reactor is being disassembled and sectioned for
examination, The disassembly work is being done
in the G-E *‘hotecell” facilities at the National
Reactor Testing Station, and the sections to be
examined will be shipped to ORNL,

Samples of the nose section of the miniature
in-pile loop which was operated for 30 hr in
vertical position C-48 of the LITR were examined.
Only slight, sporadic, corrosion penetration to a
depth of 1 mil was found, and comparisons of the
sections on each side of the center of the nose
bend and on the compression and tension sides
of the bend showed no changes in relative grain
size, Since this loop operated only 30 hr, no
definite conclusions can be drawn from these
observations. A large number of sections of the
ARE core and auxiliary equipment have been
prepared for metallographic examination,

An investigation of thermocouple errors in in-pile
loop temperature measurements was concluded,
The results confirmed the results of the previous
investigation of temperature measurements of
static in-pile capsules. Resistance-welded thermo-
couples were again shown to be superior to
discharge-welded thermocouples, in that they gave
smaller errors in temperature measurements and
were less sensitive to changes in cooling-air
flow rates.

Experiments are under way for determining
holdup times of fission gases by charcoal traps
at various temperatures with various gases in
the traps. Radiokrypton is being used to simulate
the fission gases. In the initial experiment,
nitrogen was used as the carrier gas in order to
determine whether nitrogen could be used as the
purge gas in the next MTR in-pile loop instal-
lation. The results obtained with a charcoal trap
identical to that used in the first in-pile loop
experiment indicate that a temperature of —=110°C
or colder would be necessary in the charcoal trap
to properly limit the amount of activity sent to
the MTR stack if nitrogen were used as the carrier
gas. Similar information is to be obtained for
helium and other purge gases so that the data
needed for designing the ART purge system will
be available.

Tube-burst creep-test specimens irradiated for
two weeks in the LITR are being examined. Four
tubular specimens were stressed to rupture at
a circumferential fiber stress of 2000 psi at 1550,
1500, and 1450°F. Other tube-burst rigs are being
readied for irradiation. The MTR creep-test appa-
ratus in which the specimen was tested at 1500°F
and 1500 psi has been returned to ORNL for ex-

amination,

9. Analytical Chemistry of Reactor Materials

Studies were undertaken to determine whether
positive errors in the determination of oxygen in
sodium result from the presence of alkaline-earth
metals in the alkali-metal samples. Typical
samples of sodium used in corrosion and heat.
transfer studies were found to contain approxi-
mately 100 ppm of calcium and less than 50 ppm
of magnesium. Such concentrations will have to
be taken into account, since they correspond to a
maximum error in the determination of oxygen of
about 80 ppm. In further studies of this method
it was shown that, if the reaction between sodium
or NaK and n-butyl bromide is contained in o
sealed vessel so that pure n-butyl bromide can
be used rather than a solution in hexane, the
conversion from the alkali metals to bromide salts
is complete in a matter of approximately 5 min.
With this technique it is possible to carry out the
reaction in an atmosphere of helium and to sample
sodium at temperatures as high as 1000°F,

It was demonstrated in connection with the
distillation method for the determination of oxygen
in sodium that temperature and total time of distil-
lation are important factors. Sodium oxide was
almost completely volatilized from a sample of
sodium after a 4-hr distillation period at 900°F,
while a residue of 400 ppm, which roughly corre-
sponded to the expected amount of oxygen, re-
mained after a similar period of distillation at
850°F. Reproducible results which correspond to

PERIOD ENDING DECEMBER 10, 1955

35 + 5 ppm of oxygen were obtained after distil-
lation of a purified sodium sample for 4 hr at
850°F. Essentially the same results were ob-
tained after distillation periods of 1 and 2 hr at
850°F. Analyses of the residues after distillation
show that Na,0 is the predominant residual oxide.

A method fzor the determination of the quantity
of the rare-earth elements gadolinium, europium,
and samarium in reactor construction materials
was developed. Since these rare-earth elements
have high absorption cross sections for thermal
neutrons, they must be limited to a few parts per
million. The method involves the use of spectro-
graphically pure yttrium oxide as a carrier for the
precipitation and concentration of microgram
quantities of the rare-earth elements from the
structural material, The yttrium oxide also serves
as an intemal standard for the spectrographic
determination, Concentrations of the rare-earth
elements of the order of 1 ppm were determined
in 2.5-g samples of stainless steel.

The Aluminon method for the spectrophotometric
determination of aluminum was applied to samples
of fluoride salt mixtures. Interfering elements,
such as zirconium, are removed by extraction with
cupferron.  The concentration of aluminum in
typical samples of fluoride salt mixtures was
found to be approximately 100 ppm.

A method was developed for the determination
of water relative to HF in the effluent gases from
a hydrofluorination reactor, The HF and water
were absorbed in pyridine, and the concentration
of the water was determined by titration with
coulometrically generated Karl Fischer reagent.
The concentrgtion of HF was determined by
titrating an aqueous solution of the pyridine
absorbate with standard sodium hydroxide solution.
The determinations are reproducible to about 0.5%
of water and are therefore sufficiently accurate
to permit monitoring of the reaction

ZrO2 + 4HF — ZrF4 + 2H20 .

Modifications were made to the apparatus for
the determination of oxygen in zirconium oxide by
bromination. Gold containers were substituted for
platinum, since gold is more inert to oxygen than
is platinum. When FeF; was added to the mixture
of ZrO, and graphite, complete removal of ZrQO
from the sample container was achieved after
bromination at 950°C for 2 hr. The recovery of
oxygen from the ZrO, was, however, incomplete,

A method for the determination of tantalum in
ANP PROJECT PROGRESS REPORT

fused mixtures of fluoride salts with the use of
pyrogallol was developed. The method is not
applicable to the determination of tantalum in the
presence of uranium, however.

A direct spectrophotometric determination of
trivalent iron in NaF-KF-LiF-UF, was developed
that is based on the iron(|l])-thiocyanate complex.
The fluoride salt sample is dissolved in a solution
that is 8 M with respect to H PO, in order to
prevent reduction of trivalent iron by tetravalent
uranium during dissolution of the sample.

For application to the determination of trivalent
iron in mixtures of alkali-metal fluoride salts, a
method was developed which is based on the
absorption of the ferric phosphate complex in the
ultraviolet region at 260 mu. Zirconium and
uranium interfere with this determination.

10. Recovery and Reprocessing of
Reactor Fuel

Construction of the pilot plant designed for using
the fused-salt fluoride-volatility process in the
recovery of fused-fluoride-salt reactor fuels is
scheduled for completion by March 31, 1956, The
cost, including process, project, and instrumenta-
tion engineering, is estimated at $435,000. On
the present flowsheet the product will be decon-
taminated by passage through two NaF sorption
beds.

Two types of gas contactors are being investi-
gated for the fluorinator: a percolator type and
a sieve plate. Tests have shown agitation to be
more vigorous with the percolator, but gas en-
trainment was lower with the sieve plate. The
percolator is being modified to decrease entrain-
ment. Additional tests are to be made, and tests
with wvuranium-bearing mixtures will be required
before a final decision can be made as to the
contactor to be installed in the pilot plant.

In tests of a prototype of the freeze valves
proposed for closing molten-salt transfer lines,
the valve remained sealed against a test pressure
of 20 psig in 15 of 20 cycles of melting, freezing,
and pressurizing. It is believed that slight con-
traction of the fused salt permitted gas leakage
around the downcomer. The design of the valve
is being modified.

Resistance heating was tested as a means of
heating and of maintaining the temperature of salt
transfer pipes at 1200°F. The tests showed that
an uninsulated 7-ft length of l/z-in. sched-40 Inconel
pipe could be heated to 1200°F in 5 min by passing

10

a current of 600 amp through the pipe.

Laboratory studies indicated that the NaF sorp-
tion beds can be repeatedly re-used for decontami-
nation of the UF, volatilized from the fused salt.
In a series of four runs with the same NaF in
the two-bed absorption procedure, there was no
indication of decreased decontamination resulting
from buildup of activity in the material. The
activity of the product UF, was less than the
UX,-UX, activity associated with natural uranium,
The uranium loss on the NaF beds under process
conditions was 0.1 to 0.2%. In batch tests, with
completely dry NaF, the loss was less than 0.01%
when fluorine was used as a sweep gas. It was
shown by desorbing 98.6% of the absorbed UF
with N, alone as the sweep gas that desorption
of the UF, does not require refluorination. The
UF, absorption capacity of one lot of Harshaw
Chemical Co. NaF was about 0.9 g of vranium
per gram of salt and was independent of tempera-
ture (70 to 150°C), pressure (7 to 15 psia), or
pellet size (]/8 in. to 40 mesh). The capacity of
Baker & Adamson Co. analytical-grade NaF was
1.9, which corresponds very closely to the molecu-
far ratio in the complex UF -3NaF.

PART Ill. SHIELDING RESEARCH
11, Shielding Analysis

The air-scattered gamma-ray dose rate predicted
by theory for a Co0 source at a source-detector
distance of 15 meters was compared with ex-
perimental measurements in a similar geometry,
The two results were found to be in substantial
agreement,

The coding of a Monte Carlo calculation of heat
generation resulting from the transport of gamma
radiation through shields with stratified slab
geometry has been completed. Preliminary results
are in good agreement with experimental results
obtained for lead.

A set of integral equations was derived to
determine the flux density near a thin foil. The
calculation of the flux in the foil interior is carried
out as if the flux from sources outside the foil
were constant for the small part of the foil con-
sidered. Similarly, a set of integral equations
was derived to determine the flux in air in the
presence of the ground. The flux in the ground
near any surface point is calculated as if the
surface flux were constant over the infinite plane
and equal to its value at the surface point.
S e A

12. Shield Design

Methods were devised for calculating the sodium
activation in the heat exchangers of circulating-
fuel reactors, The activation from core neutrons
was determined directly from Lid Tank Shielding
Facility (LTSF) experimental data obtained during
static source tests with mockups of a circulating-
fuel reflector-moderated reactor. The contribution
of prompt neutrons was calculated from the ex-
perimental data by the application of conventional
shielding transformations. The activation induced
by delayed neutrons emitted in the heat exchanger
was estimoted from the experimental data taken
at the LTSF by using a moving-belt source to
simulate the circulating fuel.

The total saturated activities were calculated
for four reactors. The activities of three of the
reactors, with power levels of 300 Mw and be-
ryllium reflector thicknesses of 8, 12, and 16 in.,
were 21,270, 4,880, and 1,680 curies, respectively.
The activity of the fourth reactor, which corre-
sponded to the 60-Mw ART with an 11-in.-thick

beryllium reflector, was 531 curies.

13. Lid Tank Shielding Facility

The static source tests and corresponding
analyses of the recent series of experiments on
mockups of a circulating-fuel reflector-moderated
reactor and shield (RMR.shield) have been com-
pleted. The tests included an investigation of
the effect of varying the lead thicknesses on the
dose rate and flux measurements. Various other
tests were run to determine the sources of the
secondary gamma-ray dose rate in the shield, It
appears that the unattenuated secondary dose rate
is made up primarily of capture gamma rays from
boron, lead, and hydrogen. The attenuated
secondary dose rate is made up primarily of
hydrogen and lead capture gamma rays attenuated
through a portion (1Y% in.) of the lead shield which
was spaced out in the water shield. The effect
of placing a heavy metal (bismuth or copper) in
the reflector region and eliminating an equivalent
thickness of the lead shielding was also investi-

PERIOD ENDING DECEMBER 10, 1955

gated. The production of secondary gamma rays
by the bismuth or copper appears to make this
impracticable, especially for moderate lead thick-
nesses (up to 4% in.)., The addition of boron
(1.95 wt %) to ti\e water shield was found to
reduce the gamma-ray dose rate by a factor of 2.3
and the thermal-neutron flux by a factor of about

15.

The radiation measurements behind the RMR-
shield mockups with a circulating fuel belt as
the source have been completed, but they have
not yet been analyzed. The measurements of the
activation of sodium in the heat exchanger for
various transit times of the circulating fuel belt
showed that rotation with a 1.25-sec transit time
increased the activation rate approximately 20%.
For the case of no rotation, replacement of one
of the boral sheets preceding the heat exchanger
with a 1-in. thickness of hydrogenous material
reduced the total activation from core neutrons
by 60%. There has been no correlation, as yet,
of these measurements with those taken in the
static tests because of the uncertainty of the
fission rate of the fuel in the circulating belt.

The Lid Tank Shielding Facility circulating fuel
belt and the Bulk Shielding Facility gamma-ray
spectrometer were used for the first experimental
energy-spectrum measurement ever made for gamma
rays from the gross fission-product mixture. Boral,
lead, ond water were placed inside the circum-
ference of the fuel loop to reduce as much as
possible the background caused by sources other
than the section of the belt viewed by the spec-
trometer collimator, Although a complete analysis
is not yet available, there was apparently no
variation either in the shape or in the intensity
of the spectrum as the transit time of the belt was
changed. Integration of a typical photon energy
spectrum between 0.36 and 5.8 Mev gave averages
of 4.2 photons/fission and 4.8 Mev/fission. An
estimated probable error of 120% is assigned to
these averages and to the normalization of the
experimental spectrum,

11
Part |

REACTOR THEORY, COMPONENT DEVELOPMENT,

AND CONSTRUCTION
1. REFLECTOR-MODERATED REACTOR

A. P. Fraoas

W. G. Piper

H. W. Savage

Aircraft Reactor Engineering Division

E. R. Mann

Instrumentation and Controls Division

ART FACILITY DESIGN AND CONSTRUCTION
F. R. McQuilkin

Aircraft Reactor Engineering Division

Design and construction work on the Aircraft
Reactor Test (ART) facility continued. Package 1
construction, which consists in alterations to the
original building, an addition to the building, and
installation of the reactor cell, was 10% com-
pleted at the end of the quarter; the scheduled
completion at that time was 20%. Much of the
deloy has been caused by the inability of the
contractor’'s vendors to supply reinforcing steel
and electrical items on schedule. The principal
work accomplished to date has consisted in the
demolition of portions of the existing building; the
excavation for building and cell foundations; the
provision of reinforced concrete for the cell founda-
tion, building footers and celumns, stack footers,
charcoal adsorber tank, instrument and control
tunnel, spectrometer room wall, spectrometer
tunnel, and switch-house floor; the erection of
structural steel for reinforcing the existing building
high bay; the installation of columns and girders
for 30-ton crane service; and the placement of
conduit and a 1500-kva transformer. Much of this
work can be seen in Fig. 1.1, which shows a view
of the south end of the existing building. The cell
foundation, which may be seen at the bottom center
of the picture, is at an elevation of 816 ft 6 in.
The floor of the instrument and conirol tunnel,
which is above the cell foundation, is at an
elevation of 831 ft, and the basement floor is at
an elevation of 840 ft. The building main floor is
at 852 ft, which is the elevation at the top of
the exposed grade beam above the tunnel. The
building addition columns appear on the left, with
the switch house (floor elevation, 840 ft) beyond,
Figure 1.2 is a view, looking south, taken from
atop the 852-ft-elevation grade beam, In the
right center can be seen the charcoal adsorber
tank for the off-gas piping system, and, on the
left, beyond the building columns, is the end of
the forms for the spectrometer tunnel, Figure 1.3
shows the forms for the spectrometer tunnel and

the spectrometer room wall which it joins, as well
as building addition columns and grade beams,

Steps were taken to incorporate in package 1
construction a major design change which is
required so that more space will be provided for
the NaK system piping and equipment, as well as
for a special heat dump facility for the removal of
heat from the fuel fill-and-drain tank. This ad-
ditional space is to be obtained by lowering the
radiator pit floor 5 ft 9 in,, to an elevation of
820 ft, and by constructing e vaultlike room ad-
jacent to the radiator pit.

Drawings and specifications for a modified
version of package 2 work were completed, This
unit of work consists in the installation of the
diesel generators (AEC-supplied) and facility,
the electrical control center, and the spectrometer
room electrical and air-conditioning equipment.
Release of the drawings and of specifications for
lumpesum bidding is scheduled for December 6,
1955. Bids are to be opened on January 10, 1956,
and the work is to be completed on about June 15,
1956. For the auxiliary power supply, the Allis-
Chalmers Mfg. Co. has contracted to supply four
of the 300-kva diesel-generator units, while the
fifth, a Caterpillar unit, will be transferred from
the Y-12 plant.

The piping work for nitrogen, air, cooling water,
helium, lubricating oil, and hydraulic oil drive
systems was removed from the package 2 work unit
and was designated as package A, This relatively
small unit of work will be negotiated either with
the package 1 contractor or a fixed-fee contractor,
Design work and job completion are scheduled for
January 10, 1956, and June 1, 1956, respectively.

Package 3 work, which covers the experimental
instruments, controls, process lines, and process
equipment, is in the design phase, Sufficient
design work was completed on the NaK piping and
associated equipment, which will be installed in
the radiator pit and in the special equipment room,
to permit modification and release of structural
drawings affected. Package 3 work will be done

by ORNL forces.

15
ANP PROJECT PROGRESS REPORT

oS, T [ g
G T T
“1 %

S
o

LA L Ay

 

 

 

 

 

 

 

Fig. 1.1. Aircraft Reactor Facility Construction as of December 5, 1955, Showing Cell Foundation,
Instrument and Control Tunnel, Building Addition Columns, and Switchhouse,

16
Ll

 

Fig. 1.2. Aircraft Reactor Facility Construction Showing Charcoal Adsorber Tank and End of Spectrometer Tunnel.
November 25, 1955.

"UNCLASSIFIED
PHOTO 15920

Photograph taken

S$S61 ‘0l ¥IEAWID3IA ONIANT @0I¥3d
8l

il ®

UNCLASSIFIED
PHOTO 15945

 

Fig. 1.3. Aircraft Reactor Test Facility Construction as of December 5, 1955, Showing Forms for Spectrometer Tunnel and the Spec-
trometer Room Wall,

LIdOdIY SSIJY20dd 1I23r0odd dNV
ART DESIGN

A, P. Fraas
Aircraft Reactor Engineering Division

System Flowsheets

W. B. Cottrell R. D. Schultheiss
L. A. Mann
Aircraft Reactor Engineering Division

System flowsheets and instrumentation lists have
been prepared in an initial attempt to define the
entire ART. The flowsheets include, in addition
to the various components of the particular system,
the annunciator pickup and presentation locations;
the operating conditions of temperature, pressure,
and flow rates; the control stations for electrically
operated components; the normal valve positions
and uniform valve identification; and line sizes,
Flowsheets have been prepared for the following:

Reactor Temperature Instrumentation
Fuel System

Sodium Systems

Fuel Enricher System

Fuel Recovery System

Fuel Sampling System

Main NaK System

Auxiliary NaK System

Special NaK System

Hydraulic System

Reactor Pumps Lubricating System
NaK Pumps Lubricating System
Helium System

Off-Gas System

Nitrogen System

Water System

Process Air System

Compressed Air System

The flowsheets are to be issued in preliminary
form for comments, and the necessary revisions
will be made before they are issued for con-
struction and fabrication use,

PERIOD ENDING DECEMBER 10, 1955

Fuel-to-NaK Heat Exchanger
R. |, Gray M. M. Yarosh

Aircraft Reactor Engineering Division

Layout drawings were completed for the fuel-to-
NaK heat exchanger! which indicated that a
12 by 24 array of 0.1875-in.-OD tubes on 0.2175-in,
centers would be acceptable for the heat exchanger
design; however, detail drawings have subsequently
showed that difficulties would be encountered in
overcoming interferences at the headers and that
problems might occur in the final assembly of the
heat exchangers. In addition, it was found de-
sirable to provide additional space in the region
of the headers for the beryllium support struts.
To overcome these difficulties and to provide for
uncertainties in the NaK friction factor for small-
diameter tubing (see Sec, 2, ‘‘Experimental Reactor
Engineering'’), the heat exchanger design was
modified, The revised design data are presented

in Tables 1.1 and 1.2,

 

R. D. Schultheiss, ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 19.

TABLE 1.1. DIMENSIONAL DATA FOR FUEL-TO-NaK
HEAT EXCHANGER

 

Tube dimensions, in,

Center-line spacing 0.250
Qutside diameter 0.230
Inside diameter 0.180
Wall thickness 0.025
Spacer thickness 0.020
Mean length 65.0
Equatorial crossing angle 26°20°
Inlet and outlet pipe, in.
Inside diameter 2.375
Outside diameter 2.875
Header sheet, in.
North head thickness 0.500
South head thickness 0.375
Fuel volume, £43 2.45
Number of tube bundles 12
Number of tubes per bundle, 260
13 by 20 array
Total number of tubes 2130

 

19
ANP PROJECT PROGRESS REPORT

TABLE 1.2, FUEL-TO-NaK HEAT EXCHANGER DESIGN DATA

 

Tube diameter, in,

Number of tubes per bundle
Number of bundles

Tube center-line spacing, in.
Tube wall thickness, in.
Tube array, square pitch
Mean tube length, ft

Fuel temperature range, °F

NaK temperature range, °F

0.230

260

12

0.250

0.025

13 by 20
5.42

1250 te 1600
1070 to 1500

Fuel pressure drop through heat exchanger, psi 39
NaK pressure drop through heot exchanger, psi 13
Fuel Reynolds number in heat exchanger 3740
NaK Reynolds number in heat exchanger 120,300
Fuel flow rate,* cfs 2.96
NaK flow rate,* cfs 10.45
Fuel volume in heat exchanger, 13 2.45
NaK volume in heat exchanger tubes (not including headers), £43 2.97
Heat exchanger thickness (includes 0.015-in, side wall clearance), in, 3.25
Log mean temperature difference, °F | 136
Estimated maximum tube wall temperature, neglecting secondary heating effects, °F 1535
Fuel mixture NaF-ZrF -UF, (50-46-4 mole %) heat transfer coefficient, Btu/hrft2-°F 2215
NaK (56% Na—44% K) heat transfer coefficient, Btu/hr-ft2:.°F 10,000
Heat load at design operating conditions, Mw 55

 

* At mean operating temperatures in reactor.

Fuel Fill-and-Drain System

J. Foster
Aircraft Reactor Engineering Division

An intensive study was made of the problem of
cooling the fuel fill-and-drain tank. The results
of analyses of boiling and circulating liquid and
gas cooling systems are presented in Table 1.3.
Of the fluids considered, NaK gives the most
promising system and was therefore chosen for the
ART.

Dual NaK systems have been designed in order
to provide maximum reliability of operation for
both cooling and heating the fuel fill-and-drain
tank, Each NaK system includes a pump, piping,
a radiator, an air blower, a duct, and controls and

20

is capable of removing the maximum heat load of
1.75 Mw from the fuel in the tank., The tank is a
horizontal cylinder, with each circular head being
a tube sheet fitted with U-shaped NaK tubes which
sweep through the fuel volume. Alternate rows of
these cooling tubes originate in the opposite heads
of the tank to provide adequate cooling throughout
the fuel mass even when one system is out of
service,

Reactor Shield
H. Wiard
Consolidated Vultee Aircraft Corporation

The ART lead and water reactor shield has been
designed to be essentially similar to a unit shield
4

TABLE 1.3, RESULTS OF ANALYSES OF COOLING SYSTEMS FOR FUEL FILL-AND-DRAIN TANK

 

 

 

Vapor Pressure VYapor Pipe
Coolant Temperature (°F) at Outlet Flow Rate of Coolant Diameter (in.)
_— T for Fl f Basis for Elimination
Material In Out emperature Ib/sec gpm fm or Flow o
{psia) 150 fps

Helium 1000 1400 3.8 17,500 at 20 Pipe size?
50 psig
and 1400°F

Nitrogen 1000 1400 20 13,200 at 18 Pipe size?
50 psig
and 1400°F

Boiling 1200 1200 1 1.05 8.9 48,000 at 30 Pipe size? and temperature vs pressureb

sodium 1 psia

and 1200°F

Boiling K 1200 1200 5 2.13 22 11,700 at 16 Pipe size? and temperature vs pressureb
5 psia and
1200°F

Boiling Rb 1200 1200 10 5.00 28 6250 at 12 Pipe size®®
10 psia

Boiling Hg 1200 1200 960 15 Pressure too high

Beiling H,0 705 (critical) 3206 Pressure too high

NaK 1200 1400 7.75 300 2.5 in, {liquid) Not rejected

(56-44 wt %)

 

The large pipe size is a basis for elimination for four reasons: (1) good flexibility will be required for expansion and contraction, (2) space is at a
premium, (3) large openings throughthe pressure vesselwall present serious problems, and (4) a large vapor pipe implies the need for large vapor passages

within the drain tank, where an increase in size would create a space problem and greatly increase shield weight.

bEliminafion because of low vapor pressure is based on the vapor pressure at the maximum allowable temperature not being high enough to give a well-

proportioned condenser,

“Space limitations would require condenser to be outside the pressure vessel.

SS61 ‘0l ¥39wW3D3A INIAN3 QO0I¥3d
ANP PROJECT PROGRESS REPORT

The basic mechanical
design work has been done for the support of the
7-in. layer of lead, with its cooling system, and
the 30.2-in.-thick layer of borated water, and the
procedure for the assembly and installation of the
reactor and shield has been studied. In an effort
to minimize the over-all weight, consideration was
first given to the use of a rubber container for the
borated water, but, as discussed below, rubber
proved to be inferior to metal for this application,
Of the various metals considered, 0.1875-in.-thick
61S aluminum alloy appears to be the most suitable
because it is easily formed and welded and is
sufficiently corrosion-resistant in borated water,
A weight summary of the installation is given in
Table 1.4, and the materials to be used are

described in Table 1.5.

for use in an airplane,

TABLE 1.4. WEIGHT SUMMARY OF REACTOR
AND SHIELD INSTALLATION

 

Reactor 12,000 ib
Lead shield

2 side castings {14,300 Ib each) 28,600

Bottom casting 1,500

Top casting 1,200

Lead burner material 3,700
Shield=water container 1,000
Shield water 35,000
Granulated BAC 1,000

84,000 1b*

 

*Does not include bridge support structure (™ 7000 Ib),

which is unique to the ART installation,

The feasibility of using rubber for the water
container was evaluated as to radiation stability,
weight, fabricability, method of support, and
anticipated service performance, With respect to
each of these criteria, rubber was found to be
inferior to metal for this application, The radi-
ation dose at the surface of the lead layer is
expected to be 10? rad in 1000 hr, and practically
all this dose will be in the form of fast neutrons,
This dose level is higher than was anticipated
earlier and would cause sufficient deterioration?
of the elastomer to reduce its strength to ap-
proximately 25% of its original value in 1000 hr,

22

TABLE 1,5. SHIELD AND SUPPORT MATERIALS

 

Component Material

 

61S-T6 aluminum alloy,
0.1875-in. sheet, butt-welded

Shield water con-

tainer

'l!

Lead shield

Probably **chemica

lead *

“type

Tubing for cooling Cold drawn or annealed copper;

system for lead tinned

Fiberfrax blanket; density,
6 Ib/f3

Insulation

Insulation pads Fiberfrax compressed to with-
stand approximately 100 psi

in compression

Bridge structure Boiler plate or equivalent

structural steel

4130 steel bar, heat-treated at
125,000 psi/min

Shield support studs

NaK line cover Stainless steel or Inconel

bellows and tube

 

*The composition of the lead to be used is yet to be

*“chemical’ type of

determined. At the present time the
lead appears to have the requisite structural and nuclear

properties.,

Rubber has approximately the same radiation-
absorption qualities as borated water, Therefore,
the weight penalty for the use of rubber would not
be large; that is, for every inch of rubber added,
an inch of water could be removed. However, in a
hypothetical aircraft installation study, the over-all
weight advantage appeared to favor the use of a
metal cell, It was further considered that the
fabrication and modification of the container would
be easier with metal than with rubber, since metal
can be formed readily and welded with available
tools and equipment, The problems associated
with supporting and maintaining the shape of a
rubber cell also made the metal cell appear more
desirable.

 

24, J. Stumpf and B. M. Wilner, Radiation Damage to
Elastomers, Lubricants, Fabrics, and Plastics for
Use in Nuclear-Powered Aircraft, ORNL CF-54-4-221
{April 15, 1954).
Alkylbenzene may be used instead of borated
water in an actual aircraft reactor installation
because it has a higher boiling temperature and
hence can be cooled by ram air, which will be at a
temperature near or above the boiling point of
water, In this instance, the higher temperature
would make rubber less desirable than metal as a
container, In addition, rubber deteriorates with
age and therefore would
placement.

require periodic re-

The reactor is to be supported at the flanges
of the fuel and sodium pump wells which will bear
on the cylindrical part of the bridge structure,
These pump flanges will be free to move sideways
relative to each other to accommodate the thermal
expansion of the reactor pressure shell,

The reactor bridge structure will not only support
the reactor but, through eight adjustable studs,
will alse support the lead and water shields, One
end of each stud will be screwed into a fitting
attached to the inner skin of the water cell, while
the other will be fastened to projections from the
cylindrical portion of the bridge. In effect, the
lead shield will be clamped inside the water cell,
whichwillact as a support cradle for it, Therefore,
since the reactor and the shields will be supported
from the same bridge structure, they will move as
a unit,

Both the lead and the water shields will be
fabricated in four parts: two hemispherical shells,
which will join along a longitude and leave an
opening at each pole, and two plugs, which will
effect the shielding closure at each pole. It is
planned that the two large hemispherical shells
of lead shield will be installed as prefabricated
and that the pole pieces will then be ‘‘puddled’’
on, The sides of the water shield will also be
separate units which will be joined together when
installed around the reactor., The cylindrical part
of the bridge structure will be filled with borated
water to serve both as the upper part of the water
shield and as the water expansion tank. Equipment
such as the fuel and sodium pumps, the control rod
housing, and the instruments that will penetrate
the shield in this region will be ‘‘canned’ to
prevent contact with the water. Sodium system fill
lines, shield cooling-water lines, and other un-
canned equipment will also be located in this
area, Access to these lines and equipment will
be through the open top of the cylindrical part of
the bridge structure, Almost all other equipment

PERIOD ENDING DECEMBER 10, 1955

and lines that will penetrate the shield will run
through a similar cylindrical water cell at the
bottom of the reactor.

The cold radial clearance between the reactor
and the lead shield will be 1 in., of which 0.5 in.
will be used for a layer of Fiberfrax insulation;
the remaining 0.5 in. is provided for differential
thermal expansion. A 0.25-in.-thick Johns Manville
Super X insulating blanket will be placed between
the lead shield and the water shield cell to ac-
commodate the differential expansion of the lead
and the water cell.

The vertical thermal expansion of the NaK lines
which will run through the water cell will be
accommodated by bellows (deflecting in shear)
attached to each end of a tube. The tube will be
installed over the ]/2-in.-thick Fiberfrax insulation
around the NaK lines and will be attached through
the bellows to the cell. Experiments currently in
progress ot the Lid Tank Shielding Facility (LTSF)
will determine the additional amount of gamma
shielding required around these NaK lines where
they pass through the lead.

Water-cooling lines will be provided for cooling
both the shield water and lead shield. One of the
main shield-water containers will be heated by the
NaK lines to beyond the capacity of the container
to dissipate the heat by convection, and therefore
a cooling coil will be required in the water shield,
The line from this cooling coil will be paraliel
to a line that will separate into three circuits for
cocling the lead and the ZrF, vapor trap, which
will probably be cast into the lead, Tubes em-
bedded in the lead will probably be tinned to aid
in providing a good bond with the lead.

CORE FLOW STUDIES

G. D. Whitman W. J. Stelzman
W. T. Furgerson
Aircraft Reactor Engineering Division

Further core flow studies were made on both the
full-scale aluminum and the plastic core models by
using the techniques described previously, with
the major emphasis on observing and photographing
the flow patterns obtained with various core inlet
configurations, However, some additional data
were obtained on the axial-flow system in the
aluminum model, Paired inlet guide vanes were
used that were designed on the basis of previously
obtained inlet-guide-vane and turbulence-generator
data. The reliability of the data obtained from

23
ANP PROJECT PROGRESS REPORT

these tests is in doubt, however, because of the
extreme difficulty encountered in locating the flow
direction and in measuring the total pressure in the
upper half of the core annulus. The plastic core
model will be modified to accommodate this iniet-
guide-vane configuration so that the flow patterns
can be photographed.

Photographic data were obtained with the header
designed to give a compound vortex? installed on
the plastic core model, Motion pictures were
taken at 64 and 500 frames per second, but experi-
ence showed that 500 frames per second was too
fast, lodine injections at several fixed levels
along the outer core shell indicated that the flow
along this surface had axial and rotational com-
ponents which were approximately equal in magni-
tude throughout the upper third of the annulus.
However, as the flow approached the equator,
the axial component diminished gradually until,
at the equator and throughout the fower half of the
annulus, it became, and remained, low. In general,
the peripheral velocities along this entire surface
were very high, as indicated by a very rapid
movement of the iodine. lodine injections along
the inner shell showed a pronounced upward axial
velocity component along the upper third of the
shell surface. This component diminished until,
at the equator, a good downward axial component
existed, The axial component continued to in-
crease until it equaled the rotational component
about three-fourths of the way through the core; it
again diminished throughout the lower fourth of
the annulus, and flow reversal occurred at the
outlet, The peripheral velocities along the inner
shell were very high, and there were indications of
pronounced stabilization of the boundary layer
throughout the lower third of the annulus.

Photographic data were also obtained with an
axial flow type of header without any means of
auxiliary flow guidance. lodine injection along
the outer shell indicated a flow pattern which was
similar to that obtained with the compound vortex
header, except that the peripheral velocities were
lower because of the larger pump-volute discharge-
nozzle areas. lodine injections along the inner
shell showed a flow reversal region that extended
from the inlet to slightly above the equator; how-
ever, it differed from the compound vortex flow

reversal in that the axial velocity component

 

3A. P. Fraas, ANP Quar. Prog. Rep. March 10, 1955,
ORNL-1864, p 20.

24

At the equator, the large
cross-sectional area gave a small axial velocity.
Below the equator, the axial velocity component
increased, and the direction of flow became more
stable, In general, the flow pattern showed rapid
fluctuations in local axial velocities but no major

seemed more unstable,

low-frequency irregular fluctuations.

Photographic data were also obtained with the
axial flow header and only one of the two pumps
operating. In general, the flow pattern did not
change much except that velocities were lower,
channeling in the lower region was more noticeable,
and the iodine holdup was visibly greater.

A fourth set of motion pictures was obtained with
the axial flow header equipped with a set of inlet
guide vanes (Mark |) and turbulators (5-9) attached
as illustrated previously.4 The flow pattern along
the outer shell approached an axial type of flow in
which the fluid gradually spiralled downward with
a maximum deviation of 30 deg from a direct,
downward path. The flow along the inner shell
showed no reversal in the region adjacent to the
turbulators, but, in the region extending from
slightly below the turbulators to below the equator,

it showed the same type of flow reversal as that

found when the guide vanes and turbulators were
not used, However, in the lower third of the core,
the axial and rotational flow components gradually
became equal. The velocities were about the same
as those obtained without the guide vanes and
turbulators,

Turbulators designated as type 4 were then
installed, and another set of motion pictures was
taken,
progressed generally downward but was extremely
unstable with regard to direction in the region
extending from directly below the turbulators to
about midway between the equator and core outlet,
Fromthis mid-point to the outlet the flow increased
in velocity and became more stable, The inner
shell surface again showed no flow reversal in the
region adjacent to the turbulators, but, in the
region extending from directly below the turbulators
to the equator, the flow was generally downward,
with spasmodic reversals, However, below the
equator the flow accelerated, as in previous runs.

For another set of motion pictures, the axial flow
header without inlet guide vanes or turbulators

In this case, flow aleng the outer shell

was used and simultaneous iodine injections were

 

4W. T. Furgerson et al.,, ANP Duar. Prog. Rep, Sept.
10, 1955, ORNL-1947, Fig.1.7.p 27.
made at two different points, either on the island
or shell surfaces or one on each, These injections
provided a better basis for comparing both the
relative flow directions and the twin stream
intermixing.

ENGINEERING TEST UNIT
M. Bender

Aircraft Reactor Engineering Division

An equipment layout was prepared of the Engi-
neering Test Unit (ETU), and work was started on
the design of the facilities required for operation
of the unit, A purchase order was placed to obtain
the required Inconel shells,

Specifications for the main heat exchangers, the
radiators, the boron carbide tiles, and the lube oil
pumps have been prepared for procurement, Nu-
merous design problems prevented the
issuance of firm drawings for the reactor com-
ponents, but procurement action has been started
in advance of complete design information so that
the time losses will be reduced. The advance
procurement action will not completely salvage the
time lost, and it is estimated that the ETU will
not be ready for operation before December 1956.
The previous schedule called for operation by

September 1, 1956.

have

CONTROLS AND INSTRUMENTATION

J. M. Eastman
Bendix Products Division

R. G. Affel E. R. Mann

Instrumentation and Controls Division

Instrument lists were prepared for each ART
system, and instrument sensor locations were
designated. The quantity of instrumentation
needed considerably exceeds the original esti-
mates, and the original instrument flowsheets are
being revised and expanded. Five permanent
instrument information centers will be used, in
addition to three temporary ones, and some instru-
ments will be located and read on the equipment,
The five permanent information centers will be in
the control room, the information room, the power
room, the auxiliary equipment room, and the vent
house. Temporary instrumentation will be used for
fuel enrichment, fuel sampling, and fuel recovery,
The layouts of the control and instrument panels
and boards are being prepared.

The elementary wiring diagrams have been

PERIOD ENDING DECEMBER 10, 1955

completed except for minor modifications that may
be necessary for them to be adapted to optimum
components. The information block diagrams have
been revised to include improvements worked out
in devising the elementary wiring diagrams.

The layout of the ART control rod and drive
mechanism is being prepared. Data on the per-
formance of the fuel enricher used for the high-
temperature critical experiment are being reviewed,
and layout work on the ART enricher and actuating
mechanism is expected to begin soon. A location
for the flux-sensing chambers and their actuating
mechanisms has been proposed, and a nuclear
evaluation is being made. Actuator motors and
synchronizing mechanisms have been included in
the specifications for the main radiators, the
auxiliary radiators, and the bypass air louvers.
A recommended design and the specifications have
been issued for the dump valve actuator mechanism,

The first of two high-temperature instrument de-
velopment loops is scheduled to start hot operation
late in December. It will be used for evaluating
items under actual operating conditions of pressure,
temperature, etc. Items currently being studied
include solenoid valves for liquid metal vapor
service, high-temperature pressure transmitters for
fuel and NaK service, ambient-temperature electrical
pressure transmitters, high-temperature flowmeters,
and high-temperature level-indicating devices.

PROCUREMENT OF SPECIAL REACTOR
MATERIALS AND COMPONENTS

W. F. Boudreau

Aircraft Reactor Engineering Division

The fabrication of many items of equipment for
the ART and the ETU requires the establishment
of combinations of facilities not presently available
in this country, [t will therefore be necessary
to transmit some of the very specialized welding,
brazing, and related fabrication techniques which
have been developed at ORNL to the manufacturers
of special equipment, The major projects currently
under way are described below.

Beryllium

[ndividual beryllium blocks, seven times heavier
than those which have been previously made, are
to be produced at the Luckey, Ohio, plant of The
Brush Beryllium Co. A lothe, with contour-
following equipment, has been provided at Brush’s
Cleveland plant. The facility for sintering and

25
ANP PROJECT PROGRESS REPORT

machining large blocks of beryilium, from which
spheres up to 54 in. in diameter can be machined,
is essentially complete and ready for tryout.

Shell Fabrication

The six sets of thin Inconel shells required for
the ART, which vary from 18 to 54 in. in diameter
and are fairly complex in shape, have unusually
close manufacturing tolerances. A comprehensive
discussion of this problem with a variety of manu-
facturers led to the conclusion that the newly
developed Hydrospin process is the only method
which promises to yield a solution. An order has
been given to the Lycoming Division of the Avco
Mfg. Corp. to enlarge the capacity of their
Hydrospin machine from 42 to 54 in. and to provide
the necessary annealing and pickling facilities,

CX-900 Inconel
The processing of 100,000 |b of special CX-900

Inconel is under way at the International Nickel
Company plant, and the material was released from
one manufacturing step to the next as the require-
ments gradually became somewhat firmer. For
example, a release was given during the quarter
to process the amount of redraw stock estimated to
be necessary for producing all the small-diameter
tubing for the heat exchangers and radiators.
Some of this was released for final tubing manu.
facture.

Main Heat Exchangers and Radiators

It has been recognized for some time that it
would be necessary for ORNL personnel to spend
a considerable amount of time with the manu-
facturers who have the experience and facilities
required for building the main heat exchangers and
radiators to develop welders qualified under the

26

special ORNL procedures. During the quarter,
the previously established policy of developing
this know-how by placing orders with outside
manufacturers for test heat exchangers and job
samples was continued and expanded.

When the main heat exchanger and radiator
drawings and specifications became available at
the end of September, requests for proposals to
furnish these assemblies were sent to about eight
of the sources that had been developed.

OPERATION OF ZrF“1 VYAPOR TRAPS IN THE
HIGH-TEMPERATURE CRITICAL EXPERIMENT

W. C. Tunnell

Aircraft Reactor Engineering Division

During operation of the high-temperature experi-
ment, deposits on the seats of the gas control
valves made it apparent that the ZrF , vapor was
creating a much worse problem than had been
expected. The cold traps installed in the system
removed little or no material during periods of
high-velocity gas flow, that is, during core filling
and fuel mixing operations. Large tanks were
installed in the gas lines, but these low-velocity
volumes did not reduce the deposits on the valves
or in the vent line. A cold trap consisting of
nickel Demisters packed in a tube had proved to be
successful on the fuel processing equipment
(low-velocity gas), but it was entirely inadequate
in this application, even when doubled in size.
VYenting of gases laden with ZrF, will be a con-
in the ART, as contrasted to
venting in the critical experiment

Consequently, the vapor deposition
problem will be greater,
vapor traps for reactor application is under way
(see Sec, 2, ‘“Experimental Reactor Engineering’’).

tinuous process
intermittent
fuel-loading.
Development work on
PERIOD ENDING DECEMBER 10, 71955

2. EXPERIMENTAL REACTOR ENGINEERING

H. W. Savage
Aircraft Reactor Engineering Division

An in-pile loop was inserted and successfully
operated in the HB-3 beam hole of the Materials
Testing Reactor (MTR). The difficulties en-
countered are described, and a summary of the
operating cycle is presented. Modifications that
are to be made in the nose purge system before
operation of another in-pile loop are discussed.
Fourteen forced-circulation loops were operated in
a series of corrosion and mass-transfer tests of
fused salts in Inconel and of liquid metals in
Inconel or stainless steel.

Seals are being tested for use in ART-type fuel
pumps, and water tests of a full-scale ART-type
sodium pump were completed. High-temperature
tests of an ART MF-2 fuel pump with NaK are
under way, and stands for testing models of this
pump with fuel are being fabricated. Stands for
testing the sodium and NaK pumps are also being
designed,.

The results of heat exchanger and radiator tests
are presented, and the data obtained on heat trans-
fer and pressure drop are correlated.

The status of the work on apparatus for thermal-
stability tests of the outer core shell of the ART
is given, and the results of Inconel strain-cycling
tests are discussed. Development tests on the
ART dump valve, a cold trap, a plugging indica.
tor, and a zirconium fluoride vapor trap are de-
scribed, An aluminum mockup of the top of the
ART is being fabricated for water tests.

IN-PILE LOOP DEVELOPMENT AND TESTS
D. B. Trauger

Aircraft Reactor Engineering Division

In-Pile Loop Operation
L. P. Carpenter J. A, Conlin
C. W. Cunningham P. A, Gnadt
Aircraft Reactor Engineering Division
C. C. Bolta D. M. Haines
Pratt & Whitney Aircraft
The second in-pile loop was installed in the
HB-3 beam hole of the MTR during the reactor
shutdown week of September 12, 1955. It was
brought to temperature with Calrod heaters, and
operation of the pump was started satisfactorily.

Some of the fuel mixture had been inadvertently
allowed to enter the pump during the filling of the
loop fill tank, and hence the pump could not be
operated until the loop was heated, Upon attempt-
ing to increase the temperature of the fill line to
above the melting point of the fuel mixture, the
electrical heating circuit failed, and it was impos-
sible to fill the loop. Later analyses showed that
the failure was due to electrical breakdown through
the helium blanket gas in the nose region or to
breakdown of the Sauereisen cement which pro-
tected the junctions at the Calrod ends. This
loop was returned to Ock Ridge for possible re-
building or salvage of parts.

A third loop had been assembled and shipped to
the MTR early in September. It was leak-checked
at the MTR and found to contain a small leak
through the pump bulkhead and another in the heat
exchanger. Both leaks were exceedingly small
and were considered to be tolerable. This loop
was installed in the reactor during the shutdown
period beginning October 3. |t was brought to
temperature, filled with salt (NaF-ZrF -UF ,
53.5-40-6.5 mole %), and started on isothermal
operation on October 5.

Reactor operation began on October 7 and
proceeded to full power on the moming of October
9. The loop was operated in the retracted position
for one day and then was inserted to a position
3.5 in. from the end of the beam hole, at which
point the maximum air-cooling capacity was
reached. A mixed-mean temperature differential of
approximately 150°F was reached, with a maximum
temperature of 1500°F. The coeling-air control
system for maintaining Ioop temperatures re-
sponded well, and it was possible to follow
changes in reactor power with only minor adjust-
ments of the control set point, There was some
leakage of neutrons through the concrete plug
shield, but this was easily stopped with a few
blocks of paraffin. The gamma activity at the
cubicle door, apparently originating from the fis-
sion gas lines in the cubicle, was quite high, and
considerable external shielding was required. The
water activity was slightly higher than expected,
but no other activity problem was encountered.
After four days of operation the helium purge line

27
ANP PROJECT PROGRESS REPORT

from the bearing housing was observed to plug
suddenly. Operation was continued, however, and
no increase in activity was observed except for
back diffusion of gases through the normal inlet
line to this region. At the entrance to the cubicle
this line had an activity of approximately 2 r/hr,
After four additional days of operation, it was ob-
served that the exit line for purge gas from the
pump sump was plugged. This plug caused rapid
buildup of activity in the purge-gas inlet line out-
side the cubicle, and it was deemed necessary tfo
crimp and clamp both the bearing housing and the
pump-sump inlet lines at the cubicle to prevent
high radiation levels at the instrument panel. The
loop was withdrawn to the retracted position to
minimize the activity generation and the power
density, and operation was continued. The reduc-
tion in power density decreased the hazards that
might arise from loop misoperation as a result of
the plugged lines. The activity on the pump-sump
inlet line built up to a stable value of approxi-
mately 25 r/hr at the point where it was clamped,
approximately 25 ft from the pump sump.

Irregularities in the pump operation were noted
after the purge lines had plugged. These irregu-
larities appeared as increases in pressure of the
hydraulic motor oil, although nearly constant oil
flow and pump speed were maintained. The oil
pressure would build up and decrease intermit-
tently with increasing frequency and magnitude
over a period of several hours and then retumn to a
normal value and remain there for several hours
before repeating the cycie. These deviations
continued with increased severity until pump
speed changes were observed. On September 22,
one of these cycles resulted in a pump speed
change which scrammed the reactor through the
loop control circuitry, Since the reactor was
within 20 hr of the end of its operating cycle, it
was not retumed to power. |sothermal operation
was continued for two more days before complete
shutdown and removal of the loop.

Connections were made to the clamped lines
after shutdown to facilitate purging the residual
gases to the stack. The charcoal traps were
warmed slowly, and the gases were continuously
vented to the stack. Some increase above the
normal background was observed in the stack ac-
tivity during these operations. However, there
was no fission gas released from the loop at any

28

time during the run or during the shutdown and
purging operations.

The reactor power during this operation was quite
erratic because of numerous scrams caused by
other experiments and inadvertent dropping of con-
trol rods by the reactor-operating mechanisms.
This erratic operation produced drastic power ex-
cursions, and full use had to be made of the Cairod
heaters to avoid freezing of the fuel in the loop.
Further, sizable strains must have been produced
in the loop by thermal expansion and contraction
during the rapid transitions from the operating tem-
perature differential to isothermal conditions and
during a certain amount of temperature cycling
while the loop was being returned to power opera-
tion. The excursions produced no observable
damage to the loop except for electrical leakage
of some of the Calrod heater circuits, The gas
leak in the heat exchanger did not increase notice-
ably, Following the shutdown, a much larger leak
was observed in the pump bulkhead than that which
had existed previously, but the cause for the in-
crease is unknown,

A resumé of the operating cycle is shown in
Fig. 2.1, in which high and low operating temper-
atures are plotted as a function of time. In
Fig. 2.2, the mixed-mean inlet and outlet tempera-
tures are shown as a function of time, The temper-
ature pattern in the nose coil heating section is
shown in Fig. 2.3. A marked difference can be
noted between temperatures on the inner and outer
radii of the coil.
thermocouples in this region were unstable at

Unfortunately, several of the

times, and the recorded temperatures are subject
to considerable uncertainty, It is evident that flow
asymmetry existed, but similar observations must
be made for future loops before the magnitude of
the asymmetry can be determined.

The thermocouples from which the mixed-mean
temperature measurements were obtained and which
gave the highest temperature differentials were the
most consistent during operation. Power calcula-
tions based on air flow and air temperature dif-
ferentials confirmed these values. The loop power
generation was estimated from these data to have
been 20 kw at the position of maximum insertion.
This gives an average power density of 0.6 kw per
cubic centimeter of fuel in the nose section of the
loop. Disassembly work on the loop is described
in Sec. 8, “‘Radiation Damage.’’
PERIOD ENDING DECEMBER 10, 1955

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

& ARG, M
e ORNL-LR-DWG 11258
HOURS OF OPERATION
0 100 200 300 400 500
1700
[ [ [ ‘ F ‘ \ | ! |
0 MAXIMUM TEMPERATURE
® LOW MIXED-MEAN TEMPERATURE
1600 ‘
1500 P 9 %cmfiwdyam 0 P
O of 0
¢ 0 o &=
o
of 2
1400
OE o e o, . FUN
e o . 0, .
w ¢ - o | e e o ° tu, O3 ¥ 02, 0,09 R K L N ©
g ? Py oy, ¢ .’”Omo o 2y, TeTHTemYe o 8,00,
S 1300 o) %P ot .
a- o @m:’
a .”*.ns. T
1'
2 oM
F—
1200
LOOP
LOOP INSERTED LOOP INSERTED | (RETRACTED, | LOOP RETRACTED
1100 - I
LoOP LooP ISOTHERMAL |
RETRACTED ] ISOTHERMAL INSERTED Tl
|t ISOTHERMAL —~=—" B - = REACTOR POWER
1000 | |
900 LI i | | | |
4 5 & 7 8 9 40 N 42 {3 14 15 46 (7 {8 19 20 2 22 23 24 25 26

Fig. 2.1, High and Low Fuel-Mixture Temperatures vs Time of Operation of In-Pile Loop No, 3 in MTR.

Loop Purge System
P. A. Gnadt A. A, Abbatiello

Aircraft Reactor Engineering Division

Electrical-system failures prevented the startup
of the first two in-pile loops, and therefore a study
was made of the electrical properties of loop in-
sulating materials in various gaseous atmospheres.
Because of its low neutron activation and its low
adsorption on liquid«nitrogen-cooled charcoal, he-
lium had been chosen originally as the radioac-
tivity-sensing gas for the nose section. However,
helium has poor electrical-insulation properties,
especially at elevated temperatures, whereas nitro-
gen, in comparison, is much less subject to elec-
trical breakdown. Tests were made of electrical
leakage of actual and simulated leads ond termi-

nals in atmospheres of both gases. Electrical-
insulation tests also indicated that Sauereisen
No. 1 cement, which was used at the teminals of
the heaters, tended to conduct currents at elevated
temperatures. Comparative curves for the onset of
electrical leakage as a function of voltage and
temperature for the two gases and for Sauereisen
cement insulation are presented in Fig. 2.4,

The use of nitrogen as a purge gas in the nose
section presents some additional problems, be-
cause the liquid-nitrogen-cooled charcoal trap
would adsorb large quantities of gaseous nitrogen,
and its effectiveness would thus be reduced. Also,
impurities in the nitrogen might cause the traps to
plug. As an expedient measure for operation of
loop No. 3, nitrogen was used for purging the nose
section, but it bypassed the traps during startup

29
ANP PROJECT PROGRE

$§S REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ORNL-LR-DWG 11259
HOURS OF OPERATION
0 100 200 300 400 500
1700 I ? W | ’ ‘ ‘ | ] i |
& QUTLET TEMPERATURE
© INLET TEMPERATURE
1600 B
1 _
1500 — by
. A
pon] | a5
| i Ay
85
o4 A Afi%
1400 A0 T_ ] |
|
o A ‘ Pl : s o 00
= 5 o, 0 P ‘ i | mfi% S o000y, :
& vy 3%) o0 M o oeene o 1
£ 1300 = 84— | o & | 6‘%
x < .
& ‘ 1
| aud i i
1200 ‘ - L
3 . ! :
‘ ‘ Loop | |
LOOP INSERTED || LOOP INSERTED | |RETRACTED,| | |LOOP RETRACTED
! |
oo ISOTHERMAL
LOOP . LOOP
RETRACTED =] ISOTHERMAL | INSERTED I
REACTOR
=t~ ISOTHERMAL POWER—1 REACTOR POWER
1000 | | i ’
f
900 l | | [ ] |
a 5 & 7 8 9 10 M {2 43 14 {5 16 17 18 18 20 20 22 23 24 25 26
DAYS

Fig. 2.2. Mixed-Mean Inlet and Outlet Fuel-Mixture Temperatures vs Time of Operation of In-Pile Loop

No. 3 in MTR,

and when the reactor was shutdown. Helium was
used, as originally planned, during reactor opera-
tion when the electrical heat was not needed. Ex-
periments are being conducted(see Sec.8, ‘*Radia-
tion Damage’) to determine the holdup time of
fission gases in the charcoal adsorption traps
when nitrogen is used as the nose purge gas.
Radioactive krypton is used fo simulate the fis-
sion gases, and trap temperatures of 0°C to below
~110°C are being studied. A design for using a
nitrogen atmosphere in the nose section of the
loop is being prepared, and grade A Lava sleeves
are to be substituted for Savereisen cement at the
heater terminals in future loops.

30

FORCED-CIRCULATION CORROSION AND
MASS TRANSFER TESTS

W. B. McDonald
Aircraft Reactor Engineering Division

Fused-Salt—Inconel Systems

C. P. Coughlen P. G. Smith

Aircraft Reactor Engineering Division

R. A. Dreisbach
Pratt & Whitney Aircraft

Five electrically heated and three gas-heated
fused-salt—~Inconel forced-circulation loops were
operated during the quarter., The operating condi-
PERIOD ENDING DECEMBER 10, 1955

W
ORNL~LR-DWG 14260

 

 

 

 

 

 

 

 

 

 

 

 

 

1500 o
o O |
@
i
o . ’
n !
i
» :
1400 )
o |
g PUMP SUCTION TEMPERATURE "
" |
a
E . PUMP DISCHARGE TEMPERATURE
o
&
= ®
w
|—.
1300
REACTOR POWER: 40 Mw ® MIXED- MEAN TEMPERATURE
PUMP SUCTION TEMPERATURE - 4375 °F O NOSE TEMPERATURE AT INNER RADIUS
pUMp DESCHARGE TEMPERATURE‘ ‘13’450': . NOSE TEMPERATURE AT OUTER RADIUS
NOSE POSITION: 3.5in. FROM END OF BEAM HOLE
1200
40 50 60 70 80 30

DISTANCE FROM PUMP (in.)

Fig. 2.3. Temperature Pattern in Nose Section of In-Pile Loop No, 3.

UNCLASSIFIED
ORNL-LR-DWG 11264

T T T T

4 NITROGEN ATMOSPHERE - OPEN TERMINALS
O NITROGEN ATMOSPHERE — SAUEREISEN NO.4

 

 

 

 

1200
CEMENT ON IMPREGNATED QUARTZ TAPE
_ A HELIUM ATMOSPHERE - OPEN TERMINALS
z ® HELIUM ATMOSPHERE - SAUEREISEN NO.4
‘2 1000 CEMENT ON IMPREGNATED QUARTZ TAPE
=
= NOTE: CURVE POINTS OBTAINED AT VOLTAGE
b o LEVEL WHERE THERE WAS APPROXIMATELY
o 1 mo OF CURRENT FLOW BETWEEN TERMINAL
S AND SHEATH
Q
<[
2 600
<[
= \
2 N
6 0—-—-(\ a e ——
% 400

N\
200 N \
NN

 

 

 

 

 

 

 

 

 

\.
0
0 200 400 600 800 1000 4200
TEMPERATURE (°F)
Fig. 2.4, Electrical Breakdown Across Calrod

Terminals in Nitrogen and in Helium Atmospheres.

tions are summarized in Table 2.1. The results of
metallurgical examinations of the loops are pre-
sented in Sec. 5, ‘‘Corrosion Research.”

In addition to the routine loop operations, «
study was made of temperature measurements and
heat distribution in the gas-heated loops, and the
electrical heaters of the electric-resistance-heated
loops were modified so that different ratios of
fluid-to-wall temperatures could be obtained.

Uniform wall temperatures cannot be obtained in
the heated section of loops operated in the small
gas furnaces, because of improper design of the
furnaces for this application, These furnaces
were originally designed as heat sources for loops
for testingthe corrosiveness of circulating sodium.
In order to utilize them as heat sources for the
fused-salt—Inconel forced-circulation loops, they
must be operated above rated power, Under the
operating conditions required, much of the combus-
tion takes place in the chamber around the heater
coils rather than in the combustion chamber, The
resultant hot spots in the furnace cause nonuniform
pipe-wall temperatures,

31
ANP PROJECT PROGRESS REPORT

TA?E‘EM&L SUMMARY OF OPERATING CONDITIONS FOR FUSED-SALT-INCONEL FORCED-CIRCULATION
LOOPS THAT CIRCULATED NuF-ZrF4-UF4 (50-46-4 mole %) FOR 1000 hr

 

 

 

Reynolds Te'mperatt.lre Maximum Recorded Temperatures (°F)
Foop Ne- Number D'H;:r:mm' Fluid Tube Wall
)
Loops with Direct-Resistance-Heated Straight Sections

7425-5 2,600 200 1500 1550

-6 10,000 200 1500 1565

-7A 10,000 200 1500

-41 2,750 200 1635 1700

-42 Variable 200 1300 to 1650 1700

Loops with Gas-Heated Coiled Sections

4935-6* 8,000 200 1500 1575

-7B 8,000 200 1500 1550, changed to 1700 at 617 hr
7425-71 Variable Variable Variable Variable

 

*Thermocouple failures made loop unsatisfactory for planned test. Loop now being operated as a life test. On

December 1, 1955, it had operated for more than 4500 hr with a temperature differential imposed.

The excessive cost of revamping the gas fur-
naces and the availability of electrically heated
test stands brought about the decision to discon-
tinue the use of gas furnaces for carrying out the
fused-salt corrosion program. These furnaces will
be utilized forsalt and liquid-metal life-test loops.

A series of forced-circulation corrosion-testing
loops with NaF-ZrF -UF ,(50-46-4 mole %) flowing
in Inconel tubing is being operated with the wall
temperature of the heated section held constant at
1700°F and the maximum fluid temperature held at
1300, 1500, or 1650°F; further, the temperature
differential in the circulating fluid is maintained
constant at 200°F, In order to obtain these condi-
tions without changing the surface-to-volume ratio
of the loop, it was necessary to modify the loop
design to give greater flexibility of power input to
each of the two heater sections. Four equally
spaced heater lugs were added to the entrance
section of the heater, and one lug was added to
the exit section. By shifting the power-transformer
cable connections to alter the length (and effec-
tive resistance) of the two heater sections, the
power input to each section could be varied over a
wide range, Loop 7425-41 is the only loop that
has been operated in this series, to date,

32

Liquid Metals in Multimetal Loops
C. P. Coughlen P. G. Smith

Aircraft Reactor Engineering Division

Six forced-circulation loops were operated with
sodium and noneutectic NaK. These loops were
heated electrically and had economizers in the
system. The results of metallurgical examination
of these loops are presented in Sec. 5, *““Corrosion
Research.” A summary of the operating conditions
is given in Table 2,2,

Test loop 7426-4 is the loop, described pre-
viously,! that was designed for obtaining accurate
information on the oxygen content of the sodium
during operation. Tests conducted with this loop
have shown that the plugging indicator does give
an indication of the contamination in the sodium,
However, the contaminant may not be sodium oxide
alone, since the saturation temperatures at the
plugging indicator stay consistently 250 to 300°F
higher than those at the cold trap. This loop has
operated at fluid temperatures up to 1500°F and
with system temperature differentials of 300°F,
Under these conditions corrosion and mass trans-
fer occur, and the contaminants may comprise

 

e, P Coughlen and R. A. Dreisbach, ANP Quar.
Prog. Rep. Sept. 10, 1955, ORNL-1947, p 38.
PERIOD ENDING DECEMBER 10, 1955

TABLE 2.2, SUMMARY OF OPERATING CONDITIONS FOR FORCED-CIRCULATION LOOPS THAT
CIRCULATED SODIUM OR NaK

 

 

Maximum
Loop Material of Cold Reynolds Temperature Recorded Fluid Operating Operating
No. Construction Trap Number lefe(:ennal Temperature Fluid Time
7426-2  Inconel No 59,000 300 1500 Sedium 1000
(0.05% oxide)
-3  Inconel Yes 50,000 200 1000 Sodium 1000
(as delivered)
-4*  |Inconel Yes 59,000 300 1500 Sodium Indefinite
(as delivered)
-5 Type 316 Yes 59,000 300 1500 Sodium 1000
stainless steel {as delivered)
7439-1  Inconel Yes 65,000 300 1500 Noneutectic NaK 1000
(as delivered)
-2 Incone! No 65,000 300 1500 Noneutectic NaK 1000

(as delivered)

 

*This loop includes a vacuum analyzer and a bypass loop with a plugging indicator for oxide determinations.

metal particles and metal oxides other than sodium
oxide,

The small-scale NaK-circulating loop described
previously? was also operated, The design was
found to be inadequate in that the surface area of
the small-diameter, %6-in.-OD, 0.025-in.-wall tub-
ing was insufficient for transferring, by radiation,
the amount of heat necessary to maintain the maxi-
mum design fluid temperature of 1500°F, and the
design flow of 1 gpm was not obtained. The flow
rate of 0.8 gpm obtained was probably limited by
the pressure drop, A new heater has been de-
signed to increase heat transfer, and the flow-path
cross section has been increased by adding one
tube to give a higher flow rate. The tube bundie
in the heater section is now surrounded by a NaK
bath fo increase the heat transfer from the clam-
shell heaters to the tube bundle, The pressure
drop across the heated section will be established
experimentally with water.

 

21bid., Fig. 2.2, p 41.

PUMP DEVELOPMENT

E. R. Dytko
Pratt & Whitney Aircraft

A. G, Grindell

Aircraft Reactor Engineering Division

Bearing-and+-Seal Tests

W. L. Snapp
Pratt & Whitney Aircraft

W. K. Stair
University of Tennessee

Seals of various designs are being tested and
evaluated for possible use in the ART as the
lower seal of the type MF-2 pump., The Dura-
metallic Corporation of Kalamazeo, Michigan, has
supplied, for testing, a group of so-called ‘‘Dura
seals' that consist of a Stellite-faced stainless
steel seal ring running against a gland insert of
No. @ carbon. Without exception, leakage rates
for these seals were excessive, averaging 50 to
100 cm3/hr, with extremes being as high as 200
cm3/hr.  (The specifications state that the lower

33
ANP PROJECT PROGRESS REPORT

seal is to have a leakage of oil into helium across
a pressure differential of 0 to 5 psi of not more
than 2 cm3 per 24 hr.) As a secondary shaft seal,
the Dura seals employ a stainless steel U-cup, and
it was thought that the excessive leakage was
probably due to surface irregularities on the seal-
ing surfaces of this U-cup. To smooth out the
irregularities, fixtures were made that would permit
honing of the surfaces. Also, to further verify
that the high leakage occurred at the U-cup seadl,
the U-cup of one Dura seal assembly was replaced
with a Buna N Oering, It was found that both
measures had a marked effect on the seal per.
formance. The honed U-cups reduced the leakage
to 10 to 15 em3/hr, and the Buna N O-ring reduced
the leakage rate to 1 to 2 cm®/hr.

Tests were also conducted to check the effects
on seal performance of varying the differential
pressure across the seal, the lubrication-fluid tem-
perature, and the shaft rotational speed. The
leakage was found to increase with increasing dif-
ferential pressure and with increasing lubrication-
fluid temperature, The relation of the leakage rate
to the shaft rotational speed is shown in Fig. 2.5.

UNCLASSIFIED
ORNL—LR—DWG {1262

 

   

 

 

 

3000 ~ ‘ T
|
|
2500 |— «, e
2000 [ - |
£ ;
o
& 1500 - :
L
o ! |
iz ! !
1000 - —-——--fi-—AJ e b
| | i {
MATERIAL: RUNNER, STELLITE
NOSE. CARBON NO.9
so0 L. | LUBRICATION FLUID: GULFSPIN 60 AT 200°F |
| PRELOAD: 351b i
! PRESSURE DIFFERENTIAL: O.5 psi |
: I !
O L i : J L
10 15 20 .25 30 35 40
SEAL LEAKAGE {cm3/hr)
Fig. 2.5. Leakage Rate of a Typical Dura Seal

as a Function of Shaft Rotational Speed,

34

Since the leakage rate was linear with speed, the
possibility of induced mechanical vibration being
a contributing factor in theleakage was eliminated.
The linear relationship did indicate, however, the
possibility of a hydrodynamic wedge building up
at the seal interface and forcing the seal faces
apart, Additional tests were run which completely
verified this and, consequently, pointed out a
basic limitation of this seal in its application in
the ART.

A second type of Dura seal was also tested; it
was identical with the seal described above, ex-
cept that the Stellite face was replaced with a
ceramic {zirconium oxide) face. The tests showed
that the initial leakage rates, although much lower
than the leakage rates for the Stellite-faced seals,
were still too high, averaging 5.6 ¢cm? per 24 hr,
However, by increasing the surface area of the
seal-gland carbon insert, the seal leakage rate
was further reduced, to below 2 cm3 per 24 hr,
which meets the specifications. This low leakage
rate has been obtained with only this one unit, and
it may or may not be reproducible. Further tests
are planned which will establish the performance
characteristics of this modified Dura seal.

In other tests, Fulton Sylphon seals with
Graphitar No, 14 noses against case-hardened
steel runners have proved to be quite successful,
Two such units were tested during this quarter.
One unit showed no leakage in over 200 hr before
being placed in a high-temperature pump test
stand, The other unit has been under continuous
testing for 1200 hr, with a nearly uniform leakage
rate of 0.75 cm? per 24 hr.

A spring-loaded bellows seal developedat ORNL
showed no leakage in 132 hr of testing, and then a
slow leak was noted. This seal, which has a
carbon insert, includes multiple springs located
inside the bellows to provide the seating load.
The bellows acts as the secondary static seal,
The seal is designed so that increased gas pres-
sure increases the seating load. Inspection
showed that the seal interface was perfect and
that the measured leakage was seeping through
the carbon nose on the stationary member. A
higher-density carbon nose is to be installed be-
fore evaluation tests of this seal assembly are
continued.
Sodium-Pump Water Performance Tests
M. E. Lackey

Aircraft Reactor Engineering Division

H. C. Young
Pratt & Whitney Aircraft

Full-sized models of the MN-2 impeller and the
volute for the ART sodium pump were tested with
water. The impeller was fabricated of brass and
the volute of steel. The test rig consisted of a
4.in, closed-pipe loop with a 30-hp calibrated-
drive motor, a flow-metering orifice, a throttling
valve, and the necessary manometers. A model-T
pump rotary element was used for the impeller
shaft and the bearing assembly,

The first performance test indicated that the
impeller would meet the head and flow require-
ments of the ART sodium pump at slightly lower

PERIOD ENDING DECEMBER 10, 1955

speeds than predicted Pressure measurements at
points around the periphery of the impeller indi-
cated no excessive unbalance of hydraulic forces
on the impeller, and therefore no modifications to
the volute were anticipated. The efficiency of the
pump in the first test was approximately 63%.
Maximum efficiency occurred at speeds below de-
sign speed. A series of tests was also run in
order to determine the value of inlet shroud vanes
as a method of reducing leakage flow and of in-
creasing efficiency at design point. Six perform-
ance tests and two cavitation tests have been
conducted to date; the performance-test conditions
and results are recorded in Table 2.3, The im-
peller, which had five short and five long radial
blades, remained unchanged during the tests, but
the inlet shroud, the shroud vanes, and the clear-
ances were varied, In general, the test results

TABLE 2.3. RESULTS OF WATER PERFORMANCE TESTS OF MN-2 IMPELLER AND VOLUTE
FOR ART SODIUM PUMP

Design point: 430 gpm, 90-ft head

 

Impeller Clearance (in.)

Approximate Pump

 

 

Test Efficiency at
Inlet Conditions Type of Inlet Shroud . .
No. Axial Radial Design Point
(%)

1 Volute inlet plate 3]/:'32 in, thick; 0.050 0.050 With shroud vanes 63
opening 3.5 in. in diameter, with
3]/32-in. radius at edges; 0.109-
in. spacer between plate and
bottom of impeller

2 Same volute inlet plate as above; 0.050 0.050 0.109-in. plate covering 63 +
0.109-in. spacer attached to bottom of shroud
impeller shroud

3 Same as test 1 0.050 0.050 Shroud vanes extended 61

3/32 in,

4  Same as test ] 0.039 0.050 Same as test 3 62.5

5 Volute inlet plate 3%, in. thick; 0.050 0.0625  Shroud shortened 0.258 64
opening 3.5 in. in diameter, with in. from previous tests;
31/32-511. radius at edge; 0.370- shroud vanes removed
in. spacer between plate and
bottom of impeller

6 Same as test 5 0.050 0.0625 Shroud length same as 63

in test 5; shroud vanes
reinstalled and extended

3/32 in.

 

35
ANP PROJECT PROGRESS REPORT

indicated that the shroud vanes did not increase
efficiency.

Design changes in the reactor caused the design
of the sodium impeller shroud to be shortened
0.258 in, and the radial clearance of the impeller
to be increased from 0.050 to 0.0625 in. Tests 5
and 6 were therefore conducted to mock up the
design change. Further tests will be made to de-
termine the effect of inlet conditions on perform-
ance and cavitation,

The cavitation tests performed to date indicate
that the sodium pump will operate safely with
water at inlet pressures below the pre ssure planned
for the reactor sodium pump. Some difficulty was
encountered in obtaining the original cavitation
data. Air entrapped in the water came out of solu-
tion when the pump inlet pressure was reduced,
and resulted in lowered pump performance because
of what is believed to have been vapor locking.
This phenomenon obscured the true point of cavi-
tation, The water in the test rig was heated to
remove the dissolved air, and satisfactory cavi-
tation tests were conducted,

High-Temperature Tests of ART MF-2
Fuel Pump with NaK

S. M. DeCamp, Jr.

Aircraft Reactor Engineering Division

High-temperature testing of an ART MF-2 fuel
pump (Fig. 2.6) in a short-circuit loop (Fig. 2.7)
was started on August 31, 1955, with NaK (44%
K~56% Na) as the pumped fluid. After 168 hr of
operation at 1400°F the heat source (simulated
gamma-ray heating source) in the shield plug
burned out, and the loop was shut down for re-
placement of the plug. At the same time, the
mechanical condition of the pump was examined,
In the process of disassembling the pump, it was
found to be impossible to remove the impeller re-
tainer nut, The nut was binding because of warp-
age of the impeller, An increase in the clearance
between the sides of the nut and the impeller from
2 mils to 10 mils corrected this difficulty.

It was also found that because of improper in-
dexing during assembly, the pressure-breakdown
plug had been rubbing on the main impeller hub,
Approximately 15 mils had been removed from the
top of the pressure-breakdown lands on the im-
peller hub. Examination of both the pump and the
temperature data indicated that the metallic O-ring
located at the bottom of the gamma-ray shield plug

36

had not completely sealed, and hot NaK vapors
had been able to get between the rotary-element
housing and the pump barrel. This caused a poor
temperature gradient in the pump barrel, as shown

in Fig., 2.8.

Shortly after the second test run was started,
the heat plug again shorted; also, the temperature
gradient around the barre!l was still poor (see
Fig. 2.9). The pump was again shut down and
disassembled, This time the O-ring on the shield
plug had apparently broken at the weld when it
was seated, |t was decided to replace the O-ring
and continue the tests without the heat-source
plug. The heat-source-plug tests will be resumed
later as a separate problem.

The third test run was started on September 20,
1955, and the pump was still operating on Novem-
ber 30, 1955, Figure 2,10 gives the cumulative
seal leakage, the lube oil flow rate, the lube oil
inlet temperature, and the operating NaK tempera-
ture plotted against time. The temperature gradient
along the pump barrel in test run 3 is shown in

Fig. 2.11. The operating conditions were the
following:
Speed, rpm 2700
Head, ft 46
Flow rate, gpm 600
Input power, kw 7
Cooling-oil flow rate, gpm 1.9
Cooling-oil inlet temperature, °F ~ 165
Average temperoture rise across 1
lube oil, °F
Average temperature rise across 11
cooling oil, °F
Average temperature in lower section 1100
of shield plug, °F
Average temperature of Monel section 275

of shield plug, °F

The top seal leaked 30 cm? of oil at startup, but
it did not leak thereafter, The leakage was prob-
ably oil that was used to backfill the top sedl
ptior to startup. Rotation of the pump shaft forced
this backfill oil out the drain line, Since the top
seal is lubricated by a mist of oil rather than by
submergence, it is probable that only a gross oil
leak would show up.
LE

PR
ORNL-LR-DWG 11263

O1L COOLANT LINE

UPPER SHIELD ASSEMBLY

UPPER BEARING HOUSING

DRIVE MOTOR THIMBLE

TOP SEAL RUNNER

TOP BEARING

ROTARY ELEMENT HOUSING
SPARGING IMPELLER LOCK NUT

SPARGING IMPELLER

   
  
 
 
 
   
 
 
 
 
 
  
  
 
 
 
 
   
 
 
 
 
  

PUMP BARREL

OIL LINE

LOWER BEARING

LOWER SEAL RUNNER

RETAINING NUT
LOWER SEAL ASSEMBLY

MONEL BLOCK

SHIELD PLUG

SIMULATED GAMMA-RAY HEAT SOURCE
(CALROD CAST IN BRONZE)

CALROD HEATER

METALLIC "0" RING

SLINGER AND XENON -REMOVAL.
IMPELLER ASSEMBLY

METALLIC “0" RING

MAIN IMPELLER

PUMP VOLUTE IMPELLER RETAINING NUT

PUMP VOLUTE

Fig. 2.6. ART Fuel Pump Type MF-2,

$S61 ‘0l ¥IAWIDIA ONIAN3 Aold3d
o
P
A
UNCLASSIFIED
ORNL-LR-DWG 11264

   

e RN

W

i &

Lieitr s >

LA N

2 s 23

g /{// R

/7///, - )

i B

R T RN
el L

 

 

ST
,

N
20D
SN

\\
SN
5
3
RN
s

 

 

 

 

 

 

 

 

   

 

 

i o
45'/5/%41 N
s, - |
,// ////'/ =
LIQUID-LEVEL PROBES ~<—~—____ 3 iy ke ,
-~ o Yl
3 i 3 {
A N
N e 33
3 LT v :
S L X \
5 Ao 3 s
77 " N B
L T T %3]
3 s T
SN - SHIELD
PUMP SUPPORT AND /9 : PLUG
TANK COVER- , l ErT
s
/)/f/ ‘/.//
s

 

 

 

 

 

 

 

 

 

ot

 

 

 

 

RE
PRESSURE- BREAXDOWN PLUG— y ’ SLINGER /.
. | ] MOUNTING RING—
77 - I -
L AN 7K
N . : .,\ | Y
! " TR ’ [ LW 11 © .
o X : = = | VOLUTE TOP PLATE
m: / E\ | ; \\\ /2 2 /;/ __C'i - \{ { /'
! 3 S ARV A R %
TANK SHELL = e v iy . 5 3 %0
¥ o o i o > A
5 E (oS \ | : - VOLUTE BOTTOM PLATE
L% R LA o /j,///f/’/ 5/:, R “
X Ll

 

 

AT
PO,

 

 

 

 

 

 

 

R
~ T

—_— *_L

 

 

 

 

 

 

 

ALL DIMENSIONS ARE IN INCHES.

 

D

 

 

 

——FILL AND DRAIN LINE

 

 

Fig. 2.7. High-Temperature Short-Circuit Pump-Test Loop.

39
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED

 

 

    
     
 
   

 

 

 

 

 

 

  
 

UNCLASSIFIED ORNL—-LR—DWG 11266
5 ORNL—LR—DWG 11265 16 S " T i n
B ‘ ! 1
o SN o |
, |- HEAT INPUT TO SHIELD PLUG: 375w | | T o T ‘ 1
| | OlL COOLED | ! o NO SHIELD PLUG HEATING ]
14 ‘ l OIL QUTLET TEMPERATURE: 110°F | _ | 14 - / I T _{ " OIL NOT COOLED T+
| 11 | 0 a MEASUREMENTS MADE wiTH C_p b [ OIL OUTLET TEMPERATURE: 160°F
1 \T " THERMOCOUPLES LOCATED 'L’“i [ B ! {
| 0,4 - MEASUREMENTS MADE WITH
| : ON OPPOSITE SIDES OF ‘ o DL F,l ' % THERMOCOUPLES LOCATED ON .|
12 ~—4-—r—h— PUMP BARREL T - f | A } | | OPPOSITE SIDES OF PUMP |
E ooy o ] < fill_{v b BA&R_EFTJF*#
= ; fi“fi% — w j e o
z b b | ] : | L NI P A \T Lol o e
210 ‘ T»fi—iTjfi-—l— — —] = | e ! . ' ! | T T
W l \ ! | ‘ % fi,l,,!ifi P L L d —‘I —L‘ 4‘—
L R
! - . U - 2
S —T—krlfig TJ w 8 j | ESIRED TEMPERATURE GRADIENT
| ! > | - : : : ! .
L g s
a 1 o L = ' N ]
< r ‘ r» Lt i K_,‘_ e _ . _
o b L gs| - | T
S DESIR%%JSI\QZETRATURE < . i i \
<t | gj. —t - | —— - -
; T C VN T
O

a

LN ] I
gNTAL \[:J‘J_L, . [ 0 _a BR

 

 

 

 

 

- — - "‘ ‘ | -— J—LH
2 T——\ . EXPERIMENTAL DATA-"— = ), L \l\\—k h
o t\ ' - e e e MTT \T
LN ol 1 L ‘
a — 0 200 400 600 800 1000 1200
0 200 400 600 800 1000 1200

TEMPERATURE (°F) TEMPERATURE {°F)

Fig. 2.9. Temperature Gradient Along Pump

Fig. 2.8, Temperature Gradient Along Pump Barrel in Run 2,

Batrel in Run 1.

UNGLASSIFIED
ORNL-LR-DWG 11267

HOURS OF OPERATION

 

    
    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o 96 192 288 384 480 576 672 768 864 960 1056 152 1248
1500 T T ] T
NoK TEMPERATURE
1400 ; ,
1300 i—q— — — | — -
'200 . —__r_,4t> e .. - . - p— L
1100 ’—+ - : b
200 — {000 - — - +——t — 25
w
= _ N
180 " 500 i —
160 |- £ 800 [— 2.0
&
o 140 |~ & 700 b—t 44 ]
§ = B -
< 20}~ L 600 1.5 €
w S o
g Lz =
— — —_ —_
5 100 500 B CUMUL ATIVE z
Y BOTTCM SEA ol 2
o S
60 |— 300 — + | —f — — 1
40 |— 200 — INLET OIL TEMPERATURE — | | L los
) — — e e
20 |— (00 | _CUMULATIVE TOP SEAL LEAKAGE o T - 1]
ol 0 _ 1 [ J o

 

 

 

 

 

 

 

 

 

o
o
F.
»

8 10 12 14 16 8 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52
DAYS OF OPERATION

Fig. 2.10. Performance of MF-2 Pump with NaK as the Pumped Fluid, Run 3.

40
UNCLASSIFIED
ORNL-LR-DWG {1268

* { 7 T T 1T 1 T T
AI
\‘Yfi

 

NO SHIELD PLUG HEATING
QIL NCT COOLED
OlL OUTLET TEMPERATURE: {90°F

 

 

 

— THERMOCOUPLES LCCATED
ON QPPOSITE SIDES OF
PUMP BARREL

T

| |

J
\ . O,A MEASUREMENTS MADE WITH

 

 

 

-

 

 

 

 

\

|
ot \\ ‘

! -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E ¥
X
=2
<
w
O
5
+ 8 — '
S \ { ’
g - — |
< N DESIRED TEMPERATURE GRADIENT
W & | L1
s ‘\ ‘, ‘ ]
< ! I |
b ™~

q 1 L

| ~
\ M,
J <
2 — 1 T <
EXPERIMENTAL DATA >\ 0 ~
! | | ey
. L L DS
0 200 400 600 800 1000 1200

TEMPERATURE (°F)

Fig. 2,11, Temperaoture Gradient Along Pump
Barrel in Run 3,

The original test program called for simulation
of gamma-ray and neutron heating in the shield
plug by using a Calrod heater to produce about
2.5 kw of heat. However, as mentioned before,
this part of the test program has been delayed be-
cause of troubles with the plug heater. Apparently,
several problems must be solved. A satisfactory
means must be found for insulating the elecirical
leads (close mechanical clearances complicate
this) for operation up to 1500°F in a helium atmos-
phere. Also, a heat source must be constructed
that can supply approximately 7 w/cm® without
reaching a temperature that will destroy the heater.

As a further complication, two reactor operating
conditions must be simulated by the shield plug.
First, to simulate reactor operation at zero power,
the lower surface of the plug must be maintained
at a minimum temperature of 1000°F in order to
prevent fuel from freezing on its lower face and
seizing the slinger impeller of the pump. Second,
to simulate reactor operation at full power, the
plug must be able to conduct away enough heat to
maintain the lower face at a temperature of 1500 to

PERIOD ENDING DECEMBER 10, 1955

1600°F so that the mechanical strength of the
Inconel can on the shield plug will not be impaired.
A second test stand, similar to that shown in
Fig. 2.7, will be in operation soon, This test [oop
will circulate fuel, and a hydraulic motor rather
than an electric motor will be used to drive the
pump, Data will be taken on operation of the
hydraulic drive, as well as on pump operation,

High-Temperature Pump-Performance Test Stands

R. Curry H. Young
Pratt & Whitney Aircraft

Two high-temperature loops, described previ-
ously,? for testing MF-2 ART-type fuel pumps are
being fabricated, Cavitation, performance, shake-
down, endurance, and acceptance tests on MF-2
pump rotary assemblies will be made with NaK and
with the fuel mixture I\Ic1|:-ZrF4-UF4 (50-46-4
mole %) as the circulated fluids and with fluid
temperatures of up to 1400°F,

The design layout has been completed for a
highetemperature loop for testing MN-2 ART-type
sodium pumps, Cavitation, performance, shake-
down, and acceptance tests on MN-2 pump rotary
assemblies will be made with sodium at tempera-
tures of up to 1400°F. Most features of this loop
are similar fo those of the loop for testing MF-2
ART-type fuel-pump performance.

High—temperature performc:nce-tesfing loops are
being designed for testing PK-2 and PK-A ART-
type NaK pumps. Cavitation, performance, shake-
down, endurance, and acceptance tests on NaK
pump rotary assemblies are to be made at NaK
temperatures of up to 1400°F,

HEAT EXCHANGER DEVELOPMENT

E. R. Dytko
Pratt & Whitney Aircraft

R. E. MacPherson

Aircraft Reactor Engineering Division

intermediate Heat Exchanger Tests

R. D. Pedk M. H. Cooper
J. M, Cooke L. R. Enstice
Pratt & Whitney Aircraft
Designs of the two fuel-to-NaK heat exchangers

that are to be tested in stands B and C were com-
pleted. Type IHE-3, shown in Fig. 2.12, is an

 

3R. Curry and H. Young, ANP Quar. Prog. Rep. Sept.
10, 1955, ORNL-1947, p 44.

41
ANP PROJECT PROGRESS REPORT

all-lnconel 100-tube design, for which 0,25-in,-0D,
5.6-ft-long, 0.025.in,~wall tubing is to be used.
Type IHE-8, shown in Fig, 2.13, which is also all
Inconel, is a 144-tube design, for which 0,1875-in.-
OD, 5.6-ft-long, 0.025-in..wall tubing is to be

used. These heat exchangers will be tested in a

e 02510 0D

MATERIAL : INCONEL

A

SPAGCER WIRE, 0.031 x 0.055 in. ({1}

o — ———  —  &737in-

 
 
    
   

EXCHANGER TUBE ‘TOP FILLER BAR

 

FUEL HEADER

NoK HEADER

NaK INLET

Fig. 2.12. Fuel-to-NaK

144 TUBES
01875-in. 0D
0.025-in WALLS -._

  
     

SIDE FILLER BAR -~

 

 

    
  
  
    
 

 

 

  

 

 

\n\\\\t“
SECTION A-A
T
UBE SPAGER~- - SPAGER WIRE, 0.021 x 0.045 in. ({1} A
;§§w§m@§5“m§fi§'“mg§fififiw ; // -1
T A P g e s
FUEL INLET 7 T %
|§§; ]»:i‘ it _ e =7 ,jf f.,. e
= = Y | J : \l 1 | e —¢ —

  
  
 

 

 

e

EXCHANGER

ST

 

SRR

 

 

  

 

H
e H
SR
o

4 ~-X_ FUEL HEADER

"™ NaK HEADER GOVER

NoK OUTLET ll—- S
Fig. 2.13. Fuel-to-NaK

42

  
  

 
  

f
\ ! —TUBE SPACER
\ —BOTTOM FILLER BAR
-TUBE SPACER
TUBE SHEET, g-in. THIGK, 4-in. R, HOLES APITCH 0.300 x 0.270 in.

 

 

regenerative system of the type shown in Fig.
2.14, with two tube bundles operating in series.
Figures 2,15 and 2.16 show the range of operating
conditions which can be expected from the two
heat exchangers, and Figs. 2.17 and 2.18 show
stresses which will be encountered in the test

ORNEC= g— DWG 9622

~100 TUBES

 
 
  
 
  
    

0025-in WALLS

SECTION
A-A

SIDE FILLER BAR
A=

  

   

   
    
 

 
 
 

\EXCHANGER SHELL

2 o 2 4 & e 10 12

- : . S

 

SCALE N INCHES

Heat Exchanger Type IHE-3.

L
ORNL-LR-DWG 9623

NaK INLET l

7= FILLER BAR
-~ {TOP AND BOTTOM

   
 
  
 
  
  

 

 

 

   
 
  
  
   

=

i/
SHELL -

P~ TUBE SHEET, %-in. THICK, 4-in. R, HOLES A PITCH 0.334 x 0.304 in. !

MATERIAL : INGONEL J

2 0 2 4 6
A per )
SCALE IN INCHES

|
— — 87375 — —— «l

Heat Exchanger Type IHE-8.
 

 

 

 

 

 

 

PERIOD ENDING DECEMBER 10, 1955

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ORNL- LR-DWG 1269
AIR N AIR QUT
TWO 500-kw
RADIATCR UNITS
[ NaK PUMP
|
Y NoK FLOW ™~
[ FUEL FLOW ] 1
FU PUMP
NaK-TO-FUEL EL FUEL-TO- NaK
HEAT EXCHANGER HEAT EXCHANGER
TUBE BUNDLE NO. 4 TUBE BUNDLE NO.2
NG L
i L S
FUEL FLOW
NaK FLOW
1-Mw
GAS-FIRED
FURNACE W ’

 

 

 

 

 

Fig. 2.14. Diagram of Intermediate Heat Exchanger Test Stand,

units throughout the operating range. The thermal
stresses (calculated by using Castigliano's theo-
rem) are imposed on the heat exchanger tubing by
the difference in temperature between the tubing
and the sheil of the heat exchanger. These
stresses arerelieved by creep and are the stresses
of primary interest in a thermal-cycling program.
The pressure stresses (calculated by using simple
cantilever-beam theory) are imposed on the tubing
by the drag forces of the flowing fluid and are not
relieved by creep; rather, the tubing is steadily
deformed,

Test stand A, the first experimental assembly
operated in the present series of intermediate heat
exchanger tests, was described previously.? It
was operated for 434 hr during this quarter in a

 

4R. D. Peak, M. H. Cooper, and L. R. Enstice, ANP
Quar. Prog. Rep, Sept. 10, 1955, ORNL-1947, p 45.

series of radiator and circulating-cold-trap tests,
A chronological description of the test operations
is presented in Table 2,4, The stand was op-
erated with York radiator units No. 1 and No. 2
until unit No. 1 failed. When unit No. 1 had been
removed, operation was resumed and continued
until unit No. 2 failed. York unit No. 3 is cur-
rently being installed, Examination of unit No. 1,
which failed after 140 hr of operation, revealed
severe buckling of the side plates of the radiator
core because of differential thermal expansion be-
tween the tube matrix and the side plates. With
no air flow across the radiator core, the tempera-
tures of these plates could be 125°F below the
tube temperature, and, dwing periods of heat re-
moval from the radiator, a femperature difference
of 300 to 1200°F could exist. Temperature dif-
ferentials of such magnitudes would cause severe
compressive stress in the tubes at the start of air

43
ANP PROJECT PROGRESS REPORT

ORNL -LR—-DWG 11270
1700
| | HE-3 ]f—
NakK INTO BUNDLE NQ. 1& | !

 

1600 - ! :
— T |

1500
1400

1300

TEMPERATURE (°F)

 

1200 I"FyUEL QUT OF BUNDLE NO.2*

-
INTO BUNDLE NO.B— - i
e AU ‘

  

 

 

 

 

 

 

 

 

 

   
  
  
  

 

 

 

 

1100 |——- ;
|
NaK INTO BUNDLE NO.2
1000
180 —
i
i NoK FLOW RATE
160 ‘ e — ]
G: J
o
=~ 440 b - e —_ —
I i |
o~ I
WS 120 | - - —_ ]
xZ |
L g
Lo !
8~ 400 | — - __LOG MEAN TEMPERATURE __|
w'a : DIFFERENTIAL IN HEAT
- ‘ EXCHANGER |
— a i . |
R S
w0 i
Ly |
o :
Bl
= | FUEL FLOW RATE
S ; | ~
o i - ! o
= 40 —mmm— - e -
o
2 TOTAL FUEL -
ar PRESSURE DROP -
20 [— =~  TOTAL NaK ——
,./ PRESSURE DROP
— | —_— e ——ri
0 — !
0 1000 2000 3000 4000

FUEL REYNOLDS NUMBER

Fig. 2.15. Predicted Operating Performance of
Fuel-to-NaK Heat Exchanger Type IHE-3 in Test
Stand B.

flow, and, if it is assumed that the compressive
stress is relieved by creep, equadlly severe tensile
stress would be present when the air flow was
stopped. The results of metallurgical examination
of the unit are reported in Sec. 5, ‘“Metallurgy and
Ceramics.”’

Prior to initiation of operation with York unit
No. 2, the side plates were split so that the tubes
could move freely with core temperature changes.
This unit subsequently failed during themal-
cycling tests after 435 hr of operation. Figure

44

L
ORNL—-LR--DWG 11271

 

1700

 

 

   
 
 

 

 

 

 

 

 

   
   

 

 

   
     

 

 

 

 

 

‘ HE-8 |
NoK INTO BUNDLE NO.ty | /
1600 |- p— ! —
i | !
FUEL OUT OF BUNDLE NO.1 INTQ BUNDLE NO. 2
500 |- ———-TF"= — ‘ N
: Tty | s a——
o
w 1400
o
2
& 1300 L —I —
& NaK OUT OF BUNDLE NO.{ |
=
w i
1200 I"LUEL OUT OF BUNDLE NO.2 INTO BUNDLE NO.t%y_.— ]
T--'-—-_-—__ — ey ——
1100 {-NoK INTO BUNDLE NO. 2. | -
‘ J
1000
180 —1 —‘— .
160 '— —_—
= Nok FLOW RATE
e
- 440 - ot .
-l i
T
5 5
W 3 120 —— — - - -
w © !
TS i
Lo LOG MEAN TEMPERATURE
B —~ 100 —— - DIFFERENTIAL IN HEAT —
g & EXCHANGER |
x S
£ ol o
L E ‘
g
g
£ wn
g W
Ww
= a
O
O
-t
20 b——— .
TOTAL FUEL
PRESSURE DROP J
!
0
0 1000 2000 3000 4000

FUEL REYNOLDS NUMBER

Fig. 2,16, Predicted Operating Performance of
Fuel-to:NaK Heat Exchanger Type IHE-8 in Test
Stand C,

2.19 shows the split side cover plates and the re-
sults of the NaK leak and fire. Since the location
of the leak was almost the same as the location of
the leak in York unit No. 1, that is, near the side
plate and near a horizontal stiffener plate, the
same conditions which caused York unit No. 1 to
fail undoubtedly weakened York unit No. 2. In the
case of York unit No. 2, as with ORNL units
Nos. 1 and 2, a marked increase in the radiator
friction factor (~12%) occurred prior to failure,
The exact cause of this increase, which was not
e
ORNL—LR-DWG 11272

 

 

 

 

 

 

 

 

 

 

 

 

7 x108 — . .
' IHE-3
6x10°% |——— — - —
_ TUBE BUNDLE HEAT LOAD

-~ 5X10° f—r flm_lfi
£ \
~
2 |
& axiw® % S O —
a |
Q
J 3x108 kfi —_————
b i |
o FURNACE HEAT LOAD
T 6 ! |

exwo® L ] Tfi

st ] ‘
0
18,000 _ [
16,000 l __\‘

 

THERMAL STRESS
14000 f— ——o 3

12,000 F — -

10,000 | ‘ - ' .
PRESSURE STRESS DUE TO FUEL SYSTEM QPERATION

! 1 |
PRESSURE STRESS DUE TO BIFLUID OPERATION
| PREs eos DUE TV TATION

 

 

  
   
   

 

STRESS ON HEAT EXCHANGER TUBING (psi)

 

 

 

 

 

8,000
6,000 —
4,000 |- —J
2000 o PRESSURE STRESS DUE TO
“ NaK SYSTEM OPERATION
! t
0 e | ]
0 1000 2000 3000 4000
FUEL REYNOLDS NUMBER
Fig. 2.17. Predicted Operating Heat Loads and

Stresses of Fuel-to-NaK Heat Exchanger Type
IHE-3 in Test Stand B.

encountered prior to the failure of York unit No, 1,
has not yet been established. _
During the shutdown for installation of York
radiator units Nos, | and 2, the diffusion cold trap
initially used on this stand was changed to a cir-
culating cold trap. The trap, as modified, con-
sists of a 6-in, pipe, 36 in, long, that contains
York Demister stainless steel packing in the top
30 in. The NaK flowed into the trap through an
economizer section (]/Z-in. pipe inside a 2-in. pipe,
12 in. long} mounted on top of the cold-trap body,

PERIOD ENDING DECEMBER 10, 1955

6 ORNL—LR—DWG 11273
7 X10 1

1
IHE—8
exio® - — I R

[
TUBE BUNDLE HEAT LOAD

 

   
      

5 x10°

ax10® b

 

3x10°

HEAT LOAD (Btu/hr)

 

2x10°

 

 

 

 

18,000

 

]
#R\J: M p—

\
18,000 —— - —
THERMAL STRES

S
14,000 -
S
/

10,000 ;—h —

PRESSURE STRESS DUE
TO FUEL SYSTEM OPERATION

- PRESSURE STRESS DUE
TO BIFLUID OPERATION

S5TRESS DUE I ,

12,000

8000 S

6000 | ———

 

4000 | PRESSURE

STRESS ON HEAT EXCHANGER TUBING {psi)

  

 

2000 — -

 

 

0 1000 2000 3000 4000
FUEL REYNQOLDS NUMBER

Fig. 2.18. Predicted Operating Heat Loads and
Stresses of Fuel-to-NaK Heat Exchanger Type
IHE-8 in Test Stand C.

flowed through the central ]/2-in. pipe to the bottom
of the trap, flowed back up over the packing to the
economizer, and then flowed back to the system.
The trap was cooled with forced air, and heaters
were provided for local temperature control. It
was installed in series with, and downstream from,
the plug indicator. The series arrangement was
not suitable, however, because cooling of the plug
indicator caused plugging to start in the cold trap.
A parallel installation was therefore devised, and
with this installation the trap could be operated at

45
ANP PROJECT PROGRESS REPORT

wd ABLEF 2.4, SUMMARY OF OPERATION OF INTERMEDIALE HEAT EXCHANGER TEST STAND A

 

 

Operation continued from previous quarter. NaK system repaired by the installation of new York
radiator units Nos. 1 and 2, new circulating cold trap, and plug indicator; trap and indicator

piped in series. Blower operated to measure radiator air flow at 90°F, Filled system with NaK

NaK system operated isothermally at 1500°F with flow rate of 80 gpm to check operation of cold

Plug indicator gave erroneous measurements of sodium oxide because of interaction of

Radiator tests with the NaK system operated at various flow rates and temperatures between

Radiator tests with the NaK system operated at various flow rates and temperatures; blower at
Tests terminated because of a {eak in York
This radiator was cut out for examination, and the air duct was changed to
The circulating cold trap was modified with a

throttling valve so that the plug indicator and the cold trap operated in parallel NaK streams.

NaK system operated isothermally at 1300°F with flow rate of 40 gpm to check operation of cold

Radiator tests with the NaK system operated at various flow rates and temperatures; blower at

Cold trap operated intermittently to

Hours of
Operation Remarks
690

(56% Na—44% K) at room temperature,
690 to 720

trap.

cold teap in series with it.
720 to 748

1000 and 1650°F; blower run at various air flow rates, Eleven measurements taken.
748 to 823 NaK system operated isothermally at 1300°F with flow rate of 80 gpm.
823 to 830

various air flow rates. Three measurements taken.

radiator unit No, 1.

fit the remaining York radiator unit No. 2.

Blower operated to measure radiator air flow with NaK system at 90°F.
830 to 909

trap. Sodium oxide plugging temperature reduced from 1000 to 750°F.
909 to 999

various air flow rates. Twenty-five measurements taken,

control the axide content,
999 to 1017

1017 to 1712

1112 to 1124

NaK system operated isothermally at 1200°F with a flow rate of 40 gpm. Cold trap operated to
reduce the oxide plugging temperature from 920 to 800°F. NaK system shut down to calibrate

the electromagnetic flowmeter on the main NaK system,

NaoK system operated isothermally at 1400°F with a flow rate of 40 gpm. Cold trap operated in
an attempt to reduce oxide content; plugging temperatures varied erratically between 1010 and

720°F.

Radiator thermally cycled according to the following schedule: high power for 4 hr; NaK into
radiator at 1500°F, out at 1140°F; flow, 42 gpm; air in at 90°F, out at 1100°F; low power for
2 hr; NaK circulated isothermally at 1425°F; flow, 42 gpm; no air flow. York radiator unit No. 2
failed on the start of the third cycle to high power. The radiator was cut out for examination by
the Metallurgy Division, and the cold trap was cut out for examination by the Materials Chemis-
try Division. A new circulating cold trap, a new screen filter, a duplicate plug indicator, and

York radiator unit No. 3 are being placed in the NaK system.

 

0.3-gpm NaK flow with a minimum temperature of
350°F. At the end of the test when the cold trap
was cut open for examination by the Materials
Chemistry Division, it was found that the trap had
not effectively removed oxides. The equivalent of
75 g of O, was found, which corresponds to the

 

Sibid., p 49.

46

removal of 260 ppm of O, from the 80 gal of NaK
used in the system during operation of the cold
trap.

The NaK samples taken during this quarter indi-
cate that the cleaning done to remove the fuel
constituents which leaked into the NaK system
at the time of failure of IHE No. 2, described pre-
viously,®> was not complete. Fuel constituents
LUNCLASSIFIED
PHOTO 24962

 

Fig. 2.19. York NaK-to-Air Unit No. 2 from Inter-
mediate Heat Exchanger Test Stand A after NaK
Leak and Fire.

continue to showup in miscellaneous NaK samples,
as illustrated below:

Zirconiom

Uranium Content
Coatent (wt %)
Drain from circulating cold trap 0.007 g

Leak from York radiator unit No. 1 1.4 wt % 1.9

L eak from York radiator unit No. 2 1.9 wt %
Since the sump, plug indicator, and cold trap were
opened and cleaned manually before reinstalliation
on the chemically cleaned system, it is certain
that the fuel constituents were held up in portions
of the pump and furnace which were not adequately
reached by the cleaning solutions.

The analyses of the NaK drained from the circu-
lating cold trap and of the water used to wash the
trap are also of interest:

PERIOD ENDING DECEMBER 10, 1955

Amount in NaK Amount in Water

{ppm) (9)
Nickel 10 10
Chromium 30 27
Iron 89 1.5

These results substantiate other evidence that
chromium is preferentially leached from the Inconel
in the systems and is carried to the cold regions
of the loop in the NaK stream.

Test stand B, described previously,® has com-
pleted a total of 676 hr of operation, including
247 hr of bifluid operation in a series of radiator,
heat exchanger, and circulating-cold-trap tests, A
chronological description of the tests conducted
and the difficulties encountered is given in Table
2.5, The system is now in steady«state operation
with a temperature differential imposed.

As stated in Table 2.5, considerable trouble was
experienced with operation of the cold trap and
the plug indicator, On stand B, these components
are operated in parallel, with the radiator pressure
drop as the driving force. The NaK that enters
the cold trap is cooled by a forced-air precooler.
It is convection-cooled as it passes through the
trap, and then it is reheated at the exit end before
being discharged through a throttling valve and an
electromagnetic flowmeter and being returned to
the system. The plug indicator is identical to
that on stand A,

Most of the trouble associated with operating the
plug indicator and the cold trap has been due to
plugging of the small-diameter, ?Ié-in. tubing on
the discharge sides. This tubing was utilized in
order to magnify the electromagnetic flowmeter
signal. A cold trap of the type being installed in
stand A will be installed in this loop, and the
small-diameter tubing in both the cold«trap and
plug-indicator circuits will be removed and re-
placed with larger tubing.

As stated in Table 2.5, the 2-in, IPS sched-40
pipe adjacent to the electromagnetic flowmeter on
the main NaK circuit developed a leak after 374 hr
of operation. The resulting fire destroyed the
aluminum pieces and fused the pipe surface so
that the exact cause could not be determined.
Examination of the pipe after it was cut out
showed no defects detectable by x ray or Dy-Chek.

47
ANP PROJECT PROGRESS REPORT

TABLE 2.5. SUMMARY OF OPERATION OF INTERMEDIATE HEAT EXCHANGER TEST STAND B

 

 

Hours of
Operation Remarks
a Stand assembled with Pratt & Whitney Aircraft radiator units Nos. 1 and 2 end with ORNL heat
exchangers (type IHE-3) Nos. | and 2. Filled system with NaK (56% Na—44% K) at rcom temper-
ature. Operated blower to measure radiator air flow at 90°F.

Cto9 Heated NaK system to 1200°F with flow of 80 gpm. All NaK drained into sump when remotely
operated dump valve failed to close when draining some NaK to adjust level in pump. Added
manually operated handle to dump valve,

9 to 88 NaK system operated isothermally at 1400°F with flow of 80 gpm to check operation of circu-
lating cold trap. Sodium oxide plugging temperature reduced from 940 to 660°F.

88 to 139 Radiator tests with the NaK system operated at various flow rates and at temperatures between
1000 and 1550°F; btower at various air flow rates. Cold trap valved out. Plug indicator plugged
up and then unplugged when outside resistance heat was codded. The NaK-system friction
factor® gradually increased 25% during tests.

139 to 173 NaK system operated isothermeally at 1400°F with flow of 80 gpm with cold trap on to reduce
oxide content. Faulty operation of the trop washed oxide back into the system initially; how-
ever, the trap rteduced the plugging temperature from 850 to 730°F. The NaK-system friction
factor was reduced by 13%.

173 to 175 Furnace test with the NaK system operated at flow rate of 140 gpm, a temperature of 1350°F

into the furnace and a temperature of 1500°F out. The Struthers Wells furnace burned 8100 sefh
of natural gas with a thermal efficiency of 48%. Heat transfer to the NaK was 3,69 x 10% Bu/hr
(1.68 Mw). Cold trap valved out.

175 to 350 The NaK system operated isothermally at 1400°F with a flow rate of 80 gpm. Cold trap in
operation; plug indicator plugged, unplugged, and finally plugged seolid.

204 to 206 Fuel system filled for cleaning with N«:F-ZrF"-UF4 (50-46-4 mole %) and operated isothermally
at 1400°F for 2 hr at a fuel flow rate of 30 gpm; fuel then dumped. System filled with new fuel,
NaF-ZrF4-UF4 (50-46-4 mole %). Bifluid isothermal operation ot 1400°F with NaK flow of
80 gpm and fuel flow of 30 gpm. Test stopped because of leak in 2-in. NaK pipe adjacent to
the electromagnetic flowmeter. Replaced NaK pipe and the plugged tubing downstream from the

plug indicator.

374 to 388 Heated NaK system to 1200°F with flow of 80 gpm. NaK drained into sump while repairing NaK

dump valve. Found helium line to NaK pump plugged with sodium oxide,

388 to 453 NaK system operated isothermally at 1400°F with a flow rate of B0 gpm to check cold-trap
operation. Electromagnetic flowmeter on plug indicator failed., No change in NaK-system

friction factor.

453 to 604 Heat exchanger tests in bifluid operation at various temperatures between 1000 and 1500°F;
NaK, fuel, and blower air flow rates varied. Radiator data also token. The circulating cold

trap plugged, NaK-system friction foctor increased 35%. Thirteen heat transfer measurements

taken.

604 to 628 Bifluid isothermal operation at 1300°F; NaK flow rate, 145 gpm; fuel flow rate, 30 gpm. Attempted
to unplug cold trap by using resistance heating on tubing. NaK-system friction factor did not
change.

628 to 676 Bifluid operation with a temperature differential across system; NaK flow rate, 145 gpm; high

NoK temperature, 1600°F; low NaK temperature, 1100°F; fuel flow rate, 42 gpm; high fuel
temperature, 1435°F; low fuel temperature, 1250°F., NaK-system friction factor increased an
additional 8%. Operation continuing.

 

*The friction factor was taken to be the ratio of the NaK-system pressure drop to the square of the NaK flow rate.

48
Test stand C is being fabricated and is 25%
complete, Radiator units supplied by the Cam-
bridge Comp. and heat exchangers being built by
Black, Sivalls & Bryson, Inc., will be tested first,

Small Heat Exchanger Tests
J. C. Amos

Aircraft Reactor Engineering Division

L. H. Devlin J. G, Tumer
Pratt & Whitney Aircraft

Small heat exchanger test stand B, shown in
Fig. 2.20, was placed in operation on November 3,
1955, The test units installed in this stand are a
20-tube fuel-to-NaK heat exchanger and a 500-kw
NaK-to-air radiator (ORNL No, 3), both built by
the Metallurgy Division of ORNL. The heat ex-
changer tubes, 0.25 in., in outside diameter and
0.025 in, in wall thickness, are amanged in a
4 by 5 matrix with 0,282-in. center-to-center
s pacing,

The test stand includes a circulating cold trap,
described previously® and shown in Fig., 2.21, for
removing oxides from the NaK. The NaK circu-
lating in the cold trap is cooled by an economizer
located in the inlet line to the cold trap and by
air circulated in the cold-trap cooling coil. With
this arrangement, it has been possible to operate
at cold-trap temperatures below 150°F, with sys-
tem temperatures as high as 1500°F. Relative
oxide concentration in the system is determined
by an air-cooled plugging indicator, also shown in
Fig. 2.21. Much of the initial operation of this
test system has been devoted to obtaining data on
cold-trap operation, A preliminary analysis of the
data indicates that cold-trap performance is re-
producible and that the oxide concentration of the
NaK can be repeatedly reduced,

Small heat exchanger test stand C is approxi-
mately 65% complete, This stand is identical
with stand B, with minor exceptions. The heat
exchanger is identical with the one in stand B and
was also built by the Metallurgy Division of
ORNL, but the 500kw NaKto-air radiator is York
unit No. 5. :

The design of a 25-tube heat exchanger, with
0.1875-in.-0D, 0.025-in.wall tubing and 0.2175-in,
cenfer-to-center spacing, that is to operate at
ART temperature and flow conditions was com-

 

6F. A. Anderson and J. J. Milich, ANP Quar, Prog,
Rep. Sepr. 10, 1955, ORNL-1947, p 54.

PERIOD ENDING DECEMBER 10, 1955

pleted. This unit is shown in Fig. 2,22, Negotia-
tions are in progress to obtain four of these units
from outside vendors. Present plans call for
testing these four heat exchangers, two additignal
20-tube heat exchangers, and six additional 500k w
radiators, in small heat exchanger test stand$§ 3

and C. 5
Heat-Transter and Pressure-Drop Correlations
R. D. Peak
Pratt & Whitney Aircraft
J. C. Amos

Aircraft Reactor Engineering Division

The test data on radiator and heat exchanger
heat transfer and pressure drop obtained with the
intermediate and small heat exchanger test stands
have been correlated. The parameters of the vari-
ous radiators tested are presented in Table 2.6,
and those of the several heat exchangers in
Table 2.7.

The air pressure-drop correlation for the radia-
tors is shown in Fig. 2.23, where the pressure
drop per tube row (in inches of water), comrected
from the air film temperature to 60°F, is plotted
against the ratio of air flow rate to the free-flow
area. The points for ORNL-1,-2, PWA-1,-2, and
York-1,-2 are averages for the two radiator units,
since the two radiators were usually operated
simultaneously. The air heat-transfer correlation
for these radiators is shown in Fig, 2,24, where
the air Nusselt number at the film temperature is
plotted against the air Reynolds number at the
film temperature. The air-side heat-transfer coef-
ficient was calculated by taking into account the
NaK-side and the tube and fin collar metal heat-
transfer resistances. The NaK-side resistance
was calculated by using the Lubarsky? equation:
Nu = 0.625(Re x Pr)%-4. It was found that the fin
efficiencies ranged from 88 to 94%. It is apparent
that the seven radiators built by the three sources
from the same specifications resulted in a wide
range of performance.

The NaK pressure-drop correlation is shown in
Fig. 2.25, where the friction factor is plotted
against Reynolds number. Also included in the
correlation are two water tests: one with new,

0.1875-in.-0OD, 0,025 in.-wall, Inconel tubing, the

 

’B. Lubarsky and S. J. Kaufman, Review of Experi-
mental Investigations of Liquid-Metal Heat Transfer,
NACA-TN-3336 (March 1955).

49
0¢

CONTROL PANEL

 

Fig. 2.20. Small Heat Exchanger Test Stand B.

1¥0d3Y S§53IY50¥d LI3Iroyd dNV
s

Fig. 2.21. Small Heat Exchanger Test Stand B Showing Cold Trap and Plug Indicator.

UNCLASSIF!EDB
PHOTO 24915

 

§S61 ‘0l ¥3gW3D34 ONIAN3 oIy 3d
s

ORNL-LR-DWG 11274

%~ TWENTY-FIVE, 0.1875~in.-0D,
\ 0.025-in.- WALL TUBES

  

NaK OUTLET

NaK INLET

 
 
  
 
  

SECTION A-A

mm\‘m“‘mln\m

~—TUBE SHEET

EXCHANGER SHELL

 

|
2,
FUEL HEADER

 

LTI AT
L
I

 

    
 

i
FUEL’ OUTLET ————-

 
  

 

 

 

 

\\\\\\\\\\\

 

TUBE SPACERS
(12 SETS TOTAL)

   

E;
T I T
% T
| )
i .
MATERIAL: INCONEL I
% 2 i
L
e

Fig. 2,.22. Twenty-Five Tube Fuel-to-NaK Heat Exchanger.

'
LIS TSI N ARSI TI LA o

-~ FUEL INLET =——-1

L30d3Y SSFAD0Ud 1D03Ir0dd dNV
PERIOD ENDING DECEMBER 10, 1955

TABLE 2.6. PARAMETERS OF 500-kw NaK-TO-AIR RADIATORS FROM VARIOUS SOURCES

 

 

ORNL-1,-24

Parameter PWA.1,.2 ORNL-37 York-1,-2¢
Tube material Inconel Inconel Inconel
Number of tubes 72 72 72
Inside diameter of tubes, ft 0.0115 0.0112 0.0115
Tube wall plus fin collar thickness, ft 0.003 0.003 0.003
NaK free-flow area, ft2 0.00742 0.00714 0.00742
NaK-side heat-transfer area, f2 6.92 6.72 6.92
Mean tube plus collar heat-transfer area, ft2 8.68 8.48 8.68

Fin material

Number of fins in vertical stack

{_ength of air flow passage, ft

Air-side fin heat-transfer areq, H2

Air-side tube plus collar heat-transfer areq, ft2
Air free-flow areq, ftz

Air equivalent diameter, ft

Fin average thermal cenductivity, Btu/hr§t-°F

Type 310 stainless-steel-clad copper, 0.010 in.
thick, sched 25.50-25

243 236 254
0.646 0.646 0.646
201 196 210

9 8 8

0.519 0.538 0.509
0.00884 0.00911 0.00832
109 109 109

 

“Manufactured by Oak Ridge National Laboratory Metallurgy Division.

bMflnufactured by Pratt & Whitney Aircraft.
Manufactured by York Corp.

other with the tubing from the 20-tube fuel-to-NaK
heat exchanger bundle, which showed indentations
where the wire spacers had pressed into the tubing
during 1500 hr of operation. All the data shown
were taken during isothermal operation, the range
of temperatures being 1200 to 1500°F, except for
the IHE-3 data, which were taken during the radia-
tor tests on intermediate heat exchanger test
stand B, previously described in this report,
During these tests the NaK-system friction factor
was observed to rise 25%. An intensive investiga-
tion is now being conducted on all heat exchanger
test stands to detemine the reason for the ap-
parent rise in pressure drop observed during op-
eration with a temperature differential imposed on
the system.

The fuel pressure-drop correlation is shown in
Fig. 2.26, where the sum of the skin friction and

 

8R. D. Peak and J. W, Kingsley, ANP Quar. Prog.
Rep. June 10, 1955, ORNL-1896, p 35.

the spacer friction is plotted against Reynolds
number. The spacerfriction data, which were ob-
tained by using water-test data (described previ-
ously®) corrected for the various spacer thick-
nesses and lengths by using the correlation of
Cohen, Fraas, and LaVerne,? are also plotted
separately for each heat exchanger.

The fuel-side heat-transfer correlation is shown
in Fig, 2.27 for two heat exchangers, The NaK-
side heat-transfer resistance was calculated by
using the Lubarsky equation.” The correlation for
the 20-tube fuel-to-NaK heat exchanger differs
from that presented previously'® because the

 

%G. H. Cohen, A. P. Fraas, and M. E. LaVerne, Heat
Transfer and Pressure Loss in Tube Bundles for High

Performance Heat Exchangers and Fuel Elements,
ORNL-1215 (Aug. 12, 1952), p 47.

10, c Amos, L. H. Devlin, and J. S, Turner, ANP
Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 51.

53
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
CRNL—LR-DWG 11275

 

2.0 1

| ! ‘
| ‘ ]
| NoK TEMPERATURE

ORNL —1,—2 {000~-16800°F
YORK —1,—2 1000 —1600° F

 
 
 
    

 

|
| |
05 [ —

 

 

 

 

 

0.2 [______‘L___ -

AIR PRESSURE DROP PER TUBE ROW (in. OF H,0)
(CORRECTED FROM AR FILM TEMPERATURE TO 60°F )
|
|

 

 

l
|
|
|
: | ;
3.0 4.0 6.0 8.0 100 1
RATIO OF AIR FLOW RATE TO FREE FLOW AREA (Ib/SeC‘sz)

|
0.05 L
4

06 o8 1.0 1.4 2

Fig. 2.23. Air Pressure-Drop Correlation for 500-kw NaK-to-Air Radiators,

equivalent diameter used in the Nusselt number STRUCTURAL TESTS
was
Outer-Core-Shell Thermal-Stability Test
. Cross-section area G. D. Whitman A. M. Smith
wetted perimeter Aircraft Reactor Engineering Division

rather than Work has continued on the fabrication of the one-
fourth-scale outercore-shell model and testing
4 cross-section areq assembly, The model is to be subjected to the
< tube perimeter cyclic thermal and pressure tests that were de-
scribed in the previous report. All the component
as used before. parts have been received, and the model assembly

24
PERIOD ENDING DECEMBER 10, 1955

cxowmws TABLE 2.7. PARAMETERS OF VARIQUS FUEL-TO-NoK HEAT EXCHANGERS

 

 

20-Tube Intermediate Heat Exchanger
Parameter Fuel-to-NaK Heat Exchanger Type
Heat Exchanger No. 2 IHE-3
Tube material lnconel Inconel {nconel
Number of tubes 20 100 100
Tube outside diameter, in. 0.1875 0.1875 0.25
Tube wall thickness, in. 0.017 0.017 0.025
NoK flow length, ff 5.75 6,03 6.38
NaK free-flow area, ft2 0.00258 0.0129 0.0218
NoK-side heat-transfer area, ft2 3.90 22.6 29.4
Shell inside dimensions, in. 0.910 x 1.130 2.10 x 2.10 272 x 2.72
Spacer dimensions, in. 0.032 x 0.032 0.022 x 0.040 0.020 x 0.040
Number of spacers 1 13 1
Fuel flow length, ft 4.85 5,62 5.62
Fuel free-flow areq, ft2 0.00328 0.0115 0.0173
Fuel-side heat-transfer area, 12 4.75 27.6 36.7
Fuel equivalent diameter, ft 0.0102

 

AIR NUSSELT NUMBER AT FILM TEMPERATURE

10.0

UNCLASSIFIED
ORNL—LR~OWG M2T76

 

 

 

 

5.0

 
 

 

 

 

20

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

s
4__
At;

 

—

 

o
-+

 

 

 

 

100 200

500

1000
AIR REYNOLDS NUMBER AT FILM TEMPERATURE

2000 ' 5000

Fig. 2.24. Air Heat-Transfer Correlation for 500-kw NaK-to-Air Radiators.

10,000

55
ANP PROJECT PROGRESS REPORT

QA0 1 T T ‘

SO
ORNL—LR-DWG 11277

 

  
 
    
  
 

 

 

 

 

 

| i 71 | ] i
e e e e e S Pt
ol 1Ly ® 20— TUBE FUEL-TO~NoK HEAT EXCHANGER AL
l | {1 | 0 INTERMEDIATE HEAT €XCHANGER NO.?2 R
SO S - /-T- e _L_~a__.l__ —
! ! | [ 1 ' | A INTERMEDIATE HEAT EXCHANGER TYPE IHE~—3, ORNL—1,~2 |
Af*f“‘“*‘**”'""F***fl“"7“““%“%“‘w*“*f*" '“"**“'"“"fi'”'"'"“;'"”%’W‘j"fl%’*fi“f'f”
WATER TEST OF NEW 0.4875-in.-00, ' | | | o
0.05 t————-——-—/0025-in WALL TUBING "—}’—‘i—‘T ~-—i- ~-—-~-~-~~-*—'--Mf—frmfj‘wflhfij*j‘ ]
« | i ; | WATER TEST OF 20-TUBE FUEL-TO-NaK | o
g ~~~~~~~~~ _ /’*‘""’T“"*T 4-7,42*1*/‘HEAJT EI><C1HANGER TUBING _f.-____wj _____ T___fl;_w_T-»’wer"
| | | j i
! S~ e AR
z / 1 J e | ° @ i | R
S b U — e
5 | e e T T
& L e TR
- Lo @ | o
1 %? A @ E | ; : b ; ‘,
0.2 i A e ”‘*f"“““’“"i*’””“ """ T T
! ' | b ‘ l | b
| [ !
| E Lo i BRASS TUBING | |
z ! ! Lo | Lo
‘ i i oo b ' i
| ! i ‘ | i Lo j ! { ! f by
| o | e o
» : | oo b | j ! b
00! ? Pt | b
10,000 20,000 50,000 100,000 200,000 500,000 1,000,000
REYNOLDS NUMBER
Fig. 2.25. NaK-Side Pressure-Drop Correlation for Three Heat Exchangers.
is 50% complete. Assembly of the apparatus was Inconel $train-Cycling Tests
delayed approximately one month by drfflzul;ies 3. C. Amos
enc?unfered in the bellows manufacture and by a Aircraft Reactor Engineering Division
design change in the model,
J. H, DeVan
. Metallurgy Division
The welded beliows was fabricated from hydrau- 9y
lically formed, 0.015-in~thick Inconel diaphragms. C. H. Wells

The vendor had difficulty in heat treating the
Inconel shim stock so that it would not rupture
during the forming operation or crack during the
fusion-welding process.

The design of the core shell was changed to in-
corporate o thermal barrier in the small-diometer
end. This drastically reduced the radial tempera-
ture difference in the shell wall between the test
section and the free, or bellows, end. The heat
dam consists of a rigid, evacuated container filled
with Fiberfrax insulation. This modification pro-
duces conditions which more nearly simulate the
conditions that will exist in the reactor core shell,
in that the thermal sfress induced cannot be re-
lieved by differential axial expansion in the sheil
wall at the free end.

56

Pratt & Whitney Aircroft

Two anvil bending-test assemblies were placed
in operation for obtaining basic information on the
behavior of inconel under strain cycling at ele-
vated temperatures in both inert and corrosive
atmospheres, This assembly, which was described
previously, is shown in Fig. 2.28 with the tank
removed. Tests are now being conducted in helium
at 1200, 1400, and 1600°F. Compressive and
tensile strains of 0,2, 0.4, 0.6, and 1.0% are
infroduced altemately in the outer fibers of the
specimens by bending. The time periods em-
ployed between stress reversals are from Y% to
2 hr. As the individual tests are completed, the
test specimens are forwarded to the Metallurgy
Division for visual and metallographic examina-
tion, The results for the tests completed thus far
PERIOD ENDING DECEMBER 10, 1955

L
ORNL-LR—-DWG 11278

 

 

 

  

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100
| | | | ‘
A INTERMEDIATE HEAT EXCHANGER NO.2, IHE-2 |
A\ ® 20-TUBE FUEL-TO-NoK HEAT EXCHANGER t
‘L\ - © |NTERMEDIATF HEAT EXCHANGER TYPE IHE -3, ORNL-1-2 —
|
~N | s l
KIN PLUS SPACERS |
ct |
!
\.\ J’
5 L__THE-2 |
|
50 P L
~ IHE-3
eo (7 4
z |
'_ |
2 i
L 20-TUBE HEAT EXCHANGER 28
5 l |
= | P l
8] |
& i |
.
| |
' |
l IHE-2 ! |
7-— /ZO-TUBE HEAT EXCHANGER I
SPACERS {
IHE-3 i
R oeEs SE—— __—_-
10 -
1000 2000 5000 8000

REYNOLDS NUMBER

Fig. 2.26. Fuel-Side Pressure-Drop Correlation for Three Heat Exchangers.

57
ANP PROJECT PROGRESS REPORT

L]
ORNL-LR-DWG 11279

 

Nu’/PrO'4

 

2 |~ ® FUEL-TO=Nok HEAT EXCHANGER TYPE IHE-3
[«——= 20 TUBE FUEL -TO—NaK HEAT £ XCHANGER |
| (CURVE BASED ON 70 POINTS) | | °
' |
|

 

L o
|
Ll 1 |
500 1000 2000 5000 {0 000
REYNOLDS NUMBER

Fig. 2.27. Fuel-Side Heat-Transfer Correlation

for Two Heat Exchangers,

are presented in Table 2.8, It is apparent that
temperature has a major influence on the initiation
of cracks., The number of cracks produced for a
given number of cycles was greater at 1200 than
at 1400°F and was greater at 1400 than at 1600°F,

Thermal-Cycling Test of Sodium-Inconel-Beryllium
System

M. H. Cooper R. D. Peadk
Pratt & Whitney Aircraft

The third test to be operated in the sodium-
beryllium-Inconel thermal-cycling test apparatus
described previously was terminated on Septem-
ber 30, 1955, after a total of 397 hr of operation.
Twenty-two cycles had been completed in 202 hr
of cyclic operation. Each cycle represented 4 hr
of operation at high power, during which the
sodium entered the test section at 1150°F oand
was heated electrically to 1300°F, and 4 hr of
operation at low power, during which the sodium
entered the test section at 1290°F and left at
1300°F. Of the 202 hr of cyclic operation, 26 hr

was at steady-state and low power because of

58

UNCLASSIFIED
© PHOTO 4739

 

 

 

LOWER ANVIL -

 

Fig. 2.28. 1inconel 5train-Cycling Assembly.

cycling difficulties. The 195 hr of isothermal op-
eration included a temperature range of 1100 to
1300°F.

The abbreviated test program described above
was plagued by operating difficulties which made
it necessary to drain and cool the system on five
different The first shutdown was
caused by a ledk in the electromagnetic pump cell.
The leak developed when an attempt to start flow

occasions.

by external heating resulted in overheating. The
cell had to be cut out and replaced. Three sub-
sequent shutdowns were caused by the resistance
heater. One of these shutdowns was for repairs to
the control circuit; one was for transformer re-
pairs; and the final one was for transformer re-
placement. The fifth shutdown was due to failure
of the braze joint between the bus bar and one end
of the test-piece container. It was during attempts
to rebraze this joint that the test-piece container
cracked; the resultant sodium leak caused a fire.
PERIOD ENDING DECEMBER 10, 1955

TABLE 2.8. RESULTS OF STRAIN-CYCLING TESTS ON INCONEL SPECIMENS

 

 

Cycle Specimen Results of
Specimen Strain Number of Time Temperature Mi croscopic
No. (%) Cycles (hr) CF) Examination
1F5 1 60 1 1400 Cracks observed
1E5 1 80 1 1400 Cracks observed
D7 1 100 1 1400 Cracks observed
1c7 1 120 ] 1400 Cracks observed
1F4 ] 50 ] 1600 No cracks observed
1E4 1 100 1 1600 No cracks observed
1F2 1 50 1 1200 Cracks observed
1E2 1 60 1 1200 Cracks observed
1F6 1 50 4 1200 No cracks observed
1E6 1 60 4 1200 No cracks observed
1F3 0.6 100 1 1400 Cracks observed
1E3 0.6 150 1 1400 Cracks observed

 

No further attempt was made to continue this test,
and the loop is now being cut apart for metallurgi-
cal examination. Preparations for a fourth test
are under way.

REACTOR COMPONENT DEVELOPMENT
Dump Valve

J. J. Milich
Pratt & Whitney Aircraft

J. W. Kingsley

Aircraft Reactor Engineering Division

The first, prototype, ART dump valve was
checked, and the preliminary tests were com-
pleted. This hydraulically actuated valve is of
the plug type, with an Inconel bellows seal, and
helium is utilized for the cooling of the bellows
and the valve stem. The seating material for the
plug is Kennametal 152B (64% TiC-30% Ni-6%
NbTaTiC;), and for the seat it is Kennametal
162B (64% TiC-25% Ni-5% Mo-6% NbTaTiC,).
The valve was fabricated by Black, Sivalls &
Bryson, Inc.

The preliminary evaluation testing consisted in
closing the valve with hydraulic oil pressure and
introducing helium pressure against the valve seat.
The data taken were the leak rate of helium past
the valve seat and the opening and closing hy-

draulic pressures on the actuator. The tests were
made with the valve at room temperature and at
temperatures up to 1300°F,

The room-temperature tests indicated that the
leakage rate decreased as the seating pressure
was increased, as shown in Table 2.9. The high-
temperature test indicated that there was a marked
decrease in the helium leakage rate as the tem-
perature was increased. The hydraulic pressure
necessary to open and close the valve increased
as the temperature was increased. The data from

the high-temperature tests are presented in Table
2.10,

The third portion of the evaluation test con-
sisted in keeping the valve closed for a period of
one week (168 hr) with helium pressure against
the valve seat. lLeakage rates were determined
during and at the end of this period, and at the end
of the period the opening pressures were obtained.
Two tests were made: one at 1200°F and the other
at 1400°F. At the end of each test, a loud me-
chanical noise was heard when the valve was
opened. When the valve was closed after the
noise was heard, the leakage rate was higher;
however, after the valve had remained closed for a
long time, no leakage could be detected. The re-
sults of these tests are presented in Table 2.11.

59
ANP PROJECT PROGRESS REPORT

TABLE 2.9. RESULTS OF ROOM-TEMPERATURE TEST ON ART PROTOTYPE DUMP YALVE

 

 

Oil Pressure 0il Pressure Valve Seating Helium Lecdkage
to Close to Open Pressure Pressure Rate
(psi) (psi) (psi) (psi) (em>/min)
60 80 250 10 0
20 0
40 5
60 15
80 25
60 80 300 10 0
20 0
40 3
60 15
80 20
94 25
50 B0 150 10 5
20 15
40 35
60 90
80 150

 

In future tests the valve will be evaluated ot
high temperatures in contact with fluoride fuels.
The leakage rate, the opening and closing pres-
sures, and the general reliability will be deter-
mined in order to evaluate the valve for use in the

ART.

Cold Trap and Plugging Indicator

J. J. Milich
Pratt & Whitney Aircraft

A test assembly for use in the development and
testing of cold traps and plugging indicators is
being fabricated. This test loop will include a
circulating cold trap, an economizer, and a plug-
ging indicator, and it will draw from a NaK reser-
voir with a capacity of 80 gal.

York Demister packing will be used in the circu-
lating cold trap to provide the surface to hold
precipitated oxides. Precautions will be taken to
ensure that the cold-trap temperature is the lowest
in the system so that the soturation temperature of
the liquid metal with respect to Na,O will not be
reached at any point in the system other than the
cold trap.

The plugging indicator will consist of a disk
with 19 holes 0.040 in. in diameter. The rate of
flow through the holes will be 1.0 gpm. The fluid

60

will be cooled until the oxide precipitates and
partially plugs the holes. An electromagnetic
flowmeter will give an indication of the change in
flow when the plugging indicator partially plugs.
When the change in flow rate occurs, the tempera-
ture at the plugging indicator will be recorded and
will be assumed to be the saturation temperature
of the oxide. This temperature can be translated
to oxide concentration by means of a solubility
curve. !

Zirconium Fluoride Yapor Trap

J. J. Milich
Pratt & Whitney Aircraft

J. W. Kingsley

Aircraft Reactor Engineering Division

A trap is being developed for collecting the zir-
conium fluoride vapor that will be formed during
operation of the ART, since if this vapor were
allowed to enter the off-gas system the lines would
become plugged with ZrF, and the off-gas system
would become inoperative. Various designs are

being studied, and a prototype trap is being tested

 

”C. B. Jackson (ed.), Ligquid Metals Handbook.
Sodium-NaK Supplement, TID-5277 (July 1, 1955), p 8.
PERIOD ENDING DECEMBER 10, 1955

TABLE 2.10. RESULTS OF HIGH-TEMPERATURE TEST ON ART PROTOTYPE DUMP VALVE

 

 

Test Qil Pressure Oil Pressure Valve Seating Helium lL-edkage
Temperature to Close to Open Pressure Pressure Rate
(°F) (psi) (psi) (psi) (psi) (em® /min)
400 80 100 250 10 0
20 2
40 5
60 8
80 10
90 15
600 60 80 250 10 1
20 2,5
40 7
60 10
80 13
800 40 75 250 10 0 to 0.5
20 2
40 1.5
60 2.5
80 2.5
90 3
45 90 300 10 0 to 0.5
20 1
40 2
60 3
80 4.5t0 5
90 5
1000 145 170 250 10 0 t0 0.5
20 1.5
40 1.5
60 2
80 2
90 2
120 190 300 10 0
20 0
40 0
60 0
80 1
90 1
1100 120 245 250 10 0 to 0.5
20 0.5 to 1
40 2
60 2.5
80 2.5
90 3

61
ANP PROJECT PROGRESS REPORT

TABLE 2.10 (continued)

 

 

Test Qil Pressure Oil Pressure Valve Seating Helium Leakage
Temperature to Close to Open Pressure Pressure Rate
CF) (psi) (psi) (psi) (psi) (em? /min)
1100 190 160 (started to open) 300 10 0
310 (fully open) 20 0
40 0
60 0
80 1
20 1to 1.5
1200 190 210 (start) 250 10 0
320 (fully open} 20 0
40 0.5
60 0.5
80 0.5
90 0.5
240 240 (start) 250 10 0
350 (fully open) 20 0
40 0.5
60 0.5
80 0.5 to 0.75
90 0.75 to 1
290 240 (start) 300 10 0
350 (fully open) 20 0
40 0
60 0
80 0 to 0.25
90 0 to 0.25
1300 250 265 (start) 250 10 0
400 (fully open) 20 0
40 0
60 0
80 0
90 0
300 250 (start) 300 10 0
400 (fully open) 20 0
40 0
60 0
80 0
90 0
60 80 250 10 0
20 0
40 5
60 15
80 25

62
1)

PERIOD ENDING DECEMBER 10, 1955

TABLE 2.10 (continued)

 

 

Test Oil Pressure Oil Pressure Valve Seating Helium Leakage
Temperature to Close to Open Pressure Pressure Rate

CF) (psi) (psi) (psi) (psi) (em®/min)

1300 60 80 300 10 0

20 0

40 5

60 15

80 20

94 25

50 80 150 10 5

20 15

40 35

60 %0

80 150

 

TABLE 2.11. RESULTS OF TESTS OF VALVE AFTER IT HAD BEEN CLOSED FOR 168 hr

 

 

Time Valve Temperature Valve Seating Pressure Helium Pressure
Closed (°F) (psi) (psi) Leakage Rate
168 hr 1200 250 100 1.15 em>/he
*
168 hr 1400 250 100 0.875 cm>/hr
Valve Leakage After the Above Tests
30 min 1400 250 10 12 em3/min
20 21 cm3/min
40 42 cm3/min
60 65 cms/min
30 min 1400 250 10 N em3/min
20 18 cm3/min
40 36 cm>/min
60 56 cm3/min
85 hr 1400 250 90 No detectable leakage

 

with the fuel mixture NoF—ZrF4-UF4 (50-46-4
mole %) as the ZrF, vapor source.

The trap consists of a bundle of 30 tubes ar-
ranged in a 6-in. sched-40 pipe. The tubes,
through which the gas flows, are packed with 1-in.
segments of Inconel mesh separated by 1-in. void
spaces. 1he tubes are cooled by water circulating
through the 6-in. pipe. The trap is connected to a
sump tank that is filled with the fuel. To simulate
the reactor conditions, helium is bubbled through
the sump from a dip line. The helium, which will
be present in the reactor to sweep the xenon from

the fuel, agitates the fuel mixture and increases
the surface area and thus increases the ZrF,
vaporization rate, The ZrF , vapor then passes
from the sump tank through the trap. The flow rate
of the vapor is measured by wet test meters. The
fuel is maintained at a temperature of 1300°F, and
the exit line from the sump to the trap is held at
1400 * 25°F throughout its length,

The results of the first attempts to run tests
with equipment indicate that there must be closer
temperature control of the inlet line to the trap.
[f there is a temperature differential of 25°F or

63
ANP PROJECT PROGRESS REPORT

more on the 0.5-in.-OD, 0.025-in.-wall inlet line,
the ZrF , vapor precipitates and plugs the line.

Water Test of Aluminum Mockup of Top of ART
D. R. Ward

Aircraft Reactor Engineering Division

Several problems connected with the design and
operation of the top portion of the ART are to be
investigated by using a full-scale aluminum model.
Fabrication of the model is under way.

This test unit, as shown in Fig. 2.29, will be
equipped with lnconel fuel and sodium pumps and
the external piping needed to permit the pumping

64

of water under simulated reactor flow conditions.
Operation of this system will provide means for
studying the following problems: fabricating, as-
sembling, and welding of the component parts;
controtling pump speed; maintaining flow under
both normal and unbalanced conditions; ingassing
and degassing; filling, draining, and venting; con-
trolling liquid levels; and pressurizing gases.

Drawings of the fuel-circuit components of the
test assembly have been completed, and parts are
being fabricated. The sodium-circuit components
will be fabricated later. A test room is being pre-
pared, and the major auxiliary equipment has been
ordered.
59

ORNL-LR—DH;!! 11280

 

 

 

 

 

   
  
    

   

 

 

 

  

 

 

 

 

 

 

 

FUEL. PUMP SODIUM PUMP
FUEL PUMP
(> (=
/ GATE VALVE
é((l([’l{% SR 'O + ORIFICE
4 5 ){ .+ GATE VALVE FUEL CIRCUIT
| /
— @ l[h I —
T Ty FiN | a0 [ -~
MAIN SODIUM CIRCUIT VISUAL v SODIUM ISLAND CIRCUIT r
FLOW ' '@ WINDOW ¢
! INDICATOR 2 / HEAT EXCHANGER
| L 7
-~ D
=, sopium
t = E,T_. BYPASS
f Neaa@n)Pe) e - AN
~ @47 e i
“VISUAL W e FILL
INDICATOR - || HEAT EXCHANGER
N e (== o () —
FILL uln[il'!-mi DRAIN
DRAIN
NN U SO SRR

Fig. 2.29. Assembly for Water Tests of Full-Scale Aluminum Mockup of Top of ART.

SS61 ‘0L ¥3gW3ID3A INIAONI 4Oy 3d
ANP PROJECT PROGRESS REPORT

3. CRITICAL EXPERIMENTS

A. D, Callihan .

J. J. Lynn

E. R. Rohrer

Applied Nuclear Physics Division

D. Scott, Aircraft Reactor Engineering Division

R. M. Spencer, United States Air Force

J. S. Crudele
W, J. Fader

E. V. Sandin
S. Snyder

Pratt & Whitney Aircraft

The present series of critical experiments on
the circulating-fuel reflector-moderated reactor,
including one assembly operated at about 1200°F,
has been completed. A program of critical ex-
periments on reflector-moderated assemblies of
a somewhat different design proposed by Nuclear
Development Corporation of America

initiated,

has been

ROOM-TEMPERATURE
REFLECTOR-MODERATED-REACTOR
CRITICAL EXPERIMENTS

Critical experiments were recently completed on
room-temperature reflector-moderated-reactor as-
semblies that mocked up the reactor fuel annulus,
the beryllium island, and the beryllium reflector
and included extensions that corresponded to the
entrance and exit fuel flow channels. The ex-
tensions are referred to as ‘‘end ducts.”” The
basic configuration was described earlier! as
assemblies CA-21-1 and CA.21-2, which differed
only in the uranium density in the fuel region
and therefore in the available excess reactivity,
Two modifications of the basic structure were
examined, both having been described previously.!
In one, assembly CA-22, the outer diameter of
the fuel section of one end duct was increased
from 5.28 to 6.79 in.; in the other modification,
assembly CA-23, the radius of the central beryllium
island was increased from 5,18 to 7.19 in,
additional results of these experiments are re-
ported here,

Some

Power and Neutron-Flux Distributions

The relative power distribution in assembly
CA.22 was measured along the radius at the

 

1A. D. Callihan et al., ANP Quar, Prog. Rep. Sept. 10,
1955, ORNL-1947, p 58.

66

midplane and at three locations in the large end
duct. An additional radial traverse was made in
the small end duct for comparison. The conven-
tional method of measuring the fission-fragment
activity caught on 5-mil-thick aluminum disks in
contact with uranium foils was used. The results
of these measurements, as shown in Fig. 3.1, are
similar to those reported previously? for the
assembly of basic dimensions.

Two radial flux traverses were made in assembly
CA-22, both with bare and with cadmium-covered
gold foils. One traverse (Fig. 3.2) was at the
midplane and extended from the axis a distance
of 17.84 in. The other traverse (Fig. 3.3) was
in the large end duct in a plane 11.5 in. from
the midplane and extended between points 0.31]
and 13.72 in. from the axis. It is to be remembered
in analyzing these data that the low intensity of
slow neutrons in the center of the fuel region
introduces a large uncertainty in the values of
the cadmium fractions derived from both power
and flux measurements.

Neutron Production in the Fuel-to-NaK
Heat Exchanger

The fissioning in the uranium that will be con-
tained in the fuel-to-NaK heat exchanger of the
circulating-fuel reflector-moderated reactor will
be a source of high-energy neutrons which must
be attenuated in the shield surrounding the re-
actor. A measure of the rate of this fissioning
was obtained from catcher foils placed in a mockup
of a section of the heat exchanger that was in-
corporated in assembly CA-21-2, the assembly
having basic dimensions and the lower uranium
density. The heat-exchanger-region mockup con-

 

25, D. Callihan et al., ANP Quar. Prog. Rep. June 10,
1955, ORNL-1896, p 47.
CRITICAL ASSEMBLY

 

U

PER!IOD ENDING DECEMBER 10, 1955

ORNL—-LR-DWG 11025

\\\\\\\\\\\\\\\\\\\\\\\\\

 

 

 

 

 

o ; J':f i (\Ej i BERYLLIUM REFLECTOR x
T ):E - E ; g—— :I FUEL \\\i{_-fll_-_j-jfi-_ ’LE;;,T___': ___'_1 \
7 I T,

BERYLLIUM ISLAND

END OF REACTOR—

 

80

T [

 

 

70

]

 

 

=

 

60

 

 

—

!

 

50

 

 

i 4 i
r l 1 r \ .
| © BARE FOIL . ‘
{ A CADMIUM COVERED

 

-
} DMIUM FRACTION
LT

 

 

 

 

 

J
|

B

|
|

 

 

 

 

 

 

 

RELATIVE ACTIVITY {garbitrary units )

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

*fiq‘t* 410

 

 

 

 

 

 

 

 

 

 

 

40 o TL - : —fir | ¥ L VLT 08
L w L LE
30 m&\;l l—fi-# «F v# Jr{——‘» ‘4 06 g
| B ] -
20 S fl -- LWL L %Mfi {043
‘Kc._{L [ ; ] fi | fl é
10 - - ‘L;FL : (—fi‘fi >! i lLL 1~d 02 ©
0 L f J l | 1 ‘ }t ‘ | \\ || 0
048 0 & 27 o ¢ 47 o 7 a7 o £ 17
LAYERS OF FUEL FROM ISLAND (! LAYER = 0.004 in. OF U AND 0.142 in. OF TEFLON )
Fig. 3.1. Power Distributions Across the Fuel Annulus and in the Large End Duct of a Room-Tempera-

ture Reflector-Moderated-Reactor Critical Assembly,

sisted of a layer of boral (300 mg of boron per
square centimeter) adjacent to the reflector be-
followed by a 1.,3-in.-thick fuel zone
containing uranivm metal and Teflon, and a layer
of stainless steel 1.437 in. thick. It was located
approximately at the midplane, and its dimensions
in the plane perpendicular to the flux gradient
were 8.625 by 8 in. The arrangement and the
results are shown in Fig. 3.4, together with the
fission-rate distribution across the core midplane

ryllivm,

for comparison,

Other measurements of neutron leakage from the
beryllium reflector have shown that a layer of
boral having o density of about 300 mg of boron
per square centimeter reduces the relative value
of the flux from 1000 to 3; an increase of the
boron density by 50% only reduced the flux to
a relative value of 2,

Radial Importance of Uranium in the Fuel Annulus

The effect of adding uranium at different po-
sitions across the fuel annulus was measured in
assembly CA-23, the assembly with the enlarged
central island. Two 7.1875- by 2.875-in. sheets
containing a total of 45.58 g of U235 were added
to the normal 18-sheet loading and to a reduced
loading of 16 sheets. The 16-sheet base was
established by the removal of two uranium sheets
from their respective positions at one-third and
at two-thirds of the distance across the annulus.
The test section was at the top of the core region
in the assembly and was centered ¥ in. from
the midplane. The results are given in Table 3.1

and Fig. 3.5.

Importance of Beryllium at End of Reactor

The

locations of the sodium heat exchangers

67
ANP PROJECT PROGRESS REPORT

ACTIVITY (arbitrary units)

 

 

 

 

 

 

  
 

 

 

 

 

 

" AL
v
ORNL—LR-DWG 11026
1.0 T T T T i ’ ’ 1.0
| | : | - |
i i | I l :
| f ‘ | | |
oo b — — — — A4 --F + — — T -+ ——— o9
| | i
< BERYLLIUM ISLAND - —»f=— —— FUEL — — »f= :—-m— BERYLLIUM REFLECTOR ! o
: | ‘ i i ‘ | //_..q
| | ‘
e N — - — - — % — - — - - - — ———1 08
| B
@ i | J/
< - INCONEL CORE | l / |
: SHELLS — —— = ' *L // [ J
07— — —\ — + - v’— -t o7
| . ‘ '
| ( | | | }A\CADMIUM FRACTION
| \ ; /| | |
b )LJ—‘)T—‘ o dos
| | # s | | | z
' | S ' l 5
I ‘ g
l /S | | :
e s
crimicaL assemaLy I | ’ | z
CA-22 T | | | 2
J |
- -, —'l—— L {104
| | CO~p. _ BARE GOLD ACTIVATION
| | o \
| , ]
NS N,
b | | | | | |
| ' i S | I |
| i | |
gr _ ~LA’ - 7$7L+7+7
| Y | |, CADMIUM~ COVERED
\ | | il e /GOLD ACTIVATION
! - '
\ | 144 .(
'.‘k - 0 — e
.N v 1 ‘
| \-—T/ { ’
] | | 1
6 8 10 12
RADIAL DISTANCE FROM AXIS (in.)

Fig. 3.2, Neutron-Flux Distribution Along Midplane of a Room-Temperature Reflector-Moderated-Reactor
Critical Assembly with ¢ Large End Duct.

and the swirl chambers of the pumps of the ART
have necessitated the removal
beryllium from one end of the reactor. An estimate
of the loss in reactivity to be expected by this
removal has been made in some measurements on

assembly CA-22.

68

of some of the

the resulting degreases in
from the critical
rods.

positions
In a second series

In one series of three measurements the be-
ryllium was removed from
reflector and the island in

the end of both the
1-in. increments, and
reactivity were noted
of calibrated control
the island beryllium
 

     

 

 

 

 

PERIOD ENDING DECEMBER 10, 1955

 

 

 

 

 

 

 

 

 

 

 

 

 

* AN g
v CRNL—LR—DWG 11027
1.0 [ | 1.0
i - r ‘ |
| CRITICAL ASSEMBLY
? | CA-22 !
J‘ ,
‘ 0.9 ——_*% — —wjfiflf—L— rrrrr fi .9
J
e BERYLLIUM ISLAND BERYLLIUM REFLECTOR — —— —=
| | |
| | . i =
j 08 f———— R —_—— L 7T o
1 !
? } 7
! INCONEL CORE | ~ / ‘
| J SHELLS ~ /
: | |
a o7t — — 07
i [
! } fifADMlUM FRACTION
‘ | /. |
| 0.6 — e 1 08
! » | \ | 5
P s | | / | | 5
; o / i &
§ E w
a JH N IR 0
! 2 ‘ z
> ~ ‘ o
i 5 O ~N I | ) ‘
; % o4 N R S — .._____,_Lfi_fi_i 0.4
\ { | |
| ‘ | _BARE GOLD ACTIVATION
0.3 i~—4lf RT I o =g & a—i —— 03
|
! I ‘ o
5
oel Se_ Y ] |
. | l
| |
i | o
CADMIUM ~ COVERED
® GOLD ACTIVATION
oo
~e |
|
- @
| e
; . | | 0
% . 0 2 8 10 12 14
| RADIAL DISTANCE FROM AXIS (in.}
!
| Fig. 3.3. Neutron-Flux Distribution Along a Plane 11.5 in, from Midplane of a Room-Temperature

Reflector-Moderated-Reactor Critical Assembly with a Laorge End Duct.

69
ANP PROJECT PROGRESS REPORT

s R

I 77

=
ORNL—LR-DWG 11028

N

 

  
    

   
    
  

 

 

90 = BERYLLIUM //A=——— FUEL ANNULUS ——>}«""BERYLLIUM | HEaT |
/7 I1SLAND ————m] V'///REFLECTOR / EXCHANGER
’ . 1
/ F ¢ 1.5 in. —~ Z=4\Y MOckup
FUEL

N
N

REFLECTOR —
30

20 _ . — %

 MULTIPLY ORDINATE 7]

L SCALE BY 10~2
10 L . R

/ | / /
0 22 9
LAYERS OF FUEL FROM ISLAND LAYERS OF FUEL FROM BORAL
(1 LAYER = 0.004 in. OF U AND (1 LAYER = 0.004 in. OF U AND
0.142 in. OF TEFLON)

Q473 in. OF TEFLON)

N

N\
o _ RsmanLess\
[ N\ STEEL 0NN

**\125 in. a-q

BORAL, Y in.=~-3

   

2222 72
|
|

s

SURFACE OFj§
//

E

yd

 

 

RELATIVE ACTIVITY (arbitrary units)

N

7
|

 

 

 

 

 

 

 

 

 

oz

CRITICAL ASSEMBLY
CA—-21-2

Fig. 3.4. Power Distribution Along Midplane of Fuel Annulus and Heat Exchanger Mockup of a Room-

Temperature Reflector-Moderated-Reactor Critical Assembly,

was replaced and an additional increment was
removed from the reflector. The results are re-

ported in Table 3.2 and Fig. 3.6.

Axial Importance of a Neuvtron Source

The effect of the axial position of a neutron
source (Po-Be) on the neutron level in a sub-
critical reactor {assembly CA-23) with a constant
neutron multiplication of about 600 has been
observed. The source was moved along a channel
in the beryllium island, and the change in the
neutron intensity at each of several external
detectors was observed. A fission chamber and

70

two BF3 ionization chambers were located sym-
metrically about the end of the reactor away from
the direction of source removal, and a third BF3
ionization chamber, channel A, was placed near
the axis at the other end of the reactor. The
results from the first three detectors agreed, and
their average is given in Table 3.3 and Fig. 3.7,
together with the data from channel A. Ail count-
ing rates were normalized to those observed with
the source at midplane. These results also
illustrate the importance of source-counter geome-
try on the interpretation of multiplication curves
in the approach to critical.
TABLE 3,1.

¥ g,

PERIOD ENDING DECEMBER 10, 1955

RADIAL IMPORTANCE OF URANIUM IN FUEL ANNULUS

Number of Sheets Between Island

Distance Between Island and

and Added Uranium

Change in Reactivity from

 

Added Uranium

Normal 18<Sheet Loading

 

 

(in.) Uranium Teflon (cents)
(4 mils thick) (71 mils thick)
18-Sheet Basic Loading 0.0
0.071 0 ] +5.9
0.292 2 4 +4.5
1.314 9 18 +3.0
2,265 16 31 +3.2
2.486 18 34 +3.7
16-Sheet Basic Loading —4,2
0.071 0 1 +2.5
1.310 8 18 -0.7
2.478 16 34 +0.2
o TABLE 3.2. IMPORTANCE OF BERYLLIUM

ORNL—~LR—DWG 11029

 

1 |

CRITICAL ASSEMBLY
CA—-23

T*#J

 

 

SAMPLE ADDED TO 18- SHEET

BASIC LOADING
LT
| J:

2 | — | —
SAMPLE ADDED TO 16 - SHEET
BASIC LOADING ‘ I

NDRMAL FOR 18- SHEET BASE

|

NN,
| [
|

 

 

 

 

 

REACTIVITY CHANGE (cents)

 

s
l
|
|
J T
| | -
i NORMAL FOR 16-SHEET BASE |
| e L

 

 

 

R
| |

I
| -
| .

 

-6
0 10 20 30 35 40
DISTANCE FROM ISLAND (1 UNIT = 0.073 in.)
Fig. 3.5. Radial Importance of Uranium in the

Fuel Annulus of a Room-Temperature Reflector-
Moderated-Reactor Critical Assembly.

AT END OF ASSEMBLY

 

Distance from Midplane to

End of Beryllium Reflector Change in Reactivity

(in.) (cents)

 

Beryllium Removed from Reflector and Island

20.84 0

19.84 -14.2
18.84 -34.2
17.84 -65.9

Beryllium Removed from Reflector Only

 

20.84 0

17.84 -45.8

16.84 -79.4
HIGH-TEMPERATURE

REFLECTOR-MODERATED-REACTOR
CRITICAL EXPERIMENTS

A reflector-moderated-reactor critical assembly,
which operated at about 1200°F and used a liquid
mixture of sodium, zirconium, and uranium fluorides
as fuel, was described previously.! The results
of one additional measurement are reported here.
Several disks of enriched uranium metal, 0.75 in.
in diameter and 0.004 in. thick, were placed

71
ANP PROJECT PROGRESS REPORT

i

ORNL-LR—-DWG {11030

 

 

 

 

 

 

 

 

 

 
 
  
     

 

 

 

 

 

 

 

  

 

 

 

0 T T
— T T
ONLY REFLECTOR BERYLLIUM REMOVED '
% ~20 %k — — — «‘
g |
3 {
L —30 |
- |
>
Z —40 -]
2 e
E-so b — —— — P —_—
z ISLAND AND REFLECTOR BERYLLIUM REMOVED B
‘LBI -60 S - — ii T ‘_;_T T
Z
% ;
5 -70 — — - Nd—_— ]
|
-8 L CRITICAL ASSEMBLY —| — — . B
T L ca—z2 1
~90 ' ‘
2 20 19 18 17 15
DISTANCE FROM MIDPLANE TO END OF BERYLLIUM REFLECTOR (in.)
Fig. 3.6. Importance of Beryllium at End of a Room-Temperature Reflector-Moderated-Reactor Critical
Assembly.

TABLE 3.3. VARIATION OF COUNTING RATE WITH NEUTRON-SOURCE POSITION

fonrat Al

 

Source Position Along Axis;

Counting Rate (Normalized to Midplane Rate)

 

Distance from Midplane

 

(in.) Average for Three Chambers Channel A
0 1.000 1.000
5 0.974 0.982
10 0.791 0.837
15 0.458 0.618
20 0.116 0.382

 

between somewhat thicker CaF, disks in Inconel
tubes. During the assembly® of the core shells
these tubes were mounted at three axial positions
inside the annular fuel region, and they remained
there during the entire experiment.
measures

Relative
of the power-production distribution
across the fuel region were subsequently deter-
mined from the fission-product activity which
accumulated in the disks.

A schematic drawing of a cross section of the
reactor in the longitudinal midplane is shown in
Fig. 3.8 with the positions of the foil capsules

 

3P. Patriarca, ANP (uar. Prog. Rep. Sept. 10, 19553,
ORNL-1947, p 133.

72

and the beryllium island and reflector noted.
An effective-poison-rod location is also noted.
The relative activations of the uranium foils,
normalized for decay, are also shown in the figure.
The curves represent the power production across
the fuel annulus at three longitudinal positions.

The proximity of the lowest traverse to the
borori-copper plate at the bottom of the assembly
accounts for the depression of the fission rate
there. At the other two elevations the fission
rate on the island side of the fuel region is lower
than that on the reflector side. It is not known
why the relation between these two values is
the inverse of observations made with the room-
temperature critical experiments, as shown, for
example, in Fig. 3.1.
i U

PERIOD ENDING DECEMBER 10, 1955

ORNL—LR—DWG {1039

 

 

l
0.8 -F —
J

0a ____%*F -

|

4

AVERAGE OF THREE CHAMBERS

0.6 w—_Jr* - ]

  
 

 

 

 

 

COUNTING RATE { NORMALIZED TO MIDPLANE RATE)

 

 

 

 

 

 

CRITICAL ASSEMBLY
CA—-23
. |
C 5 10 15 20

SOURCE POSITION ALONG AXIS (DISTANCE FROM MIDPLANE) (in.}

Fig. 3.7. Axial Importance of a Neutron Source in a Room-Temperature Reflector-Moderated-Reactor

Critical Assembly,

COMPACT-CORE
REFLECTOR-MODERATED-REACTOR
CRITICAL EXPERIMENTS

A reactor embodying the reflector-moderator con-
cept and consisting of solid fuel elements located
in the channels between the beryllium reflector
and the beryllium island has been proposed by
NDA.* The fuel is cooled by a stream of sodium,
A program of critical experiments has been initi-
ated, in collaboration with NDA, which is basically
similar to the recently completed ORNL program,

 

4CCR-2: A Compact Core Reactor for Aircraft Pro-
pulsion, NY0-3080 (July 30, 1954).

The fuel is formed by alternating lamince of
uranivm foil and sheets of stainless steel and
aluminum, the latter to represent the sodium of
the reactor., The core region is essentially a
20.125.in.-long cylindrical annulus 10 in. in inside
diameter and 4.312 in. in width. The central
17.25-in, section of the annulus contains the
vranium, (The bounds of the annulus are actually
octagonal, rather than circular.) A preliminary
loading indicates that the predicted critical mass
has been significantly underestimated, A 35%
increase in the uranium loading and a 23% de-
crease in the steel content of the core were
necessary in order to make the assembly critical,
The core now contains 31 kg of U235,

73
mfi'v"'-"fl.*u
v
ORNL—LR—DWG 11032

 

 

EFFECTIVE 1]
[

CONTROL ROD POSITION, || f} /
J O

BERYLLIUM ;
REFLECTOR

 

 

FUEL
ANNULUS

SAMPLE A

 

URANIUM FOILS
IN INCONEL TUBE

o

 

BERYLLIUM 9
ISLAND ;
SAMPLE A
8
CONTROL ROD THIMBLE
.
I
SAMPLE B
1 6

 

'

 

RELATIVE ACTIVITY (arbitrary units)

 

 

 

 

 

 

 

 

 

 

 

 

2 3 4 5 6

0
J DISTANCE FROM INNER EDGE
BORON - COPPER OF FUEL ANNULUS (in.)

<<(/E\\T %/// —ows 10115“”]“ .

 

Fig. 3.8. Relative Fission-Rate Distributions Across Fuel Annulus of High-Temperature Reflector-
Moderated-Reactor Critical Assembly.

74
Part ||

MATERIALS RESEARCH
e 4. CHEMISTRY OF REACTOR MATERIALS
W. R. Grimes

Materials Chemistry Division

Phase equilibrium studies were made of the sys.
tems ZrF UF ,, LiF-UF , NaF-LiF-UF , KF-UF
NaF-KF-ZrF,, NaF-LiF-BeF, NaF.LiF-BeF,-
UF ,, KF-BeF,, and NaF-KF-BeF , and of systems
containing alkaline-earth fluorides,

Additional work was done in investigating the
equilibrium reduction of FeF, by hydrogen in
NaZ:F , the reduction of UF , by structural metals,
the stability and solubility of chromium fluerides
in various molten fluorides, the reaction of UF
with alkali fluorides, and the reduction of alkali
fluorides by uranium metal. A method for preparing
pure CrF ; was devised, and x-ray and petrographic
data were obtained for compounds of ZrF , with
CrF,, NiF,, and FeF ,.

Fuel purification and production research in-
cluded further study of methods for the recovery of
contaminated fuel for re-use and the preparation of
special materials. Methods for evaluating raw ma-
terials were developed and used for evaluating
material from an outside vendor., Adequate supplies
of ZrF, are now being obtained from a commercial
source, Difficulties with vacuum pumps and
reactor cans used in production processes were
resolved.

Methods have been devised for obtaining relative
measurements of viscosity and density of molten
salts without removing the pure material from the
conventional equipment used for fuel purification,
Measurements of electromotive forces of cells in
which molten NaF.ZrF, is used as the solvent
were continued, and optical and x-ray data were
obtained for various compounds in fluoride sys-
tems.

PHASE EQUILIBRIUM STUDIES

C. J. Barton R. E. Moore
F. F. Blankenship R. E. Thoma
Materials Chemistry Division

H. Insley, Consultant

The System Z¢F .UF,

R. P. Metcalf
Materials Chemistry Division

The simple phase diagram for the system Z¢F -
UF, is presented in Fig. 4.1. Solid solutions are

the only crystalline phases which have been ob-
served in this system.! Examination of quenched
samples by x-ray diffraction and with the petro-
graphic microscope shows that the compositions
of the solid solutions vary continuously over the
entire composition range. There is no appreciable
miscibility gap in this system.

The System LiF-UF,

R. E. Moore
Materials Chemistry Division

A phase diagram based almost entirely on ther-
mal analysis was presented previously? for the
LiF-UF, system. A detailed investigation of this
binary system has now been completed, and a re-
vised diagram, based on petrographic and x-ray
diffraction examination of quenched samples, as
well as thermal data, is shown in Fig. 4.2.

A summary of the data obtained from the quench.
ing experiments is shown in Table 4,1. The
samples from which these data were taken were
equilibrated for 16 to 54 hr at temperature before
being quenched. Data obtained by thermal analy-
sis were used to determine the liquidus curves

from 0 to 23 and 60 to 100 mole % UFA.

A compound shown by petrographic examination
to be the result of quench growth was found to
have an x-ray diffraction pattern which did not cor-
respond to the patterns of any of the established
compounds in the system. This compound was
found above the solidus temperature in mixtures
containing 20 to 25 mole % UF,. A composition
and a temperature range for its stable existence
could not be found. It appears to be either a
metastable compound, a compound stable only at a
relatively low temperature, or a high-temperature
modification of Li , UF, with a very namow tem-
perature range of stability.

 

'R. P. Metcalf and R. E. Thoma, ANP Quarn Prog.
Rep. Sept. 10, 1955, ORNL-1947, p 67.

2R, E. Moore et al., ANP Quar. Prog. Rep. Dec. 10,
1950, ORNL-919, Fig. 10.4, p 245.

77
ANP PROJECT PROGRESS REPORT

1050

ORNL—LR-DWG 11524

 

t0C0

950

 

300

 

 

850

 

TEMPERATURE (°C)

 

800 f———

750 |

 

700

 

 

 

 

 

 

ZrF4 {mole %)

Fig. 4.1. Simple Phase Diagram for the System IrF ,-UF .

The System NaF.LiF-UF

H. A. Friedman R. E. Meadows
B. A. Soderberg

Materials Chemistry Division

R. E. Cleary H. Davis
Pratt & Whitney Aircraft

There does not appear to be a mixture in the
NaF-LiF-UF , system that would be a suitable fuel
for a circulating-fuel reactor. No mixtures that
contain less than 25 mole % UF , have been found
that have liquidus temperatures in the region of
500°C. A ternary compound with more than 50
mole % UF, has appeared sporadically in slowly
cooled melts, but in a preliminary survey made
with the use of quenching methods the compound
has not been found. It probably melts incongruently
and has a limited stability range.

Compatibility triangles in the NaF-LiF-UF ; sys-

78

tem are defined by the joins LiF-3NaF-UF , LiF-
2NaF-UF ,, LiF-7NaF.6UF ,, 4LiF.UF ;~7NaF-6UF,,
7LiF-6UF ,-7NaF-6UF ,, and 7NaF-6UF ~LiF.4UF ,.
Of these joins, only the LiF-7NaF-6UF, is a
quasi binary. The boundary curves and three-
phase triangles are shown in Fig. 4.3. The dotted
lines indicate tentative paths for the boundary
curves. In order for the compatibility triangle
7NaF-6UF ,-7LiF-6UF 4LiF.UF , to be valid, the
invariant point common to the primary phase field
7NaF-6UF ,, 7LiF.6UF ,, and 4LiF-UF , must be a
eutectic or its composition must be displaced be-

low the join 7NaF.6UF ,-4LiF.UF .

The System KF.UF

H. A. Friedman
Materials Chemistry Division

A preliminary phase diagram based entirely on
data obtained from thermal analysis was presented
e e

PERIOD ENDING DECEMBER 10, 1955

UNCLASSIFIED
ORNL-LR-DWG 10410

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

HOO T
1000 l - 4
900 Jfi—J*
_ UF, + LIQUID
2N |
o 800 — ]
2%\, /
t._. .
2 Li,UFg + LIQUID |
W 700 \ —t , J —
a LiU,F,, + LIQUID
wi
= _ L
600 LiF +LIQUID '\ P T
. LiU,F,> + UF
%716%1 W 447 4
500 jL/ LIQUID _ Li'?UG F31 + LiU4Fi7 ] )
) . © Li, UF L <
LIF + LiUf | 4°'8 = U
400 3 Li7zUgF3 3| 3
LiF 10 20 30 40 50 60 70 80 90 UF,

UF4 {mole %)

Fig. 4.2. Phase Equilibrium Diagram for the System LiF-UF .

previously? for the KFUF, system. This system
was recently resexamined because of its possible
similarity to the RbF-UF, system, and the results
are presented in Fig. 4.4,

In an early x-ray diffraction analysis of this
system, Zachariasen? reported the presence of
two compounds that contained more than 70% UF .
No evidence has been found in this study for com-
pounds containing more UF , than that indicated by
the formula KF:2UF ,. In addition, the compound
7KF-6UF, is stable, whereas the compound
KF-UF, apparently is not. The optical charac-
teristics of 7KF-6UF, are very similar to those of
the corresponding NaF and LiF compounds.

 

3. p. Blakely et al.,, ANP Quar. Prog. Rep. March
10, 1951, ANP-60, p 131.

4w, H. Zachariasen, Crystal Structure Studies in the
Systems KF-UF4, KF-ThF4 and KF-LaF3, MDDC-1283
(Feb. 9, 1946; dec!. Sept. 5, 1947).

The System NaF-KF.ZrF ,
H. A. Friedman B. A. Soderberg

Materials Chemistry Division

H. Davis
Pratt & Whitney Aircraft

Information on the complex ternary system Nak-
KF-ZrF, has been obtained by thermal analyses
and the careful examination of slowly cooled melts.
The ternary compounds whose compositions are be-
lieved to be certain are 2NaF:3KF.5ZrF, (con-
gruent; mp, 432°C), 3NaF.3KF.2ZrF, (incongru-
ent; mp, ~735°C), NaF-KF-ZrF, (incongruent;
mp, ~493°C), and 3NaF-K F-27_rl:4 (congruent; mp,
593°C). In the compound 3NaF-3KF-2ZrF , the
primary phase is 3KF.ZrF , and in the compound
NaF.KF.ZrF, the primary phase is 3NaF.3KF.
2ZiF ,. A fifth ternary compound exists, but its
composition has not yet been defined. |t appears

79
ANP PROJECT PROGRESS REPORT

TABLE 4.1, SUMMARY OF QUENCHING DATA ON THE LiF-UF4 SYSTEM

 

 

 

Phases
Composition Liquidus Primary P eritectic Secondary Solidus Existing
{mole % Temperature Phase Temperature Phase Temperature Below
UF,) (°C) (<) o) Solidus
20 581 LiF 498 Li4U FB 498 Li4UF8
25 525 LiF 500 Li UFg 490 Li Uy
LizUgF 3
28 503 Li7U6F3] Li4U F8 489 Lf4U FS
33.3 565 LijUFa, Li UFq 495 LiUFg
LizUgF3y
37.5 603 Li7U6F3-|
40 612 Li7U6F3]
46.2 683 LiU F,, 605 LisU Fa 605 LiyUgF g
50 718 LiU,F 588 LigUgF s,y
LIU4F~|7
60 804 UF4 775 Li4UFB
66.7 UF4 767 Li4UF8
75 >716 |..iU4F]7 >601 LTU4F]7
LizUgF 3
80 >716 LiU4F]7 >716 LiU4Fl7
to consist of approximately 24 mole % NaF, 33  at 395°C. The lowest melting eutectic below

mole % KF, and 43 mole % ZrF .

Petrographic and x-ray diffraction data show that
the compatibility triangles in the system exist as
shown in Fig. 4.5. The heavy lines indicate the
well-established triangles. The remainder of the
triangles (dotted lines) are much less certain be-
cause of the inherent difficulty in obtaining valid
data from slowly cooled melts when incongruently
melting compounds are present. The true quasi
binaries in the system are NcF-3KF-ZrF4 (mini-
mum melting temperature, ~765°C at 40 mole %
KF), 7NaF.6ZrF 4-2NaF 3KF-5ZrF , (minimum melt-
ing temperature, ~430°C ot 28 mole % KF),
2NaF3KF.5ZrF -ZrF , (minimum melting tempera-
ture, ~423°C at 27.5 mole % KF), and 2NaF.3KF.
SZrF (KF.ZrF , (minimum melting temperature,
~425°C at 33 mole % KF). The entire region bee.
tween 35 and 50 mole % ZrF , has liquidus tem-
peratures between 395 and 500°C. The lowest
melting region of the diagram is at 10 mole % NaF,
48 mole % KF, and 42 mole % ZrF ,, with a liquidus

80

35 mole % ZrF , is at 30 mole % NaF, 63 mole %
KF, and 7 mole % ZrF, and has a liquidus at
684°C. A program of quenching experiments will
be initiated in the near future to provide the basis
for a final determination of the melting points and
phase relations in this system.

The System NaF-LiF-BeF,

L.. M. Braicher R. J. Sheil
B. H. Clampitt
Materials Chemistry Division

G. D. White
Metallurgy Division

Viscosity data reported in Sec. 7, ““Heat Trans-
ter and Physical Properties,"” for the mixture
NqF-LiF-BeF2 (63.5-7.5-29.0 mole %) and previ-
ously reported data on the mixture NaF-LiF-BeF,
(64-5-31 mole %)> when compared to the viscosity

 

55. 1. Cohen, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL-1947, Table 7.3, p 157.
t

  
 

900

750

PERIOD ENDING DECEMBER 10, 1955

g
ORNL-LR-DWG 14525

UF,

Li Uy Fi7 (945)

TEMPERATURES ARE IN °C

800

 

EUTECTIC
(680)

Na; UgF3,

  

770

 

e SR
700 '
NGZUFG

flv-‘h.h

EUTECTIC

Li; Ug Fzy

LN

 

EUTECTIC &) Y ‘A‘ - 500

RS

950°C

NoF

 

650 LiF

Fig. 4.3. Boundary Curves and Three-Phase Triangles in the NuF-LiF-UF4 System,

of the binary mixture NaF-BeF, (69.8-30.2 mole
%)° show that ternary mixtures containing a small
amount of LiF and approximately 30 mole % BeF,
have a higher viscosity than that of the NaF-BeF,
mixture with the same BeF, concentration. These
data caused a shift in interest to the regions of
lower BeF, concentrations along the join between
the LiF-NaF eutectic (60-40 mole %) and the
LiF-Na,BeF , eutectic (16 mole % LiF). Thermal-
analysis data obtained by adding NaF or Li,BeF,
to several compositions along this join showed
that it is close to a drainage valley between the
two eutectics. Viscosity measurements (Sec. 7,
“‘Heat Transfer and Physical Properties’) were

made on one mixture along this join, NaF-LiF-
BeF, (53-24-23 mole %). The kinematic viscosity
found at 600°C was lower than that of any previ-
ously measured BeF ,-containing melt, except the
NaF-BeF, (69.8-30.2 mole %) mixture. However,
the small improvement in viscosity obtained by
reducing the BeF, concentration in the ternary
system to 23 mole % suggests that the point of
diminishing returns has been reached in this ap-
proach. This conclusion will be tested with one
more composition on the same join, that is, with a
mixture containing only 15 mole % BeF .

The gradient quenching technique has been ap-
plied to the study of two compounds reported to be

81
8

ORNL-LR-DWG 441526

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000 N
900 \ N I - ]
2KF-UF, + LIQUID UF, + LIQUID
3KF - UF, 3KF - UF, 7TKF-8UF, +LIQUID }
8OO +LIQUID  [+LIQUID — — —— — —
KF+
~—1——  KF- 2UF, + LIQUID
¢ 700 m%r—mi‘ B - — T
L
% a—2KF.UF4+ Q—ZKF'UF4+
= 3KF- UFy - 7KF-8UF,
ax
[TE]
o
2 600 -y - — —_t
[
o KF:-2UF,+
5 7KF-6UF, KF-2UFs+ UF,
KF+ 3KF -UF, o
®
500 — |
B-2KF-UF,+ J
TKF-6UF, 7 L
Ly 35
B-2KF-UF, + 5 2 &
3KF-UF, |- an_ ; L
N M~
400 - “JT$*‘“
- 2KF-UF, +
Y~ 2KF-UF, + et 4
3KE-UF, 2 FKF-BUF,
300 + T |
KF 20 30 40 50 60 70 80 90 UF,
UF4(m0|eWO)

Fig, 4.4. Phase Equilibrium Diagram of the KF-UF , System,

¥y e

L¥0dIY SS3¥00¥d LIO3r0¥d dNV
in this system. Quenches of the Na,LiBe,F, com-
position showed only the crystaliine phase desig-
nated by the formula. Further work will be re-
quired to determine the liquidus temperature for
this composition, which undercools strongly, and
to demonstrate that it melts congruently, as is
presently believed. A number of quenches have
been performed with a mixture having a composi-
tion corresponding to the compound Na,LiBe F,,
reported by John,® Quenches covering the tem-

UNCLASSIFIED
ORNL—LR—-DWG 11527

   

-

Rt “=-\2KF-Zrfy

-~

  

 

Fig. 4.5. Compatibility Triangles in the NaF-KF-
ZrF , System,

PERIOD ENDING DECEMBER 10, 1955

perature range 296 to 620°C have shown only
isotropic materials above 339 + 5°C and a mixture
of phases, including LiF and an unidentified
biaxial-positive crystalline phase, at lower tem-
peratures, Further studies are being made in order
to determine the composition and melting behavior
of the unidentified compound.

The System NaF-LiF-BeF -UF,

L. M. Bratcher R. J. Sheil
B. H. Clampitt

Materials Chemistry Division

G. D. White
Metallurgy Division

Investigations of mixtures in the NaF-LiF-Be F,-
UF, system containing 3.0 mole % UF, and low
quantities of BeF, were continued. These mix-
tures are considered to be promising from the vis-
cosity standpoint. Thermal-analysis and filtration
techniques are being used in this investigation.
The available data, which are summarized in
Table 4.2, demonstrate that it is possible to ob-
tain mixtures in this system containing 3.0 mole %
UF, and 20 to 22 mole % BeF, that have liquidus
temperatures of less than 505°C. Investigation of
this system is continuing in an effort to determine
the best fuel composition from the standpoint
of both viscosity and liquidus temperature.

 

W, Jahn, Z. anorg. w allgem. Chem. 277, 274 (1954).

TABLE 4.2, LIQUIDUS TEMPERATURES IN THE SYSTEM NoF.LiF-BeF -UF ,

 

 

 

Composition of Mixture (mole %) Techniae Used T::p:ii:jm
NaF LiF Ber UI:4 (°C)
61.6 7.3 28.1 3.0 Filtration 520< T < 525
54.3 15.5 27.2 3.0 Filtration 476 < T <520
53.3 22.3 21.4 3.0 Filtration <494
53.3 22,3 21.4 3.0 Filtration 440 < T < 474
56.9 17.8 22.3 3.0 Fittration 475 < T <500
55.6 21.0 20.0 34 Filtration 505< T <515
55.8 21.0 20.2 3.0 Thermal analysis 502
58.7 17.0 21.3 3.0 Thermal analysis 495
51.4 23.3 22.3 3.0 Thermal analysis 530
56.7 19.3 21.0 3.0 Thermal analysis 545

 
ANP PROJECT PROGRESS REPORT

The System KF-BeF,

L. M. Bratcher R. J. Sheil
B. H. Clampitt
Materials Chemistry Division

G. D. White
Metallurgy Division

The existence of two compounds in the KF-BeF,
system was reported ;::reviously:7 namely, KzBeF4

and KBeF3.

melts

Further studies of slowly cooled
have demonstrated the existence of two
additional compounds, which are believed to have
the formulas K;BeF . and KBe,F .. It appears from
the available thermal-analysis data that K BeF,
melts congruently at about 738°C and that the
KsBeFS-KZBeF4 eutectic, which melts at 730°C,
contains approximately 27.5 mole % BeF,. Mix-
tures containing 66%3 mole % BeF, showed only
one thermal effect on cooling curves at about
325°C, but it is not clear whether this compound
The basis

is that mixtures

melts congruently or incongruently.
for the compound formulation
of this composition were homogeneous, well-
crystallized material having optical and x-ray
diffraction properties that were different from
those of other compounds in this system. The
discovery of these compounds makes it desirable
to continue the study of this system through use
of the quenching technique, particularly in the
region of the KBe,F. compound, where thermal
analysis does not give reliable data.

The System NoF-KF-BeF,

L. M. Bratcher R. J. Sheil
B. H. Clampitt

Materials Chemistry Division

G. D. White
Metallurgy Division

All four of the known KF-BeF, compounds have
been observed in slowly cooled melts in the
NaF-KF-BeF, system, in addition to the two
NaF-BeF, compounds and one well-established

 

7¢. J. Barton et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 71.

84

ternary compound, NeKBeF,. It seems certain
that one additional ternary compound exists in
this system, and possibly there are others, but
further work will be required to establish their
identity. On the basis of the results obtained
to date from petrographic and x-ray diffraction
studies of slowly cooled melts, the following
compatibility triangles are tentatively postulated:
NaF-KF-K,BeF.; NaF-K,BeF.-K,BeF,; NaF-
KzBeF4-NaKBeF4; KZBeF4-NoKBeF4-KBeF3;
KBer-NoKBeF4-KBe2F5. On the join between
KBe2F5 and NaBeF ,

pounds, the mixture containing 57.5 mole % BeF,

which are crystalline com-

solidified to a transparent glass with no notice-
able thermal effect on the cooling curve.
sitions on both sides of this mixture contained
glass and the neighboring binary compound. Work

Compo-

on this system is continuing.

Systems Containing Alkaline-Earth Fluorides

L. M. Bratcher R. J. Sheil
B. H. Clampitt
Materials Chemistry Division
G. D. White

Metallurgy Division
The alkaline-earth fluorides MgF, and CaF,

have properties, such as low cross section, low
vapor pressure, and high-temperature stability,
that cause them to appear to be attractive for
use as fuel carriers. However, consideration of
these materials in the fused-salt research program
was previously discouraged because the published
data show only high-melting-point eutectics in
alkali fluoride~MgF , and —CaF, systems. Long-
range interest in materials for producing higher
reactor temperatures than those presently being
considered makes it desirable that pertinent data
now lacking be obtained and that some of the
published data be checked. Preliminary thermal-
analysis data are presented in Table 4.3, together
with the available published values. The data
that have been obtained for the LiF-NaF-CaF,
system are sufficient to locate the eutectic compo-
sition at approximately 53-36-11 mole %,
TABLE 4.3, EUTECTIC TEMPERATURES IN
ALKALI FLUORIDE —MgF , AND
~CaF , SYSTEMS

 

Eutectic Temperature

 

 

(°C)
System
Literature ORNL
Value Value
LiF-CuF2 773 765
Nc:F-Cc:F2 810 812
N(:ll':-MgF2 830 825
LiF-NaF-Mg F2 630 620
LiF-MgFZ-Con 672
LiF-NoF-CuF2 616
Na F-Mng-COF:Z 750
LiF-NaF-MgF -CaF, 588

 

CHEMICAL REACTIONS IN MOLTEN SALTS

L. G. Overholser F. F. Blankenship
G. M. Watson
Materials Chemistry Division

R. F. Newton

Research Director’s Division

Equilibrium Reduction of FeF, by H, in NaZtF

C. M, Blood
Materials Chemistry Division

Numerous experiments have been performed, as
reported in previous reports in this series, to study
the reaction

FeF, (d) + H, (g) &=Fe (s) + 2HF (g)
with a molten mixture of NaF-ZrF , (53-47 mole %)

used as the solvent, The apparent equilibrium
constants (mole fractions of dissolved species
assumed as activities) previously determined had,
unfortunately, only qualitative significance be-
cause of the lack of an adequate method of ap-
proaching equilibrium, With the present experi-
mental assembly and technique, however, equilibrium
can be approached effectively from either the
forward or the reverse direction of the reaction.
Furthermore, equilibration can be continued under
constant conditions for periods as long as several
weeks, that is, until it is apparent that the compo-
sition of the system is constant within the limita-

PERIOD ENDING DECEMBER 10, 1955

tion of analytical procedure and inherent experi-
mental variations,

The present experimental assembly is, in general,
similar to the one described previously for use
with the equilibration method.® However, the
limited-capacity HF generator used previously
has been replaced by a practically unlimited source
of a gaseous mixture of HF and H.,. The method
being used consists in bubbling a gaseous mixture
of H, and HF of constant composition through the
melt containing FeF,, which is contained in a
mild-steel-lined nickel reactor, The equilibration
of the melt is continued until the concentration of
FeF, is constant over a period of several days,
the remaining variables being held constant.
Periodically, liquid samples are withdrawn through
sintered-nicke| filters for chemical analysis. The
compositions of the reactor influent and effluent
gas streams are determined by continuously bubbling
accurately metered portions of these gases through
standard KOH solution,

In the earlier groups of equilibration experiments,
the attainment of equilibrium was judged by the
approximate equilization of composition of influent
and effluent gas streams of the reactor, This
criterion has been found to lack sufficient sensi-
tivity for better-than-qualitative estimates. This
is particularly true if the equilibrium is approached
from the reverse direction of the reaction.

The modified source of HF consists of a 10-1b
cylinder of liquid HF placed in a thermostat in
which the temperature is controlled to approxi-
mately £0.03°C. The tank is connected to a mixing
chamber through a copper tube held at a tempera-
ture (80°C) considerably higher than the tempera-
ture of the liquid HF eylinder. The HF vapor on
its way to the mixing chamber is made to diffuse
through a sufficient number of sintered-nickel
barriers (0.0004-in. pore size) to reduce the volu-
metric flow rate to an experimentally suitable
value and to stabilize the flow rate at this value.
The streams of H, and HF are fed to the mixing
chamber, which is also held at constant tempera-
ture (340°C). The volumetric flow rate of the
hydrogen is precisely controlled by a combination
of commercially available gas-flow regulators and
is further stabilized by ditfusion through porous
barriers. The composition of the resulting gas
mixture can be varied at will by changing the H2

 

BC. M. Blood and G. M. Watson, ANP Quar, Prog. Rep.
Sept. 10, 1954, ORNL-1771, p 66.

85
ANP PROJECT PROGRESS REPORT

flow rate or the temperature of the thermostat en-
closing the HF cylinder, These compositions can
be held constant for indefinitely long periods with
an arithmetic mean deviation of approximately 14%,

The current experiment was started during the
week ending August 19, 1955, and the apparatus
has been operating continuously and without inci-
dent to the present time. Measurements have been
made at 800 and at 700°C, The iron concentration
range studied at 800°C was approximately 300 to
1300 ppm. At 700°C the range studied was approxi-
mately 3300 to 2000 ppm. Additional measurements
at 700°C are now being made in the range from
1000 to 600 ppm. A summary of the experimental
results obtained to date is given in Table 4.4,

As mentioned previously, it was found experi-
mentally that several days of equilibration time
were required before constancy of the FeF, con-
centration was attained. This was particularly
true when sizable changes in the partial pressure
of HF were imposed on the system, During the
transition periods, the changes in composition of
the effluent gas stream were rather minute, This
may be explained satisfactorily by comparing the
extremely slow rate of the reaction (in the order
of several days) with the swift passage of gas
bubbles through the melt (in the order of seconds).

For the calculations of the mele fraction of FeF
in solution, the total number of moles in the quuié
mixture was assumed to be the same as the number
of moles of constituents added (10.0 moles per
kilogram of mixture for this composition). The
partial pressures of HF were calculated from con-
centration values, with the assumption that HF
was monomeri¢ at 700 and at 800°C.

The experimentally determined equilibrium con-
stants obtained to date are not far different from
those that may be calculated from available tabu-
lations? of free energies of formation, that is,
AF® for FeF, (s) and HF (g). By comparison of
the calculated and the experimental constants,
the activity coefficients of the dissolved FeF
may also be easily obtained. These calculations
will be made when the experiment has been com-
pleted.

 

L. Brewer et al.,, pp 65, 110 in The Chemistry and
Metallurgy of Miscellaneous Materials, Thermodynamics,

ed. by L. L. Quill, McGraw-Hill, New York, 1950.

86

Reduction of UF, by Structural Metals
J. D. Redman

Materials Chemistry Division

The apparatus and methods used to measure the
reduction of UF, by metallic chromium or iron in
molten fluorides have been described in previous
reports in this series, The data obtained for these
reactions with NaF-ZrF , (50-50 mole %), NaF-LiF-
KF (11.5-46.5-42.0 mole %), NaF-ZrF , (53-47 mole
%), or Nc:F-LiF-ZrF4 (22-55-23 mole %) as reaction
media have also been reported. During the past
quarter, studies have been made on the reduction
of UF, by various structural metals, other than
chromium or iron, in molten NaF-LiF-KF (11.5-
46.5-42.0 mole %). The earlier work showed that
chromium-containing alloys are not stable at 800°C
when in contact with UF4 dissolved in the NaF-
LiF-KF eutectic, and the possibility of replacing
the chromium with some less-active metal was
therefore considered, The other structural metals
being studied are niobium, tantalum, vanadium,
and tungsten. Molybdenum, which was studied
previously, was found to be more stable than
chromium,

The results of the studies on the reduction of
UF4 by vanadium at 600 and 800°C with NaF-LiF-
KF (11.5-46.5-42.0 mole %) as the reaction medium
are given in Table 4.5, In all the experiments 2 g
of vanadium was reacted with UF:4 (15 wt %, 2.3
mole %) dissolved in approximately 20 g of the
NaF-LiF-KF mixture contained in nickel. As may
be seen from the data presented in Table 4.5, the
vanadium concentrations found in the filtrates
were extremely low. In fact, these values are the
lowest found for any metal in equilibrium with UF4
in the NaF-LLiF-KF mixture at these temperatures,
with the possible exception of nickel. These
results indicate that vanadium metal is stable
under these conditions if it is assumed that the
vanadium fluoride formed is not appreciably vola-
tile. This assumption appears to be valid in view
of the extremely low concentrations of vanadium
fluoride present and the large excess of alkali
fluorides available to complex any metal fluoride
formed.

Data for the reduction of UF, by niobium at 600
and 800°C with NaF-LiF-KF (11.5-46.5-42.0 mole %)
as the reaction medium are given in Table 4.6. In
these experiments, as in the experiments with
vanadium, 2 g of niobium was reacted with 15 wt %
PERIOD ENDING DECEMBER 10, 1955

TABLE 4.4, APPARENT EQUILIBRIUM CONSTANTS FOR REDUCTION OF FeF2
BY H, IN NaF-ZrF, (53-47 mole %)
Reaction: FeF2 (dy + H2 (g) = Fe (s) + 2HF (g)
Average total pressure: 740 mm Hg
Initial charge: 6.0 kg of NaF-ZrF4 and 30,0 g of FeF2
Container wall: mild steel (98% Fe)

 

Partial Pressure
Sampling Time

Fe in Melt of HF in

 

Y. Sy

B -

(days at constant K_*

(ppm) Effluent Gas *

temperature) 2
(atm x 104)
At 800°C

470 7 4.28 2.3

370 8 3.92 2.5

335 10 3,53 2,2

895 18 5.70 2,2

830 19 5.70 2.4

1295 21 7.74 2.9

960 25 6.00 2,3

985 26 6.27 2.5

965 27 6.17 2.4

1015 28 6.35 2.4

955 34 6.65 2.9

1005 35 6.50 2.6

1025 36 6.55 2,6

Av 2,510.2
At 700°C

2630 4 5.43 0.69
2810 5 5.55 0.66
2880 6 5.63 0.67
3070 10 5.89 0.69
3160 11 5.82 0.66
3205 12 5.87 0.66
3285 13 5.83 0.63
3232 14 5.63 0.60
2220 19 4.88 0.64
2525 20 4,99 0.60
2430 21 4.87 0.59
2190 24 4.84 0.64
2285 25 5.02 0.67
2335 26 4.81 0.61
2390 27 4.83 0.59

Av 0.64 1 0.04

 

PO

*Kx = PaF/XFerpHZ’ where X is mole fraction and P is pressure in atmosphere.

87
ANP PROJECT PROGRESS REPORT

TABLE 4,5, EQUILIBRIUM DATA FOR THE REACTION
OF UF, WITH VANADIUM IN MOLTEN NoF-LiF-KF
(11.5-46.5-42.0 mole %) AT 600 AND 800°C

 

Conditions of

Found in Filtrat
Equilibration ound n Tiitrate

 

TABLE 4.6. EQUILIBRIUM DATA FOR THE REACTION
OF UF , WITH NIOBIUM IN MOLTEN NaF-LiF-KF
(11,5-46.5-42.0 mole %) AT 600 AND 800°C

 

Conditions of
. Found in Filtrate
Equilibration

 

 

 

Temperatore  Time Total Total Total Temperature  Time Total Total Total
°C) (hr) Uranium  Vanadivm* Nickel ©C) (hr) Uranium  Niobium* Nickel
(wt %) (ppm) {ppm} (wt %) (ppm) (ppm)
600 3 10.7 30 1 4600 3 11.2 675 20
3 10.8 15 1 3 11.3 725 65
5 10.6 25 35 5 11.1 655 1
5 10.7 15 1 5 11.2 640 6
800 3 10.4 25 i 800 3 10.8 1090 35
3 10.5 35 i 3 10.5 1200 15
5 10.3 25 20 5 1.1 1540 6
5 10.6 35 20 5 11.2 1580 40
*Blank of 15 ppm of vanadium at 800°C. 12 1.0 1830
12 10.9 1470
of UF, (2.3 mole %) dissolved in approximately 12 11.0 1930
20 g of the NaF-LiF-KF mixture contained in nickel. 12 11 1310
As may be seen from the data presented in '
Table 4.6, niobium is not stable under these con- 12 11.4 1230
ditions. Equilibrium conditions are reached within 12 12.0 1360

3 hr at 600°C and after a somewhat longer period
at 800°C. The values given for the 12-hr equili-
bration periods are not so precise as desired, but
it appears to be safe to conclude that no significant
increase in niobium concentration occurs in going
from a 5- to a 12-hr reaction period,

The reaction involved is quite temperature sensi-
tive, as evidenced by the doubling of the niobium
concentration when the temperature was raised
from 600 to 800°C. This suggests that mass
transfer of niobium would be very likely to occur
in a circulating fuel based on the NaF-KF-LiF
eutectic. No further work on this metal-fuel system
is contemplated, since all the data collected indi-
cate very strongly that niobium and UF, are not
compatible in this solvent,

Studies have been made on the UF ,-tantalum
system, but no analytical data are available at
present. Preliminary results for the UF -tungsten
system suggest that tungsten probably behaves in
a manner comparable to that reported for niobium.

 

10), D, Redman and C. F. Weaver, ANP Quar. Prog.
Rep. Sept, 10, 1955, ORNL-1947, p 75.

88

 

*Blank of 315 ppm of niobium at 800°C.

Stability and Solubility of Chromium Fluorides
in Yarious Molten Fluorides

J. D. Redman
Materials Chemistry Division

The results of studies on the stability and solu-
bility of CrF,, in NaF-ZrF , (53-47 mole %) at 600
and 800°C were reported previously, 1% and a more
exhaustive study of this system has now been
made in which the CrF, was purer than it was in
the earlier studies. The effect of varying the
ZrF -to-CrF, ratio was investigated at 600 and
700°C, since it was suspected that the solid phase
present was not CrF, but, rather, CrFyZrF . I
this was the case, some variations in the zirconium-
to-sodium ratio should be observable in the fil-
trates.

The resvits of the recent experiments, which
were carried out in nickel equipment with the
proper quantity of CrF, added to approximately

—— -
40 g of the NaF-ZrF , mixture and then equilibrated
for 5 hr, are given in Table 4.7,

An examination of the data presented in Table
4,7 shows the marked effect that the change in
the zirconium-to-chromium ratio has on the solu-
bility of CrF,, particularly at 600°C. The increase
in the solubility of the CrF, as the zirconium-to-
chromium ratio decreases must be associated with
the separation of the solid phase GrF,-ZrF . If
excess CrF, is added to this system, ZrF, is
removed, and the solvent is enriched in NaF,
which, in turn, increases the solubility of CrF .,
Evidence for this postulated process is found in a
comparison of the zirconium-to-sodiumratios present
in the filtrates with the ratio of 0.88 of the charge
material, All values for the zirconium-to-sodium
ratios in the filtrates are considerably lower than
the ratio in the starting material. Unfortunately,
the chromium values found for the tests at 600°C
in which small amounts of CrF2 were added are
not precise, and therefore those tests will be rerun,

PERIOD ENDING DECEMBER 10, 1955

However, the results indicate that the zirconium-
to-chromium ratio is not critical at the higher
CrF, values. A comparison of the Cr** values
with the total chromium values shows that at least
90% of the chromium is present as Cr**, and there-
fore the disproportionation of CrF, cannot be an
important factor in this solvent,

Data relating to the stability and solubility of
CrF, in NaF-LiF-KF (11.5-46.5-42.0 mole %) at
400 and 800°C were also presented previously, 1011
Since the values were not in agreement, additional
experiments have been performed in a further attempt
to find reliable values. A purer batch of CrF, and
a different batch of solvent were used in the tests
reported in Table 4.8.

The values given in Table 4.8 for the tests at
600°C are in agreement with the values reported
for one of the eariier experiments. However, the
values found at 800°C are approximately one-third

 

”J. D. Redman and C, F. Weaver, ANP Quar. Prog.
Rep. Dec. 10, 1954, ORNL-1816, p 64.

TABLE 4,7. SOLUBILITY AND STABILITY OF CrF2 IN MOLTEN NuF-ZrF4 (53-47 mole %)

Equilibration time: 5 hr

 

Conditions of Equilibration

 

Found in Filtrate

 

+4

 

Temperature CrF2 Added Zr-to-Cr Cr Cr Zr-to-Na
(°C) (wt % Cr) Ratio* (wt %) {wt %) Ratio*
600 1.4 17 0.79 0.67 0.80

1.4 17 0.65 1.37 0.74
4.0 5.8 0.80 1.00 0.74
4,0 5.8 0.70 0.67 0.75
6.4 3.4 3.5 3.5 0.71
6.4 3.4 3.4 3.6 0.70
9.6 2.1 5.0 5.2 0.69
9.6 2.1 5.6 5.9 0.48
700 4.0 5.8 3.6 3.5 0.76
4,0 5.8 3.4 3.6 0.77
6.4 3.4 5.4 5.8 0.76
6.4 3.4 5.5 5.7 0.77
6.4** 3.4 5.7 5.8 0.76
B.4** 3.4 5.6 5.7 0.80
9.6 2.1 8.1 8.8 0.82
9.6 2,1 8.0 8.7 0.80
800 9.6 2.1 8.1 8.3 0.79
9.6 2.1 8.2 8.2 0.78

 

*Ratio calculated from mole fractions.
**Equilibration time, 12 hr.

89
ANP PROJECT PROGRESS REPORT

TABLE 4.8, SOLUBILITY AND STABILITY OF CtF 5 IN MOLTEN NoF-LiF-KF
(11.5-46.5+42,0 mole %)

 

Conditions of Equilibration

 

Found in Filtrate

 

 

Temperature Time CrF; Added Cr++ Cr Ni
(°C) (hr) (wt % Cr) (ppm) (wt %) (ppm)
600 5 2.5 1 0.19 70

5 2.5 1 0.19 65

12 2.5 145 0.24 145

12 2.5 1 0.23 145

700 5 5.0 50 0.90 50
5 5.0 55 1.1 60

12 5.0 45 1.1 35

12 5.0 50 0.8 50

800 5 5.0 475 1.1 280
5 5.0 1 1.2 110

12 5.0 ] 1.1 240

12 5.0 1 1.3 145

 

as large as those found in the two earlier experi-
ments. No values had been determined previously
at 700°C, and hence no comparison at this tem-
perature can be made. The lack of agreement be-
tween values obtained at 800°C cannot be satis-
factorily explained at this time, Variations in
temperature can hardly account for differences of
this magnitude, and it appears to be unlikely that
variations in the different batches of CrF, could
be responsible,

Reaction of UF3 with Alkali Fluorides
B. H. Clampitt C. J. Barton

Materials Chemistry Division

Attempts to measure solubilities of UF in molten
fluorides have been complicated by the presence of
variable and unexplained quantities of U4* in the
filtered specimens. While the UF3 available con-
tains some UF,, the amount of U4* found after
the tests often greatly exceeds the amount present
from this source. This tetravalent uranium ap-
parently arises from two reactions, which appear
to be independent. The reactions are the reduction
of one of the melt constituents by UF_ and the
disproportionation of UF, to UF, and uranium.
The disproportionation reaction occurs both during
equilibration and during filtiration of the melts.

Experiments made previously showed that solu-
tions of UF, in molten alkali fluorides were more

90

stable in copper than in nickel containers, and
therefore some studies of UF3 mixtures with indi-
vidual alkali fluorides were carried out in copper
apparatus, The results obtained (Table 4.9) indi-
cate that UF, is completely stable at 900°C in
molten LiF contained in copper. However, both
NaF and KF yield the corresponding alkali metal
when heated with UF . under these conditions. The
amount of UF , in the melt after the heating period
is greater than that to be expected from the equation

UF, + MF = UF, + M°

in either NaF or KF solution. However, both the
amount of alkali metal produced and the discrepancy
between the UF, and the free alkali-metal con-
tents are much larger when KF is the solvent fluo-
ride than when it is NaF,

The effect of pressure on the reaction of UF3
with KF has been examined by heating mixtures
of the NaF-LiF-KF eutectic (15 g) with UF, (6 g)
at 650°C in copper crucibles for 2 hr under helium
atmospheres at total pressures ranging from 760
to 0.04 mm Hg and analyzing the melt to determine
the U3* concentration. The data obtained are
shown in Fig. 4.6. At total pressures above
50 mm Hg the U3% content remains essentially
constant at 12.4 wt %, about 55% of the original
value. However, when the total pressure is below
50 mm Hg, the U3* concentration decreases rapidly
e - A e -

PERIOD ENDING DECEMBER 10, 1955

TABLE 4.9. REACTION OF UF3 WITH ALKALI FLUGRIDES AT 900°C IN COPPER

 

 

 

Originloll Filtrate Analysis (wt %) Alkali Metal U4% in Filtrate
Composition 3T Produced (meq)
(mole %) u Total U (meq)
LiF-UF3 (73-27) 55.5 64.1 None None
NaF-UF3 (71-29) 41.6 58.2 4,65 5.9
KF-UF3 (85-15) 3.24 34.96 15.4 30.5

 

*Calculated by difference from filtrate analysis and corrected for U4t in original

-

ORNL-LR-DWG 11528

 

S

 

 

 

M

T
fl,. s 7,7!’7 ;"777 Y
%IL _L) — L —

d | -
T T

3

 

 

[

4o e

ol
{00 200 300 400 200 600 700 800 900
HELIUM PRESSURE (mm Hg)

 

 

 

Q

RESIDUAL ud+ {wt %)} AFTER EQUILIBRATION FOR 2 hr
»
C
|
|

Fig. 4.6. Effect of Pressure on Reaction Between

UF3 and KF in NaF.LiF-KF Eutectic at 650°C,

with decreasing pressure. This behavior suggests
that at low pressures the potassium metal escapes
from the liquid system quite rapidly, and the
reaction

UF, + KF = K° + UF,

proceeds to completion. Under these circum-
stances the melts show colors ranging from red
to olive, depending on the UF3 and UF, concen-
trations. However, when potassium metal is added
to the NaF-LiF-KF eutectic, the liquid becomes
an opaque, turquoise green and shows a metallic
luster. The green liquid is probably potassium
metal saturated with the salt mixture,

U3+,

Reduction of Alkali Fluorides by Uranium Metal
C. J. Barton B. H. Clampitt

Materials Chemistry Division

An investigation of uranium metal as a reducing
agent for alkali fluorides was initiated because
the disproportionation equilibrium for UF, involves
uranium metal, In the previous report experiments
were described in which KF and NaF in the
NaoF-LiF-KF eutectic were reduced by uranium
metal at 650°C and NaF in the NaF-LiF eutectic
was similarly reduced at 570°C. The results of
additional experiments with the individual fluo-
rides are presented in Table 4.10. In each ex-
periment, 3 g of uranium was placed in 20 g of the
fluoride, and filtered samples were withdrawn for
analysis after equilibration with uranium metal,

The reaction of liquid NaF with uranium metal
was not studied because of the high melting point
of NaF. However, since there is no reaction
between LiF and uranium metal at 920°C and,
yet, the LiF-NaF eutectic does react at 750°C,
NaF must be reduced by uranium metal. The
reaction of KF and uranium metal at 900°C pro-
ceeded so rapidly that the gas exit lines were
plugged after 1 hr, and the experiment had to be
discontinued.  The potassium collected in the
exit line accounted for 60 to 70% of the uranium
found in the filtrate. Probably not all the po-
tassium was trapped in the exit line,

Experimental Preparation of Pure Fluorides

B. J. Sturm

Materials Chemistry Division

Preparation of Pure CrF3. —~ In the past, CrF,
has been prepared by using CrF3-3‘/3H20 as the

N
ANP PROJECT PROGRESS REPORT

TABLE 4.10, REACTION OF URANIUM METAL WITH LiF AND WITH KF

 

 

 

" . Analysis of Filtrate Alkali Metal
, Equilibration
Alkali _ Temperature (wt %) Recovered as
Fluoride Time Q) " Total OH™
(hr) U Tofd' U (meq)
KF 0.75 900 0.14 14.01 26.2
KF 1.00 900 0.14 13.33 31.4
LiF 2.0 920 0.0 0.0* 0.0
*Not detectable colorimetrically.
b ™
starting material, converting this to (NH )3C1'F3 NiF,‘,.ZrF4

by treatment with excess NH,HF,, and then
decomposing this intermediate at 450°C to yield
CrF,. The principal objection to this material
is the high iron content of the hydrated chromic
fluoride used as the starting material, Recently,
the preparation of pure CrF_, has been ac-
complished by hydrofluorination of sublimed an-
hydrous CrCl, at 600°C. This CrF, will be used
as the starting material for the preparation of
CrF,; hydrogen reduction of CrF, at 750°C has

been shown to be rapid and complete.

Compounds of ZrF, with CrF,, NiF,, or FeF . -
Earlier studies showed that complex fluorides of
the type MFZ-ZrF4 were formed when ZrF, was
fused at 950°C with equimolar quantities of CrF,,
NiF,, or FeF,. The compound FeF,-ZrF, slowly
decomposes when exposed to the atmosphere, but
CrF,-ZrF, decomposes so rapidly that it is im-
possible to obtain reliable x-ray or petrographic
data if this material is handled in laboratory air.
The compound NiF,.ZrF, is relatively stable
under ordinary ofmospfiweric conditions,

Mixtures corresponding to Cer-ZrF , Fer-ZrFu
and NiF,_.ZrF, were fused in nickel capsules at
950°C, and the capsules were opened in a helium-
filled dry box in an attempt to eliminate atmos-
pheric contamination, Samples for x-ray exami-
nation were sealed in a polystyrene holder by
using dry mounting tissue to seal a strip of
polystyrene film to the holder. Samples for
petrographic examination were kept in sealed
containers until they could be studied. The
following optical and x-ray data were reported
for the compounds protected in the manner de-
scribed:

92

Optical data: Cubic; rather poorly crystallized;
probably a single phase; refractive index =
1.442; pale yellow or green-yellow color

X-ray data: Probably cubic with a few ex-
traneous lines

FeF..‘,-ZrF4

Optical data: Cubic; single phase; refractive
index = 1.432; white or colorless

X-ray data: Probably cubic; isomorphous with
NiF,.ZrF, but slight shift of lines

CtF,.ZrF,
Optical data:  Single phase; probably ortho-
rhombic; average refractive index = 1.432;

biaxial positive; 2V = 20 deg; birefringence =
0.008; pale gray or blue-gray color

X-ray data: Pattern similar to that of Fer-ZrF4
with a slight shift of lines; some double lines
present that suggest an orthorhombic structure
that approaches a cubic form

PRODUCTION OF PURIFIED FLUORIDES
G. J. Nessle G. M, Watson

Materials Chemistry Division
Removal of CrF, from NaF-ZrF ,-UF, Mixtures
F. L. Daley J. Truitt

Materials Chemistry Division

It was demonstrated previously that a purified
5-1b batch of Nt:F-ZrI=4-UF4 mixture contaminated
with added CrF, could be processed to an ac-
ceptable chromium content for use as fuel-testing
material. The procedure for the chromium removal
was found to require a treatment step that included
reduction of the CrF, to chromium metal with the
use of zirconium and then removal of the chromium
metal by filtration,

Additional experiments have now been carried
out with 5-Ib batches of used fuel materials
(NoF-ZrF4-UF4) having various amounts of CrF,,
as well as oxidizing contaminants. In these
experiments the fuel was equilibrated at 800°C
for 3 hr with zirconium metal chips, filtered, and
subjected to the standard hydroflucrination-hydro-
genation treatment. In addition, two 50-lb batches
of used material were treated under similar con-
ditions.

The data from these experiments, presented in
Table 4.11, show that the history of the material
is very important. Any handling of the fuel mixture
that increases its oxide content may greatly reduce
the effectiveness of the zirconium-reduction step.
Reduction of U4" to U3* in the melt is controlled
primarily by the amount of zirconium metal added
in excess of the amount necessary to reduce the
CrF2 to chromium. On this basis, a theoretical
amount of reduction of U4* to U3t was calculated
and compared with the experimental value so that
the percentage of theoretical reduction obtained
could be computed. It is significant that the
percentage of theoretical reduction was found to
decrease sharply with increased previous handling

PERIOD ENDING DECEMBER 10, 1955

and exposure of the fuel to oxidizing conditions.
When fuel material that contained essentially no
oxidizing impurities (no previous handling after
purification) was used, over 90% of the theoretical
reduction was obtained. When used fuels that had
been protected from exposure were processed, the
observed reduction of UF ;| decreased over fivefold;
a twentyfold decrease was obtained for a used
fuel that had not been protected from air and water.

Some rather interesting trends were noted. At
a given temperature the resulting concentrations
of CrF2 are inversely proportional to the square
of the U3t/U4t ratio, within the experimental
limitations. ~ The calculated values for K _ for
the reaction

CrI:2 + 2UF3F———'\ 2UF4 + Cr
were 2 x 10=2 and 1 x 10~¢ at 800 and 650°C,

respectively., The number of experiments was
quite limited, however, and it is felt that no
special significance should be attached to these
values,

Efforts to protect the filter from plugging by
oxides and oxyfluorides included cooling the melt
to 650°C before filtration. It was hoped that these
oxycompounds would precipitate in the reactor at
the lower temperature and thus be left there when
the filtration was made. Comparisons of the

 

 

TABLE 4,11, SUMMARY OF RESULTS OF REPROCESSING EXPERIMENTS
CrF2
Filtration Concentration ae, 4t Percentage of
Source of Material Temperature (ppm) u° ' /u Theoretical Kx
°C) Reduction
Before After
5.1b Batches
Previously purified fuel + Cr F2 800 11,500 90 0.43 100 2x107°
11,500 1400 0.13 91 3 x 10>
Used fuel stored in closed 650 2,500 440 0.045 17 1x10° 8
containers -6
500 345 0.057 14 1x10
Used fuel completely exposed 800 2,000 1540 5
to atmospheric contamination
50-1% Batches
Used fuel stored in closed 800 270 190
containers
305 110

 

93
ANP PROJECT PROGRESS REPORT

quantity of filter cake obtained at 800 and at
650°C did not indicate that any advantage was
gained by lowering the temperature. From these
data it can be concluded that CrF, removal is
feasible on a small-scale basis if proper care is
exercised during handling of the fuel mixture to
prevent excessive oxide contamination. Results
of the 50-lb-batch reprocessing experiments indi-
cate that the treatment is, in general, feasible,
but the materials studied thus far have been too
low in chromium content for really definitive data
to be obtained.

Laboratory-Scale Purification Operations

F. L. Daley F. W. Miles

Materials Chemistry Division

The standard hydrofluorination-hydrogenation
purification procedure was used in the preparation
of a 1.5-kg batch of LiF-NaF (60-40 mole %). No
difficulties were encountered, and a satisfactory
product was obtained. The transfer was nearly
complete; only 24 g of the mixture remained in the
reactor,

Four batches of RbF(KF)-ZrF ,-UF, of approxi-
mately 2 kg each were prepcreé 'Ighe first two
RbF(KF)-ZrF ,-UF, mixtures had a nominal compo-
sition of 48-48-4 mole %, and the last two, 50-46-4
mole %. Since the available RbF contained about
20 mole % KF, the compositions were calculated
by using the average molecular weight of the
RbF-KF mixture.

Unusually large residues remained in the reactor
in every case. The hydrofluorination times em-
ployed were 2, 2 , 6, and 18 hr, respectively;
in each case the composmon was cooled to 450°C
before transfer. Slight thermal breaks were de-
tected at ~ 540 to 565°C in the first three experi-
ments. No thermal break was detected in this
range on the fourth experiment. All four products
froze at ~405 to 410°C. Oxyfluorides were
present in the reactor residues from the first three
experiments, despite the rather drastic hydro-
fluorination treatments,

Petrographic, x-ray, and chemical analyses
showed the compositions of the first three products
to be rather different from those of the residues.
Consequently, these products were not issued for
viscosity studies. The composition of the last
product was found to be quite close to the theo-
retical composition, and the product was therefore
issued for viscosity and phase studies.

94

Special Preparation of NaF-ZrF ,-UF ,-UF
F. L. Daley

Materials Chemistry Division

At the request of Wright Air Development Center,
a 50-lb batch of chF-Zer-UF:;-UF4 was prepared
by a procedure previously described.’?  The
nominal, final composition was NaF-ZrF -UF
(44-53-3 mole %). A chemical analysis of the

product gave the following:

In weight percentages

Found Theoretical
Na 10.9 8.7
Zr 41.8 41.7
Total U 6.24 6.2
3+
U 5.15
F 41.1 43.4
In parts per million

Fe 260

Cr 100

Ni 90

Evaluation of Raw Materials for Fuel Preparation

F. L. Daley

Materials Chemistry Division

It has been shown that rather small amounts
(~2 wt %) of oxides or oxyfluorides in the raw
materials greatly increase the time required for
purification. The presence of these impurities
at low concentration is not, to date, readily
detectable by either petrographic, x-ray, or chemical
The more reliable index of purity
appears to be the actual preparation of small
quantities (3 kg) of fuels by using the raw ma-
terials in question. Thus samples from a batch
of ZrF and from a batch of NCIZI'FS submitted by
oufs:de suppliers were evaluated by determining
their processing characteristics. The ZrF, was
found by petrographic examination to contain a
very small amount of finely divided oxide or
hydroxide, which was not found by x-ray diffraction
and chemical analysis. The 3-kg charge of
NaF-ZrF,-UF, (50-46-4 mole %) prepared with
this material by the standard procedure was normal
in freezing-point, chemical-composition, and optical
properties,

analyses.

 

I‘2(:. M. Blood et al.,

ANP Quar. Prog. Rep. June 10,
1955, ORNL-1896, p 72.
The sample of NaZrF showed a large quantity
(20 to 25%) of a brownish material that was
identified by petrographic examination as hydrated
Na,Zr,F. .. No processing difficulties were en-
countered with this but the product
contained brownish aggregates that were tenta-
tively identified as oxyfluorides. On the basis of
these the commercially obtained ZrF

material,

results,
appears to be satisfactory, while the Nt:erF5
seems to be marginal at best.

Pilot-Scale Purification Operations

C. R. Croft J. Truitt
J. P. Blakely W. T. Ward

Materials Chemistry Division

Fluoride mixtures to be used in small-scale
corrosion testing, phase equilibrium studies, and
physical-property studies are prepared in the pilot-
scale facility. During the past quarter, 68 batches
totaling approximately 800 Ib were prepared.

Production-Scale Operations

J. E. Eorgan J. P. Blakely
Materials Chemistry Division

The shortage of ZrF,, which has prevailed for
several months because of the low-capacity pro-
duction of the Y-12 Area facility, apparently no
longer exists. The first 2500-1b shipment of a
10,000-Ib order from General Chemical Company
was received early in November, and chemical
analysis showed this material to be satisfactory.
Therefore the ZrF, supply is assured for the next
six months, This material contains hafnium and
will be used only in the production of fuels for
component tests which do not involve nuclear
activity.

The vacuum pumps originally installed in the
processing facility are beginning to wear out and
have caused three major shutdowns or delays of
approximately 48 hr each. The vacuum-pump repair
shop has recommended that these pumps be re-
placed because the parts are so worn that periodic
trouble will occur. Spare parts are not available
for these Beech-Russ pumps. Therefore they are
to be replaced with Kinney pumps, for which parts
are kept in stock,

An increasing incidence of failures of reactor
cans, which are fabricated from A-nickel, has led
to a closer investigation of the cause of such

failures. These failures consist of cracks that

PERIOD ENDING DECEMBER 10, 1955

form in the walls of the reactor vessel during
processing. Some reactors withstand the proc-
essing of 12 batches before failing; others have
failed while processing a second batch. Recently,
the failures have been occurring most frequently
after four or five batches have been processed.
Originally, it was believed that sulfur attack of
the nickel, which caused embrittlement, was the
source of the trouble. The raw materials NaF,
ZrF4, (NGF)4-ZrF4, and UF4 were therefore ex-
amined, and new specifications were set for the
sulfur content. The zirconium compounds were
the worst offenders, with sulfur contents, some-
times, of over 500 ppm. The new specifications
called for a maximum sulfur content of 100 ppm.
The incidence of failures tended to increase,
however, rather than to decrease as was expected.

Recently an ASTM specification was reviewed
that recommends low-carbon nickel for high-temper-
ature usage rather than A-nickel, which normally
has a high carbon content. At high temperatures,
carbonization occurs that causes fractures in the
parent metal,  With this mind,
another examination of past failures was under-
taken. Efforts to find evidence of sulfur attack
failed, but o grayish-black crystal growth was
found in the grain boundaries of the parent nickel
where the fractures occurred; this tended to sub-
stantiate the carbonization theory. Low-carbon
nickel reactor cans have therefore been ordered.
The delivery dates being quoted vary from three
to six months.

The long-expected increase in demand for proc-
essed fluorides has failed to materialize. As a
consequence the fuel stockpile is at an all-time
high, and the supply of storage containers is
nearly exhausted. Since no actual cancellation
of orders has occurred, there is only a small
surplus., However, if usage does not increase in
the very near future, this facility will be shut
down temporarily. A total of 5500 [b of fluoride
compositions in 22 batches was processed during
the quarter.

information in

Batching and Dispensing Operations

F. A, Doss J. P. Blakely
R. G. Wiley

Materials Chemistry Division

During the quarter, 85 batches totaling approxi-
mately 3950 Ib of fluoride mixtures were dispensed

95
ANP PROJECT PROGRESS REPORT

in batch sizes ranging from 1 to 250 lb. A material
balance for the past quarter follows:

 

 

u il Amount
ateria (ib)
On hand at beginning of quarter 6,650
Produced during quarter 5,479
Total 12,129
Dispensed during the quarter 3,939
On hand at end of quarter 8,190
Special Services
F. A. Doss J. P. Blakely
N. V. Smith W. T. Ward

Materials Chemistry Division

Filling, Draining, and Sampling Operations.
Filling, draining, and sampling operations at ex-
perimental facilities were at normal levels during
the quarter. The fuel mixture was removed from
the ART high-temperature critical experiment, for
salvage, without difficulty.

In-Pile Loops. — Two in-pile loops were filled
with an enriched-uranium-bearing fluoride mixture
during the quarter. Both loops were shipped to
NRTS for testing in the MTR. A fourth in-pile
loop is to be filled in January 1956.

Testing of ART Enrichment Apparatus. — In
order to obtain more information on the behavior
of the proposed ART enricher system, additional
tests are being made of the prototype apparatus
used for the high-temperature critical experiment.
Approximately 75% of the material originally
charged into the enricher is still available in the
equipment. Graphite cups will be used to catch
specified increments from the enricher for weighing
so that a thorough calibration of the enricher over
a larger range of plunger levels can be made. A
box has been fabricated to fit over the injection
nozzle so that a protective atmosphere can be
provided to prevent excessive oxidation of the
nozzle and the fuel during these tests.

FUNDAMENTAL CHEMISTRY OF FUSED SALTS
Relative Yiscosity-Composition Studies
of NdF-LiF-Zl’F4 Mixtures at 600°C
G. M. Watson F.W. Miles
Materials Chemistry Division

Relative measurements of viscosity and density
of molten salts can be obtained with the con-

96

ventional equipment used for fuel purification.
The experimental assembly is such that changes
in chemical composition of the molten mixtures
can be effected with ease, and the viscosity and
density measurements can be made without re-
moval of the pure material from the container,
Thus it is possible to determine whether signifi-
cant changes in these physical properties occur
with changes in chemical composition.
When such changes are found, efforts can be made
to correlate the observed changes with the for-
mation or disappearance of probable ionic and
molecular species in the liquid mixtures as the
compositions are changed.

Viscosities are being measured with a Brookfield
viscometer. The only modifications required to
the standard fuel purification assembly were the
incorporation of a support for the viscometer and a
slight change in the reactor top to permit easier
introduction of the viscometer spindle, The
viscometer assembly is calibrated before each
measurement by using an identical reactor con-

small

taining a glycerine-water solution of known compo-
sition. Since the measurements are to be relative,
no correction is introduced to account for the
effect of the thermal expansion of the spindle
between the calibration temperature, 25°C, and
the melt temperature. Measurements of viscosity
on a number of mixtures of the same composition
have indicated the precision to be better than
+10%.

The first study undertaken was an investigation
of the relative viscosities of mixtures whose
compositions lie along the join, on the phase
diagram, between NaF-LiF-ZrF , (22.55-23 mole %)
and Nc:F-ZrF4 (53-47 mole %). The experimental
results, which are summarized in Table 4.12, do
not indicate marked changes in viscosity due to
composition changes.

Some simple additions have been made to the
existing experimental assembly so that density
measurements can be made. The density measure-
ments will be attempted by noting differences
in the maximum hydrostatic heads required to blow
bubbles of inert gas into the melt at two different
depths. A milling-machine cross-feed, recovered
from salvage, was permanently incorporated into
the experimental assembly for introducing nickel
tubing to the desired depth for bubbling the gas into
the melt. The screw and micrometer vernier scale
of the cross-feed were calibrated with a cathe-
tometer. The hydrostatic heads are measured with
TABLE 4.12. RELATIVE VISCOSITY vs
COMPOSITION OF NaF-LiF-ZrFA MIXTURES AT 600°C

 

Composition
Relative Viscosity,*

 

 

 

(mole %) /
NaF LiF ZiF 7
22 55 23 1.0
25 50 25 1.0
31 39 30 1.0
37 28 35 0.94
44 16 40 0.94
50 5 45 0.97
53 0 47 1.0
M = wscoslty measurement for NaF-LiF-ZrF 4 (22-55-
23 mole %) = 12.0 centipoises at 600°C, as repor'red by

S. 1. Cohen and T. N. Jones, Measurement of Viscosity o{
Compositions 81 and 82,0RNL CF-55-5-58 (May 16, 1955
PS5

a U-tube manometer that uses water as the mano-
metric fluid. A cold trap was installed between
the manometer and the reactor to protect the melt
from water-vapor contamination. Preliminary
density determinations have been made with
apparently satisfactory precision. The apparatus
will be calibrated in the near future with molten
KCI and KCI-NaCl mixtures as standards. 3

The measurement of surface tension will be
attempted by the maximum-bubble-pressure method,
It is hoped that the same apparatus used to de-
termine densities will be applicable with no
substantial changes.

EMF Measurements in Fused Salts

L. E. Topol
Materials Chemistry Division

Measurements of electromotive forces of cells
in which molten NoF-ZrF, (53-47 mole %) is used
as the solvent were conhnued in order to obtain
activities and activity coefficients of constituents
of the melts. The measurements were made in the
temperature range 550 to 700°C in a helium
atmosphere.  The melts were contained in re-
crystallized alumina (Morganite) crucibles, and the
half cells were connected by a bridge of porous
ZrQ, impregnated with the NaF-ZrF , mixture.

The cells measured were of the general type

M|MF ,(a,), NaF(a,), ZrF (a,) || NaF(a]), ZrF (a3),

PERIOD ENDING DECEMBER 10, 1955

where M and M“may be Cr, Fe, or Ni and M and M”*
may be identical.

Previously, saturated cells with identical
saturating phases, MF2 ZrF and M'F ZrF ,,
were measured, If it is cssumed that N0+ ond
F = are the sole carriers of current, the net result
of passage of two faradays of electricity is

M+ ZrF (a,) + M'F - ZrF ,(s) + 2tNaF (a,)
= MF,ZrF ,(s) + M"+ ZrF ,(a;) + 2tNaF(a7)

where ¢ is the transference number of the Na* ion,
If the amounts of added metal fluorides are ap-
proximately equal and the solubilities of the
complexes (MF ,-Z:F, and M°F ,-ZrF ) are low,
the above equation reguces to

M+ M'Fz-ZrF4(s) = MF,-ZrF (s) + M” ,

and the emf’s of these cells, together with the
free energies of formation of the pure fluorides,
indicate the relative stabilities of the solid
complexes,

In the cells reported in Tabie 4.13 the addition
of metal fluorides was low enough to eliminate
the presence of solid salt phases at high tempera-
tures (600 to 700°C), and the net cell reaction was

M + M'F2(aé) + 2tNaF(a,)
= M’ + MF ,(a;) + 2:NaF(a;)

With small additions of MF, and M‘F,, the
activities of NaF on the two sndes are neuriy

equal, and the Nernst equation for the reaction
becomes
RT = “3 RT = Y3*3
E=E°-—-—F n —=E° -~ — In ——,
n aq nF V3%

where a,a’, x,x°, and y,y’ are the activities, mole
fractions, and activity coefficients, respectively.
From Brewer's!4 estimates of free energies of
formation, E° values of 0.33, 0.20, and 0.53 v are
obtained for the Cr-Fe, Fe-Ni, and Cr-Ni cells;
these values should be nearly independent of the
temperature over this range. From these E° values

 

13E. R. Van Artsdalen and I. S. Yaffe, J. Chem. Phys.
59, 118 (1955).

14l... Brewer et al., p 107 in The Chemistry and
Metallurgy of Miscellaneous Materials, Thermodynamics,
ed, by L. L. Quill, McGraw-Hill, New York, 1950.

 

M7F jag) M,

97
ANP PROJECT PROGRESS REPORT

TABLE 4.13. POTENTIALS OF CELLS OF THE TYPE
M MF ag), NaF(ay), ZrF (a,)|INaF(a}), ZiF (a3), M F (ad) [M"

 

Measured emf< (v)

Summation of

 

 

Temperature Cr-Fe and Fe-Ni
(°C) Cr-Fe® Fe-Ni¢ Cr-Ni? Valves (v)
550 0.325 0.420 0.751 0.745
600 0.340 0.394 0.745 0.734
650 0.341 0.399 0.749 0.740
700 0.337 0.409 0.753 0.746

 

%Mean of values from two similar cells. These emf's have been corrected for thermoelectric potentials between dif-
ferent metal electrodes and leads. The corrections are 8 to 10 mv for the Fe-Ni and the Cr-Fe cells. The Cr-Ni cells
needed no correction because the Cr electrodes were attached to Ni lead wires, with the junction at the cell temperature.

bCer concentration, 1.55 mole %; F"eF2 concentration, 0.77 mole %.

cFer concentration, 0.74 mole %; NiF2 concentration, 0.21 mole %.

dCrF2 concentration, 1.40 mole %; NiF2 concentration, 0.31 mole %.

and the values in Table 4.13, it is found that
at 650°C

YerF,
= 035 ,
YFeF,
YFeF
2 ~
= 0.002 |,
)’NiF2
yCrF2
[a V]
= 000006 ’
VNin

if the convention that a = 1 for pure, solid MF2 is
used, These activity coefficients, together with
results of hydrogen equilibrium measurements
which show that the activity coefficient of FeF

in very dilute solution is nearly unity, if solié
FeF . is used as a base, indicate that the activity
coef?icien’r of NiF, is nearly 500. Since the
saturated NiF, solution at 650°C contains about
0.25 mole % NiF ,, the saturated solution apparently
has an activity of approximately unity (500 x
0.0025), if solid NiF2 is used as a base. This
result is consistent either with having the satu-
rating solid phase NiF, instead of NiF ,-ZrF ,,
at the temperature of measurement, or with a phase
diagram in which NiF,-ZrF, has an incongruent

98

melting point that is not much above the tempera-
tures used, in which case the solution saturated
with NiF_.ZrF, is nearly the same as the meta-
stable solution saturated with NiF .

Concentration cells were measured to find the
change of the activity coefficient with concen-
tration, ldentical chromium electrodes but differ-
ent CrF_ concentrations (Table 4.14) and identical
iron electrodes but different FeF2 concentrations
(Table 4.15) were used., [f, as before, it is
assumed that the Na* and F= ions carry all the
current, the cell reaction is

MF (x) + 2(NaF (x;) = MF ,(x,) + 2tNaF(x]) .

With low concentrations of MF., the activities
of NaF are nearly equal on both sides, and the
potential is given by

 

RT @3 RT Y3*3
E=—lpn—= ~-—n T .
nF a, nF Y3%3

Duplicate sets of cells were measured at inter-
vals over a period of two days., All behaved quite
similarly, and the data were, in general, quite
reproducible, with the data from the cells con-
taining iron being more closely reproducible than
those from the cells containing chromium.,

The most reproducible readings were obtained
at 550°C, but the very low values indicate that
the solubilities of CrF, and FeF, were exceeded
PERIOD ENDING DECEMBER 10, 1955

TABLE 4.14, ACTIVITY RATIOS FOR CrF2 IN NuF-ZrF“ {53-47 mole %) FROM THE CELL

Cr|CrF (x ), NaF(x,), ZrF (xp) || ZrF (x9), NaF(x), CrF(x))[Cr

 

 

 

Temperature i E (v) al/“; Y'|/Y;

(°C) Measured ldeal

For Crf:2 Mole Fraction Ratio xl/x; = 0.0099/0.0307 = 0,322
700 6.0377 0.0475 0.406 1.26
650 0.0357 0.0450 0,407 1.27
600 0.0247 0.0425 0.518 1.61
550 0.0005 0.04M 1.0*

For CerMole Fraction Ratio x]/x; =0,0098/0,0310 = 0,316
700 0.038 0.0475 0.403 1.28
650 0.0363 0.0450 0.401 1.27
600 0.0335 0.0425 0.410 1.30
550 0.0018 0.0401 *

For CrF, Mole Fraction Ratio x /x| =0.0027/0,031 = 0.871
700 0.0827 0.1023 0.139 1.60
650 0.0791 0.0970 0.134 1.56
600 0.0682 0.0917 0.163 1.87
550 0.0307 0.0865? 0.4207?* 4.827

For CrF, Mole Fraction Ratio x /%] = 0,0027/0.0276 = 0.0963
700 0.0743 0.0982 0.170 1.77
650 0.0700 0.0931 0.172 1.79
600 0.0660 0.0880 0.173 1.80
550 0.0330 0.0830 *

 

*Solubility of CrF, is less than added concentrations at this temperature.
Y 2

in one or both half cells at this temperature. Since
the saturating phase in these systems is believed
to be MF ,-ZrF,, for which no thermochemical
information is available, the data at this tempera-
ture cannot be interpreted in terms of activity
coefficients, However, the results at 550°C
indicate that at least one half cell was saturated,
and bounds can be set on the solubilities, From
the emf values in Table 4.14 it can be determined
that the solubility of CrF, is equal to or greater
than 2.8 mole % at 600°C and is between 0.27
and 0.98 moie % at 550°C. From the emf values
in Table 4.15 it is found that the solubility of
FeF ,at 550°C is less than or equal to 0.26 mole %
and at 600°C is equal to or greater than (.74
mole %. From previous work it is known that the
solubility of FeF, at 550°C is equal to or greater
than 0.23 mole % and at 600°C is less than or

equal to 0.80 mole %. At all other temperatures
the observed potentials are less than would be
expected if the solutions were ideal, and, ac-
cordingly, the activity coefficients decrease
slightly with increased concentrations of added
salt,

Optical Properties and X-Ray Patterns for
Yarious Compounds in Fluoride Systems

R. E. Thoma
Materials Chemistry Division
G. D. White
Metallurgy Division
H. Insley T. N. McVay

Consultants

The

compounds

identifying characteristics of some new
encountered in phase studies are

99
ANP PROJECT PROGRESS REPORT

TABLE 4.15. ACTIVITY RATIOS FOR FeF, IN NaF-ZrF, (53-47 mole %) FROM THE CELL
FeIFer(x.l ) NaF(xy), ZrF  (x3) || ZrF (x5 ), NaF(x, ), FeF,(x;)|Fe

 

 

 

Temperature E (v) a/a’ y /)/’
(°C) Measured Ideal 1 17
For CrF, Mole Fraction Ratio x,/x, = 0.0027/0.0074 = 0.370
700 0.0366 0.0417 0.418 1.13
650 0.0340 0.0396 0.425 1.15
600 0.0330 0.0374 0.416 1.12
550 0.0008 (0.0353) *
For CtF,, Mole Fraction Ratio x,/x; = 0.0026/0,0074 = 0.347
700 0.0363 0.0444 0.421 1.21
650 0.0350 0.0422 0.415 1.20
600 0.0335 0.0398 0.410 1.18
550 0.0019 (0.0376) .

 

*Solubility of FEF2 is less than added concentraticns at this temperature,

listed below. The symbol d(/?\) means the distance 7.70 100
between reflecting planes measured in angstroms; 6.20 3

I/I] refers to the relative intensity as compared 5.57
with an arbitrary value of 100 for the strongest 5.15 8
line, Under optical properties, N, and N_ refer 5.01 19
to the lowest and highest indices of refraction, 4.77 8
respectively, and 2V refers to the acute angie 4.65 36
between the optic axes of biaxial crystals; 0 and 4,29 42
E refer to the ordinary and extraordinary indices of 3.86 100
refraction of unaxial crystals, 3.75 13
3.52 15
2N¢:|F°3KF'SZrF4 3.40 48
3.23 6
Optical data: 3.17 60
Probably meneoclinic 3,09 33
N, = 1.478 2.87 5
N,y = 1.488 2.80 9
Biaxial negative 2,58 6
2V, very large (™ 80 deg) 2.47 5
Marked cleavage 2.39 3
Xeray data: 2.33 8
Monoclinic 2.19 3
aO = 5.6 é 2.16 8
by = 5-7 A 2.14 48

0 o

o = 913 A 2.09 8
a = 93 deg 2.06 8
1.95 27
d(A) i1 1.94 90
1.91 10
8.51 100 1.89 10

100
PERIOD ENDING DECEMBER 10, 1955

1.88 16 1.90 15
1.84 7 1.87 20
1.81 7 1.83 20
1.76 8 1.77 30
1.70 13 1.76 30
1.68 12
RbF°ZrF4
KF-2BeF
Optical data; 4
Lath-shaped crystals Optical data:
Probably monoclinic or biclinic Uniaxial negative
Na = 1.490 % 0.002 0O = 1.319
N,}, = 1.502 * 0.002 E = 1.312
Biaxial negative
2V = 75 1 10 deg X-ray data:
Xeray data: d(z) I/I]
d(A) VI, 5.99 10
3.66 15
5.87 10 3.31 60
5461 10 3.00 100
5.04 9 2.65 4
4.77 16 2.55 4
4.50 12 0.44 4
4.29 65 2.39 13
3.98 25 2.37 4
3.82 30 2.29 15
3.78 16 2.21 15
3.72 16 2.14 15
3.63 32 2.08 4
3.59 95 2.01 5
3.52 8 2.00 6
3.44 100 .83 4
3.42 100 1.78 10
3.36 100 1.60 6
3.29 100
3.05 12 3KF-BeF,
2.60 25
2.51 30 Optical data:
2.39 12 Uniaxial positive
2.35 75 O = 1.357
2.28 1 E = 1.366
2.27 12
2.26 12 X-ray data:
2.25 12 °
2.24 8 d(A) I/I]
2.23 12
2.15 45 3.42 4
2.07 12 3.30 7
2.03 12 3.07 22
1.96 30 2.98 75
1.93 30 2.81 ' 10

101
ANP PROJECT PROGRESS REPORT

102

o
d(A)

2.64
2.54
2.51
2.39
2.27
2.04
1.96
1.85
1.78
1.73
1.72
1.69
1.60
1.56
1.54

(Nl"[4)25n|=4

Optical data:

Uniaxial negative

O = 1.570
E = 1.540
Colorless

CSZZnF4

Optical data;

Biaxial positive
2V, small

N, = 1.446

N 1.458

Colorless

it

/1

30
37
15
100
70

12
12

SOy

FleF6

Optical data:
Isotopic
Average refractive index = 1.432
Red brown

NinF6

Optical data:
Isotopic
Average refractive index = 1.442

Yellow green

KzBeF4

Optical data:
Biaxial positive
2V = 30 deg
N 1.357

a
N 1.366
X A ¢

i

v

Colorless

Probably moncclinic
PERIOD ENDING DECEMBER 10, 1955

5. CORROSION RESEARCH

W. D. Manly
Metallurgy Division

W. R. Grimes

F. Kertesz

Materials Chemistry Division

Several Inconel forced-circulation loops that
were operated with fluoride mixtures and with
sodium as the circulated fluids were examined.
The effects of loop wall temperature and of vari-
ations in temperature differential on corrosion and
mass transfer were studied for loops that circu-
fated @ zirconium-base fluoride fuel mixture. The
effect on corrosion of adding UF_ to an alkali-
metal-base fluoride fuel mixture circulated in an
Inconel loop was also investigated, The effect
of the oxide content of the sodium circulated in
Inconel loops was studied through the use of cold
traps and the addition of oxides or deoxidizers.

Additional Inconel thermal-convection loops were
examined in a study of the effects of the difference
between the wall temperature and the fluid temper-
ature and of ZrH, additions to the fuel mixture.
Thermal-convection loops fabricated from special
Inconel-type alloys and various nickel-molybdenum
alloys were also examined, after having circulated
chF-Zr[:“-UF4 (50-46-4 mole %). Two other loops
were studied in which the helium cover gas
normally used was replaced with nitrogen.

Brazed Inconel T-joints were tested
zirconium-base fuel mixture in a special thermal-
convection loop and in sodium and fuel mixtures
in seesaw apparatus. Buttons of various braze
materials were tested in static sodium, in a
zirconium-base fuel mixture, and in NaOH. In
another series of static tests, Hastelloy B T-joints
brazed with various alloys were exposed to sodium
and to NaF-ZrF4-UF4 (53.5-40-6.5 mole %). Ther-
mocouple welds with both high and low contents
of Chromel-Alumel were exposed to sodium and
to NaF-ZrF ,-UF, in seesaw apparatus. Additional
data were obtained on the effect of ruthenium on
the physical properties of Inconel.

A second boiling-sodium=Inconel loop was
operated, and a third is being fabricated. Static
tests were made in order to study the decarburi-
zation of mild steel by sodium at high temperatures
and to study the corrosion of special Stellite heats
in lithium. A differential-thermal-analysis method
and a solubility-temperature apparatus were used

in a

in attempts to determine the solubility of lithium
in NaK. Several titanium carbide cermets were
subjected to screening tests in NaF-ZrF4-UF4
(53.5-40-6.5 mole %) in seesaw apparatus. Inconel
valve parts flame-plated with a mixture of tungsten
carbide and cobalt were subjected to thermal-
cycling tests, and Kentanium cermet valve parts
brazed to Inconel were exposed to a static
zirconium-base fluoride fuel mixture.

Data were obtained on the thermal decomposition
of NaOH, and corrosion and mass transfer of
various structural materials in NaOH were studied.
Chemical studies were made of the reaction of
Inconel with sodium and with NaK and of the
reaction of NaOH with nickel. A zone-melting
method was developed for establishing the eutectic
point of binary and ternary mixtures,

FORCED-CIRCULATION STUDIES

J. H. DeVan
Metallurgy Division

Fluorides in Inconel

Several Inconel
were operated with fluoride mixtures were ex-
amined. The operating conditions of these loops,
which were operated by Aircraft Reactor Engi-
neering Division personnel,! are listed in Table
5.1.

Loops 4935-5 and 4935-7 were operated to
provide information for evaluating the effects of
loop wall temperatures on corrosion and mass
transfer. The bulk-fluoride-mixture temperatures
were the same for both loops, but the fluid
velocities and the distribution of heat to the hot
legs were varied in order to vary the maximum
wall temperature in the heated sections of the
loops. The operation of both loops was interrupted
by power failures and the resultant freezing of
the fluoride mixtures. [n attempts to restart the
loops, leaks developed in the tubing, and further

forced-circulation loops that

operation was prevented,

 

]C. P. Coughlen, P. G, Smith, and R. A. Dreisbach,
ANP Quaf. PTOg. Rep. Sept. 10: 1955! ORNL"]947| P 38-

103
vol

 

 

 

TABLE 5.1, OPERATING CONDITIONS OF FUSED-SALT-INCONEL FORCED-CIRCULATION LOOPS
Loop No.
4935-5 4935-7 7425-41 4950-6 7425-1A 7425-2
Fluoride mixture circulated NuF-ZrF4-UF4a NaF-ZrFA-UF“a NaF-ZrFA-UF“a NoF-Zer-UF4a NaF-ZrF ,-UF % NaF-KF-LiF-
4 4 b
(UF4 + UF3)

Operating time, hr 681 332.5 1000 1000 1001 550.5
Preliminary operating period at 22 24 162 24 24 15.25

isothermal temperature, hr
Maximum fluoride-mixture 1500 1500 to 1510 1650 1520 1500 1500

temperature, °F
Maximum tube-wall tempera- "~ 1582 ~1710 1700 1620 1600 1570

ture, °F
Temperature gradient of 200 200 200 300 200 200

fluoride mixture, °F
Reynolds Number 9,000 to 10,000 ~ 6000 2750 ~ 8000 10,000 10,000
Yelocity, fps 7.05 4.6 2.06 5.22 6.5 4.1
Length of heated tube, ft 23 23

First section 9.08 7.92 7.92 7.92
Second section 7.92 9.08 9.08 .08
Total length of loop, ft 53 53 55.5 51 51 51
Method of heating Gas Gas Electrical Electrical Electrical Electrical
resistance resistance resistance resistance

Shape of heated section Coiled Coiled Straight Straight Straight Straight
Ratio of hot-leg surface to 2.86 2.86 7.34 7.34 7.34 7.34

loop volume, in.2/in.3
Cause of termination Power failure Power failure Scheduled Scheduled Scheduled Pump shaft

seized

Maximum depth of attack, mils 5 8 10 8 9.5 1.5

 

Composition: 50-46-4 mole %.

bComposifion: 11.2-41-45.3-2.5 mole %, with 1.07% of the total of 11.9 wt % uranium reported to be present as u3t,

LA0dTY SSIY00A¥d LDO3r0dd ANV
Although these loops did not complete the
scheduled operating period of 1000 hr, the results
of metallographic examination qualitatively demon-
strate that, for a constant bulk-fluoride-mixture
temperature, increases in wall temperature sig-
nificantly increase the amount of corrosion. As
shown in Figs., 5.1 and 5.2, the attack in loop
4935-7 after 332 hr at the higher wall temperature
(~1710°F) was to a depth of 8 mils, and the
attack after 681 hr in loop 4935-5, which operated
with a maximum wall temperature of about 1582°F,
was to a depth of 5 mils.

The examination of loop 7425-41, in which the
fluid temperature reached 1650°F, with an ac-
companying maximum wall temperature of about
1700°F, provided further evidence? that wall
temperature is a more critical variable than bulk-
fluoride-mixture temperature. The maximum attack
in this loop, which was heated by electrical
resistance, was only 10 mils after 1000 hr, that
is, only 1 to 2 mils greater than the attack found
in loops with similar temperature differentials but
with a fluid temperature of 1500°F and a wall
temperature of 1650°F.

Two loops (4950-6 and 7425-1A) of a series

being operated to evaluate variations in temper-

 

26, M. Adamson, R. 5. Crouse, and A. Taboada,
ANP Quar, Prog. Rep. Sept. 10, 1955, ORNL-1947, p 97.

UNCLASSIFIED ]
T-8111

ES;?.ng]

T

INCH

T
i

fl; 2 E'la R B

 

 

 

Fig. 5.1. Moaximum Attack in Inconel Forced-
Circulation Loop 4935-7 After Circulating NaF-
ZrF -UF , (50-46-4 mole %) for 332 hr at a Maximum
Fluid Temperature of 1510°F and a Maximum Wall
Temperature of About 1710°F. (Smggt with caption.)

PERIOD ENDING DECEMBER 10, 1955

 

UNCLASSIFIED ] |
. T-8269 ¥ Loot
! ‘
L
L x
s
z
" juo08 |
007
100w
' ‘
L
w Sl

 

FE]

 

 

T%

Fig. 5.2, Maximum Attack in Inconel Forced-
Circulation Loop 4935.-5 After Circulating NaF-
ZrF ,-UF, (50-46-4 mole %) for 681 hr at a Maximum
Fluid Temperature of 1500°F and a Maximum Well
Temperature of About 1582°F. (&@® with caption.)

ature differential were examined following the
scheduled 1000 hr of operation. In the operation
of these loops the maximum wall and fluid temper-
atures were maintained constant, while the fluid
velocity and the hot-leg temperature pattern were
varied to achieve a specified cold-zone temper-
ature and fluid-temperature drop. Since varying
the Reynolds number, or velocity, has not had
a measurable effect on comosion results, as
evidenced by compamsons of results of thermal-
convection loop tests with results of forced-
loop tests, it was felt that any
differences in the corrosion in these two loops
would be attributable directly to the difference in
temperature drop.

The temperature differentials of the two loops
were 200 and 300°F, and the maximum hot-leg
attack was very nearly the same in each, that is,
8 to 9.5 mils. It appears, however, from data on
other loops operated with a 200°F temperature
drop, that the attack in loop 7425-1A was ex-
cessive and would normally have been about
7 mils. A leak which developed in this loop
following a pump failure and subsequent reheating
for startup necessitated the replacement of a short
section of tubing between the cooler and the pump
after 22 hr of operation. This interruption may
have caused the attack to be deeper than it had
been expected to be.

convection

105
ANP PROJECT PROGRESS REPORT

Negligible attack was found in loop 7425-2,
which circulated an alkali-metal-base fluoride
mixture, NaF-KF-LiF (11.5-42-46.5 mole %), with
sufficient UF3 and UF4 added to give 11.9 wt %
uranium, The portion of the uranium that was
present as U?* was reported to be 1.07 wt %. The
loop operated only 550 hr because of a pump motor
failure, and the maximum attack was to a depth
of 1.5 mils in the hottest section, as shown in
Fig. 5.3. A continuous metal deposit, probably
of uranium, that was found in the cold leg of the

UNCLASSIFIED
T-§509

BRI

£
o
o

 

 

T
INC!:ES

T
A

i E

g

EFEE

~
o
w

:

 

 

[o

Fig. 5.3. Maximum Attack in Inconel Forced-
Circulation Loop 7425-2 After Circulating for 550 hr
the Fluoride Mixture NaF-KF-LiF-(UF4 + UF,)
(11.2-41-45.3-2.5 mole %), with 1% of the Total
Uranium Present as U3*, (dweset with caption.)

loop indicates that the amount of UF, present
was larger than that found by chemical analysis,
since no deposit was found in the thermal-
convection loop operated with material that was
reported by chemical analysis to be the same.
The chemical analyses showed no significant
increase in chromium content of the fluoride
mixture during the test, and in this respect the
analyses confirm the observed attack.

Sodium in Inconel

A series of 1000-hr tests in Inconel forced-
circulation loops has been completed in which
the oxide level in the sodium being circulated was
varied through the use of cold traps, deoxidizers,
or oxide additions. These tests concluded the
first series of tests in a study of the effect of
oxide contamination on mass transfer in sodium-
Inconel forced-circulation systems. The results
obtained for this series of tests are presented in
Table 5.2. The total weight of the deposit, which
is a convenient parameter for comparison of mass-
transfer effects, was determined by brushing all
the deposits from the loop sections after the
sodium had been removed.

The metal deposit found in loop 7426-2, which
circulated sodium to which 0.05% O, was added
in the form of Na,0O,, weighed slightly more than
that found in control2 loop 4951-8, which included
a cold trap to remove oxides. The deposit weighed
much less, however, than that found in loop
4951-5, which circulated sodium to which 0.15% c,

TABLE 5.2. MASS TRANSFER IN INCONEL FORCED-CIRCULATION LOOPS THAT CIRCULATED
SODIUM CONTAINING YARIOUS AMOUNTS OF OXYGEN

Maximum fluid temperature: 1500°F
System temperature differential: 300°F
Operating time: 1000 hr

 

 

 

Deposit Oxide Content
Loop No. Variable Maximum Thickness Weight Initial Final
(mils) (9) (%) (%)
4951-8 Cold trap in system 14 13 0.074 0.035
(control)
7426-2 0.05% 02 added to system as Na,0, 20 15 0.035 0.025
4951-5 0.15% 02 added to system as Nc1202 30 26 0.036 0.027
7426-1 1% barium added; cold trap in system 20 - 13 0.035

 

106
A e ke e =

- b e

was added. The Na,O, additions were not de-
tectable by oxide analyses of the sodium in either
loop.

The addition of barium to act as a deoxidant in
the sodium circulated in loop 7426-1 effected no
observable decrease in the weight of the mass-
transferred deposit in comparison with the weight
of the deposit found in control loop 4951-8. The
only evidence of an effect attributable to the
barium was seen in the hot leg, where metallic
layers distinctly different from the base metal
were found in conjunction with intergranular pene-
frations,

Metallographic examinations showed intergranular
attack to a maximum depth of 2 mils in the hot
legs of all loops. The attack found in loop 4951-8
is shown in Fig. 5.4. Additional corrosion in the
form of uniform surface removal possibly occurred,
but such corrosion cannot be evaluated easily.

UNCLASSIFIED
. T«8504

EE]

o
&>
e

|'

 

1

B 1 ¥

I§ iNCHES
Lt

\

u
o
o
S

 

&
<
=

PEE

e
w

 

£l
a -
o

l

 

 

 

Fig. 5.4. Intergranular Attack on Hot-Leg Surface
of Inconel Forced-Circulation Loop 4951-8, Which
Circulated Sodium at 1500°F for 1000 hr.

The operation of a sodium-Inconel loop with a
maximum fluid temperature of 1000°F has shown
the effect of temperature on mass transfer to be
quite significant., A visuval inspection of the loop,
which was operated for 1000 hr with a temperature
differential of 200°F, showed the original oxide
film on the Inconel tubing to be undisturbed. The
loop was also shown metallographically to be free
of deposited material, and there was no hot-leg
attack.

PERIOD ENDING DECEMBER 10, 1955

THERMAL-CONVECTION STUDIES

J. H. DeVan E. A. Kovacevich
Metallurgy Division

Effect of Difference Between Loop Wall
Temperature and Fluid Temperature

A special thermal-convection loop (832) was
operated for the purpose of obtaining the wall
temperature distribution in a standard loop having
a 1500°F maximum fluoride-mixture temperature
and a 250°F fluoride-mixture temperature drop.
The fluoride mixture used was NaF-ZrF ,-UF,
(50-46-4 mole %).

Eight thermocouples were attached by four dif-
ferent methods, as shown in Fig. 5.5, to the
Inconel pipe under the first heater on the hot leg
of the loop. The thermocouples were welded or
brazed to the pipe, were placed in a single well
with the tip of the thermocouple welded or peened
into place, were placed in a double well with each
wire of the thermocouple peened into place, or
each wire was placed in a hole drilled into a weld
bead. A maximum difference of 10°F was observed
among the four types of thermocouples, and, as
expected, the well, or sunken, type of attachment
gave the lower readings.

Thermocouples were also attached by the double-
well method to other sections of the hot leg, as
indicated by Fig. 5.5. An analysis of the temper-
ature readings of all the thermocouples indicates
that a difference of 160°F existed between the
maximum fluid temperature and the hottest section
of the loop. Hence, loops described as having
operated with a fluid temperature of 1500°F had
wall temperatures as high as 1660°F in the heated
Zone,

The maximum wall temperature is known to be an
important parameter in fluoride-mixture corrosion,
and it must be considered when thermal-convection
and forced-circulation loops are compared. The
forced-circulation loops have, in general, lower
wall temperatures for a given maximum fluid
temperature — a factor which, in part, accounts
for the attack observed in forced-circulation loops
being lower than that found in thermal-convection
loops with similar bulk fluid temperatures.

Effect of Zirconium Hydride Additions
to Fluoride Mixture

A series of six Inconel thermal-convection loops
were operated with the fuel mixture Nch-ZrFA-UF4

107
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 11529

 

 

 

"Bg-in. IPS

   

RISER LEG-

 

6-in. CLAMSHELL HEATERS
— 26in.

SCH-10 {C.085~in WALL)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ay Z-Z A, By Ay g A, Ag

!?ERMOCOUPL.E DISTANCE FROM TEMPERATURE

TYPE RISER LEG (in.) (°F)

A 9 1610

A2 9 1659

A3 7 1664

A4 1 1649

Ag 14 1611

Ag 18 1644

Ap 23 1614

Ag 30 1544

B, 9 1661

Ba g 1693

B3 3 1500

Ba - 1300

Bs - 1290

A B C D Bg - 1265

{ - / i// A/_/ B, 125 1625

SPECIAL THERMCCOUPLE RING Co g i653

POSITICN Z-Z TYPES OF THERMOCCUPLE ATTACHMENT D, 9 1653

Dy 9 1641

Fig. 5.5. Thermocouple Arrangement of Special The
perature Distribution.

(50-46-4 mole %) to which various amounts of ZrH
had been added. The ZrH, reduces the UF
present in the fluoride mixture to UF,, according
to the reactions:

ZrH2—~% Ir + H2

2UF, + H,—> 2UF, + 2HF

108

 

rmal-Convection Loop (832) for Studying Wall Tem-

Since the corrosive attack of Inconel systems by
fluoride mixtures is believed to be associated with
the reaction

UF, + C° = 3UF, + GF,

it was hoped that an increase in the UF, content
would shift the equilibrium reaction to the left
and thereby considerably reduce the leaching of
chromium from the Inconel and the resulting void
formation.

Chips of ZrH, were added to the fluoride mixtures
to be used in fi'le six loops in amounts calculated
to give weight percentages of 0.4, 0.5, 0.6, 0.7,
0.8, and 0.9, respectively. The resulting molten
mixtures were then allowed to pass through o
30-4 nickel filter so that particles larger than
30 1 would be removed and so that an indication
of the solubility of the added ZrH, would be
obtained.

As may be seen from the data presented in
Table 5.3, the quantity of U3* found in the fuel
mixture increased with the increases in the
quantity of ZrH, added. However, when the U3*
in the sample taken during filling was greater than
1.5%, the U3* in the sample taken after circulating
was less than 1.5% because of the disproportion-
ation of UF, to UF, and the alloying of the

PER!IOD ENDING DECEMBER 10, 1955

uranium with the walls of the loop. The total
uranium decreased during operation of the loop
in each case, and the decrease was apparently
associated with the appearance of a metallic layer
on the loop walls. The corrosive attack of the
mixtures containing ZrH, was lower in all cases
than the attack found in loop 800. The maximum
attack (0.5 mil) in loop 789 is shown in Fig. 5.6.
The observed reductions in corrosive attack are
verified by the low chromium contents of the
fluoride mixtures after circulation,

Thin metallic layers were found on both the hot
and the cold legs of the loops with more than
0.4 wt % ZrH,, except loop 790. The thickness
of the layer, identified as metallic uranium,
increased with increased ZrH., content of the
fluoride mixture. In loop 792 Ehe layer was 0.5
to 1.0 mil thick. The absence of the layer in
loop 790 is thought to be due to the low percentage
of U3*. Unfortunately the U3* content was not

TABLE 5.3, ANALYSES OF N<:|F--1'!rF4-UF4 (50+46-4 mole %) MIXTURES TO WHICH ZrH, WAS ADDED
BEFORE CIRCULATION IN STANDARD INCONEL THERMAL-CONVECTION LOOPS

Loop operating time: 500 he

Mo .~

Maximom fluid temperature: 1500°F

 

 

 

. Uranium Content Impurities
Loop  ZrH, Added ~ D°Pth of Hot-Leg (wt %) (opm)
Attack Fluid Sampled ° PP
No, (W" %) 3+
(mils) Total U U Ni Cr Fe
787 0.4 5 During filling 8.77 0.90 15 85 95
After circulating 8.52 0.96 30 110 45
788 0.5 2 During filling 9.30 0.71 3 15 50
After circulating 7.47 0.81 270 65 85
789 0.6 0.5 During filling 8.31 1.03 15 35 75
After circulating 6.62 1.23 <1 60 65
790 0.7 1.5 During filling 6.61 * 40 25 30
After circulating 6.15 1.03 15 90 85
After circulating 6.31 1.23 <1 65 90
792 0.9 1 During filling 6.72 5.35 30 35 45
After circulating 6.30 1.56 <] 40 85
800 None 10 During filling 8.64 20 105 30
{control) After circulating 8.46 40 850 90

 

*Not reported,

109
ANP PROJECT PROGRESS REPORT

ascertained before operation of the loop, but the
relatively large amount of chromium present in
the fuel after operation of the loop indicates that
the ZrH, addition was not entirely effective in

    

UNCLASSIFIED
T-8794

  

EEE]

9

T Y
INCHE

) d

B

S
o
~

l'

3

 

s B

©

lg

. * - . .
N : "
.t / » . . ' Ly

-~ s « ¥ ’
ot . /

 

 

n
 E
> ol

 

Fig. 5.6. Maximum Attack in Inconel Thermal-
Convection Loop 789 After Circulating NaF-ZiF ,-
'JF4 (50-46-4 mole %), with 0.6 wt % Zer Added,
for 500 hr ot 1500°F. (Jwef® vith caption.)

producing U3*. As may be seen, additions of
more than 0.6 wt % ZrH2 were not effective in
further reducing the corrosive attack, and they
had the deleterious effect of causing the formation

of a metallic uranium deposit.

Loops Fabricated from Special Inconel-Type
Alloys

A group of special Inconel-type alloys were
vacuum-cast by the Fabrication Group of the
Metallurgy Division and were drawn by the Superior
Tube Co. into shapes satisfactory for thermal-
convection-loop fabrication. The compositions of
these alloys, as may be seen in Table 5.4, varied
in chromium content from the nominal [nconel
composition, and molybdenum was added to one
heat. Standard-size loops were fabricated from
the special alloys, but Y.in.-dia, 0.065-in.-wall
tubing was used rather tfion the standard ¥ -in.
IPS sched-10 pipe. The results obtained from
the operation of these loops are presented in
Table 5.4. The data show considerable reduction
in depth of attack of those alloys with chromium
contents of less than 10%. Furthermore, the
addition of 10% molybdenum had little effect in

TABLE 5.4. RESULTS OF OPERATION OF THERMAL-CONVECTION LOOPS FABRICATED
FROM SPECIAL INCONEL-TYPE ALLOYS

Circulated fluid: NuF-ZrF4-UF4 (50-46-4 mole %)
Maximum fluid temperature: 1500°F

 

Loop Ne.

Alloy Composition

Operating Time Maximum Depth of Attack

 

Heat No. (wt %) (hr) (mils)

520 14 76 Ni-7 Fe~17 Cr 1000 13
(standard Inconel)

522 15 76 Ni-14 Fe-10 Cr 823* 8
523 15 76 Ni—-14 Fe=10 Cr 1000 12
524 16 76 Ni-19 Fe=5Cr 647* 3
525 16 76 Ni--19 Fe-5 Cr 460* 3
526 17 83 Ni—~7 Fe-10 Cr 1000 13
527 17 83 Ni~7 Fe~10 Cr 1000 15
521 18 74 Ni-10 Fe—6 Cr=10 Mo 1000 2
472 18 74 Ni-10 Fe—6 Cr=10 Mo 500 3.5

 

*|_eak developed in loop.

110
an alloy with 6% chromium in comparison with the
effect of reducing the chromium content to 5%.
The attack in standard Inconel loop 520 and that
in loop 521, which was fabricated from the alloy
containing 6% chromium and 10% molybdenum, are
compared in Fig. 5.7.

Nickel-Molybdenum Alloy Loops

A series of nickel-molybdenum alloys were
prepared by the Fabrication Group to provide
material for a study of the corrosion properties
of nickel-molybdenum alloys having compositions
modified from the composition of commercial
Hastelloy B. Thermal-convection loops fabricated
from materials with the compositions shown in
Table 5.5 were operated for 1000 hr at 1500°F
with NaF-ZrF4-UF (50-46-4 mole %) as the circu-
lated fluid. The deep attack found in loop 1008
was undoubtedly caused by the air leak which
interrupted the test, since loop 1013, which was
identical, showed attack to a depth of only 0.5 mil
after operating for 831 hr. The results obtained
with the special alloys and, for comparison
purposes, with Hastelloy B and Inconel are shown
in Table 5.5. Varying the molybdenum and
chromium contents in a nickel-molybdenum alloy
does not appear to have much effect on corrosion
of the alloy in NaF-ZrF4-UF4 (50-46-4 mole %)
at 1500°F. The attack found in loop 1009, which

PERIOD ENDING DECEMBER 10, 1955

was fabricated from an 85% Ni-15% Mo alloy
is shown in Fig. 5.8.

Effect of Nitrogen Atmosphere

Two Inconel thermal-convection loops, 801 and
802, were tested with NaF-ZrF -UF, (50-46-4
mole %) under a nitrogen atmosphere to determine
the possibility of substituting this cover gas for
the currently used helium. After operation for
500 hr at 1500°F, both loops showed light surface
attack to a depth of 1 mil in all sections, with
subsurface void formation to a depth of 10 to
13 mils in the hottest section. Slight evidence of
surface-layer formation was found in the hot zone,
but no deposit or layer was seen in the cold leg.
The maximum attack in these loops slightly
exceeded that for a standard Inconel loop with
helium used as the cover gas; the attack in the
standard loop is normally less than 11 mils.

The fluoride mixture was removed from loop 801
by melting, and a gold-colored deposit, which was
believed to be CrN or Cer, was observed on the
walls of the trap. X-ray diffraction analysis of
this deposit failed, however, to show an identi-
fiable pattern other than that of the fuel mixture.
The following metallic elements were determined
to be present, spectrographically: 2 wt % Cr,
1 wt % Ni, 0.5 wt % Fe, 5.0 wt % Ti, 0.2 wt % Al

The fluoride mixture circulated in loop 802 was

 

 

Fig. 5.7. Comparison of Attack Found in (a) Standard Inconel Loop 520 with That Found in (b) Loop
521 Fabricated from a Special Inconel-Type Alloy Containing 6% Chromium and 10% Molybdenum After
Circulating NaF-ZrF ,-UF , (50-46-4 mole %) at 1500°F for 1000 hr. (Swwwe®t with caption.)

111
ANP PROJECT PROGRESS REPORT

TABLE 5,5. RESULTS OF OPERATION OF NICKEL-MOLYBDENUM ALLOY THERMAL-CONVECTION
LOOPS COMPARED WITH RESULTS FROM HASTELLOY B AND INCONEL LOOPS

Circulated fluid: NaF-ZrF4-UF4 {50-46-4 mole %)

Maximum fluid temperature:

1500°F

 

 

 

Operating Maximum Depth
Alloy Composition )
L.oop No. Heat No. (wh %) Time of Attack
(hr) {mils)
181 Hastelloy B 29 Mo—-66 Ni—5 Fe 1000 1
1004 DP1-30 24 Mo-76 Ni 791* 0
1006 -37 3 Cr-20 Mo—77 Ni 1000 1
1007 .33 15 Mo-85 Ni 1000 1
1008 -40 5 Cr-20 Mo-75 Ni 240* 10
1009 -33 15 Mo—-85 Ni 1000 1
1010 Inconel 76 Ni=~17 Cr-7 Fe 1000 12
1013 DP1-40 5 Cr-20 Mo=75 Ni 831~ 6.5
1014 -30 24 Mo—76 Ni 1000 2
1015 -30 24 Mo—76 Ni 1000 0.5
*Terminated because of a leak.
e GENERAL CORROSION STUDIES
{001
| E. E. Hoffman
vos W. H. Cook D. H. Jansen
14 Metallurgy Division
L= R. Carlander
o] Pratt & Whitney Aircraft
: :2: Thermal-Convection Loop Tests of Brazed
Inconel T-Joints in NaF-ZrF ,-UF
ST 4 4
Lo D. H. Jansen
L e Metallurgy Division
- A thermal-convection loop fabricated from ]/2-in.
i sched-40 Inconel tubing has been operated with

 

 

 

Fig. 5.8, Attack Found in Loop 1009, Which Was
Fabricated from an 85% Niw15% Mo Alloy, After
Circulating NuF-ZrF‘-UF‘4 (50-46-4 mole %) for
1000 hr at 1500°F.

removed mechanically from the trap rather than
being melted out. This trap also showed a deposit
of the type found ih the trap of loop 801, but
identification of this deposit has not yet been
completed. .k

112

NaF-ZrF ,-UF, (50-46-4 mole %) as the circulated
fluid, During fabrication of the loop, six Inconel
segments were brazed into the hot-leg section
with Coast Metals No. 52 (89% Ni—-5% Si—4%
B-2% Fe) brazing alloy. An Inconel sleeve was
welded around this section, as shown in Fig. 5.9.
Two typical Inconel segments are shown in Fig.
5.10. The duration of the test was 1000 hr, and
the hot and cold legs were maintained at 1500
and 1100°F, respectively.

The inner walls of the segments and two samples
from each brazed joint were examined metallo-
UNCLASSIFIED
ORNL-LR-DWG 9715

. INCONEL JACKET WELDED AROUND
BRAZED INSERT

 
  

=

s

 

 

 

 

BRAZED JOINTS

 

e e e e e e e

L=

 

 

 

ol

?DIF\’ECTION OF FLOW

Fig. 5.9. Diagram of Thermal-Convection Loop
with Brazed Insert in Hot-Leg Section,

HHCLASSIFIED
Y-1690%

 

Fig. 5.10. Inconel Segments of the Type Brazed
into the Hot-Leg Section of an Inconel Thermal-
Convection Loop.

PERIOD ENDING DECEMBER 10, 1955

graphically, after the test, for evidence of attack.
The results of the examinations are summarized
in Table 5.6. No evidence of mass transfer was
found in the loop. However, the brazed joints
were porous and shrinkage voids were present, as
shown in Fig. 5.11. The area where the braze
material was applied to the outer walls of two
Inconel segments is shown in Fig. 5.12, and
two brazed joints on the inner wall, which was
exposed to the fuel mixture, are shown in Fig.

3.13.

TABLE 5.6. RESULTS OF OPERATION OF A
THERMAL-CONVECTION LOOP WITH BRAZED
INCONEL SEGMENTS IN HOT-LEG SECTION

Circulated fluid: NuF-ZrFA-UF4 {50-46-4 mole %)
Hot-leg temperature: 1500°F

Cold-leg temperature: 1100°F

Duration of test: 1000 hr

 

Attack on Braze

 

 

Brazed (mils) Attack on Wall
Joint (mils)
Sample 1 Sample 2

1 1.4 1 4

2 3 2 3

3 0.6 2 4

4 1.3 3.5 3

5 9 2 4

 

Seesaw Tests of Brazed Inconel T-Joints
in Sodium and in Fuel Mixtures

D. H. Jansen
Metallurgy Division

The seesaw test apparatus has been used for a
series of tests in sodium and in fuel mixtures of
Incone! T-joints brazed with the alloys listed
in Table 5.7. The duration of each test was
100 hr, and the hot- and cold-zone temperatures
were 1500 and 1100°F, respectively.

Two T-joints brazed with the 80% Ni-10%
Cr-10% P alloy were placed in the same tube.
One of these, joint No. 1, was cut from a T-bar
on hand, and the other, joint No. 2, was a new
T-joint. The difference in attack on these two
joints, as shown in Figs. 5.14 and 5.15, is be-
lieved to have resulted from unknown variations
in the phosphorus content of the brazing alloy.

113
ANP PROJECT PROGRESS REPORT

 

Fig. 5.11. Brazed Joints from Inconel Insert in Hot-Leg Section of a Thermal-Convection Loop Showing
Porosity of Braze Material, Etched with 10% oxaiic acid. 7X.

Ni PLATE]

INCHES

.03

T
00X
i

 

 

 

Fig. 5.12. Outer Walls of Two Inconel Segments Showing Area Where Braze Material Was Applied.
Etched with 10% oxalic acid. 100X. Reduced 13%.

114
; PERIOD ENDING DECEMBER 10, 1955

 

e . A

 

. Fig. 5.13. Inner Surface of Inconel Insert Showing Portions of Two Brazed Joints After Exposure to
Flowing NaF.ZrF .UF , (50-46-4 mole %) ot 1500°F for 1000 hr. Etched with 10% oxalic acid, (et
with caption.) 100X, Reduced 13%.

UNCLASSIFIED
Y-T6389 -

 

 

 

! Fig. 5.14. Inconel T-Joint No. 1 Brazed with an 80% Ni-10% Cr~10% P Alloy After Exposure to Sodium
for 100 hr in Seesaw Apparatus at a Hot-Zone Temperature of 1500°F. Attack to a depth of 11 mils may
be seen. Etched with 10% oxalic acid. 150X. Reduced 12%.

115
ANP PROJECT PROGRESS REPORT

 
    

BLTTTEITTTEITTT BT e

£
ox
1

 

15

4

 

!

Fig. 5.15. Inconel T-Joint No. 2 Brazed with an 80% Ni~10% Cr-10% P Alloy After Exposure to Sodium
for-100 hr in Seesaw Apparatus at a Hot-Zone Temperature of 1500°F. No attack may be seen, Etched
with 10% oxalic acid. 150X. Reduced 12%.

UNCLASSIFIED UNCLASSIFIED
Y-16820 N Y-14819

 

 

 

 

Fig, 5.16. Inconel T-Joints Brazed with a 75% Ni-25% Ge Alloy After Exposure to (a) Sodium and to
(b) NaF-ZrF ,-UF , (50-46-4 mole %) for 100 hr in Seesaw Apparatus at a Hot-Zone Temperature of 1500°F.
Note attack to a depth of 2 mils on both joints. The dark network in the grain boundaries is a fine,
lamellar type of structure. (Secret with caption.) 200X. Reduced 19%. '

116
A e

- e e

b mime oo am =

This alloy will be tested again in sodium, with
the phosphorus content closely controlled. The
75% Ni-25% Ge alloy has fair resistance to both
the sodium and the fuel mixture, as shown in
Fig. 5.16. The results of the tests are presented
in Table 5.7, in which the brazing alloys are
listed in order of decreasing corrosion resistance,
as determined by the following arbitrary scale:

Resistonce Depth of Attack (mils)
Good Otol
Fair 1to3
Poor 3toé
Bad Over 6

PERIOD ENDING DECEMBER 10, 1955

Static Tests of Braze Materials

D. H. Jansen
Metallurgy Division

Buttons of six different braze materials were
corrosion tested in static sodium and in static NaF-
ZI‘F“-UF4 (50-46-4 mole %), and two palladium-
rich buttons were tested in NaOH and in the
fluoride mixture. The buttons, which were polished
on one side, were tested in the as-received con-
dition, After the tests, the buttons were cut in
a line perpendicular to the polished face in order
to ascertain the depth of attack.

Coast Metals alloy No. 52 (89% Ni-4% B-5%
Si—-2% Fe) and General Electric alloy No, 81

TABLE 5.7. RESULTS OF SEESAW TESTS OF BRAZED INCONEL T-JOINTS
IN LIQUID SODIUM AND IN FUSED SALTS

Duration of test:

100 hr

Hot-zone temperature: 1500°F

Cold-zone temperature: 1100°F

 

 

- II * .
Corrosive Medium Brazing Alloy Weight Change Cor'rOsadn Metallographic Notes
(wt %) (%) Resistance
Sodium 80 Ni-10 Cr-10 P ~0.174 Good No attack
(jeint No. 2)
75 Ni—25 Ge —0.052 Fair Attack to a depth of 2 mils
50 Ni-25 Ge-25 Mo —0.252 Poor Nonuniform attack to a depth of
3'2 mi|5
80 Ni—10 Cr-10 P -0.41 Poor Attack to a depth of 11 mils
(joint No. 1)
N"'F‘Zl’F“'UF"4 80 Ni-10 Cr-10 P -0.102 Good No attack; fillet cracked
(50-46-4 mole %)
NUF-ZrF4-UF4 50 Ni-25 Ge—-25 Mo 0.0 Good Erratic attack to a depth of
(53.5-40-6.5 mole %) 1 mil
l"~!c:F-ZrF4-UF4 75 Ni=25 Ge ~0.060 Fair Attack to a depth of 2 mils
(50-46-4 mole %)
Nt::F"ZrF"-UF4 82 Au-18 Ni +0.16 Poor Attack to a depth of 4 mils
(53-5'40'605 mole %)
Nt:IF--ZrF“-UF4 80 Au-20 Cu —-0.072 Poor Nonuniform attack to a maximum

(50-46+4 mole %)

depth of 4.5 mils

 

117
ANP PROJECT PROGRESS REPORT

(66% Ni—-19% Cr—-10% Si—4% Fe—1% Mn) showed,
respectively, good and fair corrosion resistance in
sodium, as shown in Figs. 5.17 and 5.18. The
two 93% Pd—7% Al buttons showed poor resistance
to the fluoride mixture and very poor resistance
to the NaOH. A Coast Metals alloy No. 52 button
tested in the fluoride mixture showed a very deep
attack in only one area; the remainder of the button
area exhibited attack to a depth of 0.5 mil. This

erratic attack was probably an isolated case,
since Coast Metals alloy No. 52 has exhibited
good corrosion resistance to fluorides when used
on nickel and Inconel T-joints. Buttons of the
60% Pd—-40% Ni alloy showed good corrosion
resistance to both the fluoride mixture and the
NaOH. The brazing materials tested are described
and listed in order of decreasing corrosion re-
sistance in Table 5.8.

TABLE 5.8. RESULTS OF TESTS OF BUTTONS OF VARIOUS BRAZE MATERIALS IN STATIC SODIUM,
NaOH, AND NuF-ZrF“-UF4 (50-46+4 mole %) FOR 100 hr AT 1500°F

 

Alloy Composition

Button Weight

Corrosion

 

Corrosive Medium (wt %) Change Resistance Metallographic Notes
(%)
NQF'ZrF4'UF4 60 Pd—40 Ni —-0.043 Good No attack
(50-46-4 mole %) g3 py_7 Al +0.16 Poor Uniform attack to a depth
of 3.2 mils
80 Ni—-10 Cr-10 P ~0.26 P oor Uniform attack to a depth
of 3.8 mils
66 Ni-19 Cr=10 Si—-4 Fe-1 Mn -0.37 Poor Uniform attack to a depth
(General Electric No. 81) of 4.4 mils
55 Mn-35 Ni-10 Cr ~7.3 Poor Stringer type of attack to
a depth of 9.4 mils
60 Mn—40 Ni ~8.4 Poor Stringer type of attack to
a depth of 21 mils
89 Ni-5 5i—-4 B-2 Fe -0.104 Poor Very nonuniform attack to
(Coast Metals No. 52) a maximum depth of 26
mils in one isclated area
65 Ni—=25 Ge-=10 Cr -0,22 Poor Very nonuniform attack to
a moximum depth of
37.7 mils
NaOH 6C Pd—40 Ni +1.18 Good No attack
93 Pd-7 Al Poor Button partially disselved
Sodium 89 Ni-5 Si—4 B-2 Fe -0.22 Good No attack
(Coast Metals No. 52)
66 Ni—19 Cr—10 Si—-4 Fe—1 Mn -0.047 Fair Attack to a depth of
(General Electric No. 81) 1.4 mils
65 Ni—25 Ge-10 Cr -0.083 Fair Attack to a depth of
2 mils
80 Ni-=10 Cr-10 P -0.3 Fair Attack to a depth of
3'2 mils
55 Mn=35 Ni-10 Cr -1.6 Poor Stringer type of attack to

a depth of 4 to 5 mils

 

118
—— -

PERIOD ENDING DECEMBER 10, 1955

UNCLASSIFIED
CUY16616

7 : S

Wf -
d P
w5 T T u T
i < . ‘ _}
e SERNT e

1%

1

  

 
 
  
  

    
 

   
  

   
     
 

Foi i |
Py ¥ it o'
/) m?ffl“%: )

‘ . ‘ # Pt f‘-; X ,,-x u . - 7 : :

   

 

    
 
   

   
      
  
   

T
INCHRES
1

B

;004

 

.008

 

Fig. 5.17. Coast Metals Alloy No. 52 (89% Ni=5% Si~4% B=2% Fe) After Exposure to Static Sodium
for 100 hr at 1500°F. No attack can be seen, but the second phase has been leached out at the surface.

Etched with 10% oxalic acid. 500X. Reduced 10.5%.

UNCLASSIFIED

Y:14536

 

Q

EEE

T
INCHES
A

T
i

1]
o
~

EEEEEE

 

 

Fig. 5.18. General Electric Alloy No. 81 (66% Ni=19% Cr10% Si=4% Fe=1% Mn) After Exposure to
Static Sodium for 100 hr at 1500°F. Attack to a depth of 1.4 mils may be seen. Etched with 10% oxalic

acid. 250X. Reduced 10.5%.

119
ANP PROJECT PROGRESS REPORT

Static Tests of Brazed Hastelloy B T-Joints
in Sodium and in NGF-Z!‘F4-UF4

D. H. Jansen
Metallurgy Division

A series of Hastelloy B T-joints brazed with
various alloys were tested in sodium and in
N<:|!=-Zrl:4-Ul:4 (53.5-40-6.5 mole %) at 1500°F for
100 hr. The results of these tests are summarized
in Table 5.9. It may be noted that only the
80% Ni—5% Cr—5% Fe-5% Si—5% B alloy has fair,
or better, resistance in both the sodium and the
fluoride mixture.

Seesaw Tests of Chromel-Alumel Thermocouple
Joins to Inconel Thermocouple Wells

D. H. Jansen
Metallurgy Division

A number of Chromel-Alumel thermocouple as-
semblies fabricated from 0.125-in.-OD Inconel

tubing and 0.020-in. Chromel-Alumel wires were
exposed to sodium and to NCIF'ZI‘F4'UF4 (50-46-4
mole %) in seesaw apparatus. The Chromel-Alumel
content of the weld nuggets of the wire-to-tubing
joins was varied to determine whether the quantity
of silicon, manganese, and aluminum in the nugget
would affect the corrosion rate. Portions of the
thermocouple wires were melted to form the weld
nuggets with high Chromel-Alumel content, whereas
the Inconel tubing around the wires was melted
to form the nuggets with low Chromel-Alumel
content.  The results of the corrosion tests
conducted on these thermocouple assemblies are
summarized in Table 5.10 and illustrated in Figs.
5.19 and 5.20. The Inconel tubes on which the
welds of high Chromel-Alumel content were made
were more heavily attacked in the nonweld areas
than were the tubes with the welds of low Chromel-
Alumel content. All the attack measurements were
made before the specimens were etched.

TABLE 5.9. RESULTS OF TESTS OF BRAZED HASTELLOQY B T-JOINTS EXPOSED TO STATIC SODIUM
AND TO STATIC Nt‘:F-ZrF“»UF4 (53.5-40-6.5 mole %) FOR 100 hr AT 1500°F

 

Brazing Alloy
(wt %)

Corrosive Medium

Weight Change

Metallographic Notes
(%) grap

Resistance

 

NaF-ZrF4-UF4 80 Ni=5 Cr--5 Fe=5S5i-5B

(53.5-40-6.5 mole %)

69 Ni-15 Cr-5 B.5 Si—5 Fe-1 C

69 Ni-20 Cr—11 Si

90 Ni-4 B—4 Si-2 Fe

69 Ni—-20 Cr-11 Si

Sodium

69 Ni-15 Cr-5 B=5 Si-5 Fe-1 C

80 Ni=5 Cr-5 Fe-55S5i-5B

90 Ni—4 B—4 S5i-2 Fe

-0.035 Good

No attack along surface
of fillet

+0.041 Fair No surface attack, but
several subsurface
voids to a depth of

4 mils

-0.025 Poor

Uniform surface attack
to a depth of 4 mils

-0.014 Poor

Layer of small voids 5
mils in from surface
-0.049 Good Attack along fillet sur-
face to a depth of

0.5 mil

0 Fair No surface attack, but
several subsurface
voids to a depth of

6 mils

~0.041 Fair Layer of subsurface
voids to a depth of

1 mil
~(.052 Poor

Layer of small voids 6

mils in from surface

 

120
PERIOD ENDING DECEMBER 10, 1955

TABLE 5.10. RESULTS OF SEESAW TESTS OF THERMOCOUPLE ASSEMBLIES EXPOSED TO SODIUM

AND TO NaF-ZrF ,.UF , (50.46-4 mole %) FOR 100 hr AT 1500°F

 

Corrosive Medium

Type of Thermocouple Weld

Attack {mils)

 

 

NaF-ZrFA-UF4
(50-46-4 mole %)

Sodium

High Chromel-Alumel content, ground flat

High Chromel-Alumel content, as welded

Low Chromel-Alumel content, as welded

l.ow Chromel-Alumel content, as welded

inconel Tube Weld
4.5 ' 0.5
4.5 <0.5
2.0 None
None None

 

e

0COUP

UNCLASSIFIED
Y+17029

 

INCHES

 

 

UNCLASSIFIED
o Ye17027

Fig. 5.19. Thermocouple Assemblies with Welds of Low Chromel-Alumel Content After Exposure for
100 hr at 1500°F in Seesaw Apparatus to (a) Sodium and (b) NuF-ZrF4-UF4 (50-46-4 mole %). Inconel
tube unattacked in sodium; attacked to a depth of 2 mils in fluoride mixture; no attack apparent on weld.

Etched with 10% oxalic acid. (@@= with caption.)

121
ANP PROJECT PROGRESS REPORT

[Ni PLATE ]

 

~ THERMOCOUPLE

Figo 5-200

 

 

 

UNCLASSIFIED
Y-17024
o
“m-—l
I
>
Z
L i
INCONEL TUBE
0.02
0.03
»
L O
Q

 

 

Thermocouple Assembly with Welds of High Chromel-Alumel Content After Exposure for

100 hr at 1500°F in Seesaw Apparatus to NaF-ZrF4-UF4 (50-46-4 mole %). Inconel tube attacked to a

depth of 4.5 mils; weld attacked to a depth of 0.5 mils.

caption.)

Effect of Ruthenium on Physical Properties
of Inconel

D. H. Jansen
Metallurgy Division

Additional data were obtained on the creep-
rupture properties of ruthenium-plated Inconel. An
additional nonplated specimen was tested because
it was found that the nonplated specimen tested
previously3 had not received the same prior heat
treatment as that given to the plated specimens.
The results obtained are presented in Table 5.11.
The plated specimens were tested as received and
without a spectrographic check for the positive
presence of a ruthenium plate. Therefore ad-
ditional tests of spectrographically checked speci-
mens are under way.

 

73C. F. Leitten, Jr., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 106.

122

Etched with 10% oxalic acid. (SEPemth

Boiling Sodium in Inconel Loop

E. E. Hoffman
Metallurgy Division

The results of operation of the first boiling-
sodium—Incone! loop, which was terminated after
400 hr because of a pipe failure, were previously
reported,  Another such loop has now been
operated for the scheduled test period of 1000 hr
under similar thermal conditions. A slight vacuum
was maintained in the system so that the sodium
would boil tt approximately 1500°F. In the
previous test of 400-hr duration no mass-transferred
crystals were detected in the cold trap of the
condenser lina,. but heavy intergranular cracking
to a depth of 50 mils was detected in some areas.
In this latest test, macroscopically visible quan-

 

4. E. Hoffman, ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 109.
PERIOD ENDING DECEMBER 10, 1955

TABLE 5.11. RESULTS OF CREEP-RUPTURE TESTS OF NONPLATED AND RUTHENIUM.PLATED INCONEL

Atmosphere: purified argon

Stress: 3500 psi

Test temperature: 1500°F
Heat treatment prior to test: 100 hr at 1500°F

 

Condition of Specimen

Time to Rupture (hr)

Final Elongation (%)

 

Nonplated
Plated

Plated

746 14.0
873 13.4
728 13.6

 

tities of mass-transferred crystals were detected
in the cold trap, as shown in Figs, 5.21 and 5.22.

The various sections from the condenser line
that were examined metallographically exhibited
intergranular cracking similar to that found in the
loop operated for 400 hr. In the hottest section
of the condenser pipe, there was attack to a depth
of 1 to 2 mils in the form of small subsurface
voids. No intergranular cracking was detected in
the hottest and coolest sections of the condenser
pipe, but cracks were found in the straight section
which connected the hot and cold traps. The
cracks were to a maximum depth of approximately
25 mils (see Fig. 5.23), as compared with a depth
of 50 mils in the loop operated for 400 hr. The
temperatures in the areas where cracks were
detected varied from 1150 to 1325°F.

Metallographic examination indicated that a
brittle grain-boundary phase was present in the
areas where cracks appeared. Periodic thermal
excursions caused by the condensing sodium led
to severe thermal stresses in the pipe walls,
which apparently caused the cracks to propagate.
A boiling-sodium loop fabricated from type 348
stainless steel is now in operation. With this
test system it will be possible to sample
periodically the distilled sodium and to check
its oxygen content.

Decarburization of Mild S$teel by Sodium

E. E. Hoffman
Metallurgy Division

It has been well established that various metals
may be either carburized or decarburized while in
contact with sodium at elevated temperatures. The
direction and extent of this phenomenon depend
both on the original carbon content of the sodium

UNCLASSIFIED
Y-16593

 
    

HOT TRAP A
750°C (1382°F)

COLD TRAP
560°C (1040°F)

 

Fig. 5.21. Sections Taken from an Inconel-
BoilingSodium Loop After Operation for 1000 hr.

and on the carbon content of the metal in contact
with the sodium, Tests have been made to
determine how far the decarburization of a type
1043 mild steel (0.433% C) will proceed in static
sodium at 1830°F for 100 hr. In the two tests
performed to date, the container materials were
Armco iron and type 304 ELC stainless steel.
The capsules were loaded and sealed in inert
atmospheres. The mild-steel specimens in both
capsules showed weight losses of 0.2 mg/in.2.
The extent to which carbon transferred is illus-
trated in Fig. 5.24. The decarburization of the
type 1043 mild-steel specimen is particularly
evident, The small areas of pearlite found in the

123
ANP PROJECT PROGRESS REPORT

 

Fig. 5.22. Enlargement of Section of Cold Trap from the Inconel-Boiling-Sodium Loop Showing Mass-

Transferred Crystals. 4X.

wall of the Armco iron capsule after the test
indicate that the Armco iron picked up carben,
The type 1043 mild-steel specimen tested in the
type 304 ELC stainless steel capsule was very
similar in appearance to the mild-steel specimen
tested in the Armco iron capsule; however, suf-
ficient nickel had been picked up on the surface
of the specimen in the stainless steel capsule
to cause a phase transformation which extended
to a depth of 1 to 2 mils. Table 5.12 shows the
results of the carbon analyses of the various test
components, } may be seen that the carbon
content of the mild-steel specimens decreased
and that the carbon content of the capsule walls
in contact with the sodium increased.

Static Tests of Special Stellite Heats in Lithium

E. E. Hoffman
Metallurgy Division

Two special heats of Stellite, heats B and C,
were submitted by the Union Carbide and Carbon

124

Metals Research Laboratories for corrosion tests
in various liquid metals. The results of earlier
tests of these materials in NaF-ZrF -UF, (53.5-
40-6.5 mole %) and in sodium were reported previ-
ously.” The approximate compositions of these
alloys were:

Composition (wt %)

 

Heat B Heat C
Carbon 1 0.30
Chromium 35 28
Tungsten 5
Molybdenum 12
Manganese 0.50 0.50
Silicon 0.50 0.50
Nickel 3
Cobalt 58 55.7

 

SE. E. Hoffman et al, ANP Quar, Prog. Rep. Sept.
10, 1954, ORNL-1771, p 83.
[TILEl

7] 020
2
iy —

223 |

T
IS(?X

 

 

TITET

 

Fig. 5.23. Wall of Condenser Pipe of Inconel.
Boiling-Sodium Loop Operated for 1000 hr. Tem-
perature at location shown was between 1150 and

1325°F.

PERIOD ENDING DECEMBER 10, 1955

The alloys were unattacked by sodium but were
attacked to depths of from 7 to 20 mils by the
fused salt. The results of the tests of these
materials in lithium are presented in Table 5.13.
It is apparent that in these alloys there are
isolated areas that are susceptible to very heavy
attack. The variations in susceptibility to attack
may be due to composition variations in the
specimens,

Solubility of Lithium in NaK

R. Carlander
Pratt & Whitney Aircraft

Attempts were made to determine the solubility
of lithium in NaK (44% K—-56% Na) by a differential
thermal-analysis method and by using a solubility-
temperature apparatus. In the differential thermal-
analysis method, the temperature of the bath
containing lithium was plotted on an automatic
time-temperature recorder, and the difference in
temperature between the NaK bath containing
lithium and a blank bath containing only NaK was
plotted on another automatic time-temperature
recorder., When the lithium began to solidify, heat
was given off, and a temperature difference was
noted and correlated with the time-temperature
chart to give the temperature at which solidifi-
cation occurred. Tests were performed with 5,
6, and 10 wt % lithium added to the NaK, The
results are given in Table 5.14.

In the temperature-solubility apparatus, NaK with
5 wt % lithium added was heated to 200°C, agi-

tated, and allowed to remain at that temperature

TABLE 5.12, CARBON ANALYSES OF THE VARIOUS COMPONENTS OF SYSTEMS DESIGNED FOR
STUDYING THE DECARBURIZATION OF MILD STEEL BY SODIUM

Duration of tests: 100 he
Test temperature: 1830°F

 

Carbon Content* (wt %)

 

 

Test System Material Analyzed
Before Test After Test
Type 1043 mild-steel specimen Mild-steel specimen 0.433 0.121
in Armco iron capsule Armco iron capsule 0.019
Bath zone 0.035
Vapor zone 0.019
Type 1043 mild-steel specimen Mild-steel specimen 0.433 0.100
in type 304 ELC stainless Stainless steel capsule 0.022
steel capsvule Bath zone 0.128
Vapor zone 0.022

 

*Two analyses performed on each sample.

125
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-16943

WALL OF ARMCO
IRON CAPSULE
BEFORE TEST

 

 

 

 

S20'§
T ofwen
70| |
2367 |
1
. eod
ol
80
Tio
¥i0
o
30|
e
o [ors
"~ 560
150
=
- €2
a4
m
L. ]
55T

waLL OF ARMCO
IRON CAPSULE
AFTER TEST

MILD STEEL
SPECIMEN
BEFORE TEST

e ]E IE E ]§ I 2%0)::

 

 

MILD STEEL
SPECIMEN
AFTER TEST

 

Fig. 5.24. Armco Iron Capsule and Type 1043 Mild-Steel Specimen Before and After Exposure to Static
Sodium for 100 hr at 1830°F. 200X. Reduced 18%.

126
PERIOD ENDING DECEMBER 10, 1955

TABLE 5.13. RESULTS OF STATIC TESTS OF SPECIAL STELLITE HEATS IN
LITHIUM AT 1500°F FOR 100 hr

 

Weight Change

 

terial .

Materia (g/in.z) Metallographic Netes

Heat B -0.044 Uniform attack to a depth of 2 to 3 mils plus very
heavy attack in several isolated areas

Heat B -0.059 Approximately 90% of exposed surface attacked to
a depth of 3 mils; very heavy attack on other 10%
of surface

Heat C -0.069 Uniform attack to a depth of 2 to 4 mils plus very

heavy attack in some areas

 

TABLE 5.14. RESULTS OF MEASUREMENTS OF
THE SOLUBILITY OF LITHIUM IN NaoK

 

 

Analysis Lithium

Test Methods Temperature Content

(°C) (wt %)
Differential thermal 176 10
analysis 170 6
166 5

Temperature-solubility 25 0.014

 

for 2 hr. The bath was then cooled to 25°C, and
a sample was obtained through a stainless steel
filter. The results of chemical analyses of the
samples for lithium are given in Table 5.14. it
appears from the data that a small amount of
lithium is soluble in NaK.

Seesaw Tests of Titanium Carbide Cermets

in NOF-Z!F4-UF4

W. H. Cook
Metallurgy Division

Some titanium carbide cermets with cobalt- or
nickel-base alloys as binding material were sub-
jected to screening tests in contact with NaF-
ZrF,-UF, (53.5-40-6.5 mole %) in seesaw appa-
ratus operated at 4.25 cpm. These cermets were
submitted by the Sintercast Corp. of America.
The hot and cold zones of the apparatus were at
1500 and 1200°F, respectively. Each specimen
was held in the hot zone of its capsule during the
200-hr test period. The results are summarized

in Table 5.15. The results of previous tests® of
five of the same types of Sintercast Corp.
specimens exposed to NaF-ZrF4-UF4 (50-46-4
mole %) for a 100-hr test period are also included
in Table 5.15 for comparison purposes. The
results of the earlier tests are considered to be
less significant than those obtained in the recent
tests, because at the time the earlier tests were
made there were no untested specimens available
for comparison, the specimen compositions were
not known,’
involved techniques required for metallographically
polishing these kinds of specimens.

As shown in Table 5.15, only Sintercast speci-
mens 1 and 7 were attacked by NaF-ZrF4-UF4
(53.5-40-6.5 mole %). The type of attack is
illustrated in Figs. 5.25 and 5.26. In the metallo-
graphic exomination of the specimen structures, it
was noted that the titanium carbide particles in
the cermets with cobalt-base binder were, in
general, more globular than those in the cermets
with nickel-base binder. It appears therefore that
the cobalt-base alloys may have reacted more with
the titanium carbide in the fabrication of the
specimens than did the nickel-base ailoys.

The typical extremes of the particle shapes are
i llustrated in Figs. 5.25 and 5.26, and the average
structure and typical appearance in the absence
of attack are illustrated in Fig. 5.27. As may be
noted, there was practically no mutual bonding
between the titanium carbide particles, as was

and little was known about the

 

E. E. Hoffman et al., ANP Quar. Prog., Rep. June 10,

7C0mpositions and methods of fabrication of the
Sintercast Corp. specimens are **Company Confidential
Information?’’ and not available for general publication.

127
8¢l

TABLE 5.15.

SUMMARY OF THE RESULTS OF SEESAW CORROSION TESTS ON SOME SINTERCAST CORPORATION TITANIUM CARBIDE CERMETS

Dperating cycle: 4.25 cpm
Hot-zone temperature: 1500°F
Cold-zone temperature: 1200°F
Capsule material: Inconel

 

 

 

Sintercast ' Attack (mils)
i Metal Binder _
Specimen Alloy Base In NaF-ZrF -UF In NaF-ZrF ,-UF Metallographic Notes
Ne. (50-46-4 mole %)*  (53.5-40-6.5 mole %}**
1 Cobalt 0.5t02 4108 The attack was primarily along the boundaries between the TiC particles and the
metal binder; the TiC particles were slightly smaller than those in specimens
2 and 4 and tended to be glebular in shape
2 Cobalt 0 The specimen was unattacked; the TiC particles were globular and intermediate
in size between those of specimens 1 and 4
4 Cobalt 1 0 The specimen was unattacked; the TiC particles were slightly larger than those
in specimens I and 2; the TiC particles were globular and sharply defined in
the metal
3 Nickel 0 The specimen was unattacked; the TiC particles were not so globular as those
bonded with the cobalt-base alloys, specimens 1, 2, and 4; the TiC particle
size appeared to be the same as that for specimen 1
5 Nickel 0 0 The specimen was unattacked; the TiC particies were less globular in shape than
those in specimen 3 and they appeared to be smaller
6 Nickel 0 The specimen was unattacked; the TiC particles were approximately the same
size as those in specimen 4; they were less globular than those bonded with
the cobalt-base alloys; there was more metal arec in a section than in any other
of the specimens
7 Nickel 1.5t07 The attack was primarily aleng the boundaries between the TiC particles and the
metal binder; the TiC porticles in this and specimen 10 were the most angular in
shape of all specimens
8 Nickel 0.5t02 0 The specimen was unattacked; the structure was similar to that of specimen 6
except that there was less metal area
9 MNickel 0 The specimen was ynattacked; the structure was similar to that of specimen 8
except that there was a greater quantity of small TiC particles
10 Nickel 5 0 The specimen was unattacked; the TiC particles were angular like those in

specimen 7 and were larger than those in any of the other specimens

 

*100-hr tests.
**200-hr tests,

LY¥0OJIY SSIYI0YUd L1D3rodd ANV
PERIOD ENDING DECEMBER 10, 1955

 

 

 

Fig. 5.25. Sintercast Specimen 1 (a) As-received and (b) After Exposure to NuF-ZrF4-UF4 (53.5-40-6.5
mole %) for 200 hr in Seesaw Apparatus with a Hot-Zone Temperature of 1500°F and a Cold-Zone Tem-
perature of 1200°F. Specimen held in hot zone during test. Unetched. (Gwew with caption.} 500X,
. Reduced 13.5%.

CUNCLASSIFIED
. L

 

 

 

- Fig. 5.26. Sintercast Specimen 7 (a) As-received and (b) After Exposure to NaF-Zr F4-UF (53.5-40-6.5
mole %) for 200 hr in Seesaw Apparatus with a Hot-Zone Temperature of 1500°F and a Cold-Zone Tem-
peratuwre of 1200°F. Specimen held in hot zone during test. Unetched. ({gemagywith caption.) 500X.
Reduced 13.5%.

129
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED -
Y-16889

 

 

Y-16890

 

Fig. 5.27. Sintercast Specimen 4 (a) As-received and (b) After Exposure to NaF-ZrF ,-UF , (53.5-40-6.5
mole %) for 200 hr in Seesaw Apparatus with a Hot-Zone Temperature of 1500°F aond a Cold-Zone Tem-
perature of 1200°F. Specimen held in hot zone during the test, The holes were created by removal of
TiC particles during metallographic polishing, Unetched. (Swem with caption,)

previously reported for the more common types of
cermets with low metal content.8:?

Thermal-Cycling Tests of Inconel Yalve Disks
and Seats Flame-Plated with a Mixture of
Tungsten Carbide and Cobalt

W. H. Cook
Metallurgy Division

Initial evaluation tests of valve disks and seats
coated with a mixture of tungsten carbide and
cobalt were made. The coating, which was being
tested for suitability as a hard-facing material,
was deposited by a commercial flame-plating
method developed by Linde Air Products Co.

 

8E. E. Hoffman, W. H. Cook, and C. F. Leitten, Jr.,
ANP Quar. Prog. Rep. Mar, 10, 1955, ORNL-1864, p 84.

9E. E. Hoffman er al,, ANP Quar, Prog. Rep. Sept. 10,
1955, ORNL-1947, p 104,

ey el

130

The coating, to be suitable, must be resistant
to solid-phase bonding, galling, and wearing.

The seating surfaces of four sets of Inconel
valve disk and seat blanks were flame-plated by
the Linde Air Products Co. with coatings 4 to
6 mils thick that were nominally 92 wt % tungsten
carbide and 8 wt % cobalt. Three sets of the
tiame-plated valve parts were then copper-brazed
at a temperature of 2048°F to Inconel platens.
Adequate wetting was obtained; however, cracks
were created in the flame-plated coatings, as
shown in Fig. 5.28. The cracks were thought
to be caused by the difference between the coef-
ficients of thermal expansion of the coating and
the Inconel.

Two tests were made on the fourth set of Inconel
valve seat and disk blanks to further evaluate
the usefulness of the flame-plated coating on
Inconel. The purpose of the first test was to
e e cmem e #

UNCLASSIFIED
Y-16856

 

 

Fig. 5.28. Inconel Valve Parts Flame-Plated
with a Mixture of Tungsten Carbide and Cobalt and
Then Copper-Brazed to Inconel Platens at 2048°F,
Note cracks in the flame-plated coatings.

determine whether the flame-plated coating was
satisfactory for use at the proposed operating
temperature.  The disk and seat were slowly
cycled in a vacuum from room temperature to
1500°F, which was approximately 100°F in excess
of the proposed maximum operating temperature,
There was a small amount of spalling of the
plating where it was built up on corners. It the
seating surfaces had been machined, as they
would have been for a finished valve disk-and-
seat set, there probably would not have been any
spalling. There were no cracks that could be
detected with a dye penetrant. In the second
test, the disk and seat were heated in a vacuum
from room temperature to 2012°F in 5 min, held
at 2012°F for 15 min, and then allowed to cool
to room temperature in approximately 3 hr. These
thermal conditions were similar to those used in
the brazing of the other flame-plated disks and
seats, with the exception that the thermal shock
was more severe. Ihe coatings on both the disk
and seat developed cracks like those found in
the coatings on the brazed disks and seats. The
results of these two tests are illustrated in

Fig. 5.29.

PERIOD ENDING DECEMBER 10, 1955

Static Tests in NaF-Z¢F ,-UF, of Kentanium
Cermet Yalve Parts Brazed to Inconel

W. H. Cook
Metallurgy Division

A valve seat made from the Kentanium cermet
KI151A (70% TiC=10% NbTaTiC-20% Ni) and a
valve disk made from KI152B (64% TiC-6%
NbTaTiC;~30% Ni) held in contact at a pressure
of 10,000 psi (calculated) were exposed to static
NaF-ZrF4-UF4 (53.5-40-6.5 mole %) at 1500°F
for 150 hr. The surface-roughness values for the
contacting surfaces were from 5 to 10 pin. before
the test.'® The test was terminated when a pull
rod failed. The lower valve part fell free from the
upper valve part, and thus there was an indication
that no solid-phase bonding had occurred. A
subsequent low-power microscopic examination
revealed no evidence of bonding. Visuval and
microscopic examination of the 45-deg beveled
seating surfaces of the valve disk and seat indi-
cated that there was uniform seating during the
test.

This was also the first test in which the re-
cently developed method of brazing Kentanium
cermets to Inconel'! was tested in a fused-salt
bath. There were no visible signs of attack, of
distortion of the ]/“v-in.-thick nickel shim, or of
cracking of the cermet pieces after the test.

Effect of an Air Leak into an
Inconel-Fused-Salt Test System

W. H. Cook
Metallurgy Division

An Inconel—fused-salt tube-burst test capsule
was found to be very heavily attacked above the
level of the NaF-ZrF ,-UF , mixture following a
500-hr test at 1500°F. It was suggested that the
attack (Fig. 5.30) was due to an air leak some-
where in the system, and therefore several experi-
ments were performed in which air was admitted
to similar test capsules through a small tube.
These test capsules failed within 24 hr at a
temperature of 1500°F. The reaction products

 

10The usual 1.5 to 2 pin. used for solid-phase-
bonding tests with bar specimens could not be ob-
tained because this disk-and-seat configuration pro-
hibited lapping.

b, patriarca and R. E. Clausing, ANP Quar, Prog.
Rep. June 10, 1955, ORNL-1896, p 145, and ANP Quar.
Prog. Rep., Sept. 10, 1955, ORNL-1947, p 131.

131
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-16584

  

UNCLASSIFIED
Y-16659

 

UNCLASSIFIED -
Y-16857

 
  

Fig. 5.29. Inconel Yalve Disk and Seat Flame-Plated with a Mixture of Tungsten Carbide and Cobalt.
(a) As received. (b) After being cycled slowly from room temperature to 1500°F in vacuum. (c) After
being heated in vacuum from room temperature to 2012°F, held at 2012°F for 15 min, and allowed to cool
to room temperature in about 3 hr.

132
UNCLASSIFIED
Y-16834

 

-
Wl
=
Lt
d
I
—
<I
Q0

L
—
<
=
>
O
jh
! Q-
3 a
! <L

 

Fig. 5.30. Effect of an Air Leak into an Inconel-
Fused-Salt Test System. (@OWITN with caption.)

found in the fused salt were zirconium oxide and
structural metal fluorides.

FUNDAMENTAL CORROSION RESEARCH

G. P. Smith
Metallurgy Division
Self-Decomposition of Fused Hydroxides
M. E. Steidlitz G. P. Smith
Metallurgy Division

The thermal decomposition reaction for NaOH is

(M 2NaOH == Na,0 + H,0

and it may be shown on a theoretical basis that
the partial pressure of water vapor, p, should obey
the relation:

AHC 1 As® Yh,0
+ +— log —

9152 T 9.052 2 2 yn. o
2

 

 

logp = -

PER!IOD ENDING DECEMBER 10, 1955

where AH® and AS° refer to Eq. 1 and the y's are
activity coefficients in fused NaOH, Data ob-
tained in recent measurements of the vapor pres-
sure of water over fused NaOH confirm the theory,
as shown in Fig. 5.31. The value of AH® ob-
tained from the data is 44.4 kcal. The value of
AH®, -known from other measurements, is 41.4
kcal. This agreement is considered to be good.

UNCLASSIFIED
ORNL-LR-DWG 11530

 

SLOPE = 4 .85 x103

 

 

 

 

 

 

 

 

8 10 12 14
wlr ek

Fig. 5.31. Logarithm of the Yapor Pressure of
Water Over Fused NaOH as a Function of the Re-
ciprocal of the Absolute Temperature.

Mass Transfer and Corrosion of Yarious
Materials in Fused Sodium Hydroxide

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

Inconel. — Thirty-two tests have been made of
the corrosion of Inconel by fused NaOH in the
absence of a thermal gradient. The tests were
conducted with the NaOH blanketed with an
atmosphere of helium or hydrogen, but metallo-
graphic examinations have been completed only

133
ANP PROJECT PROGRESS REPORT

for the tests in which hydrogen was used as the
blanketing atmosphere.

A semiquantitative evaluation of the extent of
corrosion of the Inconel after exposure for various
times at various temperatures is presented in
Fig. 5.32. The seven photomicrographs show
cross sections of the specimens in the region of
the fused hydroxide—metal interface. The scales
at the bottom of the figure are applicable to all the
photographs. In each photograph, the pale-gray
region at the bottom is unattacked metal. The
dark-gray region is the layer of corrosion product,
The white area at the top was originally occupied
by the fused hydroxide.

The top row of pictures shows the rather severe
attack found at 800°C. The arrow gives the ap-
proximate positicn of the hydroxide-metal interface
at the beginning of the test.
pictures shows the corrosion product formed in
tests at 700°C. [t may be noted that the thick-
nesses of the corrosion-product layers vary con-
siderably. Only one picture, Fig. 5.32¢, is shown
for corrosion at 600°C. For corrosion periods
significantly shorter than 100 hr, very little or
no attack is found at 600°C. As may be seen,
exposure for 100 hr resulted in only slight inter-
granular penetration; there is no corrosion-product
layer. The metallographic details shown in
Fig. 5.32 may be taken as typical, except that
some tests run under the same conditions resulted
in a narrow region of intergranular penetration
between and the

The second row of

the corrosion-product layer
unattacked metal.

Quantitative data on the corrosion of Inconel
by fused NaOH under a blanketing atmosphere
of hydrogen are presented in Fig. 5.33. The
circles are average corrosion values and have
the following meaning: The numerical product
of the average corrosion value times a given
area of external surface, assumed to be flat,
gives approximately the actual volume of corrosion
product

The vertical lines passing through a circle give

contained beneath the given surface.

the values of maximum and minimum corrosion
observed for the specimen.

Heavily corroded specimens were found to be
strongly ferromagnetic. Microscopic studies were
made of the migration of colloidal magnetite over
a polished cross section of a specimen in a
magnetic field. It was shown that the ferro-

134

magnetism was restricted to the corrosion-product
Samples of ground corrosion product were
leached with water for long periods, and it was

layer.

found that appreciable quantities of sodium were
removed in the form of Na,0. X-ray examination
of the corrosion product gave an abundance of
lines which could not be analyzed. Further efforts
are being made to determine the composition
of the corrosion product.

Under the influence of a thermal gradient and
at sufficiently high temperatures, Inconel under-
goes both mass transfer and corrosive attack.
The mass-transferred deposit apparently adheres
This is not
too surprising, however, in view of the metallic
matrix which the corrosion product was found to
contain. For Inconel under a blanketing atmos-
phere of hydrogen, the corrosion was slightly
worse under a thermal gradient than would have
been expected on the basis of the results from
the static tests, reported above, while the mass

quite well to the corrosion product.

transfer was about the same as that found for
pure nickel under the same conditions.

For Inconel under a blanketing atmosphere of
helium, the available data are insufficient for
Qualitatively, however,
both corrosion and mass transfer are much more
severe than they.gpe under similar experimental
conditions with a blanketing atmosphere of
hydrogen. The microstructure of the corrosion
product for tests conducted under helium is
distinctively different from that found for tests
conducted under hydrogen. The corrosion product

quantitative evaluation.

consists of branched dendritic structures, many
of which begin at the outer interface and penetrate
deeply into the base metal. Along the axis of
each dendrite is a relatively wide band of a
metallic phase.

Nickel. — Sixty-five tests have been made in an
effort to study the effect of temperature, temper-
ature gradient, and time on the mass transfer of
nickel in fused NaOH, and about one-half the
tests have been evaluated metallographically. The
maximum test temperature has ranged from 550 to
800°C, and the thermal-gradient range was between
50 and 200°C. The only results that are complete
enough to report in detail are those obtained from
tests in which a blanketing atmosphere of hydrogen
was used.

It is important to distinguish between the mass-
transferred deposits at the gas-liquid interface and
Gel

 

A~ e i ———

{a} EXPQOSED AT 800°C FOR 28 hr. (5) EXPOSED AT BOO°C FOR 54 he. (¢) EXPOSED AT 800°C FOR 10¢ hr.

  

(&) EXPQSED AT 700°C FOR 23 hr.

) EXPQSED AT 700°C FOR 100 br.

 

 

(s tEEEEERER ~<FE [zosk EE

 

(g) EXPOSED AT &M0°C FOR 100 hr.

Fig. 5.32. Corrosion of Inconel by Fused NaOH Under a Blanketing Atmosphere of Hydrogen.

EEERERER (kb

250X. Reduced 57%.

\ORIGINAL METAL SURFACE

SS61 ‘01 ¥39W3D3Q ONIAN3I d0iy¥3d
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR—DWG 11531

 

 

S

CORROSION-PRODUGCT THICKNESS {in x 10°)

 

 

 

 

 

o . 20 40 &0 80 100 120
TIME (hr)

Fig. 5.33. Corrosion of Inconel by Fused NaOH
Under a Blanketing Atmosphere of Hydrogen as a
Function of Time and Temperature.

those on the immersed surface, since deposition
occurs more readily at the interface than elsewhere
in the system, However, such interfaces would
occur at only a few places in a heat-transfer
system and should not be a source of trouble.
Therefore the evaluation study reported here deals
entirely with deposition on the immersed surface.
Usually the mass-transferred deposits were less
than 0.001-in. thick, They were also, of course,
of about the same composition as the base metal
and frequently continued the crystal orientation of
the base. Thus it was necessary to use refined
metallographic techniques to measure the thickness
of the deposited films. In order to prevent
rounding-off of the edge of a specimen during
metallographic polishing, the specimens were first
copper-plated. Despite this, very slight rounding-
off was unavoidable, and therefore the value
0.0001 in. is taken to be the minimum thickness
for which mass-transferred deposits can be posi-
tively identified. This minimum thickness corre-
sponds to a uniform deposit of 2 mg/em?2, In order
to determine the thickness of deposit, it was
necessary to establish the boundary between the
base metal and the deposited metal. It was found
that very careful etching would, in some instances,

136

yield a faint grain-boundary-like groove along this
boundary, In other instances, no such boundary
could be found. However, it was discovered that
the mass-transferred deposits oxidized at a slightly
lower rate than did the base metal, perhaps
because of the greater purity of the deposit, and
oxidation of metallographically polished specimens
tarnished the base metal but not the deposit.

A nickel mass-transferred deposit on a nickel
cold finger is shown in Fig. 5.34. The maximum
thickness of this film is 0.0007 in., and its
average thickness is about 0.0002 in. The lower
region of the photomicrograph is the base metal,
the middle region is the copper plate, and the
dark region at the top is the mounting plastic.
The irregular mass-transferred deposit may be
seen between the copper plate and the base metal.
This specimen illustrates both the etching and
the tarnishing techniques.

The mass-transferred deposits found on all the
specimens examined have been very rough. They
consisted of nickel crystals densely crowded
together over the surface of the base metal. The
bases of the crystals had grown together so that,
in cross section, the deposits appeared to be thin
and continvous and to be densely covered with
elongated protuberances of irregular dimensions.
Almost all the protuberances were of approximately
the same length, At temperatures of 750°C and
above, an occasional nickel crystal grew, as a
dendrite, to several times the length of an average
crystal, Usually, less than six such dendrites
were found on a given cross section of a cold
finger. Most frequently, the cross section did not
include either the base or the tip of these ex-
ceptional dendrites, and therefore no reliable
measure of their length could be obtained. Thus
the depth of the mass-transferred deposit is
usually described in terms of two parameters — the
length of an average protuberance and the length
of the longest protuberance of a given cross
section — and the presence or absence of the rare
dendrites is noted. In some cases, even this
description leaves something to be desired be-
cause an increase in deposition is sometimes
manifested by a greater density of long pro-
tuberances rather than an increase in the average
protuberance length. For these reasons, repre-
sentative photomicrographs provide a more ade-
quate description of mass transfer than do tables
or graphs.
PERIOD ENDING DECEMBER 10, 1955

 

Fig.5.34. Nickel Mass-Transferred Deposit on Nickel Cold Finger Immersed for 100 hr in Fused NaOH.

Maximum temperature, 700°C; temperature differential in system, 100°C; blanketing atmosphere, hydro-

gen. 330X.

lron. — A study is also being made of the mass
transfer of carbon steel (AlSI No. C1015) in fused
NaOH. The tests performed thus far have been
of 100-hr duration with a temperature differential
of 100°C. A series of cold fingers taken from
these tests are shown in Fig. 5.35. ' Heavy
deposits of iron crystals can be seen on the lower
ends of the fingers, as well as the even-heavier
skirts of crystals which formed at the liquid-gas
interface. Specimen 127, shown in Fig. 5.35,
was subjected to g maximum temperature of 550°C,
specimen 128 to 600°C, specimen 129 to 650°C,
specimen 130 to 700°C, and specimen 131 to
750°C. The scale in the photograph is in inches.
It may be noted that even at 550°C the mass
transfer of iron was significantly worse than the
mass transfer of nickel at 800°C, when the other
experimental parameters were the same. However,
the crystals deposited were roughly equiaxed in

even the heaviest deposits, whereas similar heavy
deposits of nickel were always dendritic.

Hastelloy B. — The data on Hastelloy B are
difficult to evaluate because of the uniformly poor
quality of the metal available for test. This metal
contained pits, cracks, and stringers that were
frequently of macroscopic size. It is known from
other studies that fused sodium hydroxide has an
extraordinary ability to penetrate deeply into
minute voids, and hence each flaw in the
Hastelloy B which touched the surface provided
a pathway along which the hydroxide moved and
thus caused localized corrosion deep in the
interior of the metal. Finally, the available metal
sometimes consisted of two metallic phases in
various amounts.

All tests of Hastelloy B were conducted under
a blanketing atmosphere of hydrogen for 100 hr
with a temperature differential of 100°C. The

137
ANP PROJECT PROGRESS REPORT

 

UNCLASSIFIED
Y-16247

Fig. 5.35. Mass-Transferred Deposits on Carbon-Steel Cold Fingers Tested for 100 hr in Fused NaOH

Under a Hydrogen Blanketing Atmosphere,

Test temperatures ranged from 550°C for specimen 127 to

750°C for specimen 131. System temperature differential, 100°C.

mass transfer found under these conditions was
similar in magnitude to that found for nickel under
similar conditions, with the exception that it was
detected at temperatures as low as 650°C in
Hastelloy B but not below 670°C in nickel. This
difference of 20°C is not especially significant,
however, in view of the over-all scatter in the
experimental results, Corrosion was detected at
600°C, the lowest temperature studied. The
corrosion was, in general, 1.5 to 2 times that
tound for Inconel under similar conditions, but the
scatter in the results was quite bad.

Stainless Steels, — A number of tests were made
with types 310 and 405 stainless steel exposed
to fused NaOH under hydrogen at temperatures
from 550 to 750°C for 100 hr with a temperature
differential of 100°C. In all runs scaling was so
severe that the specimens were not subiecfed to
metallographic study.

138

CHEMICAL STUDIES OF CORROSION

F. Kertesz
Materials Chemistry Division

Reaction of Inconel with Sodium and NaK

G. F. Schenck
Pratt & Whitney Aircraft

A study of the solubility and the rate of solution
of the constituents of Inconel in sodium and in
NaK as a function of temperature and of oxygen
content of the liquid metal has been initiated, in
order to clarify the mechanism of the mass-transfer
reaction observed when liquid metals are circu-
lated in Inconel loops. In view of its importance
in reactor application, the temperature range of
600 to 800°C was chosen for closer study.

The apparatus being used consists of ‘a cy-
lindrical capsule separated into two parts by a
funnel-like constriction. The liquid metal is
introduced into the Inconel reaction end of the
capsule through a side-arm, and the capsule is
sealed. The sealed capsule is placed in a furnace
in a position such that the liquid metal remains
in the Inconel end while ot the test temperature.
After the desired contact time and while the
system is still at temperature, the capsule is
inverted and the liquid metal containing dissolved
structural metal is poured through the funnel-like
constriction into a tantalum crucible which fits
snugly into the receiving end of the capsule.

After the capsule has been cooled to room
temperature, it is opened just below the funnel.
The tantalum crucible containing the liquid metal
is slipped out, and both the crucible and the liquid
metal dre analyzed for the alloy constituents of
Inconel. An additional analysis for oxide con-
tamination is performed on the liquid metal.

A high-vacuum apparatus and a NaK bubbler are
being set up to provide a purification system. To
facilitate the loading operation, o liquid-metal
dispenser is being built that is based on ideas
described by Drugas and Kelman.'2 This appa-
ratus will permit the repeated recycling of o
relatively large volume of NaK through an ultrafine
fritted-glass filter. After purification, a prede-
termined volume may be passed through a second
filter and dispensed in two portions:
oxygen analysis and one for testing in the above-
described capsule.

one for

Reaction of Sodium Hydroxide with Nickel
F. A. Knox

Materials Chemistry Division

The mass transfer of nickel in sodium hydroxide
was reported previously!? to be decreased by the
addition of chromium to the hydroxide. This work
was performed under a temperature gradient in a
tilting furnace and, in general, confirmed the
observations made by Forestieri and Lad,'4 who

 

12p ¢, Drugas and L. R. Kelman, Equipment and
Procedures for Studying the Equilibrium Solubility of
Iron in NaK, ANL-5359 (Sept. 4, 1953).

13y, 1 Buttram, R. E. Meadows, and F. A, Knox,
ANP Quar, Prog. Rep. Sept. 10, 1955, ORNL-1947,
P ]20.

YA, F. Forestieri and R. A. Lad, The Use of Metallic
Inbibitors for Eliminating Mass Transfer and Corrosion
in Nickel and Nickel Alloys by Molten Sodium Hy-
droxide, NACA RM-E54L 13 (Dec. 7, 1954).

PERIOD ENDING DECEMBER 10, 1955

used a staticcrucible test. In the current series
of tests, quartz-jacketed nickel tubes were used
to prevent the escape of the hydrogen formed
during the reaction. In addition to studying the
effect of chromium, the effects of calcium, mag-
nesium, aluminum, and sodium powders were in-
vestigated by charging 10 g of purified sodium
hydroxide and 0.1 g of the metal being studied
into the nickel crucible. The filled capsules were
crimped and welded under helium and were placed
into quartz tubes, which were subsequently
evacuated.

Visual inspection of the capsules after 100-hr
tilting-furnace tests revealed that the amount of
mass transfer was least when chromium was added
to the hydroxide and was highest in the capsules
containing added sodium. On the other hand, the
amount of nickel found in the hydroxide was
smallest when the hydroxide contained no added
metal, and it increased with the chromium, calcium,
sodium, aluminum, and magnesium additions, in
that order; that is, the presence of these metals
appears to increase the solubility of nickel in the
hydroxide.

Study of Eutectic Mixtures by Zone Melting

H. J. Buttram
Materials Chemistry Division

The establishment of the exact location of the
eutectic point of binary and ternary mixtures is
often difficult and time-consuming., Zone-melting
experiments have therefore been performed as an
aid in phase-equilibrium studies.

The apparatus used consisted of a 24-in, tube
furnace and a 4-in, tube furnace placed end to
end. The tube containing the mixture being
studied was pulled through the furnaces by an
electric motor geared to move the tube at a rate
of 1 in./hr. The long furnace melted the mixture
sealed in the tube, and the short furnace furnished
the heat for the zone melting.

After the mixture was molten, the long furnace
was turned off, and the capsule was pulled through
the 4-in. furnace. Upon leaving the long furnace,
the high-melting-point phases solidify; the low-
melting-point phases flow back into the hot zone,
and they ultimately accumulate ot the end of the
tube.  Repeated passage through the furnace
increases the separation effect.

139
ANP PROJECT PROGRESS REPORT

UNGLASSIFIED
ORNL-LR-DWG 11532

 

100

 

g0

 

 

 

l
|
80 ’

 

70

 

 

 

 

10

 

 

 

 

 

 

 

COMPOUND CONTENT (mole %}

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50
40 —
30 (—_;_

20 A
10
N
0 .
f 2 3 4 5 6 7 8 9 {0 # 2 13 14 15 16
BOTTOM CAPSULE SECTIONS TOP
Fig. 5.36. Results of Zone-Melting Experiments

on NaF-KF-ZrF , (46-44-10 mole %). Mixture passed
three times through hot zone of furnace at 900°C
at a rate of 1 in./hr.

140

The results obtained for a mixture of NaF-KF-
ZrF, (46-44-10 mole %) after three trips through
the 1urnace are shown in Fig. 5.36. The capsule
was cut into 1-in, sections for analysis. The
bottom section (No. 11) contained the mixture
which was approximately the eutectic mixture,
while the top section contained a mixture that
was 73.5 mole % NaF, 21 mole % KF, and 5.5
mole % ZrF4.
PERIOD ENDING DECEMBER 10, 1955

6. METALLURGY AND CERAMICS

W, D. Manly
Metallurgy Division

The fabrication of radiators and heat exchangers
for use in the heat exchanger development program
was continved, Welded and brazed joints of heat
exchanger job samples submitted by outside vendors
were examined and evaluated. The two 500-kw
high-conductivity-fin NaK-to-air radiators that failed
in service were examined, The results of the ex-
aminations indicated the necessary modifications
in design and fabrication techniques.

Results obtained to date in the brazing-alloy
development program were correlated with corrosion
data, and the alloys that are satisfactory for use
in the fabrication of heat exchangers and radiators
were selected,

Mechanical-property studies of nickel-molybdenum
alloys were continued. Data were obtained on the
influence of aging heat treatments on the creep
properties of Hastelloy B, ond a preliminary in-
vestigation was made of the creep properties of
Hastelloy W and some new nickel-moilybdenum
alloys.

In studies of special materials, several composite
billets were extruded successfully to form seam-
less duplex tubing; the billets were nickel-,
Inconel-, and Monel-clad type 316 stainless steel.
Methods for producing B,C tiles for use as neutron
shielding were investigated. Tests are being de-
veloped for determining the optimum diffusion
barrier for use between Inconel and niobium. A
search was made for a material to use as a gamma-
ray shield for the impeller shafts of ART pumps.
Methods for preparing rare-earth-oxide core com-
pacts for control rods were developed; and several
iron-zirconium alloys were tested to determine
their physical properties.

Nondestructive-testing equipment was developed
for the inspection of small-diameter tubing by the
eddy-current technique and by the use of ultra-
sound. An x-ray technique for the inspection of
thick sections of beryllium was also investigated.

 

1P, Patriarca et al., ANP Quar., Prog. Rep June 10,
1955, ORNL-1896, p 134 and Fig. 6,18, p 13

FABRICATION OF TEST COMPONENTS

P. Patriarca A, E. Goldman
R. E. Clausing G. M. Slaughter
Metallurgy Division

R. L. Heestand, Pratt & Whitney Aircraft
NaK-to-Air High-Conductivity«-Fin Radiators
A third 500-kw NaK-to-gir radiator with type 310

stainless-steel-clad copper high-conductivity fins
was fabricated for use in tests in connection with
the heat exchanger development program, The
design of the radiator was essentially the same as
that described and illustrated in a previous report, !
but a few modifications were made in fabrication
technique. These modifications included the con-
struction of a 36-hole punching die to obtain a
more uniform fin-punching geometry, the construc-
tion of new tube-bending dies to ensure close con-
trol of the tube-bending variables, the development
of techniques to permit the assembly of headers in
such a manner that all tubes enter normal to the
curvature of the header at the point of entrance,
the application of the presintered-ring method of
brazing-alloy preplacement, and the use of stearic
acid as a lubricant during assembly of the fins,

A photograph of the radiator taken in the early
stages of fin assembly is shown in Fig, 6.1. The
wooden ‘‘finger’’ jigs and the felt-covered wooden
support stand constructed to conform accurately to
the contour of the tubes may be seen in the photo-
graph, The tube bends and the manually inert-arc-
welded tube-to-header joints may be seen in Fig.
6.2. The presintered brazing-alloy rings shown in
Fig. 6.2 were adequate to back-braze the tube-to-
header joints, since they rested upon the headers
in an inverted position during brazing. The com-
pleted radiator, with the support members used
during brazing, is shown in Fig. 6.3. These sup-
ports were of welded-steel construction, and they
incorporated sintered aluminum oxide rods at all
mating surfaces between the radiator and the sup-
ports to minimize the possibility of brazing between
them. The completed radiater is shown in Fig. 6.4,
with the supports removed. A leck test with a
helium mass spectrometer indicated that all welded
and brazed joints were leaktight.

141
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED : ) UNCLASSIFIED
Y-16299

Y-16657

 

Fig. 6.1. NaK-to-Air High-Conductivity-Fin Radi-
ator in Early Stages of Fin Assembly.

UNCLASSIFIED
Y-16542

 

e

Fig. 6.3. Completed NaK-to-Air Radiator Show-
ing Support Members Used During Brazing.

UNCLASSIFIED
¥»16455

  

Fig. 6.2, NaK.to-Air Radiator Showing Manually
Inert-Arc-Welded Tube-to-Header Joints and Tube
Bends. Fig. 6.4. Completed 500-kw NaK-to-Air Radiator.

142
Twenty-Tube Fuel-to-NaK Heat Exchangers

Twa additional 20-tube Inconel fused-fluoride-
fuel-to-NaK heat exchangers were also fabricated
for testing in connection with the heat exchanger
The fabrication of these

development pregram.

UNCLASSIFIED
Y-16297

 

Fig. 6.5. Twenty-Tube Heat Exchanger Showing
the ‘Window'' Through Which the Brazing Alloy
Was Preplaced.

PERIOD ENDING DECEMBER 10, 1955

heat exchangers involved manual inert-arc welding
and subsequent back-brazing of the tube-to-header
joints, Back-brazing was employed to eliminate
the ‘'notch-effect’’ that results from incomplete
weld penetration and to reinforce areas of shallow
weld penetration, which might otherwise develop
leaks if they become corroded,

An evaluation of the corrosion resistance of
various brozing alloys (Sec. 5, ‘‘Corrosion Re-
search'’) has shown that low-melting Nicrobraz
(LMNB) and Coast Metals alloy No. 52 are com-
patible with both NaK and fuel. Coast Metals alloy
No. 52 was used in the fabrication of these tube
bundles because its flowability characteristics are
better than those of low-melting Nicrobraz,

Two brazing operations were required for each
heat exchanger, since the Globar pit furnace avail-
able did not have a heating zone of sufficient
length to heat both ends of the unit to the brazing
temperature, A ‘‘window’’ was removed from the
pressure shell to permit preplacement of the brazing
alloy on each end of the tube bundle, as shown in
Fig. 6.5. The window was subsequently welded
shut, and helium leak-testing indicated that all
welded and brazed joints were leaktight, A com-
pleted heat exchanger is shown in Fig. 6.6.

Intermediate Heat Exchanger No, 3

The partial fabrication of the two tube bundles of
intermediate heat exchanger No. 3 was described
in the previous report.?2 The welding and back-
brazing of these units has now been completed, and
all joints were found to be helium leaktight. A
photograph of one end of the heat exchanger in an
advanced stage of fabrication is presented in
Fig. 6.7.

2p, Patriarca, G. M. Slaughter, and R. L. Heestand,
ANP Quar, Prog, Rep. Sept. 10, 1955, ORNL-1947, p 134.

 

UNCL ASSIFIED
Y-16296

 

Fig. 6.6. Completed 20-Tube Heat Exchanger.

143
ANP PROJECT PROGRESS REPORT

g

UNCLASSIFIED i
¥-16302

 

Fig. 6.7. Intermediate Heat Exchanger No. 3 in
an Advance Stage of Fabrication.

UNCLASSIFIED
Y-16005

 

Fig. 6.8.
Heat Exchanger Welded and Brazed Joints.

Job Sample of Typical Intermediate

144

Intermediate Heat Exchanger Job-Sample
Evaluations

Job samples containing welded and brazed joints
of the types used in intermediate heat exchanger
No. 3 are being submitted by outside vendors for
examination and evaluation. A typical job sample
is shown in Fig. 6.8.

The butt welds of job samples are subjected to
radiographic examination,
graphic sections are made of typical joints for

and several metallo-
examination and evaluation, Analyses of different
areas of a joint have revealed areas of adequate
quality and areas of definitely unsatisfactory
quality. Examples of these conditions in tube-to-
header welds of typical job samples are shown in
Figs. 6.9 and 6.10. The condition illustrated in
Fig. 6.10 is extremely serious and could not be
seen by visual examination. Severe weld cracking
was found in a radiator job sample, as shown in

- U&Cl.fi‘@&ififib:
Y -

6607

 

Fig. 6.9. Tube-to-Header Weld Showing Adequate
Penetration. Etched with oxalic acid. 24X.
Fig. 6.10. Tube-to-Header Weld Showing Un-
satisfactory Porosity. Etched with oxalic acid.

24X,

Fig. 6.11. Root cracking in the welds of the heat
exchanger samples was also found to be a problem,
as shown in Fig. 6.12. Other joint defects, such
as porosity, oxide inclusions, and inadequate
brazing-alloy flowability, are also frequently found
in job samples,

Examination of NaK-to-Air Radiators That
Failed in Service

A 500-kw high-conductivity-fin NaK-to-air radiator
fabricated by the Metallurgy Division failed on
July 24, 1955, as a result of a leak. This radiator,
known as ORNL No. 1, had been in service at an
elevated temperature for approximately 600 hr at
the time of failure. During almost all the 600-hr
period the system was isothermal, with temperatures
in the range of 1000 to 1600°F. For approximately
66 hr of the total service time, the radiator operated

PERIOD ENDING DECEMBER 10, 1955

 

Fig. 6,11, Tube-to-Header Weld Showing Severe
Cracks. Etched with oxalic acid. 21X.

 

Fig. 6.12. Intermediate Heat Exchanger Root

Weld Showing Cracks.
100X. Reduced 29%.

Etched with oxalic acid.

145
ANP PROJECT PROGRESS REPORT

 

UNCLASSIFIED
Y-16065

Fig. 6.13. NaK-to-Air Radiator, ORNL No. 1, Which Failed Because of a Leak After 600 hr of Service

in the Tempetature Range of 1000 to 1600°F.

with the blower on, and the temperature difference
between the inlet and outlet NaK was 50 to 300°F.
The radiator is shown in Fig. 6.13 as it appeared
when received from the test site.

On September 6, 1955, one of the first 500-kw
high-conductivity-fin radiators fabricated according
to ORNL. specifications by the York Corp. also
failed as a result of a leak. This radiator survived
approximately 150 hr of elevated-temperature service.
During 80 hr of operation the system was isothermal,
and there was a temperature difference in the NaK
circuit duwing the remaining 70 hr of operation.
Thus the service lives of the two radiators differed
significantly only in the extent of isothermal service.

The York Corp. radiator is shown in Fig. 6,14 as
it appeared when received from the test site. Since
this failure was promptly detected and the NaK
system was emptied immediately, the destruction
caused by fire was less than that caused by the
fire on the ORNL unit., The fire damage to the

146

UNGLASSIFIED
Y«16540

 

Fig. 6.14. York Corp. NaK-to-Air Radiator No. 1,
Which Failed After Approximately 150 hr of

Elevated-Temperature Service in the Range 1000
to 1600°F.
UNCLASSIFIED
Y-16246

 

Fig. 6.15. Sections of ORNL Radiator No.
Showing Fire Damage.

 

PERIOD ENDING DECEMBER 10, 1955

ORNL radiator was so extensive that a determina-
tion of the cause of the leak proved to be impos-
sible. The sections of the ORNL radiator shown in
Fig. 6.15 were so badly damaged by the fire that
27 gross leaks were detected when the sections
were tested with water under pressure,

The partly sectioned York radiator is shown in
Fig. 6.16, and an enlargement of the photograph of
section 5 is shown in Fig. 6.17. Each tube in
section 5 was tested with water under pressure,
and the leak was located by observing the bubbles
which emerged from the intersection of one of the
tubes and the support member. The exact location
of the leak is shown in Fig. 6.18. The arrow indi-
cates the point of emergence of the water bubbles
during the pressure test, This portion of section 5
was mounted intact for metallographic examination,
and it was examined microscopically at frequent
intervals while it was being ground down to the

UNCLASSIFIED
Y-16541

Fig. 6.16. York Radiator No. 1 After Being Sectioned for Examination. Arrow in section 5 points to

location of failure.

147
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-16573

 

Fig. 6.17. Enlargement of Photograph of Section 5 Showing Location of Failure.

hole through which NaK had leaked. Although
time-consuming, this procedure was deemed neces-
sary in order ta ensure that metallographicevidence
of the failure would not be destroyed.

Longitudinal sections through the tube that failed
are shown in Figs. 6.19, 6.20, and 6.21. The
ductile fracture and the reduction in tube wall
thickness of approximately 23% indicate that the
failure was due primarily to tensile forces. A
lateral shift of the tube is also evident, however,
which indicates that shear forces may have been
instrumental in in the braze metal a
fracture that was subsequently propagated by the
tensile forces.

The tensile loading appears to have been the
result of the restraining influence of the support

initiating

members during the cooling of the fin surfaces by
forced air. Buckling of these components is clearly
evident in Figs. 6.14, 6.16, and 6.17. It may also
be noted that the tube that failed was securely

148

brazed to the support member and to the bottom
flanged plate, which were brazed to each other.
Therefore shear loading may have resulted from the
restraining influence of the bottom flanged plate
during the cooling of the fin surfaces by forced air,

A tube adjacent to the tube that failed was also
microscopically examined for evidence of fracture,
This tube had been found to be leaktight in the
pressure test, but, as may be seen in Fig. 6.22, a
fracture was being propagated in this tube also.
The reduction in tube wall thickness and the
lateral displacement characteristic of the tube that
failed may also be noted in Fig. 6.22. It is ap-
parent therefore that the failure was not due to an
isolated flaw. The examination was extended to
include the two areas shown in Fig. 6.23, which
were not near the tube that failed. A joint in area A
of Fig. 6.23 is shown in Fig. 6.24. [t may be noted
that a fracture had gone through the braze fillet and
was partly into the tube wall,
L ek

UNCLASSIFIED
Y-16574

 

Fig. 6.18. Point of Foailure of York Radiator
No. 1 As Detected by the Emergence of Water
Bubbles Under Pressure.

It is evident that York Radiator No. 1 failed be-
cause of the initiation of a fracture in a braze fillet
by shear forces and the propagation of this fracture
through the tube wall by tensile forces or combina-
tions of shear and tensile forces. The tensile
loading was caused by differences in cooling rates
between the support members and the finned tubes
when cooling air was forced across the high-
conductivity-fin surfaces. The shear forces were
caused by a difference in cooling rate between the
support member and bottom flanged plate and the
finned tubes. Modifications have been made to the
radiator design to relieve the tensile and shear
forces.

PERIOD ENDING DECEMBER 10, 1955

BNCLASSIFIED
73

Y-l

 

Fig. 6.19. Longitudinal Section Through Tube
That Failed. Note that the tube was securely
brazed to both the support member and the bottom
flanged plate.

BRAZING-ALLOY DEVELOPMENT

P. Patriarca R, E. Clausing
Metallurgy Division

Brazing alloys for use in high-temperature sodium-
to-air applications should be compatible with both
liquid sodium and air, since extensive tube wall
dilution and diffusion during brazing and subse-
quent corrosion through an area of shallow weld
penetration during service might result in a high
concentration of brazing alloys being exposed to
the liquid metal, Precious-metal brazing alloys
were known to be incompatible with sodium and

149
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
74

 

 

o o r ' Io o
]o lb o INCHES o !9.
- el » 1 w

Fig. 6.20. Enlargement of Photograph of Tube
Wall Failure Shown in Fig. 6.19. Note the typical
tensile failure,

were therefore eliminated from consideration, as
were brazing alloys with poor oxidation resistance.

An analysis of the corrosion data obtained to
date (Sec. 5, ''Corrosion Research’’) indicates
that the following alloys may be considered for
use in radiator fabrication: Coast Metals alloys
Nos. 50, 52, and 53; standard and low-melting
Nicrobraz; General Electric alloy No. 81; and an
alloy consisting of 80% Ni-10% Cr—10% P. The
Coast Metals alloy No. 52 is of particular interest
because its flow temperature of 1020°C is suffi-
ciently low to permit its use on radiators that
have stainless-steel-clad copper high-conductivity
fins. A brazing temperature higher than the melting
point of copper is required for many of the other
alloys; therefore their application is limited to
radiators that have fins of a more conventional
composition,

150

       

 

30

 

Fig. 6.21, Enlargement of Photograph of Tube
Wall Opposite the Point of Failure. Note fissure
partly propagated through the tube wall.

An analysis of the corrosion data obtained to
date for brazing alloys exposed to fluoride mixtures
and to sodium indicates that the following alloys
may be considered for use in heat exchanger fabri-
an alloy consisting of 80% Ni-10% P—
10% Cr; standard and low-melting Nicrobraz; Coast
Metals alloys Nos. 50, 52, 53, and NP; an alloy
consisting of 70% Ni-13% Ge-11% Cr—-6% Si; and
an alloy consisting of 50% Ni-25% Ge-25% Mo.
The 80% Ni-10% P—10% Cr alloy has limited appli-
cation, even though it has good corrosion resistance
to both liquids, because of its brittleness. Since
it is important for the materials used in the fabri-
cation of heat exchangers to have low nuclear
cross sections, the 70% Ni-13% Ge-11% Cr—6% Si

alloy is of particular interest,

cation:

The optimum conditions for the use of the metal-
spray process as a means of preplacing brazing
 

 

*
o
O
x |
407
23'0

 

o o * o o
[b l° INCHES l'n ]g : !
& { "~

Fig. 6.22. Fracture in Tube Adjacent to Tube
That Failed.

UNCLASSIFIED
Y-16560

 

Fig. 6.23. Additional Areas That Were Examined
for Fractures, '

PERIOD ENDING DECEMBER 10, 1955

 

 

 

v 5 S ! T
g !° o Ig INCHES }g 5

—F
: ,sox Ig

 

Fig. 6.24. Fracture Started Through Tube Wall
in Area A of Fig. 6.23.

alloy on heat exchanger and radiator components
are now being studied. Surface roughening by
sandblasting or other methods appears to improve
the bonding of the sprayed material, Tube-to-header
specimens with excellent braze fillets have been
prepared by placing short tubes in the drilled
header holes during spraying of Coast Metals alloy
No. 52. After the tubes were exiracted and traces
of brazing alloy were removed from the hole periphery
by reaming, the header plate was assembled with
tubes of the desired size and brazed. A spray
thickness of approximately 0.010 in, was found to
result in good filleting, but countersinking of the
header on the sprayed side may vary the optimum
thickness slightly,

Areas of the header which do not require spraying
can be effectively masked with glass or asbestos
tape. Conventional masking tape and aluminum
oxide powder were found to be unsatisfactory for
this purpose,

151
ANP PROJECT PROGRESS REPORT

MECHANICAL-PROPERTY STUDIES OF
NICKEL-MOLYBDENUM ALLOYS

D. A, Douglas
J. R, Weir J. W. Woods
Metallurgy Division

C. R. Kennedy, Pratt & Whitney Aircraft
Alloys

being investigated in the search for alloys with
sufficient corrosion resistance and high-temperature
strength for use in circulating-fuel aircraft reactors,
In a series of creep-rupture tests, Hastelloy B has
been shown to have good high-temperature strength
and corrosion resistance. However, over a portion
of the intended service-temperature range it ex-
hibits very poor ductility because of its age-
hardening characteristics. Although the age
hardening markedly improves the creep strength of
the material, the resultant loss of ductility restricts
its usefulness in the 1200 to 1500°F temperature
range,

in the nickel-molybdenum system are

Influence of Aging Heat Treatments on the
Creep Properties of Hastelloy B

The effect of age hardening on the creep charac-
teristics of Hastelloy B has been investigated in
tests of materials in the solution-annealed and the
age-hardened conditions at temperatures from 1300
to 1800°F,
shown marked reductions in initial creep rate and
in tensile elongation when compared with the solu-
tion-annealed specimens. Aged specimens tested
at 1500°F show an increase in time-to-rupture,
with a corresponding decrease in ductility, when
compared with the solution-annealed specimens.
Creep curves for solution-annealed and aged speci-
mens tested at 1500°F and a stress of 12,000 psi
are presented in Fig. 6.25. It would be predicted
from the nickel-molybdenum phase diagram that
Hastelloy B would be a solid-solution alloy at
1650°F and that there would be little difference in
the creep properties of aged and solution-annealed
specimens. However, as may be seen in Fig. 6.26,
the initial creep rates are lower and third-stage
creep is initiated much earlier for the aged speci-
mens than for the annealed specimens. The re-
sult is a somewhat shorter rupture life and less
total elongation at failure for the aged material,
In the tests at 1800°F, the specimens were in the

In general, the aged specimens have

152

UNCLASSIFIED
ORNL-LR—-DWG 11666

 

100

50 :

ANNEALED
20
AGED 40 hr AT 1500°F

10 ‘
o
z 5
o
b
<l
2
zZ 2
—
w

1
iLL)
0.5 :
AGED 100 hr AT 1300°F
AGED 10 hr AT 1500°F
0.2 -
0.1
10 100 1000 10,000
TIME (hr)
Fig. 6.25, Creep-Rupture Curves for Solution-

Annealed and Aged Hastelloy B Sheet Tested at
1500°F in Argon at a Stress of 12,000 psi.

UNCLASSIFIED
ORNL—-LR—DWG 11667

100 — -
50 | —- : el .
— : [ 110
SOLUTION ANNEALED, TESTED AT 40,000
20 * PR T
5 100 hr AT 1300°F
b AT 10,000 psi
10 b -

ELONGATION (%)

AGED 100 hr AT 1300°F
05 _ 7 TESTED AT 7500 psi

J - do

1
SOLUTION ANNEALED, TESTED AT 7500

 

a
12 5 10 20 50 100 200 500 1000
TIME {nr)

Fig. 6.26. Creep-Rupture Curves for Solution-

Annealed and Aged Hastelloy B Sheet Tested at
1650°F in Argon at Stresses of 7,500 and 10,000

PSio
alpha solid-solution range, and the pattern shown
in Fig. 6.26 was again obtained, as shown in Fig.
6.27. This rather peculiar behavior can be ex-
plained by considering the rates of solution of
precipitates previously formed by aging., Hardness
tests and metallographic examinations have shown
that the molybdenum-rich phase is very slow to
precipitate from a supersaturated lattice and slow
to redissolve into the matrix when heated into the
solid-solution temperature range. |t has also been
shown that strain tends to accelerate these re-
actions, Thus it would be expected that the aged
specimens would exhibit lower initial strains and
creep rates at 1650 and 1800°F than would the
solution-annealed material, since the precipitates
would not have had time to go back into solution
As the test
progresses, the precipitates in the grain-boundary
regions tend to go into solution faster than those
within the grains because of the greater strains
at the boundary regions., This results in the weak-
ening of the grain boundaries occurring at a faster
rate than does the weadkening of the bulk of the
Chang and Grant3 have observed that the

during the early portion of the test.

grains.

 

34, C. Chang and N. ). Grant, Mechanism of Inter-
crystalline Fracture, unpublished manuscript.

UNCLASSIFIED
ORNL-LR-DWG 11868

100

50 |- SOLUTION ANNEALED, TESTED AT 3500 psi—
{

 

20 -
SOLUTION ANNEALED, TESTED AT
& 10| 5000 psi
& ;
£ 5 | AGED 100 hr AT 1300°F, TESTED AT
o 5000 psi
=
O
—
)
i
05
I '
0.2 AGED 400 hr AT 4300°F, TESTED A
3500 psi \
o4
{ 2 5 10 20 50 100 200 500 1000
TIME (tr)

Fig. 6.27. Creep-Rupture Curves for Solution-
Annealed and Aged Hastelioy B Sheet Tested at
1800°F in Argon at Stresses of 3500 and 5000 psi.

PERIOD ENDING DECEMBER 10, 1955

initiation and propagation of intergranular cracking
is controlled by the ability of the grains to accom-
modate the deformation produced by grain-boundary
sliding. In the case of solution-annealed material
there are no precipitates to inhibit grain-boundary
sliding, and the rate of deformation is more rapid.
However, the grains can accommodate a consider-
able amount of deformation before intergranular
cracking begins, and the rate of propagation is
relatively slow. In the aged material there is
considerable resistance to deformation until the
precipitates in the grain boundaries begin to go
into solution. When this occurs, the rate of strain-

However, the grains still contain
and their ability to accommodate

ing increases,
precipitates,
deformation is considerably less than that of the
grains in the solution-annealed material. Thus
intergranuiar initiated in the aged
specimens fairly early in the test period, and propa-
gation of these cracks is rather rapid; the result
is the acceleration of the third-stage creep of the
aged material.

cracking is

Preliminary Investigation of Creep Properties
of Hastelloy W

Preliminary testing of Hastelloy W is now in
progress in order to evaluate the creep properties
of this alloy, which has approximately the same
composition as that of Hastelloy B, except that 3%
of the molybdenum and 2% of the nickel have been
replaced with chromium. Results obtained thus far
indicate that the age-hardening characteristics of
Hastelloy W are different from those of Hastelloy B
and that there may be an increase in ductility in
the 1200 to 1500°F temperature range in comparison
with the ductility of Hastelloy B in that temperature
range.

Investigation of the Creep Properties of Some

New Nickel-Molybdenum Alloys

New alloys of the nickel-molybdenum alloy system
are being investigated in a search for a structural
material with the excellent corrosion resistance
and high-temperature strength of Hastelioy B but
without the objectionable brittleness of Hastelloy
B in the temperature range of 1200 to 1500°F. The
results of tests of time to 1, 2, and 5% elongation
and of time-to-rupture for the various vacuum-
melted alloys studied in the past few months are
compared in Fig. 6.28. The alloys tested were
solution annealed at 2100°F. All tests were
performed in an atmosphere of argon at 1500°F,

153
1451

10,000
80C0

6000

4000

2000

1000
800

600

400

TIME (hr)

200

100
80

60

40

20

UNCLASSIFIED
ORNL-LR-DWG 11662

 

 

INCONEL HASTELLOY
B

~ NOTE:

VALUES WITHIN BARS INDICATE PERCENTAGE ELONGATION.

VALUES AT THE TCPS OF BARS INDICATE ELONGATION
AT TIME OF RUPTURE.

 

W
¥
2 g
5
=z Z
o e _ T - o

_ . 5 o o o

1
- — Bz - - L ;

— - Y - CLEAN SINGLE SOLID SOLUTION _— -
o

—_— j O —_ ——— )\
S = f

o o= _ - o _ _
w =
o9
_ - - Se—— — _ -
o=
L E
- o _ O - - - .

i
aa

T8 NI 68 % Ni 80 % Ni 76 Yo Ni 74.5 % Ni 75% Ni

20% Mo 28 % Mo 20 % Mo 24 % Mo 20% Mo 20% Mo

2% Vv 4% Fe 5% Cr 5% Cr
0.5% Nb

77 % Ni
20 % Mo
3% Cr

78 % Ni
20 % Mo
2 % Nb

OF STRUCTURE

 

TYPE

 

 

 

73 % Ni
20 %5 Mo
7% Cr

ALLOY COMPOSITION AND MICROSTRUCTURE AFTER TESTING

68 % Ni
32 %Mo

75 % Ni
20 % Mo
5 % Nb

 
 

I
0w =
-Qu=
—
iiollg 4
L — mE@
W
=z =
|
Q—— ogqflfi
= =z = =
D o = ]
o “ERg%
O ZF—ED
_® 532z ]
o “EER
1 O oo
o
U—)ii el ™
Y 5%
D
£
)
3
—
—
=

  
  

 

 

74.25 7 Ni 78 %o Ni

20 % Mo 20 % Mo
5% Cr 2 % Al
0.25% Ce

Fig. 6.28. Comparison of Times to1, 2, and 5% Elongation and to Rupture for Several Vacuum-Me lted Solution-Annealed Nickel-Molybde num
Alloys, Inconel, and Hastelloy B Stressed at 8000 psi in Argon ot 1500°F.

L¥0d3Y SSITYO0AYd LD23r0dd dNV
e e dbmh by . e

[T

with the solution-annealed specimen stressed at
8000 psi. Data on Inconel and Hastelloy B are
included for comparison.

The low ductility of some of the alioys indicates
that the vacuum-melting technique used in their
preparation did not completely remove gases from
the material, This conclusion is strengthened by
a comparison of the ductility of the alloys with
and without added deoxidants, such as cerium or
aluminum, The addition of 0.25% cerium to the
75% Ni-20% Mo-~5% Cr alloy increased the duc-
tility from 3 to 28% elongation at rupture, and the
addition of 2% aluminum to the 80% Ni-20% Mo
alloy increased the ductility from 2 to 35% elonga-
tion at rupture; the times-to-rupture for these alloys
were correspondingly increased by the addition of
the deoxidants,

In general, it appears that the chromium content
of a nickel-molybdenum alloy must be at least 7%
before there is an appreciable increase in strength
and that an alloy containing 5% niobium, because
of the resulting two-phase structure, is signifi-
cantly stronger than a similar alloy with only 2%
niobium, It should be noted that, despite the poor
ductility of these alloys, their stress-rupture
properties are equal to or superior to those of
Inconel.

Several trends may be noted in the correlation
between the microstructure and the creep properties
of these alloys. Precipitation of an insoluble
second phase takes place very slowly in these
nickel-molybdenum alloys, and therefore the micro-
structure may change considerably during a test of
long duration. Such a microstructure change may
be seen by comparing Fig. 6.29 with Fig. 6.30.
Figure 6.29 shows a 75% Ni-20% Mo-5% Cr alloy
which ruptured in 90 hr with an elongation of 3%
at a stress of 8000 psi. As may be seen, the alloy
was still in the solution-annealed condition after
rupture. However, the microstructure (Fig. 6.30)
for a specimen with the same composition, which
ruptured in 1700 hr with an elongation of 7.4% at
a stress of 5000 psi, shows precipitation of a
second phase on slip planes.

The inherently fine-grained structure that results
from adding 0.5% Nb to a 75% Ni~20% Mo-5% Cr
alloy is shown in Fig. 6.31. A five-grained struc-
ture also results from the addition of 0.25% Ce
to this alloy, as shown in Fig. 6.32, The micro-
structures of several of the other alloys after they
were tested are shown in Figs. 6.33 t0 6.40, |In

PERIOD ENDING DECEMBER 10, 1955

Figs. 6.29 through 6.40 both stressed and un-
stressed portions of the specimen are shown,
except for Fig. 6.36, in which only the stressed
portion is shown,

SPECIAL MATERIALS STUDIES

J. H. Coobs J. P. Page
H. Inouye T. K. Roche
Metallurgy Division

M. R, D'Amore, Pratt & Whitney Aircraft

Extrusion of Seamless Duplex Tubing

T. K. Roche H. Inouye
Metallurgy Division

Preliminary work in developing seamless duplex
tubing has progressed to the extent that severadl
composite billets have been extruded to tube blanks,
Little difficulty was experienced in extruding
nickel-, Inconel-, and Monel-clad type 316 stain-
less steel billets; however, the extrusion of
Hastelloy B-clad type 316 stainless steel remains
problematical, Surface cracking of the Hastelloy B
occurred, which was apparently indicative of the
hot shortness observed previously in the extrusion
of this alloy. It had been concluded from past
experiments that the successful extrusion of
Hastelloy B was largely temperature dependent and
that an extrusion temperature of approximately
2000°F was optimum for the prevention of defects.
However, the difficulties recently encountered even
at 2000°F indicate that the melting practice and
the billet lubrication must be given more careful
consideration if Hastelloy B is to be successfully
extruded,

Forged billets of Hastelloy W could not be ex-
truded as tube blanks at 2000°F. Two inconel-
canned billets of this material shattered, even at
a moderate reduction ratio, and one uncanned billet
stalled the press. It is felt that the extrusion of
the uncanned billet failed because of an excessive
drop in the billet temperature that was due to faulty
preheating practice.

Flame-spraying of Hastelloy B with type 304
stainless steel for lubrication improved the ex-
trudability slightly. Two extruded tube blanks of
this type had roughened surfaces, and, in addition,
one showed signs of surface cracking, The results
of the extrusion experiments mentioned above are
presented in Table 6.1.

155
ANP PROJECT PROGRESS REPORT

  
   

UNCLASSIFIED
L ¥-15682
\\ . fol . Lo

      

SED |

- UNSTRES |

A

o

 

 

 

 

Fig. 6.29. A 75% Ni-20% Mo-5% Cr Alloy Specimen Which Ruptured in 90 hr with an Elongation of 3%
at a Stress of 8000 psi in Argon at 1500°F., 100X. Reduced 22%.

& UNCLASSIFIED
¥-15860

 

 

 

Fig. 6.30. A 75% Ni~20% Mo-5% Cr Alloy Specimen Which Ruptured in 1700 hr with an Elongation of
7.4% at o Stress of 5000 psi in Argon at 1500°F. 100X. Reduced 22%.

156
e g e e

e e

PERIOD ENDING DECEMBER 10, 1955

L UNCLASSIFIED
Y5677

      

  

" UNSTRESSED . -

 

 

Fig. 6.31. A 74.5% Ni=20% Mo-5% Cr=0.5% Nb Alloy Which Ruptured in 77 hr with an E longation of
7.6% at a Stress of 8000 psi in Argon at 1500°F, 100X. Reduced 20%.

UNCLASSIFIED
Y-15706

 

 

= s Lo . . S . e ‘
> ‘ - 100X o ]'o INCHES ‘ - l
" } | . i

 

Fig. 6.32. A 74.75% Ni=20% Mo=5% Cr—0.25% Ce Alloy Which Ruptured in 440 hr with an Elongation
of 28% at a Stress of 8000 psi in Argon at 1500°F. 100X. Reduced 20%.

157
ANP PROJECT PROGRESS REPORT

 

 

 

TINCHES 00X — L

Fig. 6.33. An 80% Ni~20% Mo Alloy Which Ruptured in 91 hr with an Eiongation of 2% at a Stress of
8000 psi in Argon at 1500°F, 100X. Reduced 23%.

UNCLASSIFIED 1
72

 

 

 

TNCHES 106X

Fig. 6.34. An 80% Ni~20% Mo Alloy Which Ruptured in 763 hr with an Elongation of 5% at a Stress of
5000 psi in Argon at 1500°F. 100X. Reduced 23%.

158
[

~
‘\‘ {J"—-,._“ . }/ 5

.STRESSED "

 

 

     
 

. A i

PERIOD ENDING DECEMBER 10, 1955

: . . .
\'\ ] UNCLASSIFIED
‘ Y-14258

 
   

 
 

 

 

 

Fig. 6.35. A 76% Ni-24% Mo Alloy Which Ruptured in 68 hr with an Elongation of 4% at o Stress of

8000 psi in Argon at 1500°F. 100X. Reduced 22.5%.

 

 

A UNCLASSIFIED
Y3 Y1261 3
v
poe 14 —4
x
<
=z
0.02
Q.03
»
e 10
2

 

 

Fig. 6.36. A 68% Ni-32% Mo Alloy Which Ruptured in 332 hr with an Elongation of 8% at a Stress of

8000 psi in Argon at 1500°F. 100X. Reduced 22.5%.

159
ANP PROJECT PROGRESS REPORT

. UNCLASSIFIED |
Y-14706

   
 
     
    

A

 

o ! o o 1
ii: iI00X o o INCHES
X ° 1

 

Fig. 6.37. A 68% Ni-32% Mo Alloy Which Ruptured in 715 hr with an Elongation of 33% at a Stress of
5000 psi in Argon at 1500°F, 100X. Reduced 23%.

 

   

. 5 P S Nos . ‘ N UNCLASSIFIED
ro S e 5 o ‘ : NG
; LR .“(r ‘l.\\-(-«"r . 3 . ’ . g "
| STRESSED . [

 

L #‘ T

 

 

 

1 . - T
I?@CHES S

 

Fig. 6.38. A 77% Ni-20% Mo-3% Cr Alloy Which Ruptured in 674 hr with an Elongation of 8% at o
Stress of 5000 psi in Argon at 1500°F. 100X. Reduced 23%.

160
PERIOD ENDING DECEMBER 10, 1955

  

UNCLASSIFIED
Y-15689

 

| - STRESSED ' /}
AT 4
- A A\
> VI"\_‘ i }; e .. - .\"\ i
* ~ g . \t”’. s s
e T w o
e e }
14 | S LNl
- oy L
s T |
& - \\' - H"a | -
. \: . ‘..(L
. th
.V",‘ ) ‘\k ‘ "
A i i 3
. i ) ‘ - T s :; “‘ \
A \ Yoo
™ T \ |
§ e f’/ ] .
o 1
{3 le x

Fig. 6.39.

 

 

A 73% Ni=20% Mo=7% Cr Alloy Which Ruptured in 373 hr with an Elongation of 10% ata

Stress of 8000 psi in Argon at 1500°F. 100X. Reduced 21%.

 

Figo 6.40.

N WNCLASSIFIED\
, % Y-14126
J

 

 

A 75% Ni-20% Mo—5% Nb Alloy Which Ruptured in 530 hr with an Elongation of 8% at a

Stress of 8000 psi in Argon at 1500°F. 100X. Reduced 21%.

161
ANP PROJECT PROGRESS REPORT

TABLE 6.1, RESULTS OF EXTRUSION OF TUBE BLANKS

Billet size:

3 in. in diameter

 

 

Numb Soaking Reduction
umber Composition of Billet Temperature ) Remarks
E xtruded o Ratio
("F)
3 Nickel~type 316 stainless steel 2100 9:1 Good tube blanks obtained
(duplex tube)
3 Inconel—type 316 stainless steel 2100 9:1 Good tube blanks obtained
(duplex tube)
3 Monel~type 316 stainless steel 2050 9:1 Good tube blanks obtained
(duplex tube)
3 Hastelloy B—type 316 stainless 2050 9:1 Surface of Hastelloy B was
steel (duplex tube) somewhat roughened in
all three extrusions; two
extrusions showed evi-
dence of cracking
2 Hastelloy B (flame —sprayed 2050 9:1 Surfaces of both extrusions
with type 304 stainless steel) were somewhat roughened;
one extrusion cracked
2 Hastelloy W (Inconel canned) 2050 7:1 Both extrusions shattered
1 Hastelloy W (uncanned) 2050 5,5:1 Did not extrude

 

Six laboratory-prepared alloys were successfully
extruded to rod form for corrosion studies, The
compositions of the billets extruded and the re-
duction ratios used are tabulated below (the soaking
temperature used was 2150°F in each case):

Billet Composition (wt %) Reduction Ratio

80 Fe-20 Cr 13:1
80 Fe-10 Cr—10 Ni 10:1
74 Fe—-18 Cr-8 Ni 10:1
60 Fe-20 Cr—-20 Ni 10:1
90 Ni—-10 Mo 10:1
80 Ni-10 Mo=10 Fe 10:1

Neutron Shielding Material for ART
H. Inouye, Metallurgy Division
M. R. D’Amore, Pratt & Whitney Aircraft
Methods for producing B,C tiles to be used as
neutron shielding for the ART are being studied.
The Norton Company and The Carborundum Company
are cooperating in this investigation. Since the

neutron shield will be at a temperature of 1500 to
2000°F when the reactor is operating, studies are

162

also being made of the compatibility of Inconel
and B,C in this temperature range. Previously
reported studies indicate that carburization occurs
at the grain boundaries and that metallic borides
are formed. A coating, or barrier, material that is
inert both to B, C and to Inconel will therefore be
required. Tests of possible barrier materials are
under way.

Studies are also being conducted to determine
the effects of irradiation on B, C at the expected
service temperatures. Neutron capture by boron
results in the production of helium and lithium,
which are reported4 to be retained interstitially in
the B C lattice at temperatures below 1650°F.
There is evidence, however, that the helium and,
possibly, the lithium are released at some higher
temperature,

Plates of B,C with a metallic binder are also
being investigated, since they would have better
thermal-conductivity, expansion, and shock-re-
sistant properties. The plates can be fabricated

 

4p, Senio and C. W. Tucker, Jr., The Stability of
Boron Carbide After 15 Atomic Percent Burnup, KAPL-
1091 (April 5, 1954).
by wusing the picture-frame technique to hot roll
cold-pressed, sintered, powder mixtures or by ex-
truding warm-pressed billets of the powder mix-
tures, A layer, equivalent to 0,010 in. of B9,
that is thermally bonded to the inside wall of the
Inconel shell will be required if the metal-bonded
B C plates are used.

ests of the reactions between B4C, binder, and
Inconel shell at the service temperature are under
way, and possible methods for applying the thin
layer to the shell are being investigated.

Inconel-Clad Niobium

H. lnouye J. P. Pagg
Metallurgy Division

The search for a diffusion barrier to use between
Inconel and nicbium in high-temperature applica-
has been continued. Tantalum foils and
copper—stainless steel foils were selected as the
most promising of several barrier materials on the
basis of bend tests and metallographic examination
of pieces in the as-rolled condition and after 100-
and 500-hr service tests at 900°C. These types
of qualitative tests are being used to check earlier
data obtained with molybdenum and titanium as
the barrier materials and to obtain data on other
barrier materials (iron, at present). A test pro-
cedure is also being developed for obtaining quan-
titative data on the mechanical properties of Inconel-
clad niobium with tantalum and copper—stainless
steel diffusion barriers. When a satisfactory test
program has been devised, data will be obtained to
determine the optimum diffusion barrier,

tions

Gamma«Ray Shield Material for ART Pumps

J. H. Coobs J. P. Page
Metallurgy Division

A material having the following properties is
needed for use as a gamma-shield heat dam around
the impeller shafts of ART pumps: density of
over 13.0 g/cm3, thermal conductivity of less than
0.12 cal/cm2:sec:(°C/cm) at 1400°F, brazeability
to Inconel, and a coefficient of thermal expansion
similar to that of Inconel, that is, between 10 and
20 pin./ins°C, There are no known materials that
fulfill all these requirements, but it is thought that
a satisfactory two-component material can be fabri-
cated by powder-metallurgy techniques. A high-
density ‘‘filler’’ material would be used, along with
another material which would act as a binder and
would ‘“‘insulate’’ the compact. From considera-

PERIOD ENDING DECEMBER 10, 1955

tions of thermal conductivity, coefficient of ex-
pansion, brazeability, and ease of fabrication, the
nickel-base alloys appear to be the most promising
for use as binder material, especially the Hastelloys
and constantan. Calculations of the limiting cases
of series and parallel heat flow indicate that the
filler material must be either tungsten carbide or
tantalum,

Several tantalum-constantan compacts have been
successfully hot-pressed to densities of greater
than 13 g/cm3, The tantalum-base material is
far superior to the tungsten carbide—base material

from the standpoint of fabricability. However,
metallographic examination indicates a strong
tantalum-nickel reaction that results in virtual

separation of the nickel from the constantan. This
separation could cause variations in the thermal
conductivity and could cause the thermal con-
ductivity to be above the specified limit. Similar
investigations of a tungsten carbide—constantan
compact shows that the carbide is held in a homo-
geneous constantan matrix. An apparatus is being
fabricated for use in determining the thermal con-
ductivities of these and similar mixtures.

Control-Rod Fabrication

J. H. Coobs H. Inouye
Metallurgy Division

Two core compacts for control rods were pre-
pared by cold pressing and sintering 35 vol %
(Gd-Sm), 0, in copper and in iron to densities of
79 and 82%, respectively. The iron mixture was
sintered at 1100°C, and the copper mixture was
sintered at 980°C, The core compacts were then
canned in stainless steel, were evacuated, and
were hot-swaged for a total reduction of 75% in
seven passes, with the compact being reheated to
950°C between passes. The active sections of the
finished rods were ]/2 in. in diameter and 7 in, in
length.

The cores were well compacted, with calculated
densities of 97 and 98.5% for the copper and the
iron mixtures, respectively. The core was thermally
bonded to the stainless steel core, but it was
slightly irregular in cross section., Metallographic
examination showed the core components to be
compatible under the sintering and hot-swaging
conditions, No evidence was found of reaction be-
tween the iron or the copper and the (Gd-S5m),0,.
[t appears that, with slight modifications in the
techniques, control rods with a finished size of

163
ANP PROJECT PROGRESS REPORT

about 5/8 to % in. in diameter can be prepared, but
for finished rods ¥ in. in diameter, and larger,
extrusion would be preferable.

A hot-pressed body with 30 wt % (Gd-5m),0, and
70 wt % Fe was extruded in an Inconel can at
2100°F, The original compact; 3/4 in. in diameter
by 3 in, in length, with 30% porosity, was extruded
to a core 24 in. in length by 0.230 in. in average
diameter. The core-diameter variations were
+0.015 in. from the average. Examinations of the
cross section and the longitudinal section indicated
that metallurgical bonds were attained. The density
of the extruded core was 7.724 g/cm3, or 100% of
theoretical. In tensile tests at room temperature
the core elongated 3% before fracture, and the
composite 42%. At 1650°F the composite uniformly
elongated 17% before fracture. At room temperature
the tensile strength of a 0.505-in.-dia rod of Inconel
containing a 0.230-in.-dia rare-earth-oxide core
was 66,500 psi, and at 1650°F the tensile strength
was 11,800 psi.

Several iron-zirconium alloys containing between
1 and 16% zirconium were studied to determine
their physical and fabrication properties. The
zirconium in the alloy was a stand-in for hafnium,
which is o high-cross-section material that might
be useful for control rods. The alloys were melted
by induction heating in vacuum and were cast in a
split-cast iron mold. The zirconium additions were
contained in a drilled cavity in the iron melting
stock.

Chemical analyses of the ingots are presented in
Table 6.2. All the ingots (except 527-16, which
was unsatisfactory) were extruded from a 3-in.-dia
ingot to 1-in.-dia rod at 2100°F, with no difficulty.

Standard 0.505-in.-dia tensile-test bars of the
various compositions were tested at room and ele-
vated temperatures (Table 6.3). At all temperatures

TABLE 6.2, ZIRCONIUM CONTENT OF
IRON-ZIRCONIUM ALLOYS

 

Zirconium Content (wt %)

 

 

Ingot No. Top of Ingot Bottom of Ingot
52741 1.02 . 1.07
-3 3.51 3.30
-5 4.95 5.70
-8 8.36 8.93
-12 11.60 13.18
-16 14.40 14.50

 

164

TABLE 6.3. TENSILE TESTS OF IRON-ZIRCONIUM
ALLOYS AS EXTRUDED AT 2100°F

 

 

Alloy Testing Tensile Elongation
Temperature Strength
No. o ] (%)

("F) (psi)

527-1 Room 57,300 32.5

1300 10,500 80.0

1500 5,100 93.7

1650 5,800 13.8

-3 Room 64,500 21.3

1300 19,400 50.0

1500 7,900 75.0

1650 4,600 100.0

-5 Room 66,500 18.8

1300 24,900 55.0

1500 7,000 100.0

1650 5,100 102.0

-8 Room 89,600 2.5

1300 30,800 58.7

1500 11,900 68.8

1650 6,700 121.0

 

these alloys showed an increase in tensile strength
with increased amounts of zirconium. A corre-
sponding decrease in ductility accompanied the
increase in strength, Data were not obtained on
ingot 527-12 because the extruded rad was brittle
and could not be machined, The data indicate that
alloys with up to 12% zirconium can be formed
while hot, and that alloys with 8%, or less, zir-
conium can be formed at room temperature,

NONDESTRUCTIVE TESTING
R. B. Oliver

J. W. Allen K. W. Reker
R. W. McClung

Metallurgy Division

The development of the probe-coil eddy-current
equipment for the inspection of small-diameter
Inconel tubing is essentially complete. The in-
strument is shown in Fig. 6.41 in simple block-
diagram form. The exciting coil of the probe is
supplied with a constant 80-kc current by a
crystal-controtled oscillator-amplifier. The fre-
quency of 80 kc was selected as one sufficiently
high to induce a relatively large amount of eddy-
current flow in the tube wall and sufficiently low
o mm ek

 

 

 

QSCILLATOR e AMPLIFIER - CBISIESINI%E’

 

 

 

 

 

 

 

 

 

A (A—-B)
————-@—————‘ AMPLIFIER

 

 

 

PRCBE

 

 

L——

EDDY CURRENTS

 

PICKUP con_fl,[-—-—4——_——i-

j-l— EXCITING COIL

PERIOD ENDING DECEMBER 10, 1955

UNCLASSIFIED
ORNL—LR—0OWG 11542

 

 

 

 

 

 

 

 

FLAW
INDICAT!%

 

 

 

 
   

TUBING

PERSISTENT — SCREEN CATHODE -RAY QSCILLOSCOPE -

DATA POTENTIOMETER /

—— -
d TUBING

- ROTATION

 

 

 

Fig. 6.41. Diagram of System for Eddy-Current Inspection of Tubing.

b in. of

Inconel or less. The pickup-coil voltage is changed
in amplitude and phase by the changes in amplitude
and distribution of the eddy current in the wall of
the tubing. The quiescent portion of the pickup-
coil voltage is ‘‘bucked out’* by the balance
circuitry, and hence only the signal changes are

to ensure effective penetration through

measured., Since these changes are very small,
amplification of 104 to 10° is necessary before
they are fed to the cathode-ray oscilloscope for
interpretation,

It is important to note that, since a probe type
of coil is utilized, the same frequency may be used
on tubing of various outside diameters. Also,
since the operating frequency is held constant, the
sensivitity of the system remains constant even
though the outside diameter of the tubing is varied.
This system is in contrast to the system that uses
an encircling coil in which the frequency of the
encircling coil must be made to vary inversely
with the square of the outside diameter of the
tubing in order to maintain comparable sensitivities
for various diameters of tubing. The operation of
an eddy-current instrument at a constant frequency

is vastly simpler than operation of one at a variable
frequency.

The amplified changes in pickup-coil voltage
are displayed on a persistent-screen cathode-ray
oscilloscope as a function of the tubing rotational
angle. Because of the localized nature of flaws
in tubing, the flaws cause abrupt changes in signal
as they pass under the probe coil. Other variables
in the tubing which affect the coil, such as slight
changes in wall thickness, conductivity, and
permeability, are not a function of the tubing ro-
tational angle and hence are indicated by a gradual
broadening of the cathode-ray-oscilloscope picture
while the tubing is spiraling past. Eccentricity
of the tubing causes a gradual change in signal
amplitude as the tube is rotated under the probe
and produces a wedge-shaped picture on the os-
cilloscope. This type of data presentation allows
positive, rapid interpretation of the flaw signals
at high scanning speeds.

Considerable effort has gone into the develop-
ment of a probe that will allow a reasonably high
scanning rate and, at the same time, permit minute
flaws to be detected with ease. The smallest and

165
ANP PROJECT PROGRESS REPORT

most sensitive probe made to date has an effective
diameter of ¥, in,, and it was able to detect small
cracks and pin holes, With this probe it was
found that the tubing could be rotated at 300 to
400 rpm without loss in definition of signals,
Because of the small diameter the longitudinal
speed of the probe, even at 300 to 400 rpm, is
quantity Therefore
elongated probe coils are now being tested, and it
is thought that a probe coil ]/2 in, long and ]/16 in,
wide can be utilized without compromising the
sensitivity to small flaws. A probe ]/2 in. long
will allow scanning rates of 150 in./min,

very slow for inspection,

The Cyclograph is also proving to be an extremely
useful instrument for cursory inspection of small-
diameter Inconel and Hastelloy tubing., It is being
used extensively to sort out sections of Hastelloy
B tubing that contain large flaws, that is, flaws
which penetrate 10%, or more, of the wall thickness,
prior to more detailed inspection by other methods.
With the use of the Cyclograph it has been possible
to inspect all the Hastelloy B tubing and to reduce
substantially the inspection time and cost, The
Cyclograph indications obtained, to date, in the
inspection of much of the small-diameter Inconel
tubing have shown cyclic variation over the length
of the tubing. The variations were, at first, thought
to be the result of very small changes in wall
thickness that were not detected by any other
method. Subsequent metallographic examinations
of representative sections located by these varia-
tions in the indications revealed intergranular
attack on the inside wall of the tubing. The attack
penetrated as deep as 0.002 in. A similar, eyclic-
defect signal has revealed, in the tube wall, radial
cracks which had been sealed over by the grinding
operation used to produce the finished surface.

The use of ultrasound as a tool for the inspection
of small-diameter tubing also appears to be very
practical. The “B’’ scan® which presents the
defect signal on a cathode-ray tube as a function
of the tube rotation has been found to be the most
satisfactory method for interpreting ultrasonic data
(Fig. 6.42).

A pilot model of the equipment for uitrasonic
scanning, which will handle tubing in lengths up to
7 ft 6 in,, has been set up in a large tank and is
now ready for operation. The mechanical design
principles are very similar to those of the large-

 

SR. B. Oliver et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 139.

166

scale production inspection unit now being fabri-
cated, which will accommodate tubing in lengths
up to approximately 28 ft.

Experience has indicated that the position or
alignment of the ultrasonic transducer with reference
to the tubing being inspected does not critically
affect the flaw indication. Thus the misalignment
caused by the characteristic whipping and wobbling
which occur when long lengths of thin-walled tubing
are rotated at 200 to 400 rpm will not create diffi-
culties. It appears that routine ultrasonic inspec-
tion of tubing can be accomplished by laboratory-
trained personnel with a minimum of setup and
adjustment time.

For high-speed production inspection, the speed
with which the length of tubing can be scanned is
limited by the length of tubing which may be in-
spected during one revolution of the tube. The
ratio of longitudinal scan movement per
tube revolution must therefore be adjusted to ensure

pitch

complete inspection coverage. To increase the
pitch, the %-in.-diq transducer was replaced with
a 341-in.-dic| transducer. However, the %-in.-dia
transducer covered the entire cross section of the
V-in. tubing being inspected and produced a con-
fused cathode-ray-tube pattern when the full 3/4-in.
diameter In order to overcome this
difficulty and yet maintain the greater pitch ratio,
a collimator was fabricated that effectively blocked
out approximately one-third of the transducer energy

was used,

pattern along a chord parallel to the tubing axis
(Fig. 6.43). With the collimation a pitch of I/2 in,
per revolution can be maintained.

Correlation of known defects, as detected by
visual, dye-penetrant, Cyclograph, and x-ray in-
spection, has shown that all the defects are de-
tectable by ultrasound inspection, including those
those on the inside wall, as well as those on the
outside wall, of the tubing. Investigations are
continving in an effort to determine the smallest
size of discontinuity that may be detected, but it
is now believed that any detrimental defect can be
found.

Work is under way on the development of an
audible warning system to supplement the con-
ventional *“B’' scan cathode-ray-tube presentation.
The warning circuit presently available has too
slow a response rate to be adequate for the pro-
posed rotation and scanning speeds.
PERIOD ENDING DECEMBER 10, 1955

UNCLASSIFIED
ORNL-LR-DWG 11543

 

TIMING SIGNAL

REFLECTION FROM

TUBING SURFACE —f———

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

REFLECT!ION FROM FLAW VIDEO SIGNAL
TUBING SURFACE REFLECTION

r— ———————————————— _ =

I I

| |

| |

! |

I I

| <+
|

| |

' |

' |

[

| |

i |

Lo

 

 

 

 

 

 

 

 

 

 

 

 

 

I'— ______________________________ T
| FLAW REFLECTION |
|
| -
| L] }
|
TIMER |
| - {(FREE—-RUNNING |
} MULTIVIBRATORY} ||
l |
| 4 |
i |
| A |
[ |
| |
| 1
RECEIVER
THYR
‘I (TUNED TO CRYSTAL PUC\‘ST?F?N :
| RESONANT FREQUENCY) |
[ |
| |
| |
| - -— |
| |
-y A
IMMERSCOPE 1(

 

j\_ _______ J
17-in. PERSISTENT SCREEN
CATHODE-RAY OSCILLOSCOPE

"B" SCAN

ROTATIONAL SIGNAL
el

 

DATA POTENTIOMETER

 

 

 

  
  

 

 

 

 

TUBING

Fig. 6.42. Diagram of System for Ultrasonic Inspection of Tubing.

The use of ultrasound techniques for inspecting
welded Inconel plates is being investigated. To
date, difficulty has been experienced in passing
sound through the weld bead, and it appears that
conditions within the weld bead are such that the
sound energy is almost completely attenuated.
The conditions that cause the difficulty may be
excessive grain size, microporosity, or included
foreign matter,

An x-ray technique for the inspection of thick
sections of beryllium is also being developed.
Beryllium, which is a very light element, has a
very large scatter coefficient, which makes it a
difficult subject for high-resolution radiography.
It is believed that collimation of the x-ray beam
at the front surface of the object with a slit system

will reduce the scatter and increase the contrast.
The object and the film will be exposed by moving
them under the slit (Fig. 6.44).

CERAMIC RESEARCH

C. E. Curtis J. R. Johnson
J. A, Griffin A, J. Taylor
Metallurgy Division

Rare-Earth Cermet Fabrication

A rare-earth cermet has been fabricated from
Lindsay oxide code 920 and iron that is 40 vol %
rare-earth oxide and 60 vol % iron, with 50% of
the iron present as an equivalent amount of Fe,0,.
A measured density of 6.14 g/cm?® was attained,
which is to be compared with o theoretical density

167
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

  
  

 

 

 

7

 

COLLIMATOR

UNCLASSIFIED
ORNL—LR—DWG 11544

—a————— ULTRASOUND TRANSDUCER

COLLIMATOR TUBE

 

 

 

 

 

N

 

 

 

 

 

(=

ROTATICN
OF TuBE

3
~
- 7
— = = / ——

Fig. 6.43. Collimator Arrangement for Use with 3/“-in.-diu Ultrasound Transducer.

UNCLASSFIED
ORNL-LR~-DWG 11545

  
 

X -RAY TUBE

4/ !‘ /SLIT SYSTEM
IO A NS\

——
MOTION
—— BERYLLIUM OBJECT

O\gROLLERSO \-FILM O

Fig. 6.44. Slit System for X-ray Inspection of
Thick Sections of Beryllium.

168

 

 

 

 

of 7.69 g/cm3. The apparent porosity (kerosene)
was 21.4%, and the porosity obtained from the
density valve was 20.8%. Metallographic examina-
tion showed the structure to be uniform. The
ductility was estimated to be considerably less
than that of previously prepared specimens that
contained 70 vol % iron, as the metal.

Boron Carbide Shield Material

Methods for fabricating the boron carbide tiles
needed for shielding in the ART are being investi-
gated. It is planned to produce metal cans in the
shape of large tiles which would be vibration-packed
with sized B, C grains, The cans would have lap
joints at their edges to prevent neutrons from
streaming between the cans. Experiments are
under way to determine the best packing method
and the best sizing of the grains to obtain maximum
density. In addition, a study was made to determine
the tendency of B4C particles to flow under pres-
e

sure. The B,C particles were packed in a brass
sleeve (10-mil wall) which had steel plungers on
each end, Pressure was applied to the plungers,
and the bulge of the sleeve was taken as a measure

PERIOD ENDING DECEMBER 10, 1955

of the flow of the B,C particles, Bingham-type
flow was observed, with a yield point of gbout 60
to 80 psi for the least dense powder and 120 to
140 psi for the most dense powder.

169
ANP PROJECT PROGRESS REPORT

7. HEAT TRANSFER AND PHYSICAL PROPERTIES
H. F. Poppendiek

Reactor Experimental Engineering Division

Heat-transfer studies with the fuel mixture NaF-
KF.LiF-UF , flowing in a heated Hastelloy B tube
and the construction and assembly of a heat-trans-
fer system that includes a pump were completed.
The ART fuelsto-NaK heat exchanger test system
was modified for further experiments, and the fric-
tion characteristics of heat exchanger tubing were
determined. Hydrodynamic studies of the ART core
have been continued, and a study of flow in the
sodium cooling passages around the core was
initiated.

Detail designs were completed for the experi-
mental study of the temperature structure to be
expected in the ART core, and fabrication of the
apparatus was started. An analytical study is
under way of the temperature structure in a volume-
heat-source system in which the fluid is flowing
turbulently between curved channels. A general
study of the problem of heat removal from the fuel
dump tank of an ART-type aircraft reactor after
shutdown was initiated, The prediction of tem-
perature distribution in helical pipes was corre-
lated with data obtained from operation of the
MTR in<pile loop.

The enthalpies and heat capacities of RbF-ZrF
UF , were determined, and a study was made of
the heats of fusion of fluoride mixtures. Also,
viscosity measurements were made on nine fluo-
ride mixtures, and the transient-cell technique for
measuring thermal conductivities of liquids was
investigated further,

FUSED-SALT HEAT TRANSFER

H. W. Hoffman
P. E. Stover

Reactor Experimental Engineering Division

D. P. Gregory
Pratt & Whitney Aircraft

Heat-transfer studies with the fuel mixture NaF-
KF-LiF-UF4 (11.2-41-45.3-2.5 mole %) flowing in a
heated Hastelloy B tube were completed. The
results (Fig. 7.1) duplicate those previously ob-
tained with the same fuel mixture flowing in In-
conel and type 316 stainless steel tubes.? Visual

 

]H. W. Hoffman and P. E. Stover, ANP Quar. Prog.
Rep. Sept. 10, 1955, ORNL-1947, p 149.

170

examination of the inside surface of the Hastelloy
B tube showed no deposits. These heat-transfer
data, which fall approximately 40% below the gen-
eral turbuient~flow heat-transfer correlation, further
strengthen the opinion that the observed deviation
from the general correlation is caused by charac-
teristics of the fluid itself. The presence of
particulate matter in the fluid has not yet been
conclusively established. To check this, further
experiments are in progress. A fresh fuel mixture
and a newly constructed system are being used,
The experimental unit was treated with hydrogen
after fabrication to ensure the removal of oxides
from all surfaces to be exposed to the fuel mixture.

Construction and assembly of a heat-transfer
system containing a pump developed by the Experi-
mental Engineering Department were completed.
The salient features of this system are indicated
in Fig. 7.2. Operating experience will be gained
by using NaNO -NaNO ;-KNO (40-7-53 wt %) as
the test fluid. The heat-transfer data obtained will
be used as a check on the performance of the unit.
Since this nitrite-nitrate salt mixture is noncorro-
sive and water soluble, it can be used without harm
to the experimental apparatus. Foilowing this
preliminary operation, studies will be made with
the fuel mixture Ncn"—'-Zer-UF4 (50-46-4 mole %) at

higher Reynolds numbers than previously covered.

ART FUEL-TO«NaK HEAT EXCHANGER
J. L. Wantland

Reactor Experimental Engineering Division

The ART fuel-to-NaK heat exchanger test sys-
tem? has been modified for further experiments so
that the heat-transfer and friction characteristics
of the heat exchanger can be determined with dif-
terent numbers of tube spacers in the tube bundle.
The results of these tests will give definite in-
formation on the effect of tube spacers on heat
transfer from the fuel side of this heat exchanger.
A report is being prepared in which the heat-
transfer and friction data obtained with the ART
fuel-to-NaK heat exchanger will be compared with
the heat- and momentum-transfer data for somewhat

 

2J. L. Wantland, ANP Quar. Prog. Rep. June 10, 1955,
ORNL-1896, Fig, 7.4, p 152.
—— e

PERIOD ENDING DECEMBER 10, 1955

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

= )
CRNL—-LR—DWG {0817
0.010 — |
0.009 |- - 2 Co ‘
0008 f—— -+ e ——
| ; !
0.007 ‘ - - —
O IN HASTELLOY B
™ 0.006 — - ® IN INCONEL -
. A IN TYPE 316 STAINLESS STEEL
® 0.005 '
& ‘ f |
1 ! ’
~ 0.004 | - ; : Can -
= T T GENERAL CORRELATION, /= 0.023 N %%
e ~ o
g 0003 b— | - . } e e T
z |
@ 4 | |
z 3 i
x / 1 ; ‘ ° . ;
g / e ¢ ]
8 0002 /. ; g_(gl& e e
IO ! ;
L . A L/NGF—-KF~LEF (11.5-42-46.5 mole Yo}
w ~<® N INCONEL i
F o ‘ ! ‘
| ; |
| |
!,
0.001 ; ‘ ‘ ‘
1000 2000 5000 10,000 20,000 50,000

REYNOLDS MODULUS, NRe

Fig. 7.1. Comparison of Heat-Transfer Measurements on NuF-KF-LiF-UF4 (11.2-41-45.3-2.5 mole %)
with the General Turbulent-Flow Correlation and the Measurements on NaF-KF-LiF (11.5-42-46.5 mole %)

in Inconel.

similar configurations studied by ofher investi-
gators.

The friction characteristics of two, small, drawn,
Inconel tubes, such as those to be used in the
ART fuel-to-NaK heat exchanger, were determined
in the Reynolds modulus range 50,000 to 200,000.
One tube, which was clean and new, had an inside
diameter of 0.01163 ft and a length of 5.84 fi. The
other tube, which had been in service for 1500 hr
in an experimental fuel-to-NaK heat exchanger, had
an inside diameter of 0.01305 ft and a length of
2.24 ft. Results of these experiments (Fig. 7.3)
indicate that service in the heat exchanger will
not appreciably vary the friction characteristics of
these tubes; the data obtained for each tube cor-
related to within +2% with the empirical equation
f = 0.149/NR°°'18 within the Reynolds modulus
range tested and fell about 6% above the smooth-
pipe friction-factor curve. Photomicrographs (Fig.
7.4) prepared by the Metallurgy Division show that
neither hydrogen firing nor service in a NaK heat
exchanger appreciably changes the inside surface
characteristics of small, drawn, Inconel tubes.

ART CORE HYDRODYNAMICS

F. E. Lynch L. D. Palmer
C. M. Copenhaver

Reactor Experimental Engineering Division

G. L. Muller
Pratt & Whitney Aircraft

Hydrodynamic studies of flow through quarter-
scale models of ART cores have revealed several
ways in which the problem of poor flow distribution
may be corrected. The studies of flow through the
scale models of the 18- and 21-in. ART core with
vortex generators at the entrance have shown that
there is some transient flow but no flow separation
in the model of the 18-in. core and that there is a
small amount of flow separation in the model of
the 21l-in. core. Since vortex generators with
larger vane angles might eliminate flow separation
in the 2l.in. cores, such vortex generators are
being made by Pratt & Whitney Aircraft for testing
in the flow-visualization system,

Surveys of literature on flow through diffusers,
and observations made at ORNL of flow through

171
ANP PROJECT PROGRESS REPORT

FLUID COOLER

 

UNCE ASSIFIED
PHOTO 25016

Fig. 7.2. Fused-Salt Heat-Transfer System.

two constant-gap core models with diffusers, indi-
cate that diffusers with smaller expansion ratios
than those used thus far would eliminate flow
separation. |t would also be possible to eliminate
the problem by making the core into a straight
annulus. Qualitative evaluations of flow through
two constant-gap core models and through the
model of the 21-in. core with several entrance
systems produced the results presented in Table
7.1

Preliminary measurements on the models with
vortex generators indicated that the pressure loss
through the vortex generators was of the order of
1.5 psi for a 600-gpm fuel flow rate. It is expected
that the loss will be greater for vortex generators
with increased vane angles. It has also been esti-
mated (since the flow through the core is extremely
difficult to analyze) that the pressure drop through

172

the core for a spiral type of flow, such as would
exist with the entrance system illustrated pre-
viously,? would be 800 times the pressure loss for
straight-through flow. For a redesigned entrance
system, also illustrated previously,? in which the
rotational component would not be so great as in
the previous case, the pressure loss through the
core was estimated to be 35 times the straight-
through pressure loss. The pressure loss for
straight-through flow was roughly estimated from
an equivalent parallel-piate system to be about
0.02 psi. This would make the loss for the first
case about 16 psi and, for the second case,

 

3A. P. Fraas, ANP Quar. Prog. Rep. March 10, 1955,
ORNL-1864, Fig. 2.3, p 20.

4a. P, Fraas, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL-1947, Figs. 1.2 and 1.3, p 17-19.
M

PERIOD ENDING DECEMBER 10, 1955

UNCLASSIFIED
ORNL—-LR—-DWG 10818

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

0.03 I
| |
| | |
| O NEW INCONEL TUBE |
® INCONEL TUBE USED N
! AN EXPERIMENTAL HEAT EXCHANGER
F=0149 /08 |
- -0‘.\4 |
E 002 - : | |
E *‘—nu&-nfihifihiiii
&
5 .
\“: “ o ‘ i \
\/—17 = -08+4 2 Iog(/VRe\/?) !
| {SMOQTH TUBES)
0.01
3 x40% 5 10° 2 3x10°
REYNOLDS NUMBER, Mg,
Fig. 7.3. Friction Characteristics of Small, Drawn, Inconel Tubes.
TABLE 7.1, RESULTS OF EVALUATION STUDIES OF YARIOUS CORE MODELS
Core Flow Entrance Conditions Flow Features
Constant gap, inlet-to=equator Axial No vanes or screens Very thin separation layer;

area ratio 1:1.443

position of separation
changed with Reynolds

number

Constant gap, inletsto-equator Axial
area ratio 1:2.133
Model of 21-in. ART core Rotational

No vanes or screens

Turning vanes designed to give
filow angles relative to axes
of 30, 25, 20, and 15 deg

(four separate cases)

Separation layer thicker than

in first case

Large amount of separation and
transient flow in all cases;
separation shifted from island

wall for 30-, 25., and 20-deg

cases to shell wall for 15«deg

case

 

0.70 psi. An analysis of experimental measure-
ments® on the full-scale core model for the first
case shows a pressureloss of approximately 15psi,
Pressure taps have been installed in the ‘?22-scole
model of the 21-in. core to determine the pressure
loss through the core. More accurate measure-
ments will also be made of the pressure loss
caused by vortex generators.

 

5Personu::i communication, W. T. Furgerson.

Attempts have been made to photograph the
velocity profilesobserved with the phosphorescent-
particle technique. A high-output electronic flash
tube and its power supply have been received, but
there has been a delay in obtaining the proper
shutter mechanism for the camera. The excitation
of the phosphorescent particles by the tube ap-
pears to be much more intense than the excitation
obtained with a mercury-arc source., A much nar-
rower excited-particle band is obtained because of

173
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED |
Y1129

 

UNCLASSIFIED
S YT

 

UNCLASSIFIED

 

Fig. 7.4.

Inside Surfaces of Three, Small,
Drawn, Inconel Tubes. (@) New tube. (4) Hydrogen-

fired tube. (c) Tube used in fuel-to-NaK heat
exchanger for 1500 hr. (Secret with caption.)

174

the short flash duration.

A study of flow distribution of the sodium which
will circulate in an annulus to cool the outer core
shell of the ART has been initiated. It is con-
ceivable that poor flow distribution could occur in
the annulus because of secondary flow patterns
caused by nonuniform flow at the entrance or by
nonuniform gap spacing. A §2z-scule model of the
annulus system has been designed, and its fabrica-
tion is nearly completed. Although the model does
not incorporate the actual reflector cooling-hole
distribution, provisions have been made for ex-
tracting sodium, in amounts representative of cool-
ing-hole flow, at various sectors of the inlet
header. The reflector spacers and ridges will be
included in the annulus in the scale model. The
flow through the various sectors of the annulus
will be studied by the phosphorescent-particle
technique, and pressures will be measured by
static pressure taps and hypodermic impact tubes,
These measurements will be made for uniform and
nonuniform annulus gap conditions, as well as for
one- and two-pump operating conditions.

ART CORE HEAT TRANSFER

N. D. Greene J. A, Russell
H. F. Poppendiek

Reactor Experimental Engineering Pivision

G. L. Muller J. M. Cornwell
Pratt & Whitney Aircraft

Much progress has been made in preparing for
the experimental study of the temperature structure
in a half-scale model of the ART core, within
which the volume heat source is to be generated
electrically. The supporting research for the ex-
periment, the conceptual design, and the detail
designs of the test section, entrance sections, exit
section, and mixing chambers have been completed.
The detail designs were developed for ORNL by
Pratt & Whitney Aircraft. The layout is shown in
Fig. 7.5.

Two entrance sections were designed. One will
simulate the scrolls of the reactor pumps to yield
a rotational-flowentrance condition, and the second
one will yield purely axial flow at the entrance.
Provisions have been made at the entrance to the
test section for the addition of vortex generators
and flow vanes. The test section, which is com-
posed of Bakelite and platinum, is currently being
fabricated by the ORNL Machine Shop at the X.10

Area.
SL1

ORNL-LR-DWG 1081
‘ FLOW IN . LoLR- 9

  
    
     
    
     
    

ENTRANCE TO MIXING CHAMBER

THERMOCOQUPLE THERMOCOUPLE

 

 

 

PERFORATED PLATES

BOLTARCN VALVE

 

 

 

  

 

 

 

) /- |

HONE YCOMB L§V B LN

STRAIGHTENER— |

 

 

 

 

 

 

 

PUMP AND HEADER
MOCKUP

 

 

 

 

 

 

 

 

 

 

ALTERNATE ENTRANCE SYSTEM

:/\/— THERMOCOUPLES LOCATED ON
ISLAND AND SHELL WALLS

ELECTRODES (24)

ISLAND

SHELL

EXIT MOCKUP

THERMOCCUPLE

]

e

FLOW QuT

THERMOCOUPLEJ

EXIT MIXING CHAMBER

Fig. 7.5. Layout of ART Volume-Heat-Source Experiment for Studying Temperature Structure in Core.

§S61 ‘01 ¥I3IW3ID3A ONIANZ GOId¥3d
About 40 copper-constantan thermocouples will
be posifidned in a nearly uniform manner over the
outer core and the island walls at about 0,030 in.
below the surface; these thermocouples will meas-
ure the steady-state temperature structure of the
uncooled system. About 24 platinum—platinum-
rhodium thermocouples will be located near the
wall-to-fluid interfaces; these thermocouples will
determine the transient nature of the temperature
structure at the wall surface. Mixed-mean fluid
temperature measurements will be made in inlet
and outlet mixing chambers.

The electric power for end electrodes will be
supplied by saturable reactors, and power for the
remainder of the electrodes will come from variable
autotransformers. Currents and voltages for this
multielectrode system will be measured by amme-
ters and voltmeters. The thermocouples located
near the fluid-toswall interface will probably be in
electrical contact with the fluid (acid solution).
The problem of measuring such a variable d-¢
voltage (having a relatively low frequency) in the
presence of a high, 60-cycle, a-c voltage is cur-
rently being studied. Several possible solutions
to this problem are believed to exist.

VOLUME-HEAT-SOURCE CONVECTION
ANALYSES

H. F. Poppendiek

Reactor Experimental Engineering Division

G. L. Muller
Pratt & Whitney Aircraft

An analytical study is under way of the tempera-
ture structure in a volumesheat-source system in
which the fluid is flowing turbulently between
curved channels. Such an analysis will be useful
in estimating the temperature distribution in the
ART fuel channel under rotational-flow conditions.
The experimental hydrodynamic data of Wattendorf
are being used to obtain the temperature solution.
However, certain apparent discrepancies in the
Wattendorf® data are currently being checked and
reconciled before the analysis is completed.

A temperature solution for the case in which a
fluid with an internal volume heat source flows
laminarly and suddenly encounters a flat plate is
being used to calculate the error that results when
a thermocouple is used to measure the temperature

 

6F, L. Wattendorf, Proc. Roy. Soc. (London) 148A,
565-598 (1935).

of a flowing fluid having a volumetric heat source.
The problem involves the study of the development
of thermal and hydrodynamic boundary layers over
a flat plate.

HEAT REMOVAL FROM FUEL DUMP TANK

C. M. Copenhaver
Reactor Experimental Engineering Division

A general study of the problem of heat removal
from the fuel dump tank of an ART-type aircraft
reactor after shutdown has been initiated. Boiling
water, boiling NaK, air, and circulating NaK cool-
ing systems are being considered.

The temperature rise to be expected in the ART
fuel during dumping was investigated by a tran-
sient, one-dimensional conduction study. It was
found that the temperature of the fuel would rise
from 1100 to 1700°F at the center of the stream
within 3 min after shutdown. Thus the maximum
temperature of the fuel will be well below the
2247°F boiling temperature of the fuel. The wall
temperatures of the Inconel core shells will remain
essentially constant during dumping if the sodium
coolant system is operating at full capacity, since
the required heat-removal rate through the Inconel
will be about 20% of the full-power heat-removal
rate.

HEAT TRANSFER IN HELICAL PIPES
N. D. Greene L. D. Palmer

Reactor Experimental Engineering Division

A study of the temperature asymmetry in helical
pipes which have volume heat sources within the
pipe walls, as well as within the fluids flowing
through the pipes, was made previously,” and it
was predicted that the wall-temperature asymmetry
in the ANP in«pile loop operated in the MTR would
be as great as 70°F at full power. The recent
temperature measurements at NRTS on the in-pile
loop, which was operated below full power, indi-
cate that the prediction was satisfactory.

Experimental measurements of the temperature
asymmetry in a glass model (3.7 times larger than
the actual system) of the in-pile locop were made
for the case of heat generation in the fluid only, at
a series of Reynolds numbers. A typical set of
data at a Reynolds number of 5100 is shown in

 

7. D. Palmer, G. M.* Winn, and H. F. Poppendiek,
Investigation of Fluid Flow in Helical Bends of the
ANP In-Pile l.oop, ORNL CF-55-6-183 (June 27, 1955).

177
ANP PROJECT PROGRESS REPORT

ORNL -~ LR*D&E 10820

 

 

 

 

 

 

130 | | ]‘ | \
| | |
126 : - - e - —_
: ‘ , |
INNER WALL TEMPERATURE
2 b — - — 7\2
o

 

)

 

 

118

 

14

 

TEMPERATURE (°F

 

 

 

 

 

 

 

 

 

l- Re = 5160

 

 

 

 

 

 

 

 

 

    
 

 

 

 

 

 

 

 

' . Pr=63
|
o ‘ —
106 - - - — —_—— —_—
MIXED-MEAN FLUID TEMPERATURE
102 —- |‘ —_— ]
1 ;
\ : .
4 | 1 :
| \ ; 1
98 : :
o 10 20 30 40 50 60 TO 80 90 100 11 120

DISTANCE FROM BEGINNING OF SPIRAL (in.})

Fig. 7.6. Typical Temperature Structure in a Model of the ANP-MTR In-Pile Loop with No Heat Source

in the Wall,

Fig. 7.6, The inner wall temperatures (convex
side) were significantly higher than the outer wall
temperatures (concave side), which fell very near
the mixed-mean fluid temperature. These results
are in satisfactory agreement with the limiting
case predictions made prior to the experiment. A
report on these calculations and the experimental
measurements is being prepared.

HEAT CAPACITY MEASUREMENTS ON
FLUORIDE MIXTURES

W. D. Powers

Reactor Experimental Engineering Division

The enthalpies and heat capacities of two
rubidium-bearing fluoride mixtures were studied,
and the following results were obtained.

178

RbF-ZrF ,~UF , (48-48-4 mole %)

Solid (142 to 398°C)
Hp = Hygoe = =7.7 + 0.1490T + (3.2 x 10~3)T?2

Cp = 0.149 + (6.5 x 10~ 7T

Liquid (458 to 880°C)
Hp = Hygoe = ~14.3 +0.2844T - (5.4x10~3)T?2

c, =0.284 - (10.8 x 10=9)T
AH =35 at 425°C

LiF-RbF (43-57 mole %)

Solid (134 to 420°C)
Hp = Hagoc = =6.7 +0.1849T + (2.45 x 10~ 5)T2

c, =0.185 + (4.9 x 10=9T
Liquid (497 to 878°C)

. Hp = Hagoo = =28.5 +0.39697 - (8.1 x 10~ 72
¢, =0.397 - (16.1 x 10=9)T
. AHf = 55 at 475°C

In these expressions,
H.. - Hggo. = enthalpy in cal/g,

¢, = heat capacity in cal/g-°C,
T = temperature in °C,
AH}, = heat of fusion in cal/g.

A method used to correlate heat capacities of
fluoride mixtures was described previously.® It
has been found that the heat capacity per gram
atom was 8.1 for the zirconium fluoride=bearing
mixtures and that it was 9.5 for the alkali fluoride
mixtures not containing ZrF ,. The heat capacity
per gram atom was 8.13 for the rubidium<bearing
mixture containing ZrF, and was 9.29 for the
rubidium-bearing mixture without ZrF ,, the results
being in agreement with the general correlations.

HEATS OF FUSION OF FLUORIDE MIXTURES
W. D. Powers

Reactor Experimental Engineering Division

The heats of fusion of the fluoride mixtures c¢an
be determined from the enthalpy-temperature rela-
tionships of the solid and liquid mixtures., It was
found that the heats of fusion per gram atom of ali
non-beryllium-bearing fluoride mixtures varied from
about 1200 to 2000 cal. Two mixtures containing
BeF , had the lowest heats of fusion determined,
probably because of their glasslike nature. The
heats of fusion that have been determined, to date,
are listed in Table 7.2, Two beryllium<bearing
mixtures, NaF-BeF UF, (76-12-12 and 25-60-15
mole %), were studied, but the heat of fusion de-
terminations could not be made because the mix-
tures melted over wide temperature ranges.

VISCOSITY MEASUREMENTS ON FLUORIDE
MIXTURES
S. 1. Cohen
Reactor Experimental Engineering Division
. Viscosity measurements were made on nine

fluoride mixtures; the results are shown in Fig. 7.7
and listed in Table 7.3. The table presents abso-

 

8w. D. Powers, ANP Quar, Prog. Rep. June 10, 1955,
ORNL-1896, p 157.

PERIOD ENDING DECEMBER 10, 1955

TABLE 7.2, HEATS OF FUSION OF YARIOUS
FLUORIDE MIXTURES

 

Fluoride Mixtures Heat of Fusion

 

 

Composition {(mole %) In cal/g=atom In cal/g

NaF-Zer-U F4 1780 56
(50-46-4)

NaF—Zer 1820 61
(50-50)

NCIF-ZI’F4-UF4 1970 63
(53-43-4)

NaF-ZrF4~U F, 2040 63
(53.5-40-6.5)

NaF-ZrF4-U F4 1790 57
(56-39-5)

RbF-ZrF4-U F4 1370 35
(48-48-4)

NaF-ZrF ,-U Fu 1700 42
{50-25.25)

NOF'ZFF4-UF4 1570 42
(65-15-20)

NaF«K F«LiF 1970 95
(11.5-42-46.5)

NaFK F-LiF-UF 1940 88
(10.9-43.5-44,5-1.1)

NaF-LiF-UF4 1160 56
(38.4-57.6-4)

KF-LiF 1960 93
(50-50)

K F-LiF-U F4 1690 68
(48-48-4)

RbF-LiF 1800 55
(57-43)

NoF-K FUF, 1290 30

(46.5226+27.5)

 

lute and kinematic viscosities for the mixtures and
equations for the data in the form

= AeB/T

’

where T is in °K, No equation is listed in the
table for salt g because of the slight curvature of
the data, which may be noted in the figure. Meas-
urements were made on salts a, 4, f, g and b with

179
ANP PROJECT PROGRESS REPORT

TABLE 7.3.

SUMMARY OF CURRENT VISCOSITY MEASUREMENTS

 

Composition

Absoclute Viscosity

Kinematic

 

Mixture Viscosity i Ref
{mole %) (centipoises) (centistokes)
a NaF-K F-LiF At 500°C, 9.2 4.25 0.03914219/T (a)
{11.5-42-46.5) At 800°C, 2.0 1.03
b NaF-BeF ,.LiF At 600°C, 8.0 3.9 0.0934,°881/T (a)
(63.5-29-7.5) At 800°C, 3.5 1.8
c KF-BeF, At 750°C, 2.6 1.3 0.07773599/T (5)
(79-21) At 850°C, 1.9 0.96
d NaF-ZrF -UF, At 600°C, 8.1 2.4 0.1045¢3798/T (c)
(56-39-5) At 800°C, 3.6 1.15
e NaF-K F-BeF, At 600°C, 8.0 3.0 0.0837398%/T (b)
(49-15-36) At 850°C, 2.9 1.45
/ LiF-NaF At 700°C, 3.2 1.6 0.1145¢3239/T ()
(60-40) At 850°C, 2.05 1.05
g NaF-ZrF UF, At 700°C, 8.5 2.0 (a)
(50-25-25) At 900°C, 3.5 0.9
b RbF-ZrF LUF, At 540°C, 10.0 2.9 0.05794197/T (d)
(50-46-4) At 800°C, 2.9 0.9
i NaF-LiF-BeF, At 600°C, 5.9 2.8 0.1853018/T (&)
(53-24-23) At 800°C, 3.0 1.5

 

9Previously unpublished data.

bs. 1. Cohen and T. N. Jones, Measurement of the Viscosity of Compositions 90 and 96 and (KF-BeF'z; 79-21 mole

%), ORNL CF.55-11-28 (to be issued).

€S. 1. Cohen and T. N. Jones, Measurement of the Viscosity of Composition 70, ORNL CF-55.9-31 (Sept. 6, 1955),
43, |, Cohen and T. N. Jones, Measurement of the Viscosity of Compositions 87, 95, and 104, ORNL CF«55-11-27

(to be i ssued),

The composition shown here i s the nominal composition; the

used in pteparing this mixture con-

tained about 20 mole % K F; therefore the actual composition is more nearly RbF.K F-ZrFA-UF4 {40+10+46+4 mole %).

Brookfield and capillary viscometers; values ob-
tained by these two different instruments were in
satisfactory agreement (deviations from an average
line through the data did not exceed +10%). Salts
b, c, e, and i, which contained BeF,, were studied
in a separate dry box used only for work with
beryllium. Measurements were made on each of
these mixtures, and at least two calibrated capil-
lary viscometers were used to furnish a check.

The density values used for the salts which did
not contain BeF , were calculated from the correla-
tion based on experimental data on other mixtures
described earlier.? Density values for the mixtures

 

%S. 1. Cohen and T. N. Jones, A Summary of Density
Measuremenis on Molten Fluotide Mixtures and a Cor-
relation Useful for Predicting Densities of Fluoride Mix-
tures of Known Composition, ORNL-1702 (May 14, 1954).

180

containing BeF, were calculated from a similar
correlation based on experimental data taken at
Mound Lcuborcn‘ory]0 on mixtures containing BeF2,

An investigation!! was carried out to ascertain .
the optimum conditions and procedure for obtaining
inert dry box atmospheres by purging or sweeping
with an inert gas. It was found that heavier-than-
air gases should be introduced at the bottom of the
dry box and exited at the top, that lighter-than-air
gases should be introduced at the top of the dry
box and exited at the bottom, that purging effi-
ciency is decreased with mixing, and that purging

 

IOPersonal communication to W. R. Grimes,

S. I. Cohen and J. M. Peele, Determination of the
Optimum Procedure for Obtaining Inert Atmospheres in
{I\(qosrz-)Vac‘uum Dryboxes, ORNL CF-55-10-132 (Oct.” 25,

5.
UNCLASSIFIED
ORNL-LR—CWG 10824

TEMPERATURE (°K)
750 80C 850 900 950 1000 1100 1200

20

VISCOSITY (centipoises)

 

1
450 500 550 600 650 700 800 900 1000

TEMPERATURE (°C)

Fig. 7.7. Summary of Current Yiscosity Measure-
ments. See Table 7.3 for compositions and
equations.

efficiency is increased as the flow rate is in-
creased. '

THERMAL CONDUCTIVITIES OF LIQUIDS
W. D. Powers

Reactor Experimental Engineering Division

Three different devices are currently being used
to measure the thermal conductivities of liquids:
the variable-gap system, the constant-gap cell, and
the transient cell. Recently a great deal of empha-
sis has been placed on the transient technique.
The apparatus consists of a 4-in. section of
3/32--in. tubing which can be heated electrically
while in contact with a liquid, The temperature
rise of the tube depends upon the time of heating
and the thermal properties of the liquid. The
exact solution of the differential equation that
ideally represents the experimental apparatus has
been obtained. This ideal system is defined by an

PERIOD ENDING DECEMBER 10, 1955

infinite cylinder that is at a constant temperature
for time less than zero and is bounded internally
at r = a by a medium of infinite conductivity; heat
is added at a constant rate for all times greater
than zero. In this system r is the radial position
in the region bounded extemal to a. At r = a the
temperature of the liquid is the same as the tem-
perature of the medium of infinite conductivity,
The temperature solution is of the form

TK at 2cp
o/ "N\e?/" \cp?/|’

where the quantities in parentheses are dimension-
less moduli and

= temperature at r = q,
thermal conductivity of the liquid,
heat added at r = a per unit length,
= thermal diffusivity of liquid = K/cp,
= time greater than zero,
= radius of tube,
cp = volumetric heat capacity of liquid,
c’p” = volumetric heat capacity of material of
infinite conductivity.

Valves for (TK/(Q) have been found for various
values of (at/a?d) and (2cp/c’p?). All quantities
are known except K and «, and, since a = K/cp, the
thermal conductivity may be found from the tem-
perature rise at one given time, The thermal con-
ductivity of water was determined by this method
and was found to differ from the literature value by
10%. The conductivity of molten sodium hydroxide
was also determined by this technique; a value of
0.55 Btu/hr-ft2(°F /ft) was obtained over the tem-
perature range 700 to 900°F.

il

i

Since it is difficult to completely fill the cell of
the constant-gap apparatus with liquid, a rocking
mechanism has been devised which forces the mol-
ten fluoride to flow from one riser through the cell
to the other riser and back again. This rocking
technique must often be used for long periods of
time before the cell is completely filled. Recently,
some preliminary measurements were made on
RbF.ZrF -UF , (48-48-4 mole %), and the conduc-
tivity was found to be about 1.2 Btu/hr-ft2(°F/ft).

181
ANP PROJECT PROGRESS REPORT

8. RADIATION DAMAGE

D. S. Billington
Solid State Division

The in-pile loop recently operated in the MTR by
the Aircraft Reactor Engineering Division is being
disassembled in the G-E facilities at NRTS for
shipment to ORNL. Samples of the miniature in-
pile loop that was operated in vertical hole C-48
of the LITR were examined, and a large number of
ARE components have been made available for
examination,

A study of thermocouple errors in in-pile loop
temperature measurements was completed. Experi-
ments are under way for determining holdup times
vs charcoal-trap temperatures for fission gases
carried with nitrogen, helium, and other purge
gases. The first in-pile, tube-burst, creep-test
specimens were irradiated in the LITR, and an
altemate, in-pile, stress-corrosion apparatus was
designed.

DISASSEMBLY AND EXAMINATION OF
IRRADIATED EQUIPMENT

M. J. Feldman
Solid State Division

MTR In-Pile L oop

The in-pile loop recently operated in the MTR is
being disassembled and sectioned in the General
Electric Company hot-cell facilities at the National
Reactor Testing Station (NRTS). The plans for
the disassembly and the design of the necessary
equipment were coordinated with General Electric
Company personnel and with ORNL Aircraft Reac-
tor Engineering Division personnel.

The loop was moved into the G-E hot cell on
November 9, about one week after the forced termi-
nation of operation in the MTR. The interim period
of about one week had been spent in setting up and
testing the disassembly equipment. The initial
insertion of the loop in the hot cell immediately
revealed that equipment alterations were required,
and so the loop was removed. On November 10 the
ioop was reinserted, and the initial cut was made
to remove the nose section, Most of the loop
tubing enclosed in the heat exchanger was removed
in two l4-in. sections obtained with two additional
cuts made on November 14, During attempts to
make the fourth cut, the saw became inoperable,
and arrangements were made to clear and decon-

182

During the de-
contamination procedures the molybdenum boot,
the fill tank, the nose section, and the two sec-
tions of loop tubing were placed in a cask for
shipment to ORNL.

Revisions to the scheduled disassembly proce-
dure have been caused by the request for early
retum of the MTR carrier, which necessitated
changing the order of cutting, and by the desire to
inspect the pump and about 3 ft of auxiliary loop
parts associated with the off-gas system, which
failed during operation. The disassembly equip-
ment was designed to handle only the active por-

taminate the cell for repair work.

tion of the loop, and the ideal place to cut the
loop for off-gas line inspection could not be
brought under the saw because of a change in tube
diameter. Two attempts to make this cut through
the large, steel shield blocks of the loop were
interrupted by equipment failures, and efforts have
been shifted to other portions of the loop until a
more suitable method is worked out.

Disassembly operations were again interrupted
on November 22, when a saw blade broke. In at-
tempting to replace the saw blade, the gripping
mechanism of the General Mills manipulator be-
came inoperable. The extent of the damage to the
manipulator has not yet been determined. Disas-
sembly of the loop will be continued when repairs
have been made,

LITR Miniature In-Pile Loop

M. J. Feldman A. E. Richt
Solid State Division

Samples of the nose section of the miniature in-
pile loop which was operated for 30 hr in vertical
position C-48 of the LITR were examined. Opera-
tion of this loop was terminated, as described
previously,! because of faulty behavior of the
pump motor. Observations of four samples of the
nose section of the loop in the unetched condition
showed only slight, sporadic, corrosion penetration
to a depth of less than 1 mil. Comparisons of the
sections on either side of the center of the nose
bend and on the compression and tension sides of

 

1G. W. Keilholtz et al., ANP Quar. Prog. Rep. Sept.
10, 1955, ORNL-1947, p 164.
the bend showed no changes in relative grain size.
Since 30 hr of operation would not be sufficient
time for the development of pronounced radiation
effects, no definite conclusions should be drawn
from these results.

ARE Components

Work is continuing on the examination of various
components of the ARE. A large number of sec-
tions of the ARE core and auxiliary equipment
have been transferred to the Solid State Division
canal. Melt-out facilities have been readied for
the removal of fused salts from the sections,
Metallographic examinations of the various sec-
tions are to be made.

THERMOCOUPLE ERRORS IN IN-PILE LOOP
TEMPERATURE MEASUREMENTS

G. W. Keilholtz
W. R. Willis M. F. Osborne
Solid State Division

Extensive investigations of errors in metal-
surface temperature measurements obtained from
thermocouples placed in high-velocity air streams
have shown that the results are dependent upon
the velocity of the air stream and the type of
thermocouple installation. Results obtained in
recent experiments with models of themocouple
installations on the miniature in-pile loop have
verified the results of previous tests of thermo-
couple installations on static in-pile corrosion
capsules.? As before, resistance-welded thermo-
couples were superior to discharge-welded thermo-
couples in that they gave smaller errors in tem-
perature measurements and were less sensitive to
changes in cooling-air flow rates. As yet, no
thermocouple installation has been found that is
not affected by the air flow rate in a wide air
annulus such as that used in the in-pile loop.

HOLDUP OF FISSION GASES BY
CHARCOAL TRAPS

G. W. Keilholtz W. E. Browning
Solid State Division

C. C. Bolta
Pratt & Whitney Aircraft

Experiments are in progress to determine holdup
times of fission gases by charcoal traps at various
temperatures with various carrier gases in the
traps. Radickrypton is being used to simulate the

PERIOD ENDING DECEMBER 10, 1955

fission gases, and, in the initial experiments,
nitrogen and helium are being used as the carrier
gases. Helium was found to be unsatisfactory for
use in the MTR in-pile loop installation because of
its poor electrical insulation properties, and there-
fore nitrogen is of more immediate interest as the
carrier gas because of its better electrical insula-
tion properties. In order to use nitrogen in place
of helium as the nose purge gas in the in-pile loop,
information is needed concerning the holdup time
for gaseous fission products in the carbon trap vs
temperature of the trap, with nitrogen gas used as
the carrier. Similar information for helium and
other carrier gases is needed for use in designing
the ART purge system and in designing purge sys-
tems for use in other reactors.

The apparatus being used is shown in Fig. 8.1.
The charcoal trap is the same as the ones used in
the MTR in-pile loop, being 2}; in, in diameter and
14 in. long and containing 340 g of activated char-
coal, The flow rate of the carrier gas, 5.05 cth, is
the same as that used in operation of the loop at
the MTR. The trap temperatures are controlled by
‘“freezing’’ mixtures, which are kept at the freezing
point with liquid nitrogen. The temperature of
—50°C isobtained with a eutectic mixture of CaCl,
and water; a Freon mixture is used to obtain a tem-
perature of ~110°C. The data obtained thus far
are presented in Figs. 8.2 and 8.3. These data
show that radiokrypton is absorbed and contained
in the charcoal traps at a given temperature for
longer periods of time with helium as the carrier
gas than with nitrogen. It appears that, if nitrogen
were used as the carrier gas for the in-pile loop-
nose purge system, temperatures of —110°C or
colder would be needed to properly limit the
amount of activity sent to the MTR stack in case
of a fuel leak. Efforts are being made to correlate
these data into generalized analytical expressions
relating trap properties, temperature, gas proper-
ties, and holdup time.

Investigations are planned in which the radio-
krypton will be diluted in larger quantities of nor-
mal krypton to determine the effect of krypton con-
centration on holdup time. Additional tests are
planned in which other carrier gases will be used,

 

2W. E. Browning, G. W. Keilholtz, and H. L. Hemphill,
ANP Quar. Prog. Rep. March 10, 1955, ORNL-1864,
p 146.

183
ANP PROJECT PROGRESS REPORT

 

/ KRYPTON TRAP

 

 

 

 

   
 

N, OR He

 

[

 

A P
| u—— VACUUM PUM

 

 

 

 

 

 

|1
CHARCOAL TRAP —

¢ /KRYPTON CHAMBER
[ f
——><

UNCLASSIFIED
ORNL-LR-DWG 10822

 

v
] —

—
GM COUNTER /

FLOWMETER

 

 

 

 

;fl

~\

 

 

 

 

 

|.>fl

|_—— FREEZING SOLUTION

 

 

 

 

 

 

 

 

 

Fig. 8.1. Apparatus for Investigating Holdup Times of Fission Gases by Charcoal Traps.

CREEP AND STRESS CORROSION

J. C. Wilson
Solid State Division

In-Pile Tube-Burst Creep Tests

W. W. Davis N. E. Hinkle
J. C., Zukas
Solid State Division

The first in-pile, tube-burst, creep-test specimens
were irradiated for two weeks in the LITR. In
these tests four tubular specimens were stressed
to rupture at a circumferential fiber stress of
2000 psi at temperatures of 1500°F for two and
1550 and 1450°F for the other two. Postirradia-
tion measurements of the total strain in the uni-
formly bulged regions are under way in the hot
cells. Corresponding bench tests under the same
conditions are in progress. A second in-pile rig
is complete and ready for LITR irradiation.

Additional fuel-containing tube-burst rigs, pre-
viously described, have been assembled and are to

184

be filled with fused-salt fuel. Several stress-
corrosion rigs in which static sodium is used as a
coolant for a tubular fuel-containing specimen that
is stressed in bending are ready for welding.
These rigs will be used for control tests of a simi-
lar apparatus already irradiated in the LITR,

The MTR creep-test apparatus, in which the
specimen was tested at 1500°F and 1500 psi, has
been returned to ORNL, and the water jacket is
being removed for postirradiation measurements of
specimen elongation.

Alternate In-Pile Apparatus

W. E. Brundage C. D. Baumann
Solid State Division

An altemate apparatus for in-pile stress-corro-
sion experiments has been designed. In this rig,
as in the rigs now being used,® the molten fuel

 

3). C. Wilson et al,, ANP Quar. Prog. Rep. Sept 10,
1955, ORNL-1947, p 165.
ORNL~-LR-DWG 10823
1000 .

500 TOTAL TIME TRAP WAS ACTIVE ~

 

200

100

N
O

TIME (min)
™
o

TIME TO PEAK
ACTIVITY IN TRA

<

e ——

TIME TO BREAK-THROUGH OF ACTIVITY

 

1 ‘
-130 —-110 -90 -70 —-50 -30 —~f0 +10

CHARCOAL BED TEMPERATURE (°C)

Fig. 8.2. Holdup Time vs Temperature of Char
coal Trap for Radiokrypton Carried in Nitrogen.

contacts one wall of the annular container, and
solidified fuel contacts the other wall. This
permmits the heat generated to be removed and, at
the same time, a small surface-to-volume ratio to
be maintained and the maximum fuel temperature
to be limited to that of the sample.

The heat for maintaining the fuel in the liquid
state will be supplied by fission heating during
normal operation, but supplemental heating can be
supplied by a cylindrical heater around the outside
of the container if necessary. Heat will be re-
moved by flowing water being passed through the
central tube of the fuel container. Stress loading
witl be supplied by gas pressure over the fuel,
with the outer portion of the fuel container used as
the specimen in hoop tension.

PER!IOD ENDING DECEMBER 10, 1955

ORNL—LR-DWG 10824
10,000

5000

2000 :
TOTAL TIME

TRAP WAS ACTIVE
1000 S

500

200

TIME TO BREAK-
THROUGH OF ACTIVITY

[ B

100

TIME (min)

50

20 ———— e
TIME TC PEAK ACTIVITY IN TRAP
10 . s =

 

1
-130 {10 =30 =70 =50 —-30 —10 +10
GHARCOAL BED TEMPERATURE (°C)

Fig. 8.3. Holdup Time vs Temperature of Char-
coal Trap for Radiokrypton Carried in Helium.

Calculations have been made of the flux depres-
sion outside the sample, the flux distribution
in the fuel, and the resulting heat distribution
through the fuel. The maximum specific power is
generated in the outer portion of the annulus, with
the power generated at the inner surface being ap-
proximately 40% of the maximum, The flux.depres-
sion calculations are being checked by flux meas-
urements, in the LITR, with the use of a cadmium-
magnesium alloy of the same size as the fuel
volume and with a thermal-neutron cross section
the same as that of the fuel to be used.

185
ANP PROJECT PROGRESS REPORT

L. 9. ANALYTICAL CHEMISTRY OF REACTOR MATERIALS

C. D. Susano

J. C. White

Analytical Chemistry Division

An apparatus was modified so that the r-butyl
bromide method could be applied to the determina-
tion of oxygen in sodium sampled at operating
temperatures. Studies were continued on the dis-
tillation method for the determination of oxygen
in sodium, A method for determining the concen-
tration of rare-earth elements in stainless steel
was developed. Aluminum was determined in
NaF-ZrF4-UF4 by a modification of the Aluminon
method, In connection with hydrofluorination
studies a rapid method of analysis was developed
for the determination of hydrogen fluoride and
water in effluent gases. Studies were continued
on the determination of oxygen in metallic oxides
by the bromination method, A method was de-
veloped for the determination of tantalum in
NaF-KF-LiF. Methods were developed for the
determination of ftrivalent iron
fluoride salts.

in mixtures of

DETERMINATION OF OXYGEN IN SODIUM

A. S. Meyer, Jr. G. Goldberg
W. J. Ross
Analytical Chemistry Division

n-Butyl Bromide Method

Studies were undertaken to ascertain whether
positive errors are inherent in the determination of
oxygen in sodium by the n-butyl bromide method,!
Investigations of the errors introduced by the con-
tamination by oxygen during sampling and by traces
of water in the organic reagents were reported
previously.2 Additional positive errors may result
from the presence of alkaline-earth metals in the
alkali-metal samples.

The n-butyl bromide method is based on the con-
version of the electropositive metals to neutral
bromide salts by their reaction with butyl bromide
and the subsequent acidimetric titration of the
alkali-metal oxides. In the pure state the alkaline-
earth metals calcium, barium, and strontivm are
essentially inert to n-butyl bromide, while mag-

 

‘J. C. White, W. J. Ross, and R. Rowan, Jr., Anal
Chem. 26, 210 (1954).

2p. S, Meyer, Jr., W. J. Ross, and G. Goldberg, ANP
Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 172.

186

nesium may react with n-butyl bromide to form a
Grignard reagent, The alloys of these alkaline-
earth metals with alkali metals are probably more
to n-butyl bromide than are the pure
alkaline-earth metals, and any residual metal or
Grignard reagent will be converted to hydroxide
upon dissolution of the reacted sample in water
and will thus be titrated as oxygen.

Since the lower-atomic-weight alkaline-earth
metals in concentrations below the range of spec-
trographic determination may be sufficient to intro-
duce significant errors in the oxygen determination,
chemical methods for their detection were tested.
Calcium and magnesium were determined by flame-
photometric measurement after separation from
the bulk of the sodium metal. Successful separa-
tions were carried out either by precipitation of

reactive

the calcium and magnesium from ammoniacal solu-
tions with 8-quinclinel or by precipitation of the
sodium as sodium chloride by saturating aqueous
solutions of the samples with HCl gas. After
separation by either of these methods, the calcium
and magnesium concentrates from a 10-g sample
contained less than 10 mg of sodium,

Typical samples of the sodium used in corrosion
and heat-transfer studies were found to contain
approximately 100 ppm of calcium and less than
50 ppm of magnesium. These concentrations corre-
spond to a maximum error in the determination of
oxygen of about 80 ppm, and appropriate correc-
tions will have to be determined. Since the
alkaline-earth elements that are alloyed with
sodium are at least partially converted to bromide
salts during dissolution of the samples, the error
introduced by them is not significant if the con-
centration of oxygen in the sodium is high. The
increase in the reactivity of the alkaline-earth
metals upon dissolution in sodium was demonstrated
by the determination of oxygen in sodium which
contained relatively high concentrations of altkaline-
earth metals, for example, 0.2% barium. Since the
molar concentration of oxygen as determined by
the n-butyl bromide method was less than that of
the alkaline-earth metal, a portion of the metal
must have reacted with the n-butyl bromide.

In further studies of the n-butyl bromide method
it has been shown that, if the reaction is contained
in a sealed vessel, the conversion of the alkali
metals to bromide salts can be carried out by the
reaction of the sample with pure n-butyl bromide
rather than with the moderated solution of n-butyl
bromide in hexane, which is used in the usual ana-
lytical reagent. The reaction time is reduced from
about 4 hr to less than 5 min when the pure reagent
is used. The procedure is applicable both to NaK
and to sodium,

When the reaction is carried out with the pure
reagent under pressure, the possibility of con-
tamination during reaction is materially reduced
and it is possible to effect the direct transfer of
the test portion of the sample from the bulk sample.
A sampler and on apparatus for the defermination
of oxygen in sodium by the amalgamation pro-
cedure have been modified, as shown in Fig, 9.1,
for use with n-butyl bromide, In this apparatus a
sample of molten metal is added, from a transfer
line, to approximately 150 ml of n-butyl bromide
reagent contained in a 3 x 18 in., glass reaction
vessel. The transfer is carried out under helium
at a pressure of 1 atm. Samples have been taken
from sodium at temperatures as high as 1000°F.
A pressure below 10 psi was developed even when
samples which contained 9 g of sodium were re-
acted. This procedure is now being used to com-
pare the results of determinations of oxygen by the
n-butyl bromide method and by the distillation
method,

Distillation Method

The distillation method for the determination of
oxygen in sodium, which was described in the
previous report,® was applied to the determination
of the oxygen in sodium contained in forced-circu-
lation corrosion-testing loops and in sodium-transfer
reservoirs. Samplings have been made at tempera-
tures from 600 to 1400°F.

For the oxygen determination the sample is
introduced through the top of the distillation appo-
ratus into a calibrated, hemispherical cup which
is fitted with a preshaped liner of 1-mil nickel foil.
A sufficient quantity of the liquid sodium is intro-
duced to flush the transfer line and the sample cup
thoroughly. When the temperature of the sodium is
less than 600°F, the distillation apparatus is
evacuated before the sample is transferred. At
higher temperatures, a pressure of 1 atm of helium
must be maintained to prevent flash distillation of

 

31bid., p 173.

PERIOD ENDING DECEMBER 10, 1955

the sodium during transfer. After the cup has been
filled with the test portion of liquid metal, the
pressure in the distillation chamber is reduced to
less than 10 y, and the sample is heated to distill
off the metallic sodium. After the distillation is
completed, the nickel liner is transferred to a
beaker, in which the residual oxides are titrated
with a dilute solution of standardized acid.

In a check of the effect of time and temperature
made by the distillation method,
analyses were made of sodium which had been
deliberately contaminated by the addition of an
amount of Na O, equivalent to 700 ppm of oxygen
before introduction into the corrosion-testing loop.
When the distillation was carried out at above
900°F, that is, near the maximum of the recom-
mended4 range of distillation temperatures of 350
to 550°C (660 to 1030°F), the oxygen values were
extremely low., The effect of distillation time was
also evident; for example, after a 1-hr distillation
period at 900°F, the oxygen concentration of the
residue was less than 100 ppm, and after 4 hr the
oxides had been almost completely volatilized.
Distillation of another sample of the contaminated
sodium at 850°F for 4 hr gave a residue which
corresponded to 400 ppm of oxygen. Since part of
the oxygen could have been trapped in cold legs
of the loop, the 400-ppm value represents a reason-
able analysis of the oxygen in the sample that was
originally contaminated with 700 ppm of oxygen.

on analyses

The accuracy of the distillation method is being
studied by comparing the results with those ob-
tained by the n-butyl bromide method and by
testing the quantative recovery of added oxygen.
A distillation apparatus and an apparatus for
analyses by the n-butyl bromide method have been
attached to a reservoir which contains sodium that
was deliberately contaminated by exposure to air
during locding. Repeated determinations after the
dissolution of samples of the sodium in the n-butyl
bromide reagent have established the concentrgtion
of oxygen as 600 * 50 ppm. An analysis by the
distillation method showed 620 ppm of oxygen,

In the study of the quantitative recovery of added
oxygen, reproducible results which correspond to
35 = 5 ppm of oxygen were obtained after dis-
tillation of as-received sodium for 4 hr at 850°F.
Essentially the same results were obtained after

 

4). Rr. Humphries, personal communication, June 24,

1955.

187
ANP PROJECT PROGRESS REPORT

distiflation for periods of 1 and 2 hr at this tem-
perature. Analyses of the residues indicated that
Na,O is the predominant residual oxide. The effect
of metallic impurities in the liquid metal on the

 
 
  
   

CALROD HEATER.

~ OVERFLOW TRAY>/

e Wt Gy

"O" RING

VACUUM

GLASS PIPE FLANGE

VALVE

    
  

 

     

 

 

~ GLASS PIPE— | A
_ww
OVERFLOW REACTION
RECEIVER VESSEL

distillation method is to be studied in detail, After
the oxygen content of a measured quantity of sodium
has been established, a calculated weight of NaOH
will be added to raise the concentration of oxygen

UNCLASSIFIED
ORNL-LR-DWG 10825

SODIUM INLET

   

 

 

   
   

HELIUM
VIEWING
WINDOW
/A
I
S
E
BUTYL
BROMIDE
ANHYDRONE
GLASS FRIT

TEFLON GASKET

1 Q 1

2

 

 

\_/

 

SCALE IN INCHES

Fig. 9.1. Apparatus for the Determination of Oxygen in Sodium by Use of the n-Butyl Bromide Method.

188
ek e o e

to 300 ppm, and the determinations wiil be re-
peated to test the accuracy of the distillation
method.

DETERMINATICN OF TRACES OF RARE-EARTH
ELEMENTS IN STAINLESS STEELS

A, S. Meyer, Jr. B. L. McDowell
Analytical Chemistry Division

The concentrations of the rare-earth elements
gadolinium, europium, and samarium must be
limited to a few parts per million in reactor con-
struction materials because of the high absorption
cross sections of these elements for thermal neu-
trons. Therefore a method was developed for de-
termining the quantities of the rare-earth elements
in stainless steel, Inconel, and nickel,

In order to concentrate the rare-earth elements in
a sample, 25 mg of spectrographically pure yttrium
oxide is added to a 2- to 5-g sample of stainless
steel. The ytirium oxide is used as a carrier for
the precipitation of the rare-earth elements and
also serves as an internal standard for their
spectrographic determination. The sample, fo-
gether with the Y 0,, is dissolved in HCI and
then fumed with Hé|04. After the hydrated oxide
precipitate of silicon, niobium, tantalum, etc,,
has been removed by filtration, the major con-
stituents of the sample, iron, chromium, and nickel,
are separated by electrolysis at a mercury cathode,
The volume and acidity of the solution are reduced
by evaporation, and the rare-earth elements, to-
gether with the yttrium carrier, are precipitated as
hydrous oxides by saturating the solution with
ammonia, The alkaline-earth elements remain in
solution, together with a portion of the amphoteric
elements, The precipitate is dissolved in HCI,
and the rare-earth concenfrate is precipitated by
the addition of an oxalic acid. The precipitate is
filtered and then ignited to the oxide,

The concentration of rare earths in the precipi-
tated Y, 0, is determined spectrographically, and
the concentration in the stainless steel sample
is calculated from the ratio of carrier to sample.
When 2.5-g samples of stainless steel are taken,
concentrations of the rare-earth elements of 1 ppm
can be determined, Between 75 and 90% of the
added carrier is recovered. The final precipitate
is essentially pure Y203, with SiO2 as the prin-
cipal contaminant,

The method offers several advantages over the
procedure of Spitz et al.,® in which the rare-earth

PERIOD ENDING DECEMBER 10, 1955

group is precipitated with an iron carrier and
measured against uranium as an internal standard.
Fewer precipitations are required, and, since the
carrier is added before the separation step, quanti-
tative recovery of the rare earths is not essential,
The Y, 0, provides a much less complex spectrum
than that provided by the uranium and iron,

The method has been used for the determination
of gadolinium in types 347 and 304 stainless steel,
Quantitative recovery of a standard addition of
gadolinium was obtained. The method will also
be tested with other types of stainless steel,
Inconel, and nickel. With minor modifications the
procedure should be applicable to all ferrous and
nonferrous alloys that do not contain as major
constituents elements such as zirconium, vanadium,
and other metals which would not be deposited at
the mercury cathode,

SPECTROPHOTOMETRIC DETERMINATION OF
ALUMINUM IN FLUORIDE SALTS WITH
AURIN TRICARBOXYLIC ACID

A. S. Meyer, Jr. C. R, Williams
Analytical Chemistry Division

The Aluminon (aurin tricarboxylic acid) method
for the spectrophotometric determination of alumi-
num was applied to the determination of aluminum
in fluoride salt mixtures such as NaF-ZrF ,-UF ,.
Aluminon forms a stable, red-colored lake with
solutions of pH 5 to 5.4. While this reaction is
utilized as one of the more sensitive colorimetric
methods for the detection and determination of
aluminum, it is subject to interference by many
elements,® particularly those which form hydrous
oxide precipitates at pH 5.

In the determination of aluminum in samples of
NaF-Zer-UFA‘, the principal interfering element,
zirconium, is removed by extracting it as zirconium
cupferrate (nitrosophenylhydroxylamine) from a 4 M
solution of H_SO, with chloroform containing hy-
drogen cupferrate.  Other interfering elements,
such as iron, which is present in trace concen-
trations in these samples, are also removed by
this extraction. Any uranium which may be present
in the tetravalent state is quantitatively extracted,
while hexavalent uranium remains in the aqueous
phase, Since only a small part of the uranium is

 

5E. W. Spitz et al., Anal. Chem. 26, 304 (1954).

6F. D. Snell and C. T. Snell, Colorimetric Methods of
fl‘lnalysis, 3d ed., vol Il, p 248, Van Nostrand, New York,
949.

189
ANP PROJECT PROGRESS REPORT

oxidized to the hexavalent state during dissolution
with H2504, its residual concentration is not suffi-
cient to interfere in the determination of aluminum.
When cupferron is added to acidic solutions which
contain zirconium, a very slightly soluble precipi-
tate of zirconium cupferrate is obtained. Since this
precipitate is not readily extractable by organic
solvents,” the zirconium cupferrate is formed in
the organic phase by extracting the aqueous solu-
tions with a solution of hydrogen cupferrate in
chloroform., Aluminum is not extracted from a 4 M
H,S0, solution if the concentration of the hydrogen
cupferrate in the chloroform phase does not exceed
3%.

An aliquot of the extracted solution that con-
tains 6 to 50 ug of aluminum is taken for color
development. The color is developed by heating
the solution with Aluminon for ¥ hr at a tempera-
ture of 80 to 90°C; gelatin is added to stabilize
the lake. Mercapto acetic acid is added to complex
any iron that is introduced in the reagents which
are added after the cupferron extraction. The
absorbance of the final solution of 50-ml volume
is measured at 530 my in l-cm cuvettes 2 hr after
the addition of the chromogenic reagent.

The number of determinations made by this
methed is not yet sufficient to establish the pre-
cision, but, on the basis of calibration datq, it
appears that the coefficient of variation of the
results is approximately 5%. The concentration of
aluminum in typical samples of NaF-ZrF , eutectic
mixtures and NaF-ZrF -UF, (50-46-4 mole %) fuels
was found to be approximately 100 ppm.

DETERMINATION OF WATER IN HYDROGEN
FLUORIDE GAS

A. S. Meyer, Jr. W. J. Ross
Analytical Chemistry Division

A method was developed for the determination of
the concentration of water relative to hydrogen
fluoride in the effluent gases from a hydrofluorina-
tion reactor. The measurement of the relative
water content was needed to ascertain when the
hydrofiuorination reaction of oxides in zirconium-

base fuels, that is,

ZrO2 + 4HF —> Zrl":4 + 2H20

was complete. Pyridine was found to be an effi-
cient absorber for both hydrogen fluoride and water,

 

’N. H. Furman, W. B. Mason, and J. 5. Pekola, Anal.
Chem. 21, 1327 (1949).

190

The gases from the reactor were bubbled through
approximately 5 cm of pyridine contained in a
simple absorber constructed of 30- by 200-mm test
tubes. Two absorbers were connected in series
to measure the efficiency of the absorption. After
a quantity of gas had been passed through the
pyridine, the test tubes were removed from the
assembly, stoppered immediately with serum-type
bottle stoppers, and replaced with tubes filled
with fresh pyridine, Test portions of the absorber
solutions were taken with 1-ml| tuberculin syringes.

The concentration of water in the pyridine solu-
tions was determined by titration with coulometri-
cally generated Karl Fischer reagent.® The con-
centration of hydrogen fluoride was determined
by titrating an aqueous solution of the pyridine
absorbate with 0.1 N NaOH solution to a phenol-
phthalein—thymol-blue mixed-indicator end point.
Since pyridine is a relatively weak base (K, =
1.8 x 10-9), the hydrogen fluoride can be titrated
directly in aqueous solutions which contain pyri-
dine.

The hydrogen fluoride and water were almost
completely removed from the gases in the first
absorber.,  When the concentration of hydrogen
fluoride in the first absorber did not exceed 50
mg/ml, the concentration of hydrogen fluoride in
the second absorber was negligible. Increases
in the concentration of water in the second ab-
sorber were limited to a few tenths of a milligram
per milliliter and may have been introduced by
contamination with atmospheric moisture during
loading of the pyridine into the test tube. The
determinations made by this method are reproducible
to about +0,5% water in hydrogen fluoride. Since
in the concentration of water from 1%
in commercial hydrogen fluoride to 10 to 30% in
the effluent gases were observed during test runs,
the reproducibility of the determinations is suffi-
ciently high to permit monitoring the reaction.
The over-all precision of the resulis is limited
primarily by the precision of the determination of
the water, which is approximately 2%. The pre-
cision of measurement of small increases in the
concentration of water is much lower because of
a relatively high original concentration of water
in the pyridine. The precision of the method can
be improved by the use of anhydrous pyridine.

increases

 

84, s. Meyer, Jr., and C. M, Boyd, The Determination
of W)ater by Coulometric Titration, ORNL-1899 (June 5,
1955).
DETERMINATION OF OXYGEN IN ZIRCONIUM
OXIDE BY BROMINATION

J. P. Young M. A. Marler
Analytical Chemistry Division

The study of the methods of determination of
combined oxygen in metal oxides and fluoride
salts by the bromination method of Codell and
Norwitz? was continued. The reactions involved
in the process include the reaction at approximately
1000°C of bromine vapor in a carrier gas of helium
with a mixture of a metal oxide and powdered
graphite to form CO and the metal bromide. After
removal of the excess bromine and the metal
bromide in cold traps, the CO is oxidized to CO
by reaction with hot CuO. The resultant CO2
is absorbed in Ba(OH), and estimated quantita-
tively by titration with standard acid.

Several additions and modifications were made
to the apparatus, which is described in previous
reports. 1011 A tube containing copper and copper
oxide, which is maintained at 400°C, was inserted
between the helium tank and the apparatus in order
to remove any oxygen or hydrocarbons that might
be present in the gas. Also, the Tygon tubing
that had been used to comnect some parts of the
apparatus was replaced with glass and short
pieces of gum-rubber tubing. As recommended by
Codell and Norwitz,? granular zinc maintained at
350°C was inserted between the cold traps and
the CuO tube to remove the last traces of bromine
that might escape the cold traps. Further, the
thin, platinum liner, described in the previous
report,12 which is inserted inside the reaction
tube during the analysis of fluoride salts, was
replaced with a geld liner that extends into the
first cold trap. Although gold is attacked by
bromine, it is more inert to oxygen than is platinum,
and therefore lower blank corrections for the appa-
ratus are required. The sample container has also
been fabricated from gold.

For this determination, mixtures of ZrO, and
graphite are ground together in a mortar by use of
a pestle and then allowed to react with bromine

 

M. Codell and G. Norwitz, Anal. Chem. 27, 1083
(1955).

10y C. White, J. P. Young, and G. Goldberg, ANP
Quar. Prog., Rep. March 10, 1955, ORNL-1864, p 161.

1y, p, Young and G, Goldberg, ANP Quar. Prog.
Rep. june 10, 1955, ORNL-1896, p 178.

12, P Young and M. A. Marler, ANP Quar Prog.
Rep. Sept. 10, 1955, ORNL-1947, p 176.

PERIOD ENDING DECEMBER 10, 1955

vapor under varying conditions. The residual
amount of ZrO, is determined by weighing the
residue after ignition of the graphite. Bromination
of Zr0,, according to the equation

ZrD2 + QBr2 + 20— ZrBr4 + 2CO

was incompiete after 4 hr of reaction time at tem-
peratures as high as 1100°C. Approximately 20%
of the ZrO, remained unreacted under these con-
ditions.

When 30 mg of FeF, was added to 10 mg of
ZrO, and 50 mg of graphite and the intimate mix-
ture was brominated at 950°C for 2 hr, complete
removal of ZrO, from the sample container was
achieved, The order of mixing was found to be
critical. In order to achieve the complete removal
of ZrOz, it was necessary to mix the ZrO, with
FeF, in a mortar with a pestle, and then mix the
graphite with the ZrO -FeF, mixture in the same
manner., |t is assumed that the reactions involved
are:

3Zr02 + 4FeF3—> 3ZrF4 + 2Fe203
F3203 + 3Br2 + 3C — 2FeBr3 + 3CO

The amount of CO formed during the bromination
of the sample was, however, much less than the
expected theoretical amount, The CO was de-
termined by oxidizing the gas to CO, and absorbing
the CO, in standard Ba(OH), solution. In all
determinations of the resultant CO, the amount
was only 30 to 40% of that expected. In order to
check the possibility that the method of absorbing
CO, was not efficient, several samples of CaCO,
were thermally decomposed in the apparatus and
the resulting CO, was determined. Quantitative
recovery of CO, was achieved in all cases.

The study on the bromination of ZrO, and FeF,
in the presence of graphite is still in progress,
but, at present, it appears that the reactions
postulated do not adequately describe the mechanism
of the removal of ZrO_ from the sample container,
In order to establish the quantitative recovery of
the CO produced in the bromination of the oxide,
the determination of oxygen in ferric oxide by
bromination was investigated. Quantitative re-
covery of the oxygen as CO, was obtained, but
it was necessary fto mix the ferric oxide and
graphite in a mortar with a pestle. Bromination of
a 15-mg sample of ferric oxide was complete in
4 hr at a temperature of 950°C.

191
ANP PROJECT PROGRESS REPORT

DETERMINATION OF TANTALUM IN
FUSED FLUORIDE SALTS

J. P. Young J. R. French
Analytical Chemistry Division

A search was made for a suitable colorimetric
method for the determination of trace amounts of
in NaF-KF-LiF and NaF-KF-LiF-UF .
Dinnin'3 has reported the use of pyrogallol, which
forms a colored complex with tantalum, but uranium
and fluoride ions are known to interfere in this
determination, A satisfactory modification of
Dinnin's method has been developed, however,
for application to the fuel solvent NaF-KF-LiF.
Preliminary tests revealed that the usual method
of removing fluoride ion by heating in an H,SO,
solution was not satisfactory, since tantalum was
volatitized as TaF_. In order to avoid this diffi-
culty, the sample was digested in 0.1 M H_SO
until almost all the water had volatilized; another
volume of 0.1 M H2SO4 was then added, and the
solution was evaporated to dryness, |n this method,
TaF, is hydrolyzed to Ta,0, before the tempera-
ture of the solution exceeds the volatilization
temperature of TaF _.
tion, including the

tantalum

The residue of the evapora-
020 , is then treated according
to the procedure described by Dinninl3 for the
colorimetric determination of tantalum with pyro-
gallol.

The residue is fused with K25207, and the melt
is dissolved in ammonium oxalate. The pyrogallol
complex is formed in a solution that is 0.175 M
ammonium oxalate and 4 M HCI, and its absorbance
is then measured at 330 mp. This method was
found to be satisfactory for the concentration range
of from 0.15 to 0.8 mg of tantalum in a 50-ml
volume. It is essential that the H 504 evaporation
and the K,$,0, fusion be made in Vycor crucibles
rather than platinum, since even very small quanti-
ties of platinum interfere with the pyrogailo} method
for determining tantalum. The attack of fluoride
ion on Vycor has not been of any consequence.

The method of Milner, Barnett, and Smales, 4
which was developed for the separation of tantalum
in tantalum-uranium alloys, was examined for pos-
sible application to samples of fluoride salts which
contain uranium, |n this method tantalum is ex-
tracted with hexone from an aqueous solution

 

134, 1. Dinnin, Anal Chem. 25, 1803 (1953).

s, w. C. Milner, G. A, Barnett, and A, A, Smales,
Analyst 80, 380 (1955).

192

which is 10 M in HF, 6 M in H2$O4, and 2,2 M in
NH4F. The sample is first fumed with H2SO4,
and then HF and NH ,F are added. The tantalum is
extracted into hexone and then extracted from the
hexone phase with a 15% solution of H,0,. The
efficiency of this extraction was evaluated by
determining the amount of tantalum which had been
extracted in the solution of H,O,. The pyrogallol
method, described above, was used for these
analyses. Quantitative recovery of small amounts
of tantalum has not been achieved, as yet.

DIRECT DETERMINATION OF TRACES OF
Fe(lll) IN NuF»KF-LiF-UF4

J. C. White Jo W, Miles!?
Analytical Chemistry Division

A direct spectrophotometric determination of
Fe(lll) in Nc:l"'-KF-LiF-UF4 was developed in
which the sample is dissolved in a solution of
H3P04, H,50,, and H,BO,. Potassium thiocyanate
is added to form the iron(}l)-thiocyanate complex,
which is then extracted with isobutyl alcohol.
The H,PO, concentration must be maintained above
8 M to prevent reduction of the Fe(ill) by U(IV)
during dissolution of the sample. The coefficient
of variation for the method was found to be 3% for
the range 0.25 to 1.5 mg of Fe(lll) per gram of
sample. Further details of this investigation are
available in a separate report.!é

DETERMINATION OF TRACES OF Fe(lil) IN
MIXTURES OF ALKALI-METAL
FLUORIDE SALTS

J. W. Miles J. C. White
Analytical Chemistry Division

A method, which is based on the absorption of
the ferric phosphate complex in the ultraviolet,
has been developed for the determination of Fe(lll)
in the presence of Fe(ll) in mixtures of alkali-
metal fluoride salts. Iron(l}l}) in concentrations of
1 to 5 pug/ml in 0.5 M H2504 and 1 M H,PO, solu-
tions may be determined accurately by observing
the absorbance at 260 mu. The alkali metals,
Ni(il), GCe(Hl), Fe(ll), ond the fluorides do not

interfere, but zirconium and uranium interfere

 

]SResearch participant, University of
Medical School, Louisville, Ky.

16, c. White, Determination of Traces of Iron(lll) in
NaF-LiF-KF-UF4 with Thiocyanate, ORNL CF-55-9-96
(Sept. 20, 1955).

Kentucky
seriously, The details of this method have also
been reported separately,17

ANP SERVICE LABORATORY

W. F. Vaughan
Analytical Chemistry Division

Uranium analyses were performed for the high-
temperature ART critical experiment. This service
was provided so that other experimenters could
follow the increase in uranium concentration as
criticality was approached by the addition of the
fuel concentrate, N02UF6, to the fuel carrier,
NaF-ZrF, (50-50 mole %). The samples, weighing
approximately 60 g each, were ground in a plastic
dry box prior to analysis. Test portions (1 g) were
fused with K;5,0,, and the melt was dissolved
in 5 vol % H2504. The entire portion was passed
through a zinc reductor, and the vranium was
titrated with standard oxidant. Appropriate correc-

 

17J. C. White, Determination of Iron(Ill) in Mixtures
of Alkali Metal Fluoride Salt, ORNL CF-55-7-103
{July 22, 1955).

PERIOD ENDING DECEMBER 10, 1955

tions were made for the iron present, the concen-
tration of which was determined colorimetrically.
The elapsed time for analysis of each sample
was of the order of 30 to 45 min.

During the period, 1276 samples were analyzed
for the ANP project, and a total of 6609 determina-
tions was made. The backlog consists of 46 sam-
ples. A breakdown of the work is shown in

Table 9.1.

TABLE 9.1. SUMMARY OF SERVICE
ANALYSES REPORTED

 

Number of Number of

Samples Determinations

 

 

Reactor Chemistry 746 3963
Experimental Engineering 487 2503
Miscellaneous 43 143

1276 6609

 

193
ANP PROJECT PROGRESS REPORT

10. RECOVERY AND REPROCESSING OF REACTOR FUEL

F. R. Bruce

D. E. Ferguson
M. R. Bennett

F. N. Browder
G. |. Cathers

W. K. Eister

H. E, Goelier

J. T. Long
R. P. Milford
S. H, Stainker

Chemical Technology Division

PILOT PLANT DESIGN

Engineering design of the pilot plant for re-
covering fused-salt fuels is .to be completed by
December 31, 1955, and construction is to be
completed by March 31, 1956, The ORNL Engi-
neering Department has estimated that the facility
will cost $346,000,
Process, project, and instrumentation engineering
is expected to raise the estimated over-all cost
to $435,000.

The single-bed absorber containing granular NaF
which was to have been operated at 650°C has
been replaced on the engineering flowsheet by a
pair of absorbers! containing NaF, These ab-
sorbers will both be at 100°C during absorption of
UF , and some fission-product fluorides in the first
bed, The beds will then be raised to 400°C, and
the UF, and a few fission-product fluorides will
be desorbed from the first bed with F_ and passed
to the second absorber, where the fission products
will remain. The large two-section furnace re-
quired to heat the single-bed absorber was returned
to the manufacturer for conversion into two separate
furnaces.

excluding all engineering.

ENGINEERING DEVELOPMENTS
Contactor

A percolator type of contactor was found to pro-
duce more violent agitation than that produced by
a sieve-plate contactor during gas contacting of
the molten fused salt in tests made in the fluori-
nator designed for use in the pilot plant. In these
tests, 270 b of molten NaF-ZrF, was sparged
with nitrogen, and the percolator, or air-lift, type
of contactor was operated at gas rates of 1 to
3.5 cfm. Visuval observations made during these
tests showed violent pumping of the salt and a
large amount of salt collecting on the cooler upper
wall. The entrainment and subsequent deposition

 

1F. R. Bruce et al.,, ANP Quar, Prog. Rep. Sept. 10,
1955, 0RNL"]947, Figu }0.2, P ]800

194

of salt caused by percolation was greater than that
caused by sieve-plate contacting. An inverted
funnel-type splash plate installed over the perco-
lator discharge contained the salt splash and thus
reduced salt entrainment. An unpredictable in-
crease in sparging-gas pressure during percolator
tests was a major problem., Replacement of the
]/4-in.-dia tubing, which dispersed the gas through
]/16-in.—dic| holes, by a i/4--in.-dic:| open-end tube
did not affect the pressure buildup. The sieve-
plate contactor was operated at gas rates of | to
7 cfm. Salt agitation was gentle and there was a
slight splash of salt on the vessel wall. Tests
of these contactors with uranium-bearing fuel
mixtures will be required in order for a decision
to be made as to which to install in the pilot plant.

Freeze Yalves

A 4-in.-0D, 8-in.-long cone-body prototype of
the freeze valves designed for the pilot plant re-
tained a pressure of 20 psig in 15 of 20 cycles of
melting (350°C), freezing, and pressurizing. Ex-
amingtion of the salt seal around the downcomer
revealed a porous structure but no open gas channels
(Fig. 10.1). It is believed that slight contraction
of the fused salt permitted gas leakage around the
downcomer. The design of the valve is being

modified.

Resistance Heating of Transfer Pipes and
Waste-Discharge Nozzle

Resistance heating was tested as a means of
heating and of maintaining the temperature of salt
transfer pipes at 1200°F. An uninsulated 7-ft
length of Y-in. sched-40 Inconel pipe was heated
to 1200°F in 5 min by passing a current of 600 amp
through the pipe. The voltage drop through the
pipe was 1.2 v/ft,

Resistance heating will also be used to heat the
nozzle which will discharge salt from the fluori-
nator waste pipe to the waste carrier., The nozzle
is designed with a hood section to contain and
   
  
 
   
 
  
 
 
  
 
 
  
 

 
  
 

LOWER CONE BODY :
NICKEL BARSTOCK -

  
 

NO DIRECT GAS
CHANNEL HERE

 
  

 
   
 
 

METAL @
CUTTINGS

POROUS SALT SEAL, a4
t—in.~dic DOWNCOMER -
AND DIAPHRAGM REMOVED =~ -

  

  

THERMOWELL
1 Ya-in-dia NICKEL &

SALT OUTLET PIPE
!/o=in. SCH-40
INCONEL

S%ST

i

PERIOD ENDING DECEMBER 10, 1955

  
   

UNCLASSIFIED |
PHOTO 15686

 

UPPER DISENGAGING SECTION
NICKEL PLATE

 

  
   
 
  
  
 

     
 
 
  
  

PART OF
_ DIAPHRAGM

 

 

~ ENTRANCE LIQUID »
" LEVEL PROBE

  
 

_ e

SALT INLET PIPE
Yo—in. SCH-40
INCONEL

  
      

  
  
   
 
 

  

Fig. 10.1. Sectioned 4-in.-0OD, 8-in.-Long Cone-Body Freeze Valve After 20 Cycles of Melting (350°C),

Freezing, and Pressurizing.

remove any gases evolved from the waste (Fig. 10.2).
The hood vacuum connection is to be nickel, and
the salt transfer pipe and hood section are to be
Inconel. These materials and the large cross-
sectional area of the hood section will minimize
self-resistance heating in the hood section and in
the vacuum connection and will permit a tempera-
ture of 1200°F to be reached by the salt transfer
pipe. The voltage drop across the entire unit will
be 4.25 v with a current of 600 amp.

PROCESS DEVELOPMENT

Fused-Salt Fluorination Studies

Additional evidence of the induction period? in
UF, volatilization from fused-salt fuel was ob-
tained (Fig. 10.3) in further work on the fluorina-

 

2p, E. Ferguson et al., ANP Quar. Prog, Rep. March 10,
1955, ORNL-1864, p 164,

tion step. This work was directed toward full
evaluation of the effect on the fluorine efficiency
of the fluorine-to-nitrogen mole ratio, the fluorina-
tion rate, and the method of contact of the gas
with the fused-salt phase. In Fig. 10.4 the same
data are plotted to show how the efficiency of
fluorine utilization varies during the course of
fluorination and to illustrate the deviation from
the ideal case, in which there would be no solu-
bility effect to cause an induction period.

A comparison of sampling methods showed that
agreement to within 3% was obtained in uranium
analyses of samples taken during the course of
fluorination experiments with the use of a dip
ladle or by immersion of a solid rod into the fused
salt to obtain a quick-freeze sample and samples
taken after fluorination by grinding and sampling
the entire batch of salt, This study was made

195
ANP PROJECT PROGRESS REPORT

SALT TRANSFER, Ya-in. SCH 40

INCONEL PIPE

Yg-in. INCONEL‘L

UNCLASSIFIED
ORNL-LR-DWG 10826

   
  

Yo-in. SCH-10 NICKEL PIPE

VACUUM CONNECTION

 

N &

2in. HOOD SECTION

T

 

 

 

 

 

_ 1/8 in.

 

o 5-in. DIA

 

with three different uranium concentrations in the

Nc:F-ZrF4 salt, namely, 8, 2, and 0.5 wt %.

NaF Absorption Capacity and UF6 loss
on Desorption

The absorption capacity of NaF for UF6 was
determined to be about 0.9 g of uranium per gram
of salt in the case of one lot of Harshaw Chemical
Co. material under a variety of conditions (Table
10.1) The capacity of Baker & Adamson Co.
reagent-grade NaF was 1.89 ¢ of uranium per gram
of salt. Absorption of UF‘5 on NaF under an initial
vacuum indicated a capacity range of 0.80 to 0,90
for different mesh sizes, while absorption values
obtained when the excess UF6 was being removed
with fluorine as a sweep gas varied from 0.72 to
0.86. Practically the same capacity values were

196

 

 

]

Fig. 10.2. Waste-Discharge Nozzle.

obtained at 70, 100, and 150°C and at two different
pressures, The results indicated that the capacity
of 0.9 in the case of the Harshaw Chemical Co.
material is independent of the particle size to
which it is degraded. The 1.89 value for the
capacity of the Baker & Adamson Co. material
corresponds closely to the molecular ratio in the
complex UF6-3N0F.

Fluorine was much more satisfactory as a sweep
gas than was nitrogen in the desorption of UF
from NaF, perhaps as a result of some moisture
in the nitrogen, Two runs were made with 12- to
40-mesh NaF which had been well conditioned by
being alternately evacuated and exposed to fluorine
at a pressure of 15 psia before sufficient UF , was
introduced to completely saturate the NaF.
run the UF , was desorbed with a stream of nitrogen

In one
L]
ORNL—LR~DWG 10827

100

50

20

 

 

 

 

 

URANIUM [N FUSED SALT (% OF INITIAL)

 

05

 

0.2

 

 

0 1 2 3 4 5
MOLE RATIO OF FLUORINE TO TOTAL URANIUM

 

o4

Fig. 10.3. Amount of UF6 Remaining in 375 g of
NaF-ZrF .UF , (50.-46-4 mole %) Fluorinated at
600°C at a Rate of 100 ml/min as a Function of
Amount of Fluorine Introduced.

while the temperature of the NaF was being raised
from 100 to 400°C; 1.4% of the uranium stayed on
the NaF, The total desorption period covered 1 hr,
although the temperature of the bed had reached
400°C in gbout 15 min. In the second run the UF6
was desorbed with fluorine, and only 0.0075% of

the initial uranium load remained on the NaF.

UF‘5 Decontamination in NaF Absorption Step

In each of four consecutive fluoride volatilization
runs (20 to 40 g of uranium in each) in which the
two-bed NaF decontamination system was used,
the activity of the product UF, was less than the
UX,-UX, activity associated with natural vranium.
There was no evidence of decreased decontamina-
tion with repeated re-use of the bed (Table 10.2).
The over-all beta or gamma decontamination factor

PERIOD ENDING DECEMBER 10, 1955

-rev

ORNL-LR~—DWG 10828

 

 

 

 

 

 

 

 

 

1.0
B P\ “ ‘
Lt
\ | | |
0.8 - *— — _ T
") T
s _______1__\ [ R
= \6 |
R
= 08 .LJKJ | ]
" \ - |
Lol N
a i \\ ‘ |
S oa N
5 * N
2N
5 ] 7\.! .
Hop o’ \
T
S *
0 L i

 

o ! 2 3 4 5 6
MOLE RATIO OF FLUORINE TO TCTAL URANIUM

Fig. 10.4. Efficiency of UF Volatilization as a
Function of Amount of Fluorine Introduced.

was about 10°, with 102 being attributable to the
fused-salt fluorination step and 103 to the NaF
absorption-desorption step. The process flowsheet
used ! provides for volatilizing UF‘5 from the molten
fused-fluoride-salt fuel with fluorine, absorbing
the UF . on an NaF bed at 100°C, desorbing the
UF6 at 100 to 400°C in a stream of fluorine, and
passing the desorbed UFé through a second NaF
bed. As in previous work, much of the decontami-
nation from ruthenium occurred during the absorption
period; that is, the UF , and niobium were absorbed
on the first bed ot 100°C, while the ruthenium
passed through (Table 10.3). Almost all the de-
contamination from niobium occurred in the de-
sorption period, when the UF , was being regener-
ated from the NaF at 100 to 400°C; the niobium
was left on the NaF.

For all four runs the average uranium loss in the
first cold trap of the process was 0.06%. This was
perhaps due to the low NaF-to-uranium weight
ratio of 2/1 in the first bed. The uranium loss in

197
ANP PROJECT PROGRESS REPORT

 

 

&
- TABLE 10.1, CAPACITY OF NeF .EOR UF6
Conditions
i Capacity
Material UF6 Pressure Temperature Removal of (g of U per g of NaF)
(psia) (°C) Excess UF6

 

Harshaw Chemical Co.

12 to 40 mesh 15* 100 Not removed 0.86
15* 100 By vacuum 0.86

7.5% 100 By vacuum 0.80

15+ 70 By vacuum 0.88

15* 150 By vacuum 0.90

T5* 100 By fluorine 0.86

15%* 100 By fluorine 0.86

8 to 12 mesh 15%* 100 By fluorine 0.81
%-in. pellets 15%+ 100 By fluorine 0.72
15* 100 By vacuum 0.81

Baker & Adamson Co.

Powder 15* 100 By vacuum 1.89

 

*UF, introduced to NaF under vacuum,

**UF6 introduced at atmospheric pressure to displace fluorine.

TABLE 10.2. SUMMARY OF FOUR CONSECUTIVE RUNS IN TWO-BED FUSED-SALT
FLUORIDE-YOLATILITY PROCESS

Conditions: Total of 128 g of uranium in Nch-ZrFA-UF4 (52-44-4 mole %) with gross beta activity per
milligram of uranium of 5 X 10° counts/min. Each run fluorinated with 1:1 F:;_)N2 mixture
for 1.5 hr and then with pure l'_'2 for 0.5 hr; UF6 in F2N2 gas stream absorbed on NaF, with
some volatile activity passing into cold trap; UF6 desorbed at 100 to 400°C into a second
cold trap

Average F2/U mole ratio in absorption pericd: 4/1
Average F2/U mole ratio in desorption period: 1/1
Absorbent beds: 60 ml of 12- to 40-mesh NaF in 1-in.-dia tubes

NaF/U weight ratio after four runs: 1/1

 

Uranium Loss (%) Product Gamma
Product Yield Waste Activity per Milligram
Run First Cold First NaF Second NaF

(%)
Trap Bed Bed

 

Salt of Uranium*

(counts /min)

 

1 83 0.0 0.02 3.6
2 35.2 0.10 0.05 3.1
3 151 0.07 0.08 1.0
4 43.8 0.08 0.02 2.1
Over-all 70.1 0.06 0.5 5.1 0.04 2.5

 

*Gamma activity per milligram of natural uranium is 8 counts/min.

198
PERIOD ENDING DECEMBER 10, 1955

TABLE 10.3. DISTRIBUTION OF VOLATILIZED ACTIVITY

 

Amount* (% of total)

 

 

Activity Fission-Product First Second
Cold Trap NaF Bed NaF Bed
Gross beta 81 18 0.85
Gross gamma 11 89 0.14
Ru gamma 97 1.6 1.1
Zr-Nb gamma 2.2 98 0.04
Total rare-earth beta 4,3 92 3.6

 

*Collected and actually found by analyses of cold trap and NaF beds.

the fused-salt fluorination step was about 0.04%,
but losses in the two NaF beds were 0.5 and 5%,
respectively, These losses were due to some back-
pressure buildup, which was evident in all runs,
as the result of either partial plugging of the NaF
during the absorption period (probably because of
a volume change) or plugging of the UF  cold-trap
inlet. Occasional tapping of the absorption bed
prevented the NaF-plugging problem from being

very serious in all runs except the second. Here
the plugging was so severe that excessive nitrogen
pressure had to be used, with considerable loss of
product. Plugging of NaF upon absorption of UFé
was completely avoided in two subsequent runs,
without activity, by using a sieve plate at the
enfrance to the first bed for dispersion of the gas
instead of just a 1/4-in. inlet line as in the above
four runs.

199
Part lli

SHIELDING RESEARCH
11. SHIELDING ANALYSIS

F. H. Murray

C. D. Zerby

Applied Nuclear Physics Division

S. Auslender

H. S. Moran

Pratt & Whitney Aircraft

AIR SCATTERING OF Co%® GAMMA RAYS:
THEORY vs EXPERIMENT

H. S. Moran

The air-scattered gamma-ray dose rate predicted
by theory' for a Co%® source at a source-detector
distance of 15 meters was compared with experi-
mental measurements in a similar geometry.? After
an appropriate correction® for ground scattering was
applied, the two results were found to be in sub-
stantial agreement. The details of the comparison
were published in a separate report.?

ENERGY ABSORPTION RESULTING FROM
GAMMA RADIATION INCIDENT ON A
MULTIREGION SHIELD WITH SLAB GEOMETRY

C. D. Zerby S. Auslender

fiémzoding of a Monte Carlo calculation of heat
generation resulting from transport gamma radiation
through shields with stratified slab geometry has
been completed.’®  Preliminary results are in
good agreement with experimental results obtained
for lead.’

The code has been extended to produce data on
the gamma-ray dose and energy flux as well as

energy deposition. Parameter studies on a water-

 

'E. P. Blizard and H. Goldstein (eds.), Report of the
1953 Summer Shielding Session, ORNL-1575 (June 14,
1954), p 170-203.

28, L. Jones, J. W, Harris, and W, P. Kunkel, Air and
Ground Scattering of Cobalt 60 Gamma Radiation,
CVAC-1707T (Marcfi 30, 1955).

3M. L. Coffman and B. T. Kimura, Gamma Ray Ground
Scattering for Co60 and GTR Sources, NARF-55-16T
(May 30, 1955).

14, s. Moran, Air Scattering of €080 Gamma Rays:
Theory vs Experiment, ORNL-2019 (Jan. 6, 1956).

S5c. D. Zerby and S. Auslender, ANP Quar. Prog. Rep.
Mar, 10, 1955, ORNL-1864, p 173.

6¢. D. Zerby and S. Auslender, ANP Quar. Prog. Rep,
June 10, 1955, ORNL-1896, p 192,

7€, S, Kirn et al., Oblique Attenuation of Gamma-Rays

from Cobalt-60 and Cesium-137 in Polyetbylene, Con-
crete, and Lead, NBS5-2125 (Dec. 23, 1952).

Data
from these studies will be presented as dose and
energy buildup factors for monodirectional, iso-
tropic, and cosine monoenergetic sources. Also,
curves on heat deposition will be obtained for all
slabs in the parameter study.

lead slab and an iron slab will be made.

INTEGRAL EQUATIONS FOR THE FLUX

DENSITY NEAR A THIN FOIL AND FOR

NEUTRON SCATTERING IN AIR IN THE
PRESENCE OF THE GROUND

F. H. Murray

In calculations for a thin foil it appears that
very often it is possible to carry out calculations
of the flux in the foil interior as if the flux due to
sources outside a part of the foil were constant
for the small part of the foil considered. With this
assumption, known formulas may be employed fora
slab so that upper and lower surface source densi-
ties for any part of the foil can be determined in
terms of given sources and other parts of the foil.
The sum of the upper and lower surface densities
becomes the effective surface source density for
each small part of the foil, and a set of integral
equations is obtained for the coefficients of a
harmonic expansion, giving the outward flux from
any part of the surface.

The analysis needed for neutron scattering in
air with ground present is very similar to that
employed for a foil. In this case there is no lower
surface source density to be considered, and the
flux in the ground near any surface point is calcu-
lated as if the surface flux were constant over the
infinite plane and equal to its value at the surface
point. This assumption leads to a set of integral
equations for the harmonic coefficients of the sur-
face source density, as before.
have been published separately.®

These equations

 

8. H. Murray, Integral Equations for the Flux Density
Near a Thin Foil and for Scattering in Air in the Presence
of the Ground, ORNL CF-55-9-108 (Sept. 23, 1955).

203
ANP PROJECT PROGRESS REPORT

12, SHIELD DESIGN

J. B. Dee
C. A. Goetz H. C. Woodsum
Pratt & Whitney Aircraft
R. M. Davis
The Martin Company
D. R, Otis

Consolidated Yultee Aircraft Corp., San Diego

The neutron-induced activation of sedium in the
heat exchangers of circulating-fuel reactors has
been calculated. In this calculation, both core

neutrons and delayed neutrons were considered.

CALCULATION OF THE SODIUM ACTIVATION
IN THE HEAT EXCHANGERS OF
CIRCULATING-FUEL REACTORS

J. B. Dee D. R. Otis

The activation of sodium caused by core and
delayed neutrons in the heat exchangers of circu~
lating-fuel reactors was calculated for a range of
reactor dimensions and a power of 300 Mw. Calcu-
lations for a 60-Mw reactor that corresponds to the
ART were also made,

Calculation of Activation by Core Neutrons

The activation from core neutrons was deter-
mined directly from LTSF experimental data ob-
tained during static source tests with mockups of
a circulating-fuel reflector-moderated reactor
(CFRMR)'=3 and various reactor-shield configura-
tions. The resuvits for each configuration were
averaged over the thickness of the heat exchanger
to obtain the average specific activation. The
specific activation for each configuration was then
plotted as a function of the thickness of various
reactor regions (Figs. 12.1 through 12.5). The
data were corrected for minor differences between
the experiment and the reactor, such as cladding
thicknesses and air gaps.

The wvariation of the sodium activation with
respect to the beryllium reflector thickness (Fig.
12.1) appears to be independent of the boron cur-
tains, and it has an apparent relaxation length of

 

. T. Chapman et al.,, ANP Quar. Prog. Rep. June
10, 1955, ORNL-1896, p 194.

ZG, T. Chapman et al., ANP Quar. Prog. Rep. Sept.
10, 1955, ORNL.-1947, p 197.

3R. VW. Peelle et al., Sec. 13, this report.

204

5.7 em. The small variation with respect to heat
exchanger thickness (Fig. 12.2) implies that self-
shielding is offset by neutron moderation in the
heat exchanger, so that a sodium layer between the
reflector and heat exchanger would not be effective
in reducing the activation. Increasing the thick-
ness of the first boron curtain beyond 2 in. (0.66 g
of B'® per square centimeter) indicates that
further reduction of the sodium activation (Fig.
12.3) may be obtained by increasing the thickness

 

 

   

 

 

4 in. OF HEAT EXCHANGER, Y in. OF BORAL
5 [ CONFIGURATION: Be, 2in OF BORAL,

 

e
. 2—01—059—27
10 - T
5[ & CONFIGURATION: Be, Ypin. OF BORAL, N
4in. OF HEAT EXCHANGER, 2 in. OF BORAL —— |
i © CONFIGURATION: Be, 2in. OF BORAL, -
2 Bin OF HEAT EXCHANGER, 2 in. OF BORAL
o5 L & CONFIGURATION: Be, 2 in. OF BORAL,

 

4in. OF HEAT EXCHANGER, 2 in. OF BORAL .

|

 

 

 

4'—

 

 

 

 

 

 

 

SPECIFIC ACTIVATION PER g OF Na (activations/min}

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N T
W A=57cm
¥ 40? - N
= — | —— - —
< - \\
i 5 ~
2
N 'NOTE: DATA CORRECTED FOR BORAL THICKNESS
o 2 T AND COMPOSITION B -
LS
s 0%
8 10 12 14 16
BERYLLIUM THICKNESS (in )
Fig. 12.1. Mean Sodium Activation from Core

Neutrons in the Heat Exchangers of LTSF CFRMR
Mockups — Effect of Beryllium Thickness.
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

@’""“'
105 2—01—-059—-28
{ : | ] [
= ) | J I
= © CONFIGURATION: 8in. OF Be, 2in. OF BORAL, __ |
E HEAT EXCHANGER, 2 in. OF BORAL
w
§ 5 % A CONFIGURATION: 12 in. OF Be, ‘% in. OF BORAL, |
B HEAT EXCHANGER, 2 in. OF BORAL
g | ® CONFIGURATION: 12 in. OF Be, 2 in. OF BORAL,
5 HEAT EXCHANGER, 2 in. OF BORAL
=z | T
: |
O
U" .
e 2 ! ]
wl !
o.
z \
Q :
= I
2 T i ; ! -
A . L
E ‘\":"--—— [
G . I
ul t
o
0 5 —
km "
. . |
2]
«® -
3 o
< — “
| 1
N 27 S \
S NOTE: DATA CORRECTED FOR BORAL THICKNESS
— . AND COMPOSITION |
O I . |
I ‘ | :
5 I
0 { 2 3 4 5 6 7
HEAT EXCHANGER THICKNESS (in.)
Fig. 12.2. Mean Sodium Activation from Core

Neutrons in the Heat Exchangers of LTSF CFRMR
Mockups — Effect of Heat Exchanger Thickness.

of the enriched boron curtain in present designs.
The variation with respect to the thickness of the
second boron curtain (Fig. 12.4) is small. The re-
placement of natural boron with B'? has still to be
investigated. This would reduce the moderation
(largely by B for equivalent absorption.

The decrease in activation occurring with the
introduction of a hydrogenous layer (polyethylene)
between the beryllium reflector and the first boron
curtain is shown in Fig. 12.5 in terms of its beryl-
lium replacement effectiveness. Since polyethyl-
ene cannot be used in the reactor, the computed
effectiveness of zirconium hydride, which has high
thermal stability at reflector operating tempera-
tures, is shown for comparison.

The contribution of prompt core neutrons to the
sodium activation in the heat exchanger of a CFR
was calculated from the L.TSF experimental data
by application of conventional shielding transfor-

PERIOD ENDING DECEMBER 10, 1955

2—-01-059-29

1C T l I | a
jo CONFIGURATION: 8in. OF Be, BORAL, *t‘
4in, OF HEAT EXCHANGER, % in. OF BORA

& CONFIGURATION: 8in. OF Be, BORAL,
4in. OF HEAT EXCHANGER, 2 in. OF BORAL

‘\\ L]

 

 

 

 

NOTE: DATA CORRECTED
FOR BORAL THICKNESS
AND COMPOSITION ~ =~~~
! ! .

 

 

  

 

 

 

 

 

 

 

 

 

 

 

f
8°8 SPECIFIC ACTIVATION PER g OF Na(activations/min)

 

 

 

 

 

 

 

 

 

5
2]
O
= |
=
~ :
; ~
I {
N 2 [~ A CONFIGURATION: 12 in. OF Be, BORAL, Co
E 6in. OF HEAT EXCHANGER, 2 in. OF BORAL |
- ® CONFIGURATION: 12 in. OF Be, BORAL,
. 4in. OF HEAT EXCHANGER, 2 in. OF BORAL
~ ' .
w3 ! | 1 | i
0 0.5 1.0 1.5 2.0 2.5 3.0 35
BORAL THICKNESS (in.)
Fig. 12.3. Mean Sodium Activation from Core

Neutrons in the Heat Exchangers of LTSF CFRMR
Mockups — Effect of First Boral Curtain Thickness.

mations.? This implies that the neutron captures
in sodium in the heat exchanger may be expressed
in terms of a kernel involving approximately ex-
ponential neutron attenuation through the material
thicknesses along the roy connecting a small fis-
sioning region with a sodium volume element. In
terms of these transformations the prompt-neutron
contribution was expressed, as follows:

 

(1)

o o z k ZBrB
dp.c =—H(a,z)g ayr.c i e ,
where

@p.c = core-neutron-induced activation per
gram of sodium in a CFR heat ex-
changer (atoms/min},

 

4E. P. Blizard, Introduction to Shield Design, ORNL
CF-51-10-70, Part | rev. (Jan. 30, 1952) and Part |}
rev. {March 7, 1952).

205
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 2-01-059-30
10 :
E E T — — —JP
- - - — [ —
g A CONFIGURATION: 8 in. OF Be, 2 in. OF BORAL,
5 | 4 in, OF HEAT EXGHANGER, BORAL
= S 1 —t—1 — 4t
8 ¥ CONFIGURATION: 12 in, OF Be, 2 in. OF BORAL,
5 P 4 in. OF HEAT EXCHANGER, BORAL
z | : | l |
L ® CONFIGURATION: 2 in. OF Be, % in. OF BORAL,
o> ’ 6 in. OF HEAT EXCHANGER, BORAL
§ ! ‘
5 2 ) | _ ; ‘ : _ |
g |
= _-_‘—-—.}-_____
2 T
q |
= i
5 10t R |
<L - -
——
9 - 0 ——— T [
i - L - I
O I
a L — ‘ ' -
“ 5 NOTE: DATA CORRECTED FOR BORAL THICKNESS - 4
27 | AND COMPOSITION
w‘b
QO
* — —_
- —
N
| | i |
N 2 ' T
= :
N
—
g o |
ol Lt
k t
oY 40° -
0 1 2 3

Fig. 12.4.

BORAL THICKNESS {in.}

Mean Sodium Activation from Core

Neutrons in the Heat Exchangers of LTSF CFRMR
Mockups — Effect of Second Boral Curtain Thick-

ness.,

206

arr.c T

Tt = surface

core-neutron-induced activation per
gram of sodium in an LTSF mockup
(atoms/min),

source strength of LTSF
source plate,

0p = equivalent surface source strength

for CFR core
= A_p(1 )

A_ = relaxation length in CFR core for
activation neutrons (11.6 cm, with
the assumption that neutron removal
cross sections are valid for this
correction),

p = power density in CFR core {w/cm?),

t. = thickness of CFR core annulus (cm),

 

     
 

 

 

 

 

 

 

 

 

 

 

 

. s ey
e wm 2—01-059-3
10 [ T B — |
- __T, ——— e i -
R | i .
; # j |
S S— N
50 . .. —— S - e - 1
| BERYLLIUM THICKNESS, 12 in.
|
o | .
} /'VH[ Zer]= 5.8 X 10%2 atoms /cm®
|
|
W[ (eHy),] =775 X 1072 atoms/om”
- 20 b— — - Sa—— i
o : ‘ :
= .
s 1 | ‘
Q |
< 0 S . e !
g — e .. [
= - ZIRCONIUM HYDRIDE ———
<7 o . o ! : . .
. ST T -
@ \‘\-.\ ‘
5 - —_— < . \ —_— —_—
~ ~ |
- - ™~ . |
| RSN
! | ~ ~
L L NS N
POLYETHYLENE™ ™~ ™~
~ ~
o | - IS
1 .
R
i i ~
NOTE: DATA FROM LTSF TESTS | N
WITH CFRMR MOCKUPS ‘ ™~
1 | | | |
0 05 1.0 1.5 2.0 2.5 3.0 35

POLYETHYLENE OR ZIRCONIUM HYDRIDE THICKNESS (in.)

Fig. 12.5.

Mean Sodium Activation from Core

Neutrons in the Heat Exchangers of LTSF CFRMR
Mockups — Estimated Effect of Replacing Beryl-
livm in the Reflector with Polyethylene or Zir-
conjum Hydride.

Hla,z) =

r /Ty =

ratio of activation from an infinite
plane source to that from LTSF disk
source

2
/(1 = ™8 /225,

= LTSF source-plate radius (35.5 cm),

= average neutron relaxation length

(7.5 em),

distance from source plate to sodium
LTSF heat exchanger mockup
(em),

ratio of activation from spherical
CFR core of radius . (¢m)} and heat
exchanger of radius THx (em) to ac-
from infinite plane

in

tivation
source,

an
k
(—z————> = correction allowing for the sodium
z - Az displacement outward in the experi-
ment because of the bulk of alumi-
num in the boral and the air gaps
which are not present in the CFR
(plotted in Fig. 12.6),%

Az = thickness of boral and air gaps {cm),

e = allowance for the difference between
the neutron attenuation through boral
in the LTSF mockup and that through
B,C-Cu in the CFR,

EB = activation-reducing cross section for
the aluminum in the boral
= 0.0164 ¢cm~!, based on multigroup
calculations at Pratt & Whitney,
tg = thickness of boral (cm).

Calculation of Activation by Delayed Neutrons

The activation induced by delayed neutrons
emitted in the heat exchanger was estimated from
the experimental data taken at the LTSF with the
use of a moving fission belt (see Sec. 13). Since
the delayed-neutron emission from any point on
the belt is proportional to the time of exposure to
the circular opening (core hole) of the ORNL
Graphite Reactor, and since the emission is con-
stant at any point during the belt rotation (be-
cause the belt transit time is less than the aver-
age half life of the delayed-neutron emitters), the
distribution of delayed-neutron emission for the
belt can be expressed, as follows:

(2) nLT()’) = ”LT(O) VAR (y/d)2 K,

where

nLT(y) = number of delayed neutrons emitted
per second from a square centimeter of
the belt at a distance y (ecm) from the
center line (neutrons/cm?.sec),

”LT(O) = emission rate on the belt center line
(neutrons/cm?-sec),

 

PERIOD ENDING DECEMBER 10, 1955

UNCLASSIFIED
2-01-052-38
30 : ‘ | | \ !

 

2.0

 

o4

 

 

 

 

 

 

 

 

 

 

0 1 20 30 40 5C 60 70 80 90 100
z— Az

Fig. 12.6. Plot of £ vs (z — Az).

a = half width of the emitting area of the
belt (equal to the radius of the core
hole of the ORNL Graphite Reactor),

K = ratio of emission from belt completing

circuit in finite time to emission in
zero transit time.
decay of delayed-neutron emitters that
complete one-half the circuit is K ~

0.96.)

(The correction for

Normalizing the emission rate over the total belt
area to the total power from fissions in the belt
gives

2P, ..d
3 ny p(0) = —=L
®) LT Tal
where
P, + = total power from fissions in the belt

(5.2 w obtgined by normalizing to the
LTSF static tests),

d = number of delayed neutrons emitted per
watt of fissions (neutrons/w),
! = length of belt (cm).

5Under the assumption of exponential ray attenuation, the following expression was used to determine &:

In

E,q [(z - Az)/?\n} - E-‘{(z - Az)/)tn V1t [az/(z - Az)z] }

£,z =82/ ] - By {(z = Dap/\ V1 + (Y22}

 

In [z/(z - Az)]

207
ANP PROJECT PROGRESS REPORT

An approximate value for the age of delayed neu-
trons moderated from 500 to 3 kev in the heat ex-
changer was calculated, by means of the following
expression, to aid in interpreting the experimental
results in terms of the CFR geometry.

(4) T =f50 ) ,
E 1 fET E
which was approximated by
In (Eo/E )
3 2 No, ZS‘ ENO;

7

(4a) T = = 216 cm?

F = neutron energy (kev),
D = diffusion coefficient (1/3X N7 ),

1

2, = fotal neutron macroscopic cross section
for the heat exchanger {assumed to be
homogeneous),

Nz. = number of atoms per cubic centimeter of
the ith kind,

&, = average logarithmic energy decrement for
atoms of the ith kind,

O . = total neutron microscopic cross section
for atoms of the ith kind averaged over the
lethargy range.

The average composition {in g/cm?) of the LTSF
heat exchanger mockup was: nickel, 2.23; sodium,
0.65; fluorine, 0.54,

Since /7 (14.7 cm) is small compared with the
belt length in the heat exchanger (120 cm) but
significant compared with the belt half width
(35.5 cm), the ratio was obtained of the slowing-
down density from a source of infinite length but
finite width (with sources distributed according
to Eq. 2) to the slowing-down density from an
infinite plane source representing the uniform re-
The slowing-down density
from an infinite plane was taken to be
e_,02/11'7'

— = 0.0192n
4T

actor heat exchanger.

(5) qmp[ = no:)p[ opl !

where

ep) = delayed neutron emission per square cen-

timeter from an infinite plane (neu-
trons/cm?.sec),

p = distance from infinite plane to sodium

sample (2.54 cm).

208

The slowing-down density from the LTSF fis-
sion belt was taken to be

“ ~V(p? —y%)/ar

e
(6) = 2 n (y) dy
LT LT 4T 1
0
since
q; = slowing-down density from an infinite line
source
2
= nle—r /47/4777',
n; = line source strength, taken to be n; .(y) dy
(neutrons/cm?.sec),
r = distance from the line to the sodium sample
2 (czm)' 2
r = p — y -

The integral was evaluated numerically to ob-
tain the following resuit:

 

(7) qor = 0.01937, .. ,
where
P d
— LT
(8) LT = 2al

For equivalent source strengths the ratio of slow-
ing-down density (‘prl/qLT) at 3 kev for the LTSF
belt compared with an infinite plane is thus seeo
to be approximately unity.

The fuel in the heat exchanger of a CFR is
approximated by an infinite plane with a delayed
neutron source strength of

PRd
(9) Popl T, Taxtux
where !
P, = reactor power (watts),
V, = total fuel volume of system (em3),
fyx = fraction of heat exchanger volume con-

taining fuel,
tx = heat exchanger thickness (cm).

If it is assumed that the sodium activation is
proportional to the slowing-down density of 3-kev
neutrons, then the specific activation in the heat
exchanger of a CFR is given by

0.019210,; iy

0.019372'LT Ty

4R-D Topl
(10) L

AL T-D LT

 

 
It

sodium in
{atoms /min),

delayed neutron activation per gram of

a CFR heat exchanger

delayed neutron activation per gram of

sodiumin a LTSF mockup (atoms/min),

(For

the calculations presented in

FPERIOD ENDING DECEMBER 10, 1955

Calculated Total Activations for Several Reactors

The total activation of sodium in the heat ex-
changers from core and delayed neutrons was
caleculated for four circulating-fuel reactors (Table
12.1). The power for three of the reactors was
300 Mw, and the fourth reactor corresponded to the
ART, with a power of 60 Mw. In addition to the
total saturated activities, the total activity after

Table 12.1,a; +.p = 1500 atoms/min.)

30 hr of operation was determined.

TABLE 12.1. SODIUM ACTIVATION IN THE HEAT EXCHANGERS OF CIRCULATING-FUEL REACTORS

 

Configuration

 

 

] 2 3 4*
Island radius (in.) 4 4 4
Core radius (in.) 8.59 10.16 9.74 10.5
Beryllium thickness (in.) 8 12 16 1
First B'0 curtain (atoms/cm?) 3.52 x 1022 3,52 x 1022 3,52 x 1022 *x
Heat exchanger thickness (in.) 8 6.5 5 2.1
Second B0 curtain (atoms/cm?) 1.12 x 1022 .12 x 1042 1,12 x 1022 *
PEWai density (kw/cm>) 5.5 2.75 2.75 0.72
Weight of sodium in heat exchanger (g) 118,000 144,000 137,000 42,000
Reactor power (Mw) 300 300 300 60
Saturated activity from core neutrons (curies) 19,300 3,350 430 470
Saturated activity from delayed neutrons (curies) 1,970 1,530 1,250 61
Total saturated activity (curies) 21,270 4,880 1,680 531
Total activity after 30 hr (curies) 16,060 3,680 1,270 400

 

*Configuration 4 comresponds to the ART configuration,

**First and second B]0 curtains consist of 0,375 in. of natural B4C.

209
ANP PROJECT PROGRESS REPORT

13. LID TANK SHIELDING FACILITY
R. W. Peelle

J. M. Miller
Applied Nuclear Physics Division

W. J. McCool

J. Smolen

Pratt & Whitney Aircraft

Analysis of the static source tests in the second
series of experiments with mockups of a circulating-
fuel reflector-moderated reactor and shield (RMR-
shield) in the Lid Tank Shielding Facility (LTSF)
was completed. A summary of the tests is pre-
sented. The dynamic source tests were completed,
and the tests are being analyzed. Two tests with
the dynamic source are described here.

STATIC SOURCE TESTS WITH MOCKUPS CF A
REFLECTOR-MODERATED REACTOR
AND SHIELD

H. C. Woodsum!

The static source tests in the recent series of
experiments with RMR-shield mockups were sub-
divided into sections according to the parameters
being investigated. These parameters, descrip-
tions of the various mockups tested, and an index

J. Smolen

to the results of the measurements are given in
Table 13.1. Much of the information has been
reported previously,?:3 as indicated in the table,
Most measurements not previously published are
presented in this report,

Effect of Varying Lead Thickness

The thickness of lead was varied in mockups
having 8-, 12-, and 16~in.-thick beryllium reflectors.
Gamma-ray dose-rate measurements for the 8- and
16-in,-thick reflectors are reported here (Figs. 13.1
and 13,2), but the data are not directly comparable
because the ‘‘core shell”’ materials differed. With
the 8-in. reflector, Li-Mg (¥ in, thick) was sub-
stituted for the Inconel (% in. thick) used to mock
up the core shell. Inconel gives a hard capture
gamma ray, approximately 9 Mev. Thus, in order
to correlate the gamma-ray data from these two
beryllium thicknesses with data from 12 in. of
beryllium,2 two configurations were measured to

 

1Shield Design Group, Pratt & Whitney Aircraft at
ORNL.

2, T. Chapman, J. B, Dee, ond H. C. Woodsum, ANP
Quar. Prog. Rep. June 10, 1955, ORNL-1896, p 194.

3G. 1. Chapman et al., ANP Quar. Prog. Rep. Sept, 10,
1955, ORNL-1947, p 197.

210

determine the effect on the gamma-ray dose rate of
interchanging Inconel and Li-Mg (Fia. 13.3).

The effects on the thermal-neutron flux of varying
the thickness of lead in mockups with 8 and 16 in.
of beryllium are shown in Figs. 13.4 and 13.5,
respectively. Measurements beyond mockups with
12 in, of beryllium were previously published.?

Fast-neutron dose-rate measurements beyond
mockups with various lead thicknesses and 16 in.
of beryllium (Fig. 13.6) appear to be self-consistent
and to be consistent with data for 12 in. of beryl-
lium.2 The measurements for 8 in. of beryllium
(Fig. 13.7), however, do not appear to be self-
consistent,  This was apparently due to inter-

mittent instrument malfunctioning. The results
Sunane
3 2-01—-057—-66—-169

GAMMA-RAY DOSE RATE {mr/hr)

o CONFIGURATION 2-A, NG LEAD

" a CONFIGURATION 2-B, 1'% in. OF LEAD
® CONFIGURATION 2-C, 3n.OF LEAD
A CONFIGURATION 2-D, 4% in. OF LEAD

 

60 7O 80 90 100 HC {20 130 140 150 160
z, DISTANCE FROM SOQURCE (cm)

Fig. 13.1. Gammo-Ray Dose Rate Beyond RMR-
Shield Mockups — Effect of Varying Thickness of
Lead Behind 8 in. of Beryllium,
TABLE 13.1. SUMMARY OF REFLECTOR-MODERATED REACTOR AND SHIELD MOCKUP TESTS AT THE LTSF

Note: The water region behind the lead was borated to 1,95 wt % unless otherwise indicated

PERIOD ENDING DECEMBER 10, 1955

 

Configu-

Components and Thicknesses

 

Index to Plots of Data (&)

 

 

) Nickel Lead a
Investigation ration S(.I:orlel se:r”ium Boral HNQF';' :anks Boral (Pressure (Gamma T‘:TGI Other Deviar ;:2,(.")) Gamma-Ray Thermal- Fast o
No. ‘e {(Re ‘ector) in.) (Heat 'xc anger) (in.) Shell) Shield) ir ther Deviations Dose Rate Neutron Neutron Activations
(in.) (in.) (in.) (in.) (in.) (em) Flux Dose Rate
Effect of varying water bora- 1-A Plain water in Lid Tank (i. e., no RMR tank instalied) o 0 ORNL-1896 ORNL-1896
tion 1-B RMR tank installed and filled with plain water 4,55(€) 5.50(4) " " ORNL-1896
1-C RMR tank installed and filled with 0.45%-borated water 4.55(6) 5.50(d) x X X
1.D RMR tank installed and filled with 1.95%-borated water 4.35(°) 5.3  ORNL-1896 ORNL-1896 X
Effect of varying lead thick- 2-A Li-Mg, Y% 8 2 4 1 0 6.47 52.2 Fig. 13,1 Fig. 13.4  Fig.13.7
ness behind 8 in, of beryllium 2-B i " " " n " ],1/2 7.36 56.9 " " "
9.C 1 " " I 1" 1" 3 6.55 59.9 ' " "
2.D 1" "t n 1 n " 41/2 6.84 64.0 1" 1 n
Effect of varying lead thick- 3-A Inconel, ]/8 12 2 4 2 1 0 4.30 59.6 ORNL-1896 ORNL-1896
ness behind 12 in. of 3-B " " y " ! " 1% 4.89 64.0 " §
beryllium 3-C " " " " I " 3 4,78 67.7 " n ORNL-1896
3-D " n rn n 1 1 41/2 9.83 76.1 x x x
3-D° " " " " " " " 6.07 72.8 ORNL-1896 x x Au, ORNL-1896
3-D” " " " " " " " 4,77 71.5 " ORNL-1896 ORNL-1896
3.E " ] 1 Ii " 1 6 5.06 75.6 n 1"
3-F L I d y iy d 7% 555 79.9 t " ORNL-1896
1-G 1 " " 1 " " 9 6.14 84.3 I 1" x
Effect of varying lead thickness 4-A Inconel, 1/8 16 2 4 2 1 3 6.45 79.8 Fig. 13.2 Fig. 13.6 Au and Na
behind 16 in. of beryllium 4-B " " " ! " " al, 6.0 82.9 ! Fig. 13.5 .
4-C " | " 1 I 1 6 6.93 87.7 I n
Inconel core shell and capture 5-A Inconel, ]/8 8 2 4 2 1 41/2 5.03 61.6 Fig. 13.3 x Au,Ng, ORNL-1947
gamma-ray dose 58 LiMg, % 12 " " " ' " 5.15 72.2 I x
Effect of Inconel cladding be- 6-A Inconel, }’8 12 2 4 2 1 4]/2 5.72 0.020 in. of Inconel behind beryllium 72.5 ORNL-18%6
tween beryllium and first 6-B " " i " " " " 5.85 0.125 in. of Inconel behind beryliium 72.9 "
boral curtain
E ffect of boral thickness 7-A Inconel, }é 8 !/é 4 2 1 4],6 4,24 57.0 X X Na
behind 8 in. of beryllium 7-B Li-Mg, ) n 1%, " 1Y E n 7.68 62.3 x x x
7.C 1 1" 2 1" 1,2 1 " Na
7.D n 1 3]/2 1 I it 1" 7.44 64.6 Ng
Effect of heat exchanger thick- 8-A Li-Mg, 1/‘" 8 1 6 1 1 3 x Na, LiCl
ness and boral distribution 8-B " It n " " " n 6.11 I/2--in.-thick boral slab behind first and 62.0 x Na, ORNL-1947,
behind 8 in, of beryllium second NaF-Ni tanks (total of three tanks) MgCI2
8-C Inconel, '/8 " 2 " 2 " 41/2 Ma, ORNL-1947

211
ANP PROJECT PROGRESS REPORT

TABLE 13.1 (continued)

 

Components and Thicknesses

 

Index to Plots of Dafu(b)

 

 

Configy- Core Beryllium NaF-Ni Tank Nickel Lead Total 2,
Investigation ration Shell (Ref:ecfor) Boral (Hec:lf E |chc: sr) Boral  (Pressure  {Gamma :i: Other Deviations (2'“) Gamma-Ray Thermal- Fost .
Na. (in) (i (in.) (ixn )u ge (in.) Shell) Shield) (cm) ' Dose Rate Neutron Neutron Activations
: . . (in.) (in.) Flux Dose Rate
Effect of boral thickness be- 9.A Inconel, !/B 12 0 4 ) ] 4}'2 6.55 68.2 ORNL-1896 x
hind 12 in. of beryllium 9-B " " ‘,5 n n " I 5.98 68.9 " x Na, ORNL-1896
9-C " " ] " " " " 5.81 70.0 x x Na, ORNL-1896
9-D " " 1% " 1Y " " 8.57 72.3 x x x
9-E  Li-Mg, Y " 2 " % " " 5.16 68.4 x Na
Effect of heat exchanger thick- 10-A Inconel, }8 12 ]/2 2 2 1 4]/2 4.56 62.4 x x x Na
ness and boral distribution 10-B " " 2 6 " " " 6.09 77.9 x x Na, ORNL-1896
behind 12 in. of beryllium 10-C ' i Y " " " " 5.50 73.5  ORNL-1896 x x "
10-D " n n n ]é n 1 5.71 69.9 X X n
Effect on secondary gamma-ay 11-A Li-Mg, ]& B ]]/2 4 11/2 1 4]/2 7.74 ]/z-in.-thick boral slab on each side of last 64.9 Fig. 13.8 X
dose of distributing boral in 11/2-in.-thick lead slab
lead
Effect on secondary gamma-ray 12-A Li-Mg, }:‘ 8 2 4 2 1 4!/2 7.05 Lead-water spacing as shown in Fig., 13.9 60.4 Fig. 13.9
dose of lead spacing 12-B " " " " " " " 8.65 I 62.0 3
12-C " " I I 1 " " 8.75 I 62.1 1"
12-D " 1 E t 1 1" 1" 8.55 " 61.9 n
12-E n u 1" 1" i " 6 7.94 n 65.1 t
12-F n 1 n n 1" " " 7.05 1 60.4 "
Capture gamma-ray dose from 13-A Inconel, }é 12 2 4 2 ] 9 6.40 z-in.-fhick Li-Mg slab behind lead 85.2 x
shield and effect of gamma- 13-B " " " " ” i n 6.00 k‘-in.-thick boral slab behind lead 84.8 x(€) x{€} Na, GORNL-1896
ray streaming 13-C n " " " " " I H Same as configuration 13-B with bismuth on 84.8 x(€) x
sides of RMR tank
13-D n " 1" 1 " n " " Same as configuration 13-C with !fi-in.-fhick 84.8 x
boral slab just inside water bag
13-E n " " " " " I 6,14 Bismuth and lead shields on sides of 84.3 x
RMR Tank
Effect of varying pressure shell 14-A Li-Mg, !:1 8 2 4 2 0 41/2 7.48 62.1 x
thickness 14-B Inconel, ‘/8 12 " " " 2 i 4.93 74.2 ORNL-1896 x
Effect of heavy metals in re- 15-A Li-Mg, ]fi 8 2 4 2 1 41/2 7.42 3-in.-thick bismuth slab behind beryllium 72.2 Fig. 13.13 Fig. 13.14
flector 15-B " " i " ) " " 7.56 2-in.-thick copper slab behind beryllium 69.8 " " Na, ORNL-1947
Effect of replacing borated 16-A inconel, I/é 12 1]/2 4 'Ilé 1 4]/2 9.27 Borated water replaced by plain water 73.0 Fig. 13.15 Fig. 13.16

water with plain water behind

configuration

 

 

 

(@)Thickness of configuration measured from swface of source plate to first water layer.

(8)y indicates that z traverse measurements have been taken but not published to date. Configuration numbers cited in
ORNL-1896 should be disregarded because the numbering system has been changed.

212

() Air bag inserted between source plate and aluminum window of RMR tank.
({)Represents total thickness of air and tank wall of RMR tank.

(€)|ln configuration 13-B, x and y traverses were also made; in configuration 13-C, y traverses were also made.
—
2—01-057—66-161

**Jw

PERIOD ENDING DECEMBER 10, 1955

Wrs. kW

2—01—-057—-66—-176

 

e S S — j:i;’
£I—]:;f T e
- | TS ]
5 _ o CONFIGURATION 2-A, NO LEAD

 

—

 

+— & CONFIGURATION 2-B, ii/z'!n‘ OF LEAD — -

GAMMA —RAY DOSE RATE (mr/hr})
o

© CONFIGURATION 4-A, 3in. OF LEAD
A CONFIGURATION 4-8, 4l5in. OF LEAD
O CONFIGURATION 4-C, & in. OF LEAD

 

-2
10

80 20 100 Ho 120 {30 140 150 160
2, DISTANCE FROM SOURCE (c¢m)

Fig. 13.2. Gomma-Ray Dose Rate Beyond RMR-
Shield Mockups =« Effect of Varying Thickness of
Lead Behind 16 in. of Beryllium.

S
2-01-057—66-175

2 o

10" -0 CONFIGURATION 5-A, INCONEL, 8 in. Be
o CONFIGURATION 2-D, Li-Mg, 8in. Be —

GAMMA-RAY DOSE RATE {(mr/hr)

o

CONFIGURATION 3-D INCONEL, 12 in. Be
& CONFIGURATION 5-8, Li-Mg, 12in.Be

 

70 80 90 100 HO 120 130 140 150 160
z, DISTANCE FROM SOURCE {cm)

Fig. 13.3. Gamma-Ray Dose Rate Beyond RMR-
Shield Mockups — Effect of Inconel Core Shell.

 

 

0 CONFIGURATION 2—C, 3in.OF LEAD —— —

 

 

\\ ® CONFIGURATION 2-D, 4% in OF LEAD
1ot D\ : ! . | | Lo
C— )T - ; - — t

 

 

 

 

 

 

 

 

 

 

THERMAL - NEUTRON FLUX (nvm)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50 60 70 80 30 100 1o 120 130
2, DISTANCE FROM SOURCE (cm}

Fig. 13.4. Thermal-Neutron Flux Beyond RMR-
Shield Mockups = Effect of Varying Thickness of
Lead Behind 8 in. of Beryllium,

for 0 and 3 in. of lead are thought to be in error by
a factor of 2, but the remaining data are probably
more accurate, |t is hoped that this set of meas-
urements can be made again. In the meantime, the
data obtained to date are presented in lieu of more
exact measurements.

Study of Secondary Gamma-Ray Production

In order to analyze the gamma-ray dose rate from
secondary gamma rays produced in the shield,

213
ANP PROJECT PROGRESS REPORT

L
2-01-057-66-162

S

n

n

 

CONFIGURATION 4-8

THERMAL-NEUTRON FLUX (nvfh)

 

10
5
2
1072
70 80 90 100 1o 120 130
7, DISTANCE FROM SCURCE {(cm)
Fig. 13.5. Thermal-Neutron Flux Beyond RMR-

Shield Mockups — Effect of 41/2 in. of Lead Behind
16 in. of Beryllium,

two sets of tests were run, The first was a study
of the effect of distributing boral in the lead region,
and the second was a study of the effect of spacing
a portion of the lead out into the borated water.

In a configuration that had a 4'/2-in.-thick lead
region, a Y-in.-thick boral slab was inserted on
each side of the last l'/z-in.-thick lead slab. No
large decrease in the gamma-ray dose rate resulted
(Fig. 13.8). Thus it appears that the secondary

214

Sy

2-01-057-66-163

FAST-NEUTRON DOSE RATE (mrep/hrl

® CONFIGURATION 4-A, 3-in. OF LEAD
A CONFIGURATION 4-B, 4% -in.OF LEAD
O CONFIGURATION 4-C, 6-in.OF LEAD

 

8O 20 100 110 120
Zz, DISTANCE FROM SQURCE (cm}

Fig. 13.6. Fast-Neutron Dose Rate Beyond RMR-
Shield Mockups « Effect of Varying Thickness of
Lead Behind 16 in. of Beryllium.

gamma-ray dose rate is not mainly due to thermal-
neutron captures in the lead. This is confirmed by
calculations of the lead capture gamma-ray dose
rate based on the thermal-neutron flux measure-
ments immediately behind the lead and on the
thermal-neutron capture cross section in the lead.

Spacing the last 1Y%-in.-thick lead slab out into
the borated water showed that the total gamma-ray
dose rate decreased as the water thickness pre-
ceding the slab was increased from 13 to 7 in.
(Fig. 13.9). However, as the water thickness was
increased beyond 7 in., there was no detectable
change in the total gamma-ray dose rate. This is
Gonrtr i
2 2-04~-057-66-172

10

5

2

10

5
<
i
.
o
2
€

w2
W
<I
x
[
W)

O i -

O
Z
]
18
5

o 5
=z
l
—
W
I
b

2

To

O CONFIGURATION 2-A, NO LEAD

5 " A CONFIGURATION 2-8, 1%—in. OF LEAD
® CONFIGURATION 2-C, 3-in.OF LEAD ~
A CONFIGURATION 2-D, 4% —in. OF LEAD

|
2 e [

 

107¢
50 60 70 80 30 100
z, DISTANCE FROM SOURCE {e¢m)

Fig. 13.7. Fast-Neutron Dose Rate Beyond RMR-
Shield Mockups — Effect of Varying Thickness of
Lead Behind 8 in. of Beryllium.

shown by a comparison of configurations 12-C and
12-D, which correspond to spacings of 7 and 11 in.
of water, respectively. Consequently, the gamma-
ray dose rate for 7 in. of water was taken to be the
“orimary’’ dose rate (that is, the dose rate not
caused by secondary gamma rays generated in the
shield itself). The secondary gamma-ray dose
rate from the shield then was obtained for each
water spacing less than 7 in. by subtracting the
dose rate for each water spacing from that for the

7-in. water spacing. The secondary-dose-rate

PERIOD ENDING DECEMBER 10, 1955

2-04-057-66-166

GAMMA-RAY DOSE RATE (mr/hr)

® CONFIGURATION 7-8, NO BORAL IN LEAD REGION
O CONFIGURATION 41-A_ {in. OF BORAL IN LEAD REGICN

 

70 BO 20 100 MO 120 130 140 150 {60
2z, DISTANCE FROM SOURCE (cm)

Fig. 13.8. Gomma-Ray Dose Rate Beyond RMR-
Shield Mockups — Effect of Boral in Lead Region.

curves were then plotted as a function of the
distance behind the last lead siab (Fig. 13.10),

The secondary gamma-ray dose rate for a
constant distance ¢ from the back of the lead
was then plotted as a function of the lead-water
spacing, t_, and was compared with the fast-
neutron dose rate and thermal-flux attenuation
in water (Fig. 13.11). It was found that the
secondary gamma-ray dose rate for small water
spacings, ¢ _, and small water thicknesses, ¢,
decreases at nearly the same rate as the thermal-
neutron flux., This can be explained by the fact
that the gamma-ray dose rate from boron capture
gamma rays makes up the major portion of the
secondary gamma-ray dose rate at small water
thicknesses, and the boron capture gamma-ray
dose rate is, of course, directly proportional to
the thermal-neutron flux in the borated water. For
large water spacings, however, there was enough
scatter of experimental points to mcke the com-
parison questionable,

Another experiment was performed to clarify
some of the uncertainties present in the previous
investigations of the secondary gomma-ray dose
rate. |t was thought that by placing an additional
lead slab in the water behind the shield it would

215
ANP PROJECT PROGRESS REPORT

o

no

GAMMA-RAY DOSE RATE (mr /hr)
on

CONFIGURATION 42-A

CONFIGURATION 12-B

5 CONFIGURATION 412-C

S

2-01-057-66-168

 

 

 

CONFIGURATION {12-D

 

2 CONFIGURATION 42-E

CONFIGURATION 42-F

50 60 70 80 90 100

 

110 120 130 140 150 160

z, DISTANCE FROM SOURCE (cm)

Fig. 13.9. Gamma-Ray Dose Rate Beyond RMR-Shield Mockups — Effect of Lead Spacing.

be possible to tell something about the energy of
the secondary gamma rays by their attenuation in
the water and lead. In configuration 12-E (8 in. of
beryllium), a lead slab ('ll/2 in. thick) was placed
]2]/2 in, behind the 4]/2-in.-fhick lead gamma shield.
Gamma-ray dose-rate measurements behind the
configuration (Fig. 13.9) should be compared with
those behind configuration 12-F, where the extra
lead slab was placed behind a mockup in which
the lead region was spaced in such a way as to
virtually eliminate the secondary gamma-ray dose.
These, in turn, can be compared with an extrapo-

216

lation of measurements behind configurations 2-A
through 2-D, which represents the total gamma-ray
dose rate beyond a 6-in.-thick lead region not
spaced out. This extrapolated curve (Fig. 13.9)
was designated as configuration 6-0. It should
be noted that configurations 12-E, 12-F, and 6-0
all have 6 in. of lead, and thus the ‘‘primary"’
gamma-ray dose rate is attenuated by the same
amount in every case. Since the lead in configura-
tion 12-F was spaced in such a way as to eliminate
the secondary gamma-ray dose rate, subtracting
the dose rate for configuration 12-F from that for
—
2-01-057-66-177

  
 

SECONDARY GAMMA-RAY DOSE RATE (mr/hr)
S

% =LEAD-WATER SPACING

0 20 30 40 50 60 70 80 20
t, THICKNESS OF WATER BEHIND LAST LEAD SLAB (cm)

Fig. 13.10. Secondary Gamma-Ray Dose Rate
from RMR-Shieild Mockups as o Function of the
Wate TRickness (t) Behind the Last Lead Slab.

configuration 12-E gives the attenuated secondary
gamma-ray dose rate. Similarly, the unattenuated
secondary dose rate can be obtained by subiracting
the gamma-ray dose rate for configuration 12-F
from that for configuration 6-0. The attenuated
and unattenuated secondary dose rates were then
plotted as a function of the distance from the
secondary source (Fig. 13.12), which was con-
sidered to be at the back surface of the ‘‘dry’
lead, that is, at the back surface of the lead pre-
ceding the borated water. By inspection of these
two curves, it may be seen that there is a “*harden-
ing'' of the secondary gamma-ray spectrum caused
by attenuating the secondary gamma rays by ]'/2 in,
of lead. By comparing the slope of the attenuated
secondary dose rate with the calculated attenua-
tion of hydrogen capture gammas (2.2 Mev) in
water, it was found that the attenuated secondary
gamma rays had a much shallower slope than the

PERIOD ENDING DECEMBER 10, 1955

g

S
2-01-057-66-174

SECONDARY GAMMA-RAY DOSE RATE (mr/hr)
THERMAL-NEUTRON FLUX (nvy,)
FAST-NEUTRON DOSE RATE (mrep/hr)

} = WATER THICKNESS BEHIND LAST LEAD SLAB

 

0 5 10
ts, LEAD-WATER SPACING (cm)

Fig. 13.11. Secondary Gamma-Ray Dose Rate
from RMR-Shield Mockups as a Function of the
Lead-Water Spacing (t): Comparison with Thermal-
Neutron Flux and Fast-Neutron Dose Rate.

217
ANP PROJECT PROGRESS REPORT

 

2-01-057-66-173
T e -

T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SECCNDARY GAMMA-RAY DOSE RATE (mr/hr)

 

 

 

 

 

 

 

 

 

20 30 40 50 60 T0 80 20 100
DISTANCE FROM SECONDARY SOURCE (cm)

Fig. 13.12. Effect of Attenuating Secondary
Gamma-Ray Dose from RMR-Shield MockupsThrough
11/2 in, of Lead.

2.2-Mev gamma ray would indicate, Thus it
appeared that lead capture gamma rays (7.4 Mev)
must also contribute to the secondary gamma-ray
dose rate. A simple calculation of the attenuation
of 0.48-Mev boron capture gamma rays through
]'/2 in, of lead showed that they would contribute
negligibly to the attenuated secondary dose rate,
Therefore, apparently the total unattenuated
secondary dose rate is made up of three component
secondary sources: boron capture gammas, lead
capture gammas, and hydrogen capture gammas.
The attenuated secondary dose rate is made up of
the hydrogen and lead capture gamma-ray dose
rate attenuated through ]]/2 in. of lead.

Effect of Heavy Metals in the Reflector

In a study to determine the effect of repiacing
the outer region of the reflector with a heavy metal,
3 in. of bismuth or 2in. of copper was added behind
8 in. of beryllium. Since in a spherical reactor
these heavy-metal layers would be at a smaller
radius than the lead gamma-ray shield, it seemed
possible that some weight saving might result by

218

repkicimg part of the lead gamma-ray shielding
with an equivalent thickness of some heavy metal
in the reflector. The results (Fig., 13.13) show
that on a thickness basis the bismuth and copper
are both much less effective than lead, and they
are, of course, much less effective than their
attenuations (neglecting secondary
gamma-ray production) would predict. This is due
to the fact that the additional attenuation gained
by use of the heavy metal is offset by the produc-
tion of secondary gamma rays. A calculation of
the weight savings which might result from intro-
ducing 3 in, of bismuth into the reflector and
removing an equivalent thickness of lead from the
outer region of the gamma shield indicates that
for the moderate lead thicknesses (up to 4]/2 in.)
characteristic of the divided shield, small or
negligible weight savings ate made. However, for
unit shields where the outer region of lead is at
a large radius, some weight saving may result by
replacing lead with a heavy metal in the reflector.

It is not obvious from these tests which of the
two metals would be superior. Unfortunately, only
a 2-in.-thick slab of copper was available, and,
for this thickness, the self-absorption of secondary

calculated

{02 2-01-057-66-165

— — b S

CONFIGURATION 5-A, 3in. OF BISMUTH
5 _ CONFIGURATION 15-8, 2in. OF COPPER
CONFIGURATION 2-D, NO HEAVY METAL, 8in. OF Be
CONFIGURATION 5-B, NO HEAVY METAL, 12in. OF Be

GAMMA-RAY DOSE RATE {(mr/hr)

 

1072
80 9C 100 10 120 130 140 150 160 170
z, DISTANCE FROM SOURCE {cm)

Fig. 13.13. Gamma-Ray Dose Rate Beyond RMR-
Shield Mockups - Effect of Heavy Metals in Re-

flector Region.
gamma rays would be small, Hence, on a thickness
basis, the copper would appear to be less effective
than tests with a larger thickness would probably
show,

The copper did appear to be as effective in
shielding fast neutrons (Fig. 13.14) as was the
berytlium which it replaced, while the bismuth was
considerably less effective than beryllium. On a
weight basis, the copper and bismuth were equally
effective in shielding gammas.

L
2 2-01—057-66-164
e CONFIGU‘RATION 154Aj 3in. OF BISMIUTH, 8in. OF lEn:,_—j-—j
7.42 -cm AIR SPACE.
—O CONFIGURATION 15-B, 2 in. OF COPPER, 8in. OF Be,
5| 7.56-cm AIR SPACE.
A CONFIGURATION 3-D”, NO HEAVY METAL, 12 in. OF Be,
4.77-cm AIR SPACE.
- A CONFIGURATION 5-A, NO HEAVY METAL, 8in. OF Be,
5.03-cm AIR SPACE.
2 [ .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FAST-NEUTRON DOSE RATE (mrep/hr)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10-1
70 80 a0 100 10 120
z, DISTANCE FROM SQURCE (em}

Fig. 13.14. Fast-Neutron Dose Rate Beyond
RMR-Shield Mockups — Effect of Heavy Metals in

Reflector Region.

Effect of Borating the Water Shield

Borating the water of the shield to 1.95 wt %
decreased the gamma-ray dose rate by a factor of
2.3 (Fig. 13.15) and the thermal-neutron flux by a
factor of about 15 (Fig. 13.16).

PERIOD ENDING DECEMBER 10, 1955

2-04-057-66-170

|
B

CONFIGURATION 2~D, {.95 % BORATED WATER

GAMMA-RAY DOSE RATE (mr/hr)

 

80 90 100 110 20 130 140 150 160
z, DISTANCE FROM SOQURCE (¢m)

Fig. 13.15. Gamma-Ray Dose Rate Beyond RMR-
Shield Mockups ~ Effect of Replacing Borated
Water Behind Mockup with Plain Water.

DYNAMIC SOURCE TESTS WITH MOCKUPS OF A
REFLECTOR-MODERATED REACTOR
AND SHIELD

For the dynamic source tests the source plate of
the LTSF was removed, and a mechanism that
simulated the circulating fuel of a reactor was
installed. The mechanism consisted of an electric-
motor-driven continuous chain belt (Fig. 13.17) to
which MTR-type fuel plates were attached. The
plates moved continuously past the exit end of a
hole in the ORNL Graphite Reactor (LTSF source
plate removed) to the heat exchanger region of the
RMR-shield mockups. The time required for one
complete rotation of the belt, called the transit
time, T, varied between 1 and 2'/2 sec, During
these tests, radiation measurements were made
behind two basic configurations {with 11 and 3 in.
of lead), but the data have not yet been analyzed.
Measurements of the sodium activation in the heat

219
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 201 —057—66—171
G Ff’, ;r [ ] ! e
T e T - ‘T
1 T
fi_l—l—__ i 1 e
2l e S A
'04 oo T \ T . —_— - : i
sl N T
| . : | |
2 0\ L-CONFIGURATION 16-A, '
3 : BORATED WATER . ‘
IO /.. ._..._k_______':‘ti_:::fifirii,ii:“ - ] %, .-
® _‘L .
T
- . T -

 

 

 

 

 

 

THERMAL-NEUTRON FLUX (nv,, )

 

 

o |- | :

CONFIGURATION 9-D
] PLAIN WATER——7 77 . N e
T \ L x\____,,,__
sp— - | ;S\ N —
A ] 8 )

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

60 70 80 20 100 110 120 130 140 150
2, DISTANCE FRCM SQURCE {cm)

Fig. 13,16. Thermal-Neutron Flux Beyond RMR-
Shield Mockups — Effect of Replacing Borated
Water Behind Mockup with Plain Water,

exchanger are reported below. In addition, a de-
termination of the fission-product gamma-ray spec-
trum is presented, For this determination the
RMR-shield mockup was removed and borated
water and lead were placed inside the perimeter
of the fuel loop.

Sodium Activation in the Heat Exchanger
W. J. McCool

Sodium fluoride pellets were exposed at approxi-
mately the center line in the sodium fluoride—filled

220

nickel tanks that were used to simulate a circulating-
fuel-reactor heat exchanger (Fig. 13.18). The other
components used in the mockups to simulate dif-
ferent regions of the reactor and shield were:
(1) fuel plates on the motor-driven belt, (2) 1/S-in.-
thick Inconel “‘core shell,”’ (3) 12-in.-thick (three
4-in. slabs) beryllium reflector, (4) first boron
curtain (four Y-in.-thick boral slabs), (5) 2-in.-thick
heat exchanger, () fuel plates on the belt, (7) 2-in.-
thick heat exchanger, (8) second boron curtain
(four I/‘,-in.firhick boral slabs), (9) 1-in.-thick nickel
presgyre shell, and (10) 3-in.-thick lead (two ]]/2-
in.-thick slabs) gamma-ray shield. Most of the
regions are shown in Fig., 13.18; however, the
fuel plates, the second boron curtain, and the
nickel pressure shell are not visible. In addition
to the components shown in Fig. 13.18, an alumi-
num tank containing borated water was placed in
the void behind the lead slabs to simulate the
neutron shield.

Measurements of the sodium activation were
made for various transit times (Fig. 13.19) in
mockups that represented a reactor with approxi-
mately 50% of the fuel in the core and 50% in the
heat exchanger. In test 1 the fuel belt was not
rotated; that is, the uranium plates were held in a
stationary position both at the exit end of the hole
in the ORNL Graphite Reactor shield and in the
heat exchanger. This corresponded to a circu-
lating fuel with an infinite transit time. The re-
sulting activations were due to neutrons born in
the region simulating the core of the reactor. In
tests 2, 3, and 4, made while the fuel plates were
circulated with various transit times, the measured
activations were due to delayed neutrons emitted
in the heat exchanger, in addition to the neutrons
emitted in the core of the simulated reactor. It is
assumed that the activation rate of the sodium
increases monotonically with decreasing transit
time, although a comparison of the results of
tests 2 and 3 would not appear to verify this. The
discrepancy s atiributed to experimental error,
In test 5, a '/2-in.-thick boral sheet in the first
boron curtain was replaced with a l-in.-thick
plastic sheet to determine the effect of a hydroge-
nous material preceding this curtain,

A comparison of the results of test 1 (infinite
transit time) with the results of test 4 (1,25-sec
transit time) shows an increase in the total acti-
vation by approximately 20% (1.2 x 103 d/min).
Since the transit time for test 4 was considered to
 

PERIOD ENDING DECEMBER 10, 1955

R

PHOTO 14969%

Fig. 13.17. Circulating-Fuel Mechanism for LTSF Dynamic Source Tests with the RMR-Shield,

be typical of a circulating-fuel reactor (CFR), it
is indicated that, in the heat exchanger of a CFR
with a 12-in.-thick reflector, 20% of the total acti-
vation in the heat exchanger is due to delayed
neutrons. A comparison of test 1 with test 5 shows
that, for an infinite transit time, replacing one of
the boral sheets with a hydrogenous material re-
duced the total activation due to core neutrons to
40% of the value obtained when no hydrogenous
material was used.

At present there has been no correlation of the
sodium-activation measurements in the static and
dynamic phases of the RMR-shield mockup tests
because of the uncertainty of the fission rate of
the fuel on the circulating belt. However, the

static source tests? can be compared with test 1
in this series, in which the fuel plates were
stationary between the two heat exchangers, If
the space between the heat exchanger tanks occu-
pied by the frame of the circulating mechanism is
subtracted from the distance, z, between the fission
source and the second heat exchanger tank of
test 1, the resulting curve will be found to be
parallel to the curve for mockup 1 of the static
tests.4 This indicates that for a static source
the vuranium in the heat exchanger affected both
heat exchanger tanks proportionally, Most of the
sodium activation takes place from neutrons

 

4G. T. Chapman et al., ANP Quar. Prog. Rep., Sept. 10,
1955, ORNL-1947, p 198.

221
[£44

 

FUEL PLATES

“HE AT

 

¢

it . . s

prioT o971

      

o

L¥0dIY SSIYO0¥d LOIrodd dNV
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i
2—01—057—66—180
10 ,
] | l o |
! i J | !
5 L _L\‘ | . |
N {
= \.< ; i.g: ]
E ‘ ‘ &4
% n 2
§ ~3 |
g’ 2 ” T v T1 ™
@ i .
=
@
s e
> \
=
> |
E 10° A R , L |
< — - : .
o ; . o .\LT_ ]
L | i
§ TEST.  TRANSIT TIME {sec}  CONFIGURATION )
a 1 @ L o S = ] 50
Etf S 2.00 R |
o 3 150 B B R |
= 4 1.25 o I = I 2020
@ 5 ® B X
i ® ' -in- THICK BORAL SHEET '
2 - [ 1 2-in-THICK NaF TANK o
‘ 1-in- THICK PLASTIC SHEET
‘ B FUEL PLATES
| |
S |

35 40 45 50 55 60 65
z, DISTANCE FROM FISSION SOURCE (cm)

Fig. 13.19. Sodium Activation in the Heat Ex-
changer Region of the RMR-Shield Mockups for
Yarious Fuel Transit Times,

originating in the core, so that if the stationary
uranium piates between the heat exchanger tanks
affected sodium resonance neutrons at all, the
effects would tend to be more pronounced in the
second heat exchanger tank, Therefore the com-
parison between the two test implies that the
uranium in the heat exchanger had little, if any,
effect in the static tests.

Fission-Product Gamma-Ray Spectrum
R. W. Peelle W, Zobel3
T. A. Love’

For a circulating-fuel reactor (CFR) a large
fraction of the gamma-ray flux outside the reactor

 

3Bulk Shielding Facility.

PERIOD ENDING DECEMBER 10, 1955

shield originates from fission-product decays® in
The first ex-
perimental energy spectrum has now been measured

the primary heat exchanger region.

for the gamma rays from the gross fission-product
mixture,  For these measurements a multiple-
crystal 78 was used in
conjunction with the circulating solid fuel belt
(described above).

Although the spectrometer was installed in the
LTSF, the associated electronic instrumentation
remained at the Bulk Shielding Facility. Neutrons
from the ORNL Graphite Reactor were allowed to
impinge upon the solid fuel belt shown in Fig. 13,17,
Figure 13.20 is a sketch of the configuration used
in the LTSF in combination with the spectrometer
and fuel loop. The boral, lead, and water placed
within the perimeter of the loop were provided to
reduce as much as possible the counting rate in
the spectrometer caused by sources other than the
section of the fuel belt viewed by the spectrometer
collimator, A photograph of the experimental
assembly with the spectrometer region partially
dismantled is shown in Fig. 13.21. The fuel
loop (1) is at the left of the picture, hidden by
the remainder of the configuration. The water
tank (2) and the lead shielding (3) within the loop,
as well as the major portions of the spectrometer
assembly, show plainly. The scintillation spec-
trometer (4) is surrounded by 6 to 8 in. of lead,
except for a ]],é-in.-dic gamma-ray collimator
pointing at the fuel loop. Bricks (5) made from a
combination of LiF and paraffin were stacked
around the lead spectrometer shield to a thickness
of 8 to 12 in. During operation the top of the
assembly was plugged with lead and LiF-paraffin
shields (not shown in the photograph). Lumber
was stacked in the regions between the lithiated
paraffin and the plywood wall (6) protecting the
rotating fuel belt.

Measurements of the gamma-ray spectrum were
made while the fuel loop was operated in such a
manner that any one fuel plate completed a revo-
lution every T seconds. The observed counting
rate was a function of time after the start of acti-
vation of the rotating belt,

gamma-ray spectrometer

In principle, it would

 

6R. W. Peelle, T. A. Love, and F. C. Maienschein,
ANP Quar. Prog. Rep. June 10, 1955, ORNL-1896, p 203.

’F. C. 'Maie,nschein, Multiple Crystal Gamma-Ray
Spectrometer, ORNL-1142 (April 14, 1952).

BT, A. Love, R. W. Peelle, and F. C. Maienschein,
Electronic Instrumentation for a Multiple-Cry stal Gamma-
Ray Scintillation Spectrometer, ORNL-1929 (Oct. 3, 1955).

223
ANP PROJECT PROGRESS REPORT

t—-¥17’/2in. Afl

Y/2in. OF BORAL

Y4—in. BORAL SHUTTER

WALL OF TANK

 

9 5}/3 in.
#

* o, iy
" 2-01-057-66-179

73/, in, OF BORATED WATER

   
  

Yo in. OF BORAL

GAMMA-RAY

/ COLLIMATOR

 

 

SPECTROMETER SHIELD

8in. OF LEAD

FRAME SUPPORTING
ROTATING FUEL BELT

Fig. 13.20. Expetimental Arrangement for Measurement of the Fission<Product Gamma-Ray Spectrum of

the Circulating Fuel.

also be a function of the transit time T. The

spectrum obtained under the conditions described

contained three components of importance:

1. the desired photon spectrum from the section
of the fuel loop directly in front of the spec-
trometer collimator;

2. the photon spectrum originating within the
Graphite Reactor and the prompt gamma-ray
spectrum associated with fission events occur-
ring in the side of the fuel loop near the re-
actor;

3. the random background events associated with
the seepage of radiation through the spectrom-
eter shield.

Conftribution 2 was made as small as possible by
the lead and water shield within the fuel loop.
Contribution 3 was reduced not only by the lead
and lithiated paraffin shield around the spectrom-
eter but also by the insertion of a boral iris be-
tween the reactor and the fuel belt sothat a sec-

224

tion of the fuel only 4 in. high was activated by
the neutron flux incident from the Graphite Reactor,

In order to measure the undesirable contributions
2 and 3, spectral measurements were made with
the fuel stationary and also with a large lead plug
filling the collimator opening. Spectral measure-
ments were not started until some time after the
rotating rig had been operating at full power., This
allowed the fission-product activity to approach its
equilibrium with the saturation of a large share of
the fission-product gamma-ray emitters,

Transit times 7 between 1.2 and 2.5 sec were
used for a total of four spectral measurements,
Although a complete and detailed analysis is not
yet available, there is no apparent variation either
in shape or in intensity of the spectrum as the
transit time is changed. A typical spectrum is
shown in Fig. 13.22. This spectrum is repre-
sentative of the gamma radiation from the fuel
loop about 3 hr after the beginning of the activa-
tion,
 

PERIOD ENDING DECEMBER 10, 1955

onensh
PHOTO 15129

mFUEL PLATES

ON ROTATING BELT

LiF BRICKS

 

   

225
 

ey
2-01-057- 66181

O COMPTON DATA
® PAIR DATA

phofons/Mev- fission

 

3.0 4.0 5.0 6.0

ENERGY (Mev)

Fig. 13.22. Fission-Product Gamma-Ray Spectrum of the Circulating Fuel.

The shape of the spectrum is based upon the
experimental data and upon calibrations of the
efficiency of the detector as a function of photon
energy. The normalization of the spectrum de-
pends on additional factors such as the solid
angle of the collimator used, the thermal-neutron
flux impinging upon the reactor side of the loop,
and the percentage utilization of this flux for the
fission process. The curve presented in Fig. 13.20
represents a rapid evaluation of these factors. A
more complete evaluation of the data will also
include the calculation of the statistical errors on
the points shown in Fig. 13.22. These errors are

considerable at the higher energies, and it is ex-
pected that they will adequately explain the
experimental scatter of the data points.

The fission-product photon energy spectrum
shown in Fig. 13.22 was integrated between 0.36
and 5.8 Mev, and averages of 4.2 photons/fission
and 4.8 Mev/fission were obtained over the inter-
vening energy interval. An estimated probable
error of +20% is assigned to these averages and
to the normalization of the experimental spectrum.
This estimated error is largely caused by un-
certainties in the fission rate in the fuel plates.
 

 

THE AIRCRAFT NUCLEAR PROPULSION PROJECT

THE OAK RIDGE NATIONAL LABORATORY

AT

DECEMBER 14,1955

 

 

 

 

 

 

 

PRATT & WHITHEY AIRCRAFT

E. R. DYTKQ, ASST. PROJ. ENG.
A. GIANGREGORIC, ADM. ASST.
H. C. GRAY, DESIGN ENGINEER

 

 

 

 

 

ANP PROJECT DIRECTOR W. H. JORDAN RD
CO-DIRECTOR S. J. CROMER RD
ASSISTANT DIRECTOR A. J. MILLER RD
D. HILYER, SEC. RD
REPORTS
A. W. SAVOLAINEN ARE
PWA N. J. HARPER ARE
PWA uT
PWA ERATURE SEARCHES
A.L.DAVIS ARE

 

 

i

 

 

 

AIRCRAFT REACTOR ENGINEERING DIVISION

S. J. CROMER, DIRECTOR RD
A. B. LIVINGSTON, SEC. ARE

 

 

 

 

 

ASSISTANT TO DIRECTOR

R. S. CARLSMITH

ARE

 

 

 

 

 

 

 

EXPERIMENTAL ENGINEERING
H. W. SAVAGE ARE

 

 

 

 

 

POWER PLANT ENGINEERING
A. P. FRAAS ARE

 

 

 

 

INSTRUMENTATION AND CONTROLS
E. R. MANN Ic

 

 

 

 

 

 

ENGINEERING DESIGN
H. C. GRAY Pwa

 

 

 

 

 

REACTOR PHYSICS
A. M. PERRY ENR

 

 

 

 

 

 

7503 AREA CONSTRUCTION
W. G. PIPER ARE

 

 

 

 

 

ANP STEERING COMMITTEE

W.

PEmMITIPEEPEnOMOT

TroEn-Um DR

. JORDAN, CHAIRMAN
. 5.
P.
E.

BILLINGTCN
BLIZARD
BOYD

J. CROMER

ERGEN
FRAAS
GRIMES
MANLY
MILLER
POPPENDIEK
SAVAGE
SHIPLEY
SWARTOUT
WEINBERG

 

 

 

l

 

 

SUPPORTING RESEARCH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NOTE: THIS CHART SHOWS ONLY THE LINES OF TECHNICAL COORDINATION OF THE ANP PROJECT. THE
YARIOUS INDIVIDUALS AND GROUPS CF PEOPLE LISTED IN THIS AND THE FOLLOWING CHARTS ARE EN-
GAGED EITHER WHOLLY QR PART TIME ON RESEARCH AND DESIGN WHICH 15 COORDINATED FOR THE
BENEFIT OF THE ANP PROJECT IN THE MANNER INDICATED ON THE CHART. EACH GROUP, HOWEVER,
IS ALSO RESPONSIBLE TO ITS DIVISION DIRECTOR FOR THE DETAILED PROGRESS OF THE RESEARCH
AND FOR ADMINISTRATIVE MATTERS.
THE KEY TO THE ABBREYIATIONS USED IS GIVEN BELOW.

AC ANALYTICAL CHEMISTRY DIVISION - ORNL

AP APPLIED NUCLEAR PHYSICS DIVISION — ORNL

ARE  AIRCRAFT REACTOR ENGINEERING DIVISION — ORNL

BMI  BATTELLE MEMORIAL INSTITUTE

BP BENDIX PRODUCTS, DIVISICN OF BENDIX AYIATION CORPORATION

C CHEMISTRY DIVISION — ORNL

cT CHEMICAL TECHNOLOGY DIVISION — ORNL

cv CONSOLIDATED VULTEE AIRCRAFT CORPORATION

ENR  ELECTRONUCLEAR RESEARCH DIVISICN — ORNL

GLM  GLENN L. MARTIN COMPANY

IC INSTRUMENTATION AND CONTROLS DIVISION - ORNL

LAC LOCKHEED AIRCRAFT CORPORATION

M METALLURGY DIVISION — ORNL

MC MATERIALS CHEMISTRY DIVISION - ORNL

MH MINNEAPOLIS-HONEYWELL REGULATOR COMPANY

MP MATHEMATICS PANEL — ORNL

CRINS OAK RIDGE INSTITUTE OF NUCLEAR STUDIES

PWA  PRATT & WHITNEY AIRCRAFT, DIVISION OF UAC

RD RESEARCH DIRECTOR'S DEPARTMENT — ORNL

REE REACTOR EXPERIMENTAL ENGINEERING DIVISION — ORNL

Si STABLE ISOTOPE RESEARCH AND PRODUCTION DIVISION — ORNL

SS SOLID STATE DIVISION ~ ORNL

USAF UNITED STATES AIR FORCE

*PART TIME

 

 

 

 

 

 

¥. H. JORDAN
A, . MILLER
METALLURGY
W. D. MANLY, STAFF ASSISTANT M
CHEMISTRY
W. R. GRIMES, STAFF ASSISTANT MC
SHIELDING
E. P. BLIZARD, STAFF ASSISTANT AP
RADIATION DAMAGE
D. S. BILLINGTON, STAFF ASSISTANT SS

 

 

 

 

 

HEAT TRANSFER AND PHYSICAL PROPERTIES
H. F. POPPENDIEK, STAFF ASSISTANT REE

 

 

 

 

 

REACTOR PHYSICS
W. K, ERGEN, STAFF ASSISTANT ARE

 

 

 

 

 

 

 

FUEL REPROCESSING
F.R. BRUCE* CT

 

 

 

 

 

 

 

 

 

227
 

 

AIRCRAFT REACTOR ENGINEERING DIVISION
THE OAK RIDGE NATIONAL LABORATORY

 

DECEMBER 1,1955

 

5. J. CROMER, DIRECTOR
A, B. LIVINGSTON, SEC.

RD
ARE

 

 

 

 

ASSISTANT TO DIRECTOR

R. S. CARLSMITH ARE
R. E. THOMPSCM ARE

 

 

 

 

 

 

Ty e R

 

 

 

 

 

 

 

INSTRUMENTATION AND CONTROLS

POWER PLANT ENGINEERING

A.P. FRAAS ARE
P. HARMAN, SEC. ARE

 

 

 

 

REACTOR PHYSICS
A. M. PERRY ENR
M. WILSOM, SEC. ARE
H. W, BERTIN! ARE
A. FORBES ARE
W. £. KINNEY ARE
M. TSAGARIS ARE

 

 

 

E. R. MANN IC
SYSTEM DEHGN

E. VINCENS ic
M. C. BECKER IC
C. F. HOLLOWAY Ic
6. C. GUERRANT Ic

T. HERRELL Ic

INSTRUMENT ATION

R.G. AFFEL i<
G. H. BURGER Ic

R. F. HYLAND i<

J. W. KREWSON IC

H. METZ s

W. R. MILLER Ic
A, L. SOUTHERN i<

J. W STARKEN IC

K. M. YOUNG I

NUCLEAR AND MECHANICAL DESIGN

C. 5. WALKER I
J. M. EASTMAN EP

5. C. SHUFORD PWA
€. B, THOMPSON MH

 

 

ADMINIZTRATION
w. i, 5COTT ARE

HYDRODYNAMICS AND THERMODYNAMICS

 

 

¥. T. FURGERSON ARE

M. E. LACKEY ARE

G. SAMUELS ARE

M. W, HINES, COMPUTER ARE

J. 1. TUDOR, DRAF TSMAN ARE
APPLIED MECHANICS AND

STRESS ANALYSIS

R, Y. MEGHREBL IAN ARE

L. E. ANDERSON PWA

T. J. BALLES Pwa

F. A_FIELD USAF

B. L. GREENSTREET ARE

D. M, MILLER PWA

S. E. MOORE PWA

D. H. PLATUS USAF

D. L. PLATUS USAF

C. H. WELLS PWA

HEAT TRANSFER

R. D. SCHULTHEISS ARE

J. W, ADERHOLDT GLM

J. FOSTER ARE

R. 1. GRAY PWA

M. M. YAROQSH ARE

SPECIAL PROBLEMS

W. B. COTTRELL ARE

H. C. HOPKINS PWA

H. WIARD v

L. A. MANN ARE

¥. 1. PRICE USAF

CONSULTANTS

A. H. FOX, UNION COLLEGE

¥. LOWEN, UNION COLLEGE

R.L.MAXWELL, K UNIVERSITY OF
TENMESSEE

A. 5. THOMPSON, STUDEBAK ER-PACKARD
CORPORATION

G. F. WISLICENUS, PENNSYLVANIA STATE
COLLEGE

 

 

 

 

228

 

 

 

EXPERIMENTAL ENGINEERING
H. W. SAVAGE

D. ALEXANDER, SEC.

 

 

 

 

 

 

DEYELOPMENT

E. R. DYTKO
L. RUSSELL, SEC.

PUMPS

. G. GRINDELL
R. CURRY
S. M. DECAMP
W. L. SNAPP
J. J. SIMON
H. C. YOUNG

>

HEAT EXCHANGERS AND
SPECIAL EQUIPMENT

E. MACPHERSON
J. C. AMOS

J. W, COOKE

M. H. COOFER
L. H. DEVLIN
L. R. ENSTICE
F.
R,
J.
D.

o

FERRIGNO
D. PEAK
G. TURNER
R. WARD

W. F. BOUDREAU
J. PARKER, SEC.

GENERAL EQUIPMENT

%. R. C580RN

F. M. LEWIS
D. T. STOREY

1. R. EVANS

D. L. GRAY

E. MAEYENS

T. H. MAYES
5. R.ASHTON
W. 0. CHANDLER
G. C, JENKINS
E. B. PERRIN

WELDED EQUIPMENT

R.B. CLARKE
B. E. BLACK
E. B. EDWARDS
H. W. HOOVER
E. A. JAGGERS
T. A, KING
1 W.TEAGUE

SPECIAL EQUIPMENT

W. D. GOOCH
L.w. LOVE
T. K. WALTERS
D. R, WESTFALL

IN-PILE LOOP TESTS

B. TRAUGER
C.C.BOLTA
L. P. CARPENTER
J. A, CONLIN
C. W. CUNMINGHAM
P. A GNADT
D. M. HAINES

T

. BENDER
M. L. OVERTON, SEC.

CORE STUDIES

WHITHAN
M. SMITH
J. STELZMAN

Ero

. PEACH
. TUNMNELL

o
O

 

PROCUREMENT CODRDINATION

REACTOR CONSTRUCTION

ENGINEERING TEST UNIT

PWA
ARE

ARE
PWA
ARE
FWA
FWA
PWA

ARE
ARE
PWA
PwWA
PWA
PWA
PWA
PiA
PWA
ARE

ARE
ARE

ARE
ARE
TCARE
ARE
ARE
ARE
ARE
ARE
ARE
ARE
ARE

ARE
ARE
ARE
ARE
ARE
ARE
ARE

ARE
PWA
ARE
ARE

ARE
PWA
ARE
ARE
ARE
ARE
PWA

ARE
ARE

ARE
ARE
ARE

ARE
ARE

 

 

 

 

 

ENGINEERING DESIGN

 

 

ARE
ARE
TEST SERVICES
W. B. MCDONALD ARE
D. P, HARRIS, SEC. ARE
A. N. MONTGOMERY ARE

TEST FACILITIES ASSEMELY
COORDINATION

E. $STORTO ARE
J. L. CROWLEY ARE
W, H. KELLEY ARE
W. ). MUENZER ARE
H. C. SANDERSON Pwa

CORROSION LOOPS

C. P. COUGHLEN ARE
R. A, DREISBACH Pwa
P, G, SMITH ARE

ELECTRIC SERVICES

E. M. LEES ARE
D. L. CLARK ARE
C. E. MURPHY ARE

TEST ENGINEERING

4.1 MILICH Pwa

.
1. W KINGSLEY ARE

TECHNICIAN GROUP

R.HELTON ARE
J. 5. ADDISON ARE
T. ARNWINE ARE
G. 5. CHILTON ARE
J. M. COBURN ARE
1. R. CROLEY ARE
J. M. CUNNINGHAM ARE
R.E.DIAL ARE
J. R. DUCKWORTH ARE
W. H. DUCKWORTH ARE
W. K. R. FINNELL ARE
H. FOUST ARE
R. H, FRANKLIN ARE
W. D. GHORMLEY ARE
C. ). GREEN ARE
T.L. GREGORY ARE
F. M. GRIZZELL ARE
R. A. HAMRICK ARE
P. P, HAYDON ARE
€. G. HENLEY ARE
B, L. JOHNSON ARE
J.R. LOVE ARE
D. E. MCCARTY ARE
G. E. MILLS ARE
B. H. MONTGUMERY ARE
C. €. NANCE ARE
J. 1. PARSONS ARE
D. G. PEACH ARE
H. E. PENLAND ARE
R. RED ARE
F. J. SCHAFER ARE
J. R. SHUGART ARE
C. E. STEVENSON ARE
A. G, TOWNS ARE
C. A. WALLACE ARE
B. T, WILLLAMS ARE

OFFICE SERYICES
J. P. LANE ARE

 

H. C. GRAY Pwa
1. ZASLER ARE
L. E. FERGUSON, SEC, ARE
REACTOR DESIGN
D. C. BORDEN PWa
J. M. CORNWEL L PWA
R. C. DANIELS ARE
H. J. DICKINSON Pwa
J. Y. ESTABROOK ARE
W. E. LEUTLOFF Pwa
C. MANTELL PWA
¥. ). NELSON PWA
1%, TUMAVICUS PwA
PUMP DEWGN
W. G. COBB ARE
I. T. DUDLEY ARE
W._ S, HARRIS PWA
R. E. HELMS ARE
A, D, JARRATT PWA
F. L. MAGLEY ARE
W, H. POND Pwa,
L. ¥. WILSON ARE
W. C. GEORGE ARE
ELECTRICAL DESIGN
T. L. HUDSON ARE
A. H. ANDERSON ARE
J. KERR ARE
B. C. GARRETT PwA
J. 0, HICHOLSON ARE
GENERAL DESIGN

A.A. ABBATIELLOD ARE
E. F. JURGIELEWICZ Pwa

J. T. MEADOR Pwa

W. A, SYLVESTER Pwa

N, E. WHITNEY PWA

Wl GALYON ARE

G. R HICKS ARE

J.R. LARRABEE Pwa

G. G- MICHELSON ARE

C. A, MILLS ARE

C. F. SALES ARE

RECORDS AND PRINTING
§. J. FOSTER ARE
J. 1. PLATZ ARE

 

 

 

CONSULTANTS
F. A, ANDERSON, UNIVERSITY OF
MISSISSIPPI
L. F.BAILEY, UNIVERSITY OF TENNESSEE
R. L. MAXWELL, UNIVERSITY OF
TENNESSEE
W. K. STAIR, UNIVERSITY OF TENNESSEE

 

 

 

 

 

7503 AREA CONSTRUCTION

 

 

W. G. PIPER ARE
S, M. JANSCH, SEC. ARE

R. CORDOVA ARE
W. F, FERGUSON ARE
V. J. KELLEGHAN ARE
F. R. McQUILKIN ARE

A M. MILLS, JR. ARE
G. C. ROBINSON ARE
R. D. STULTING ARE
C. F, wEST ARE

 

 

 

 

 
 

THE AIRCRAFT NUCLEAR PROPULSION PROJECT
AT
THE OAK RIDGE NATIONAL LABORATORY

DECEMBER 4, 4955

 

 

SUPPORTING RESEARCH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

W. H. JORDAN
A, J. MILLER
STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT
W. K. ERGEN ARE E. P. BLIZARD AP . R. GRIMES MC W. D. MANLY M D. 5. BILLINGTON S5 H. F. POPPENDIEK REE
REACTOR PHYSICS SHIELDING CHEMISTRY ANALYTICAL CHEMISTRY METALLURGY METALLOGRAPHY RADIATION DAMAGE HEAT TRANSFER AND PHYSICAL
A. SIMON* AP W. R. CRIMES e M. T. KELLEY* AC . D. MANLY M ) GRAY " b 5. BILLINGTON s PROPERTIES RESEARCH
SHIELDING RESEARGH D. R. CUNEO NC C. D. SUSANO* AC W. H. BRIDGES M M. J, FELDMAN ] H. F. POPPENDIEX REE
. D. E. CALDWELL, SEC. NC B. E. YOUNG,* SEC. AC J. THOMAS, SEC. M G. W. KEILHOLTZ 55 T. K. CARLSMITH, SEC. REE
C L. BRADSHA\: MP E. P. BLIZARD* AP 1. C. WHITE AC R. S. CROUSE M J. C. WILSON 5§
R. R. COVEYOU AP R. RICKMAN, SEC. AP E. M. ZARZECKI,* SEC. AC T. M. KEGLEY M HEAT TRANSFER
L. DRESNER AP L. 5. ABBOTT AP SPECIAL PROBLEMS D. F. STONEBURNER
T H. MARABLE® “P GENERAL CORROSION GROUP - F. M
- 1 - AP 5. AUSLENDER Pwa R. F. NEWTON RD E.R. BOYD M MTR LIAISON C, M. COPENHAYER REE
. IIS. ;g;gfi A2 L. B. HOLLAND AP TE MO R RESEARCH E. E. HOFFMAN M B. F. DAY M N. D. GREENE REE
&b : g' bg;ifi“ Pfiz R. E. MINTURN USAF A. S. MEYER, JR. AC 5' fiAESSEDER Pwfi B. J. REFCE M .V KLAUS s D.P. GREGORY PA
F. H., MURRAY AP G. GOLDBERG AC D. H. JANSEN M HOT LAB FACILITY ;' : E%Ex"“ EEE
CONSULTANT R. B, MURRAY AP PHASE EQUILIBRIUM STUDIES B. L. MCDOWELL AC I E POPE " & NULLER A
L. A, NCHEL, GEORGIA TECH. g. ZE%:;BEY :Ip: C. 1. BARTON e W. J. ROSS AC L R. TROTTER " M. é-.’;?_t?SMAN gg b BALNER e
- B L. M. BRATCHER MC DEVELOPMENT E. J. MANTHOS 55 J. L. WANTLAND REE
SHIELD DESIGN R. 1. SHEIL MC PYNAMIC CORROSION GROUP CERAMIC RESEARCH W. B. PARSLEY 55 R. M. BURNETT REE
R. E. THOMA ve 1. P. YOUNG AC J. H. DEVAN M M. WARDE* " R. A RAMSEY 55 J- LONES REE
1. B.DEE PwA R. E. CLEARY PWA J.R. FRENCH AC E. A, KOVACEVICH M - . E. 5. SCHWARTZ s R. L. MILLER REE
M. A. MARLER AC A. HOBBS,™ SEC. M G. M. WINN REE
CRITICAL EXPERIMENTS R. DAVIS GLM H. A. FRIEDMAN MC G. D. BRADY M
C. A, GOETZ PWA B. A. SODERBERG MC E. J. LAWRENCE M RADIATION METALLURGY
A, D, CALLIHAN® AP 0.R.OTIS cv R. E. MOORE MC SERVICE M. A. REDDEN M C.E. CURTIS® M J. C. WILSON* 55 PHYSICAL PROPERTIES
M, L. RUEFF,* SEC. AP H. €. WOODSUM PwaA H. DAVIS PWA W. F. VAUGH T .
« P~ YAUGHN AC FABRICATION GROUP L. M. DONEY M C. D. BAUMANN 5 . 1. COHEN REE
LID TANK SHIELDING FACILITY R B MEADOWS e R.F.APPLE AC J. A. GRIFFEN" M ¥. E. BRUNDAGE 5 W. D. POWERS REE
R. P. METCALF MC D. E. CARPENTER AC 1. H. CooBS M A.J. TAYLOR* M W. W. DAVIS 58 " 5. €. BLALOCK REE
M. K. ALBRIGHT AP R.W.PEELLE AP FUEL PREPARATION RESEARCH R C. BRYANT AC M, R. D'AMORE PWA G. D. WHITE* M N E. HINKLE 55 5. J. CLAIBORNE REE
W Faobe oA wiecoor o o . waTsON w L F. oo a e ouh £ ALEXADER: ¥ b5 oLy, 5 7% Jones
R. GWIN AP 1. MILLER AP M SLotD e A. H. MATTHEWS AC R. MCDONALD PWA J. C. ZUKAS 58
V. G. HARNESS* AP ). SMOLEN PWA o C. E. PRATHER AC 3. P, PAGE W T. PRICE PWA
1. J. LYNN AP " E. BECKHAM AR E &. zfl}EESY :g A. D. WILSON AC T. K. ROCHE M CONSULTANTS
E. R. ROHRER* AP 1. FRANCIS AP - C. M. WILSON AC R W. JOHNSON M RADIATION CHEMISTRY
E. V. SANDIN PwA D. R. HENDRIX AP CHEMICAL EQUILIBRIA D LN L CoMPANY
D. SCOTT, JR, ARE H. JARVIS c NONDESTRUCTIVE TESTING GROUP SRfNZ'l'ETf SPECIALTIES CORPORATION G g- EFNI;E?LTZ ig FUEL REPROCESSING
5. SNYDER . .
5.V, P WILLIAMS Fup R TAYLOR A “E . KETCHEN e R.B. OLIVER M T-N. Mcvay b T BROVNING 5 F- R BRUCET o7
BUL 3. D. REDMAN MG J.W. ALLEN " 5. T. ROBINSON, SANDERSON & PORTER D. E. GUSS USAF
K SHIELDING FACILITY 3. 1 STURM e SPECTROGRAPHIC ANAL YSIS R, W, MCCLUNG M D. SCHLUDERBERG, BABCOCK & WILCOX H. L. HEMPHILL 55
F. C. MAIENSCHEIN ap s K. W. REBER " T. S. SHEVLIN, OHID STATE EXPERIMENT J. KRAUSE PWA CHEMICAL DEVELOPMENT
€. BOUNDS, SEC. ap CORROSION STUDIES 4R MCNALLY" sl 0. E. CONNER M STATION M. F. 0SBORNE s D. E. FERGUSON* T
1. ANNO BMI J. A. NORRIS* AND OTHERS sl W, J. MASON M H. E. ROBERTSON 5 G. I. CATHERS o1
T.V. BLOSSER AP F. KERTESZ G M. T. ROBINSON 5 M. R. BENNETT cT
W. R. CHAMPION LAC ? i' i‘d&“" mg WELDING AND BRAZING GROUP CONTRACTOR 5: g: :EBEITSER gg R. M. DUFF cF
G. ESTABROOK AP - A P. PATRIARCA M NORTH CAROLINA STATE UNIVERSITY
J.D. FLYNN AP G. F. SCHENCK PWA UNIT OPERATIONS
25 e b. FUCKER e R. E. CLAUSING M ENGINEERING PROPERTIES
T R PWA 1 V. DIDLAKE e A. E. GOLDMAN M W. K. EISTER® T
M. P. HAYDON* AP - M. MASS SPECTROMETRY G. M. SLAUGHTER M 0. SISMAN* 55 <
K. M. HENRY AP : BA MCDOWELL M R. M. CARROLL 55 J. T. LONG CcT
£. B. JOHNSON AP PRODUCTION OF PURIFLED FUELS C. R. BALDOCK* Sl C' E. SHUBERT M 1. G. MORGAN s S. H. STAINKER CcT
T. A LOVE AP J.R.SITES sl C-E o " T MORGAN = G. IONES, IR. cT
E G SILVER AP G. j EEgiLAEKELY ::g R g ififfl%? " METALLURGY CONSULTANTS s D. W. LEIGH cT
W. ZOBEL AP C.R.CROFT NG T N. CABRERA, UNIVERSITY OF VIRGINIA
D. J. KIRBY AP A DOSS e PHYSICAL CHEMISTRY OF N. 3, GRANT, MASSACHUSETTS DESIGN
C. O. McREW c J. E. EORGAN MC LIQUID METALS GROUP INSTITUTE OF TECHNOLOGY H. E. GOELLER* cT
J. SELLERS ic v SHITH s J. L. GREGG, CORNELL UNIVERSITY "R, P. MILFORD T
R. M. SIMMONS AP 1L TRUITT e G. P.SMITH M W. D. JORDAN, UNIVERSITY OF ALABAMA F. N. BROWDER cT
D. SMIDDIE IC W. T. WARD C. R, BOSTON M E. F. NIPPES, RENSSELAER o
G. G. sTOUT [« "R K. BAGWELL mg J. V. CATHCART M POLYTECHMIC INSTITUTE PILOT PLANT
H. WEAVER AP 1 P EUBANKS o H, W, LEAVENWORTH PWA W. F. SAVAGE, RENSSELAER
J. W. WAMPLER AP B F. HITCH o J. ). McBRIDE M POLYTECHNIC INSTITUTE H. K, JACKSON* cT
W, JENNINGS NG G. F. PETERSEN M G. SISTARE, HANDY & HARMON W, H. CARR* cT
TOWER SHIELDING FACILITY G A PALMER s M. E. STEIDLITZ M P. C. SHARRAH, UNIVERSITY OF W. K, LEWIS* cT
B C. THOMAS e L. L. HALL M ARKANSAS
D e o R. G WILEY Me J. L. GRIFFITH M F. G. TATNALL, BALDWIN-LIMA-HAMIL TON
i E-HELSE- . e E. C. WRIGHT, UNIVERSITY OF ALABAMA
E e s MOLTEN SALT THERMODYNAMICS MECHANICAL PROPERTIES GROUP
F. J. MUCKENTHALER AP E.F. BLANKENSHIP Me D. A. DOUGLAS M
G. RAUSA GLM M. BLANDER MC C. W. DOLLINS M
F. N. WATSON AP 5. CANTOR MO C. R. KENNEDY PWA
E. C. CARROLL ic B. H. CLAMPITT MC J.R.WER, JR. M
|. D. CONNER AP 5. LANGER NC 1. W. WOODS M METALLURGY CONTRACTORS
J. N, MONEY AP M. B. PANISH M K. W. BOLING M BATTELLE MEMORIAL, INSTITUTE
g‘ g ‘L\Jf::JGEHRTwooo Alg L E. TOPOL NG j g. 5GSD;ON M BRUSH BERYLLIUM COMPANY
L E. .D. M FERROTHERM
SUPPORTING STUDIES V. G. LANE M GLENN L. MARTIN COMPANY
B. MCNABB, JR. M METAL HYDRIDES, INC.
CONSUL TANT P. A, AGRON E B. C. STOWERS, JR. M NEW ENGLAND MATERIALS TESTING
J. E. SUTHERLAND C. K. THOMAS M LABORATORY
H. A, RNELL LNIVERSIT
BETHE, COl UNIVERSITY C. W. WALKER M RENSSELAER POLYTECHNIC INSTITUTE
SUPERICR TUBE COMPANY
CONTRACTORS CONSULTANTS INSPECTION GROUP UNIVERSITY OF TENNESSEE
METAL HYDRIDES, INC. H. INSLEY A. TABOADA M
NUCLEAR DEVELOPMENT CORPORATION T. H. Mcvay R. L. HEESTAND PWA
OF AMERICA R. M. EVANS M
CONTRACTORS

 

AMES LABORATORY

BATTELLE MEMORIAL INSTITUTE

CARTER LABORATORIES

A, R, NICHOLS, SAN DIEGO STATE
COLLEGE

UNIVERSITY OF ARKANSAS

 

 

 

*PART TIME

 

 

 

229
‘.&fl Seo W nwu.f..s;‘

xR, nfi.fln“dt.