ORNL-2061.txt

 

fi ORNL-2061, Part |, 11, 111 z 133 -4
C-34 Reactors-Special Fuch.lres of Aircraft Reactors

AEC RESEARCH AND DEVELOPMENT REPORT-

   

  
 
 

TR

¢

  
    
   
 
       
 

    

  
 
 

 

      

34
O
=S
E-.‘-'_.':f ;g‘ 3 4456 0251032 §
o - i
E!"E' ( "‘3}"
g -
ey J“"‘ rfi: AIRCRAFT NUCLEAR PROPULSION PROJECT
e iy -
R J QUARTERLY PROGRESS REPORT
o™
g gjhfi FOR PERIOD ENDING MARCH 10, 1956
5
-
55|
£ 5]
3 s
5 & &
Iy
| iopdl OAK RIDGE NATIONAL LABORATORY |
OPERATED BY

UNION CARBIDE NUCLEAR COMPANY

A Division of Union Carbide and Carbon Corporation

ucc) -
POST OFFICE BOX P * OAK RIDGE, TENNESSEE 4o
150 5

   
 

ORNL-2061, Part |, 11, I}
C-84 — Reactors-Special Features of Aircraft Reactors

This document consists of 272 pages.

Copyl 3301‘ 330 copies. Series A,

Contract No. W-7405eng-26

AIRCRAFT NUCLE AR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT

For Period Ending March 10, 1956

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

DATE ISSUED

mAY 23 1958

 

OAK RIDGE NATIONAL LABORATORY
Operated by
UNION CARBIDE NUCLEAR COMPANY
A Division of Union Carbide and Carbon Corporation
Post Office Box P ' : S

dge, Tennessee

MARTIN MARIETTA ENERGY SYSTEMS L1

i

3 4456 0251032 5

 

 

 

 

 

 

 

 

 

 

 

 

  

 

{
i
 

C-84 — Reactors-Special Features of Aircraft Reactors

INTERNAL DISTRIBUTION

1. Rn G. Affel
2. C. R Baldock
3. C. J. Barton
4. M. BenderA
5. D. S. Billington
6. F. F. Blankenship
7. E. P.Blizard
8, C., J. Borkowski
9. W. F. Boudreau
10. G. E. Boyd
11. M. A. Bredig
12. W. E. Browning
13. F. R. Bruce
14, A, D. Callihan
15. D. W. Cardwell
16, C. E. Center {(K-25)
17. R. A. Charpie
18. G. H. Clewett
19. C. E. Clifford
20, J. H. Coobs
21. W. B. Cottrell
22. D. D. Cowen
23, S. Cromer
24, R. S. Crouse
25. F. L. Culler
26. J. H. DeVan
27. L. M. Doney
28. D. A. Douglas
29. E. R. Dytko
30. L. B. Emlet (K-25)
31. D. E. Ferguson
32, A. P. Fraas
33. J. H. Frye
34, W, T. Furgerson
35, H. C. Gray
36, W. R. Grimes
37. E. E. Hoffman
38. A. Hollaender
39. A.S. Householder
40, J. T. Howe
41, H. K. Jackson
42. W. H. Jordan
43. G. W. Keilholtz t
44, C. P. Keim ' Y
45, M. T. Kelley
46. F, Kertesz

 

ORNL-2061, Part 1, 11, Il

47.

48-49.

50.
51.
52.
53.
54.
55.
56.
57.
- 58.
59.
60.
61.
62,
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
/3.
74.
75.
76.
77.

78.
79.

80,

81.
82.
83,

84.
85,

86.
87.
88.
89.
90,
91.
92.
93.

 

R. Van Artsdalen "
C. VonderLage
M. Watson

E. M. King

J. A. Lape

R. S. Livingston

R. N. Lyon

F. C. Maienschein

W. D, Manly

E. R, Mann

L. A, Mann ¢
W, B, McDonald

F. R, McQuilkin

R. V. Meghreblian ?
R. P. Milford

A, J. Miller

R. E. Moore

J. G, Morgan

K. Z. Morgan

E. J. Murphy

J. P. Murray (Y-12)

G. J. Nessle

R. B. Oliver

L. G. Overholser .
P. Patriarca

R. W. Peelle

A. M. Perry '
J. C. Pigg

H. F. Poppendiek

P. M. Reyling

A. E. Richt

M. T. Robinson

H. W. Savage ’
-A, W, Savolainen

‘R. D. Schultheiss

é D. Shipley L
A Simon

0. §isman

M. Ji. Skinner

G. P, Smlth

A, H. $ne||

C.D. %sono

J. A, Swflrfout

E.

R.

D.

E.

F.

G.
 

 
 
 
   
  
  
 
 
  

%. A. M. Weinberg 100.
9%, J. C. White 101-110.
963G, D. Whitman |
97. “E. P. Wigner (consultant) 111-131.
98. GAC. Williams 132.
99, J. 133-135.

Wllson

F Plant Representative, Seattle

55
. Air
. Air
. Air Red .
. Air Tech’iycal Intelligence C*enfer
146. Aircraft L&bora’rory De5|gr,| Branch
147-149. ANP Projeck Office, Forf Worth
150. Argonne Non_‘jal Lobom‘rory
151. Armed Forces.!
152. Assistant Secref
153-159, Atomic Energy-
160. Battelle Memor
161, Bettis Plant
162.
163,
164.
165.
166,
167-168.

ateriel Area

   
  
  
      
  
 

¥

Chicagg .percmons O
Cthdgo ‘Patent Group © :
Chlefd‘ff Naval Research

EXTERNAL DISTRIBUTION

C. E. Winters

ORNL - Y-12 Technical Library,
Document Reference Section

Laboratory Records Department

Laboratory Rec¢ords, ORNL R. C.

Central Research Library

. AF Plant Representative, Baltimore
/. AF Plant Representative, Burbank

AF PlantsRepresentative, Marietta
“/AF Plant Representative, Santa Monica

. AK Plant Representative, Wood- Rldge

search and Development. Command (RDGN)
earch and Developmenf 'Command (RDZPA)

(WADC)

ecml!Wedpons Project, Sandia
ary of the Air Force, R&D

Amission, Washington

169. Cor}vg ir-General Dynamics G ,rporcmon

170.

171.

172,
173-1754,

Dl,rét'ror of Laboratories (WC‘”-:.(
Dl’récfor of Requirements (AFD
{3irector of Research and Develé‘_'

  
 

 
 
 

Pirectorate of Systems Manageme

“"‘f‘(RDZ ISN)

176-178./ Directorate of Systems Management {RDZ-1SS)

179
180-,1’
/}‘34.
#1185,
£ 186,
187.
188.
189.
190.
191.
192.
193.
194,

Equipment Laboratory (WADC)
General Electric Company (ANPD)
Hartford Area Office

ldaho Operations Office

Knolls Atomic Power Laboratory
Lockland Area Office

Los Alamos Scientific Laboratory

 

Mound Laboratory

Naval Air Developmenf Cem‘er

 

Headquarters, Air Force Special Wec:pons Can’rer

7\",‘:4?‘3
AN
T
.
T
o

 

Materials Laboratory Plans Office (WADC) “\

National Advisory Committee for Aeronautics, Cieveiahd
National Advisory Committee for Aeronautics, Wcshlngtol}&
195.

196, R

197.
198,
199-201.
202-205.
206,
207,
208.
209.
210,
211.
212-214,
215-329.

330. §

 

 
     
 
 
   
 
  
  
   

  

e York Operations Office #F
7

American Aviation, |g. (Aerophysics Division)

ion (Fox Project)
Rhations Office

 

f i"'ry of California Radia™® Laboratory, Livermore
Air Development Center (WCOSI-3)
nical Information Extension, Oak Ridge

lvision of Research and Development, AEC, ORO

 
Reports previously issued in this series are as follows:

ORNL-528
ORNL-629
ORNL-768
ORNL-858
ORNL-919
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
ORNL-2012

Period Ending November 30, 1949
Period Ending February 28, 1950
Period Ending May 31, 1950
Period Ending August 31, 1950
Period Ending Decembker 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
Period Ending December 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 at the Laboratory under its Contract W-7405-eng-26, The report is divided into
three major parts: |. Reactor Theory, Component Development, and Construction,
Il. Materials Research, and Ill, Shielding Research.

The ANP Project is comprised of about 550 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
$
A

 

CONTENTS

FOREWORD
SUMMARY

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

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

PART I. REACTOR THEORY, COMPONENT DEVELOPMENT, AND CONSTRUCTION

1. REFLECTOR-MODERATED REACTOR
ART Facility Design and Construction ... sieies s stiene e s er s e s esbesasesssnen e
AR T DS igN ettt ete s e bbbt e e et e s s et et b b eanteeates

SIS S ANGIYSES oo e e bbb et en e s ee e ne e st te et e taerestaebtenens
SOdIUM SYSTOM ettt e etk et bR b e e nae et s et et ensenrens
REACOr Shield ..ot et ba e s bbb b eab e ny b e e s e e et sranebe e b beeteetnas

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

ART Component Development, Procurement, and Testing ..c.ccccoeoereininnineencne e
Water Test of Aluminum Mockup of North Hedd........o.vveioii s
Core Flow SHUIES .ottt e e eas bbbt s v e sae b aree e st e nnsbesraseasaetnsrerens
Engineering Test Unit. ittt
Procurement of Special Reactor Materials and Components ........cccoeveeiiiiniiicinceececicinreena,
Inspection of Materials and Fabricated Parts ..o

ART Instrumentation Gnd Comtrol oo eooeeeeeeereterssesessaeesssseesssesansssereessseesssserssasassssasassresssesesnsareess

REACIOE PRYSICS oottt ettt v et e et e e ses it ek aeaeeaeesebbeseebenenseae e e ae e e enssane
NaK Activation in the Fuel-to-NaK Heat Exchanger of the ART ..o,
Activity of Mass-Transferred Material in the ART Radiators ...,

Radiation Heating in the ART ..t sttt s
Self-Absorption of the Decay Gamma Rays in the ART Fuel Dump Tank ..coocovviiicincnneinncennnn,
Activation of the Sodium in the ART ..ot e sr s s aenen
Shutdown Reactivity of Lithium in the ART Reflector Coolant ..o

Forced-Circulation Corrosion and Mass-T ransfer Tests vttt ee e es e s sbae
Fused Salts inm IReonel .o e et e s e s estabtsssase st assserasaeseaentasntasaes s eaasnsasnrsseeeesrrnnannees
Liquid Metals in Inconel and in Stainless Steel ......cccocvccvvvrviienneen. e eeeiee et ee e as i nr ot aeteeteannenes

PUMP DEVEIOPMENT .oiieiiieeeeeeee ettt s et b et st e ab s ee e ses b ae bbb e e e enr s eneenens
Bearing-and-Seal Tests ...t e b
Sodium-Pump Performance Tests with Water ..ottt
NaK-Pump Performance Tests with Water ..ot aene e
High-Temperature Tests of ART Fuel Pumps oo
High-Temperature Pump-Performance Test Stands ...,

Heat Exchanger Development ..o e
Intermediate Heat Exchanger Tests ..ot e
Small Heat Exchanger Tests ..ottt s e
Heat-Transfer and Pressure-Drop Correlations ......ocoiiiienerriinieineisriece e

 

vii

15
15

21
22
23
23
23

23
23
24
24
25
26

26

28
28
30

35
37
39
39

41

41
41
41
42

42
42
43

43
43
45
48
48
51

52
52
52
54
Structural Tests
Outer-Core-Shell Thermal-Stability Test ..oocoviiiirec e e
Inconel Strain-Cycling Tests .otk s e s
Swagelok Tubing-Connector Tests ... s e

Reactor Component Development ...
DUMP Y alV@ et ettt e e e s e a e bR e e e e e ea et et eeae b e e e s bt
Cold Trap and Plugging Indicator ..ottt e s
Zirconium Fluoride Vapor Trap .ot ente e i ece et rear et sa b e ees e

3. CRITICAL EXPERIMENTS Lt e

Compact-Core Reflector-Moderated-Reactor Critical Experiments......ccoooiiiiiinniiiniiinn
N EUIEON=F LUX DiStribUIOM o viviuiireee ettt ee et e sr e sb e s b et s e b b rme e s men e s e emrnesas o neeenansseres seneeens
Mass-Reactivity Coefficients of Reactor Component Materials .....cccoooiiivviiiiiii

PART Il. MATERIALS RESEARCH

4, CHEMISTRY OF REACTOR MATERIALS ..ot

Phase Equilibrium StUdies . ..ottt
The System UF j-UO, ot
The Systern NGF-UF4 ............................................................................................................................
The System KE-ZrF ;.o
The System RBF-ZIrF 4. b s
The System NaF-ZrF (-UF ;oo
The System NaF-KF-ZrF ;o
The System NaF-RBF-ZrF ; oo s
The System RBF-BelF , o s
The System NaF-BeF ,-UF ;s
The System NaF-LiF-BeF , o
The System NaF-LiF-BeF j-UF | oo
The System NaF-LiF-CaF , oo
The System NaF-MgF ,-CaF .o
The Systems N(JF-CeF3 and RBF-CeF 5 oo
The System LiF-NiF , oo
Optical Properties and X-Ray Patterns for Compounds in Fluoride Systems ...ccocovviciiiiccnnnnenn.

Chemical Reactions in Molten Salts.. ..ot
Equilibrium Reduction of FeF , by H, in NaF-ZrF , ...,
EMF Measurements in Fused SQIts .ot ev e et e e e ene s
Activity of Chromium in ATOys ..o sttt en st seees
Reduction of UF, by Structural Metals........coooiiiiiiii s
Reaction of UF ; with NaF-KF-LiF Eutectic i
Experimental Preparation of FIuorides ...ttt st s rasnees
Color Studies of Alkali Metals in Equilibrium with Alkali Halides at

High Temperatures

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

Physical Properties of Molten Materials ..........ccceeiiiieiiciinec s e
Vapor Pressures in the System KF-ZrF ;oo
Vapor Pressure of FeCly, o)
Surface Tensions of Molten Salts ..o e eab e

Density and Surface Tension of Molten KCI at 800°C
Physical Chemistry of Fused Salts

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

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

 

 

58
58
60
60

60
60
62
63

64

64
64
64
Production of Purified Fuel Mixtures .....oo.oiiieieeriicrrcec et 107
Use of Copper-Lined Equipment for Fuel Purification ..., 107
Preparation of ZrF , by Gas-Phase Hydrofluorination of ZrCl, .cccovmiceriiiiiiis 109
Laboratory-Scale Purification Operations ... s 109
Pilot-Scale Purification Operations ...t sss e 109
Production-Scale Operations .........c.oeieiineeeceeree st ens s arab bbb 110
Quality Control of Raw Materials and Products .......cccooiviiriciiiiniiiic 111
Batching and Dispensing Operations ... iciicnceicins st ss e m
Filling, Draining, and Sampling OPErAtiONS .v.veivieierteeeteeeeeceietetete et st bet bbb bbbt enae e 112
Calibration of ART ERFICher ..ottt e st et s bbb s 112

. CORROSION RESEARCH ...ttt ettt ottt st e et sa b 114

Forced-Circulation STUIies ..ottt bbb 114
Fluoride Fuel Mixtures in INComel ..ot et eener e e 114
Liquid Metals in Inconel and Stainless Steel ... 117

Thermal-Convection StUdIs ...cciiiicirciencr sttt 118
Alkali-Meta!l Fluoride Fuel Mixtures in Hastelloy B ..cocoooeiiiii 118
Effect of Condition of Inner Surface of Hastelloy B Tubing on Depth of Attack .....c.cccoevrieneen. 119
Screening Tests of Special Fuel Mixtures ..., 120

General Corrosion STUdIES ..oiiiire ittt ettt ee e sttt bbb bbbt 122
Low-Cross-Section Brazing Alloys in Liquid Metals and in Fluoride Fuel Mixtures .................. 122
Cathalloy A-3T in SOAIUM ceiceeicicii e b 122
Rare-Earth Oxides in Sodium .ottt 124
Sodium in Type 316 Stainless Steel Thermal-Convection Loops ......cccccccovniiiiiiienierciecicce, 125
Boiling Sodium in Type 348 Stainless Steel Loops ..o 127
Compatibility of Hastelloy B and Beryllium in Sodium ... 128
Solid-Phase Bonding of Cermets oot ses e b e 129
Solubility of Lithium in NaK ..o 130

Mass Transfer and Corrosion in Sodium Hydroxide ......ccooviiniie 131
SHAtic CaPSUIE TeSES ittt bbb 131
Thermal-Convection Loop Tests .. ettt e et st 133

Chemical Studies of CorroSion ..o.iceiriieenreeiere s ereseseeee st sssise st st e r st s bbb s s snsbsnes 134
Physical Properties of Elastomers Exposed to Attack by Liquid Metals ...cocoevineiiiiincin, 134
Resistance of Possible Moderators to Fused Fluoride Mixtures ..o, 137
Diffusion of Chromium in AlIOYS «eveieeeieireice ittt snens 138

. METALLURGY AND CERAMICS .....oiiiitieieiere ettt ias s et et 139

Welding and Brazing STUdies ... 139
Examination of NaK-to-Air Radiators After Service at High Temperatures.....cooceoiiinniinii 139
Crack Susceptibility of Back-Brazed Tube-to-Header Joints ..o, 139
Measurement of Weld Shrinkage ...t 140
Dimensional Control During Fabrication of Pump Volutes ..o 142
Carmet-t0mMetal JOTNES troeoiii ittt st sttt et s s b r e s v e e st 143
Brazing Alloy Development ...t e 144

Mechanical Properties of InConel .....cc.cccviiiiiniiii s 146
Creep-Rupture Design Data ..ot 146
Low-5tress Creep Data .o iiieiccciii i s 146
Effect of Biaxial Stress on Creep Ductility at High Temperatures ....cccovroninniniiininnins 151

- Xi
 

Special Materials StUies ... e 151

Neutron Shield Material for High-Temperature Use ... 151
Fabrication of Boron-Containing Materials ..ot e 152
Inconel-Boron Compatibility ..o e et e 154
Nickel-Molybdenum-Base Alloys .........cccccioniiiiii e 155
N O UM F OB O I ON e et a s e s rar e st sbany e seee e e raae s ben e e rmnase s baneanses 160
Seamless Tubular Fuel Elements ... cenie s cesee s st s eeaeseasane s e sbnanineaane 161
Control Rod FObriCtion ciieeeeieieivveeteeimeteise s eeer s nseesesaaestsentsteasssesesnesassrerssesasessmestesarasesssesssransesanases 163
Shield Plugs for ART PuUmps ..o ittt st 163
Lithium-Magnesium Aloys ..o s s 163
Fabrication of Ceramic Materials ......ocivireiiiiieiee e e e 164
NONAESTIUCTIVE T@STING oottt st s em e s s b e s s s s e s sn e et s nresaneneens 164
. HEAT TRANSFER AND PHYSICAL PROPERTIES ...ttt ssranees 171
Fused-Salt Heat Trams er. ettt ettt e e bbb s e bba b ee s e se e ba e be e e sr e e s bbabba st e e sabaesmaese e 171
ART Fuel-to-NaK Heat EXChanger .........cocoiiiiiiiiiiire it et 172
ART Core Hydrodynamics ..occeoveiii it e bbbt e e 174
Temperature Structure in Region Beyond the ART Reflector ..o 174
Temperature Structure in an ldealized ART Core...oooiiiii e 176
ART Core Heat-Transfer Experiment ..ottt e e 176
HEAE CAPACIHTY ooviviiieieere st ee s er et s s e e e s aeese s s en e ar s e na et e e st eat nreaneneessenreereranes 176
VESCOSITY Lovrreainniiieieietetes et s e c bt e om e s b s bbb bbb e b s bt des Sha b a4 b b eat et bR e b e S b e b e e R bR e et e b e n Rt en e 178
Performance Comparisons of Several Fluoride Fuels ..o, 179
RADIATION DAMAGE ...ttt ettt s eaesteare s re et at e e e esteebeemeeesennas 181
Review of Tests of Corrosion of Inconel and Stability of Fuel Under Irradiation.........cccceevvrennnn.e. 181
Examination of the Disassembled MTR 1n-Pile Loop No. 3.t 183
Effects of Radiation on the Mechanical Properties of Structural Materials .....cccooooivieneee 187
SHPES S COMTOSTON T@STS 1oviviiieieieeees et ee it e e e e e ettt e st e e eee e eee e e e e eeaeesarestneesten e snsebtsstesesseeseenssens seneornaeons 187
Alternate Stress-Corrosion APParatus ... i ce e ieie et es e e s sse s s s e s e e s sresessssase s smeeeesnes 188
MTR Tensile Creep TeStS . airiairieseee e e e erreess e e srrsbeesessesasmateseressasssseneessarsesssstessasarens 190
Preumatic SHressing DeVice ettt et st ene e ean 191
Ductility of Nickel ATTOYs .ottt et esne e st st aocarasr s e nnees 191
Experimental Studies of Reactor Materials and Components .......ccocccoieiiniiiiciinnreseee e 192
Effect of Radiation on Stotic Corrosion of Structural Materials by Fused Salts ... 192
Holdup of Fission Gases by Charcoal Traps......ccciioieieiicicceicrteee sttt e 193
LITR Vertical In-Pile Loop ottt e 196
Instrumentation for ART Off-Gas Analysis .ciiriiciiiiiiecrecenec et 196
Inconel as a Thermal-Neutron-Flux Monitor.......ccoe ittt e 197
Measurement of MTR Flux near Tip of In-Pile Loop No. 3 et 198
Fast-Flux Measurements in Hole 19 of ORNL Graphite Reactor ......cccoevmeevvevcviinicveicieceeiee, 198
Chemical Effects of Nuclear Reactions ... it eva st s ra e 199
Effects of Fission Products on Properties of Fluoride Fuels ..o 199
Use of Natural Lithium in Fluoride Fuels Circulated in In-Pile Loops .....coocoiiiciiciiiien 203

Radiation Damage to Boron Carbide......cccuuiieriinicriniieriii i ete e e saen et e e ceaeaees 204

 
10.

11.

12.

13.

14.

 

ANALYTICAL CHEMISTRY OF REACTOR MATERIALS ..ottt ene e 207
Detection of Traces of NAK in Al et cr e et e e sae e e e 207
Detection of Microgram QUantities ........oociciiiiiiiiiiicciiece et et et s 207
Detection of Submicrogram QUAantities .....cccoiciieriieiiieiieeitcie et et 209
Spectrophotometric Determination of Titanium in Mixtures of Fluoride Salts with Tiron................ 210
Determination of Tantalum in Fused Mixtures of Fluoride Salts ...ccooveviieieiieccce 211
Determination of Oxygen in ZrF 4 by Bromination .......coeovmmieieniisciiiiicstr s 212
Determination of Micro Amounts of Boron in Fused Fluoride Salt Mixtures......c.ccccoeveccvivnmrineeene. 212
Spectrophotometric Determination of Bismuth in Fused Mixtures of Fluoride Salts ......ccoceenvennnnn. 213
Determination of Dissolved Oxygen in Lubricating Fluids....cccocoovveviriiriireece e, 213
Determination of Zirconium by the Compleximetric-Versene Method.........cccccocvierevieriniceeeiieceien. 214
Determination of Rare-Earth Elements in Stainless Steels and Inconel ....oooveviiniencicic, 214
Determination of Oxygen in Metallic Sodium ..ot s e 214
ANP Service Laboratory ...ttt sttt s a e 215
RECOVERY AND REPROCESSING OF REACTOR FUEL ..o 216
Pilot-Plant Design and Construction ......cciiereieeicicieie sttt esissss et s ns s smses s sesesesesssessssssnns 216
Engineering Developments .. ...t e 216
OMEACTOR .ttt ettt ettt e ettt e e be et e e s eaa b et ettt et esneae et s nenters e st ere et et et sesanreneseees 216
Freeze Valves .t ee saeee sae bbb ebb e b e eb et e beeseae e aeen s 216
Process Development ...ttt bbbttt e erae s 216
FIUOrination SUIes ... ee et bt et eb bbb e ee e sens 216
Vapor Pressure of the UF (-NaF Complex .......ccoiiiiimieiic st 219
Uranium Losses on Desorption of UF‘S From NaF oo eeene 219
PART Ill. SHIELDING RESEARCH
SHIELDING ANA LY SIS ettt e e sttt et saea b sne bt st saesbe e 223
Energy Absorption Resulting from Gamma Radiation Incident on a Multiregion Shield
With SIab GeOMEIIY ...ceeieie ettt e n bt st resanarebmesasebeaee s 223
REACTOR SHIELD DESIGN ..ottt ettt ettt s et s st b ase st e sesmnanesas s s s e e e enens 227
Gamma-Ray Heating in a 300-Mw Circulating Fuel Reactor ........ccoevveeicennivieceicicecescee e 227
Primary Gamma-Ray Heating ..ottt ettt s s ne st s ne s 228
Fission-Product Gamma-Ray Heating .....ccccvvimriiriiierseee et st sev e essean 233
Heating by Thermal-Neutron Captures in the Shield ... 234
Dose Rate Outside the ART Shield ..ottt eee st s s 236
LID TANK SHEELDING FACILITY oo iee e saeseesresasne s sersessasssasssss srsensssesessnnsseanens 237
Analysis of the Dynamic Source Tests on Mockups of the Reflector-Moderated Reactor
oTaTe 25T T 1 I O TSROSO 237
BULK SHIELDING FACILITY ettt e saeb et s e 249
Gamma-Ray Streaming Through the NaK Pipes That Penetrate the ART Shield ... 249
Decay of Fission-Product Gamma Radiation.........c.ccocoiiiiciiiiciicii e 250

 

; T vas
xiii
 
ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART |. REACTOR THEORY, COMPONENT
DEVELOPMENT, AND CONSTRUCTION

1. Reflector-Moderated Reactor

Construction work on the building additions,
building alterations, and cell installation (package
1) for the Aircraft Reactor Test facility was on
schedule and at the 50% completion point at the
end of the quarter. A contract for $58,400 was
awarded to Rentenbach Engineering Company,
Knoxville, Tennessee, for package 2 construction
work, which consists in the installation of the
diesel-generators and facility, the electrical
confrol centers, and the spectrometer-room elec-
trical and air-conditioning equipment,

Design was completed and contract negotiations
are under way for package A work, which consists
in auxiliary-services piping. Design work on
package 3, which includes process equipment,
process piping, etc., is currently under way,

Detailed layouts have been completed on all
major subassemblies of the ART, Typical weld
joints are being fabricated to determine shrinkage
allowances for the final weld designs. Also,
extensive calculations of stress distributions in
major structural components and in piping are
under way,

Recent calculations have indicated that the heat
flux through the core sheils of the ART will be
higher than previously estimated. The design
conditions for the reflector-moderator cooling
system were therefore modified accordingly. The
design sodium-temperature rise was increased
from 150 to 200°F and the design sodium-system
operating-temperature range became 1050 to 1250°F
to take the estimated heat load of 6.2 Mw. Minor
changes were also made in the dimensions of the
circuit, to assure that the pressure drop will not
exceed 30 psi, and in the arrangement of the
auxiliary radiators, to accommodate the increased
heat load.

The thickness of the lead portion of the reactor
shield was reduced from 7 to 4.3 in. to obtain
space for increased neutron shielding around the
pumps, This decrease in lead thickness will in-
crease the gamma-ray dose rate by a factor of 10

 

{from 7/8 r/hr at 50 ft to ~9 r/hr), but the dose rate
will be substantially less than that proposed by
most of the airframe companies for nuclear aircraft.

A detailed one-sixth-scale model of the facility
is being constructed to assist in layout, fabrication,
and installation work., Also, the one-half<scale
plastic model of the top portion of the reactor
(**north head'") is being modified to reflect current
design. This model is used for studies of flow
and for layout design. A full-scale aluminum
mockup of the north head is nearly complete; this
mockup will be used for simulated service tests of
the fuel and sodium systems.

Studies of flow in a full-scale plastic model of
the core were continued. A guide-vane system has
been developed which generates a flow pattern
that contains no flow reversal throughout the core
annulus and provides good surface scrubbing, good
midstream mixing, and fairly high velocities
throughout the upper half of the core. Further
modifications are being made to improve the flow
in the lower half of the core.

Flow diagrams and instrumentation lists were
prepared for the Engineering Test Unit (ETW),
which is a nonnuclear mockup of the ART, Present
design and shop schedules indicate that the ETU
will be ready for operation in January 1957. Con-
siderable progress has been made in the establish-
ment of contracts and facilities for the procure-
ment of major ART and ETU components, and an
inspection group has been set up for quality control
of materials and fabricated parts.

Modifications have been made in the proposed
ART control system on the basis of information
obtained with the reactor simulator, which indi-
cates that in the event of g fuel-pump failure at
design power the maximum fuel core-outlet tempera-
ture can be limited to 1700°F by inserting the
control rod at the rate of ]/8% (Ak/k)/sec, The
magnetic, emergency, 1% rod drop was abandoned
because of the thermal-shock conditions that would
develop at the core outlet in the event of a spurious
rod drop at design power, and an emergency rod
insert rate of ]/8% (Ak/k)/sec was adopted. Per-
formance with U233 fuel was also investigated on
the simulator. The lower proportion of delayed

) 1
ANP PROJECT PROGRESS REPORT

neutrons, in comparison with U233 fyel, made the
fuel core-outlet temperature less sensitive to a
fuel-pump failure. It was also found that the core-
outlet fuel temperature would not exceed 1700°F,
even with no corrective control-rod action.

Three separate calculations of the activity that
will be produced in the NaK in the fuel-to-NaK
heat exchangers of the ART were normalized to
the current ART design. The normalized values
of activity from core neutrons for the three calcu-
lations were 463, 600, and 970 curies. The values
of activity from delayed neutrons could be normalized
for only two of the calculations, and thus total
activities were obtained for only those two. The
total normalized NaK activities were 650 and

1057 curies.

An estimate was made of the activity to be ex-
pected in the ART radiators because of the depo-
sition on the radiator tubes or header walls of
constituents of Inconel that will become activated
in the heat exchanger and will be carried to the
radiators in the NaK stream. Based on specific-
activity and weight-of-deposit estimates, the total
activity to be expected from mass-transferred
material in the ART radiators is about 1 curie,
The dose level 1 ft from the hot surface of the
ART radiators is thus expected to be about 1 r/hr
and to be due mostly to Co69,

Information regarding all possible sources of
gamma rays in the ART and the factors that will
affect their transmission and absorption is being
collected for calculations of the radiation heating
in the ART. Preliminary calculations have been
made to validate the calculational techniques and
to determine the gamma-ray sources that can be
neglected in calculating the heat-deposition rate.

A calculation of the self-absorption of decay
gamma rays in the ART fuel dump tank indicated
a value of 90%, and therefore cooling facilities
must be provided to remove almost all the decay
heat. The activation of the sodium to be used as
the reflector coolant was recalculated to correspond
to the present design, and a value of 1,45 x 106
curies was obtained. In onother calculation it was
found that it would be necessary to add 0,13 fi3
of lithium to the sodium in the reflector cooling
circuit to shut down the ART from isothermal
operation at 1200°F to room temperature; from

1400°F, 0.15 13 of lithium would be required.

2. Experimental Reactor Engineering

Disassembly and sectioning of in-pile loop No. 3,
which operated in the MTR, were completed. This
Inconel forced-circulation loop circulated the
fluoride fuel mixture NoF-ZrF4-UF4 (53.5-40-6.5
mole %) at a maximum temperature of about 1500°F,
There was a temperature differential in the fuel
system of 155°F for 103 hr and 100°F for 168 hr.
Metallographic examinations of various portions of
the loop have revealed corrosion penetration to a
maximum depth of 1 mil,

In-pile loop No. 4 was completed and shipped to
NRTS. It was inserted in the MTR HB-.3 beam hole
on February 6 and removed February 29. [t circu-
lated fuel for approximately 500 hr, and there was
a temperature differential in the fuel system for
approximately 390 hr. Loop No. 5 has been
partially assembled, but it will not be completed
until the operation of loop No. 4 can be analyzed
sufficiently to indicate any required changes.

Ten Inconel forced-circulation loops that were
electrical-resistance heated and two that were
gas heated were operated with fused salts. Six
forced-circulation loops were operated with sodium
and NaoK in Inconel and in stainless steel.

Synthetic lubricants in the UCON LB series were
investigated to determine their suitability as pump
lubricating fluids in the ART. Since an operating
temperature of 200 to 240°F is anticipated, pe-
troleum products, which ‘‘coke'' at temperatures
above 180°F, might not be satisfactory from a
heat-removal standpoint. Unfortunately, the UCON
fluids pick up copper from system components and
plate it out on mechanical-seal interfaces. Al-
though this problem could be circumvented by
eliminating all copper-brass parts, equally difficult
design problems would be introduced. Other
synthetic fluids are being investigated.

The lower seal of the ART fuel pump has been
modified to assure that any oil leakage will ac-
cumulate in a catch basin, from which it can be
continuously removed. For the modified design
the seal will be a bellows type with a stationary
Graphitar 39 nose piece running against a
hardened-steel ring clamped to the rotating shaft,

Additional tests of the performance of the ART
sodium pump with water were made, Estimates of
the performance with sodium, kased on the water-
test data, show that the cavitation characteristics
of the pump will be satisfactory, The tests also
indicate that the balance is satisfactory and that

 
the pump will produce the required design head of
90 ft and flow rate of 430 gpm at an efficiency of
63% at 2860 rpm, It will require 15.6 bhp., Water
tests of the NaK pump have shown that the design
point can be reached at considerably below the
maximum motor speed of 3550 rpm, The efficiency
is approximately 75% at design speed, and the
volute unbalance is within satisfactory limits,

Difficulties were encountered in high-temperature
tests of the fuel pump with NaK, The lower seal
oil-leakage removal system did not function
properly on the first pump tested, and the lubri.
cating-oil pump on the second pump was unsatis.
factory. Two test stands are being prepared for
high-temperature petformance and endurance tests
of the fuel pump with fuel. Similar stands for
testing sodium, NaK, and special pumps were
designed.

Operations were continued on heat exchanger
test stands. York radiator No. 3, Pratt & Whitney
radiator No. 2, and ORNL radiator No, 3 failed
during the quarter, These failures occurred on the
air~upstream side at or near the base plate. An
intermediate heat exchanger, ORNL No. 1, which
had operated for 1825 hr, also failed. These units
are being examined. York radiator No. 9 was put
into operation, This radiator, which is the first
fabricated according to the revised design, has
no side plates, support plates, or base plate.
Nickel plates with oversize holes provide top and
bottom air seals.

The additional air-pressure~drop and heat-transfer
data obtained were in substantial agreement with
the previous data, All the 80-gal heat exchanger
test loops are equipped with circulating cold traps
that are capable of maintaining a contamination
level of 100 ppm O,. The 20-gal systems are
maintained at a contamination level of below 34
ppm O, with cold traps.

A stand for testing the themmal stability of the
outer core shell of the ART is in opetation. A
cycle time of 21/2 hr is being used, with a T-hr
hold time under temperature differential and under
isothermal conditions, and a transient time be-
tween conditions of approximately 15 min, The
core shell will be examined after 100 thermal
cycles have been completed.

Anvil-bending tests of Inconel specimens in a
helium atmosphere are under way. Metallographic
of the specimens tested thus far
initiate at the outside

examinations
have shown cracks that

PERIOD ENDING MARCH 10, 1956

fibers and propagate inward. |t appears from the
data obtained that the number of cracks produced
for a given number of cycles at 1% strain is higher
at 1400°F than at 1600°F,

Tests of Swagelok fittings were made to determine
their suitability as replacements for welded joints
These fittings proved
to be satisfactory in isothermal tests in contact
with fused salts, and tests of the reliability of the
joints under thermal cyeling are under way.

Tests of the first ART prototype dump valve
indicated the need for a design modification to
provide a conically shaped seating surface. A
test stand was designed for development testing of
cold traps and plugging indicators, and the ZrF,
vaporstrap test system was put into operation,

in forced-circulation loops.

3. Critical Experiments

The study of the sodiumecooled reflector-
moderated reactor with solid fuel, proposed by the
Nuclear Development Corporation of America (NDA),
was continued. The loading of the mockup, which
was found to have a room-temperature critical mass
of 31 kg of U239, was increased to 33 kg to provide
excess reactivity for measurements of neutron flux
distribution and the mass-reactivity coefficients of
various reactor component materials. The measure-
ments have been reported in detail and correlated
with the predicted reactor parameters by NDA,

PART Ii. MATERIALS RESEARCH

4, Chemistry of Reactor Materials

Additional phase-equilibrium studies were carried
out in several fluoride systems in order to gain a
better understanding of the structure and to aid in
devising improved fuels. Extensive low-melting
areas became apparent in the NaF-RbF-ZrF,
system, and it appears likely that a useful fuel
may be developed upon the addition of UF ;. New
data indicated that the viscosity of NaF-BeF ,-UF
(70-26-4 mole %) was lower than that indicated by
previous measurements and that further investi-
gation of this system might lead to a desirable
fuel.

A material containing 63.5 mole % NaF, 18.0
mole % ZrF4, and 18.5 mole % UF, was found to
show promise as a lower melting, nonsegregating
substitute for Na,UF  as a fuel concentrate for
ANP PRQJECT PROGRESS REPORT

““enrichment’ of reactor systems. |t appears to
melt at 607°C, about 50°C lower than Na,UF .
Studies were made on the UF ,-UO, system which
led to a value of 1034°C for the melting point of
pure UF ,, a value which is in good agreement with
the published data. With amounts of UO, on the
order of 2 wt %, there was a tendency for the
mixtures to undercool. The compound NaF:2UF ,,
which has appeared as a primary phase field in
several ternary systems, was found to be a sub-
solidus binary in the system NaF-LiF-UF .

 

in molten NaF-ZrF , (53-47 mole %) was concluded.
At the temperatures at which the measurements
were carried out, there was no change in the values
of the equilibrium constant

2
Ky = PHF/XFaF2PH2

large concentration ranges. From the ex-
perimental data and from published values of free

over

energy of formation of the compounds involved,
activity coefficients for FeF, were computed by
using crystalline FeF, and pure supercooled liquid
FeF., as standard states.

From the emf of concentration cells of the type

M|MF ,{a,), NaF(a,), ZrF ,(a;) || NaF(a3), ZrF ((ag), MF (a])|M ,

 

Experimental work was also carried out in the
following zirconium-bearing systems: KF-ZrF ,
RbF-ZrF,, NaF-KF-ZrF,. A newly revised, but
still tentative, phase diagram of the RbF.ZrF,
system was prepared from the data obtained by
thermal analysis and by petrographic and x-ray
analyses of quenched and slowly cooled samples.

Thermat-analysis data were obtained on the
RbF-BeF2 system, Phase relationships were
found to be quite similar to those in the KF-BeF,
systems, and four compounds were tentatively
identified. Quenching studies were continued in
the NaF-LiF.BeF, system, particularly in that
portion having LiF, NaF, and NaBeF, at the
apexes of the triangle. In the Nc:F-LiF-Ber-UF4
system many compositions containing moderately
high amounts of UF, and small amounts of BeF,
were shown to have acceptable liquidus tempera-
tures, but on the basis of available viscosity data
it appeared that the viscosity would not be com-
petitive with that of the best ZrF -base fuels.

Additional information was obtained in the NaF-
LiF-CaF, and NaF-MgF ,-CaF, systems. Explora-
tory studies in the NaF-CeF , and RbF-CeF, sys-
tem gave no indication of the presence of genuinely
low-melting compositions., In connection with the
electrochemical studies involving structurai-metal
fluorides, an investigation of the LiF-NiF, system
was begun,

The diversified program for studying chemical
reactions in molten fluorides was continued. The
study of the equilibrium

FBFQ(d) + H2(g);‘. Fe(s) + 2HF(g)

activity coefficient ratios at various concentra-
tions were derived for FeF, in molten NaF-ZrF
(53-47 mole %) in the temperature range 550 to
700°C, Sharp changes in emf were observed when-
ever the solubility of a metal fluoride was exceeded
in such a half cell, and by means of such observa-
tions solubility data were obtained on Fer, NiF,,
and CrF, in the NaF-ZrF , solvent. From cells con-
taining CrCl, and electrodes of pure Cr° and Cr-Ni
alloys, emf measurements gave some indication of
the chromium activity in the alloys, The data on
inconel seemed to be reasonably close to those
for an ideal solid solution,

Studies of the reactions of tantalum and tungsten
with UF, in NaF-LiF-KF eutectic showed that
metal was stable and that they offered
little or no advantage over chromium, Niobium, in
marked contrast to its instability toward UF4 in
the NaF-LiF-KF system, was shown to be quite
stable when the UF , was contained either in NaF-
ZrF , (50-50 mole %) or in NaF-LiF-ZrF , (22-55-23
mole %).

Studies of the reaction

KF + UF3——>UF4+K°

neither

in NaF-KF-LiF eutectic indicated that UF, is
quite unstable, possibly because of complexing
of UF, by fluoride ions. This produces not only
unexpected quantities of K® but also g high activity
of uranium metal to satisfy the

4UF , —> 3UF, + U

disproportionation reaction,
Many special structural-metal fluorides were
prepared for experimental studies, including CrF,,
Fer, CeF,, LaF,, complexes of ZrF4 and metal
fluorides, CuF,, NaCrF,, and NaNiF,. Color
studies of alkali metals in equilibrium with alkali
halides at high temperatures are under way. In
general, alkali-metal phases in contact with
alkali-metal salts are highly and strikingly colored,

In preparation for thermodynamic studies of
molten salts, vapor-pressure, density, and surface-
tension measurements were made, Vapor-pressure
equations for ZrF, obtained by various investi.
gators were correlated, and measurements of the
system KF-ZrF, were made., The vapor pressure
of pure F:e(:l2 was measured in preparation for a
study of ihe activity of FeC[2 in melts of alkali-
metal chlorides,
are being studied by the sessile-drop technique.
Apparatus added to the conventional equipment for
fuel purification is being used for the study of
variations of density and surface tension of molten
salts with changes in chemical composition,

It was established that, in the solid, CaCl,
forms 1:1 complexes with KCl, RbCl, and CsCI.
Density and preliminary conductance data were
obtained for several molten rare«earth halides, and
fused alkaline-earth halides and their mixtures
with some alkali halides are being investigated.

Experimental studies were made with the copper-
lined stainless steel reactor cans proposed for

Surface tensions of molten salts

use in large-scale fuel production to replace the
unsatisfactory nickel reactor cans, It appears
that the copper-lined stainless steel reactor cans
will be satisfactory with ZrF ,-bearing mixtures.
Tests with alkali-metal and BeF ,-bearing materials
are yet to be made.

The equipment for conversion of ZrCl, to ZtF,
by direct hydroflucrination of the solid is being
rebuilt to overcome mechanical difficulties, and
an alternate process for the conversion is being
studied on a small scale,

Experiments carried out with the fuel-enrichment
apparatus used during the high-temperature critical
experiment on the ART showed that the equipment
will be satisfactory if usage is restricted to addi-
tions greater than about 500 g.

5. Corrosion Research

The effects on corrosion of varying the ratio of
hot-leg surface area to volume of fluoride fuel
mixture in a forced-circulation Inconel loop were
investigated. In a loop in which the ratio was

decreased by a factor of 2 in comparison with the

PERIOD ENDING MARCH 10, 1956

ratio for a standard loop, the maximum depth of
attack in 1000 hr was 6 mils, which is comparable
to the usual attack in a standard foop. The addi-
tion of sufficient fuel mixture to decrease the
ratio by a factor of 4 resulted in attack to a depth
of 9 mils in 1000 hr, Further, a very thin gald-

colored deposit was found in the cooler portions

of the loop. The significance of the increase in
depth of attack with increased volume cannot be
judged until the source of the deposit is found, but,
even if the increase in depth of attack can be
attributed to the increase in volume, it appears
that wvarying the system volume in the range
studied does not have a significant effect on
corrosive attack over a 1000-hr period.

Examinations of two forced-circulation loops
that had operated with the fuel mixture NaF-KF-
LiF-(UF, + UF,) clearly demonstrated that the
presence of U3* in such a mixture is very effective
in reducing hot-leg attack. Continuous layers of
metallic uranium, formed by the dispropertionation
of UF3, were found, however, along the walls in
the cooled sections of the loops. The maximum
depth of attack in 1000 hr was 2 mils, in contrast
to attack to a depth of 35 mils in a loop operated
under similar conditions but which circulated an
alkali-metal fluoride mixture with
present only as UF

the uranium

Two Inconel forced-circulation loops in which
NaK was circulated completed 1000 hr of operation
with a temperature gradient of 300°F and a maxi-
mum fluid temperature of 1500°F, These loops
differed only in that one included a bypass cold
trap. As in the case of Inconel-sodium loops,
metallic deposits were found in the economizers
and in the cooled sections of the loops. Analyses
of the deposits showed them to be identical to
those found in Inconel-sodium loops., The simi-
larity of the results for the two loops indicated
that the cold trap was not effective in reducing
mass transfer. A stainless steel loop operated
under the same conditions with sodium showed
only very slight mass transfer in the economizer
and in the cold leg.

Examinations of two Hastelloy B thermal-
convection loops operated for 500 hr with an alkali-
metal fluoride mixture containing UF4 at a maximum
temperature of 1500°F showed the maximum attack
to be less than 2 mils. The attack appeared as
heavy surface pitting, A similar loop that was
operated for 2000 hr to determine the effect of time
ANP PROJECT PROGRESS REPORT

of operation showed no increase in depth of attack,
but the surface pits were more concentrated.

Thermal-convection loops constructed of Hastel-
loy B tubing that had been reamed to ensure uni-
formity of the inner surface were operated with
sodium and with NaF-ZrF ,-UF , for 500, 1000, and
1500 hr to determine the effect of increasing the
operating time. The depths of attack were similar
to those found in loops constructed of as-received
tubing, and the variations in aftack as a function
of operating temperature were slight,

Screening tests of special fuel mixtures are under
way. In the preliminary Inconel thermal-convection
loop tests, it was demonstrated that NaF.LiF-
ZrF ,-UF , mixtures show corrosion behavior typical

of alkaliemetal fluoride mixtures rather than
zirconium-base mixtures.
Six standard thermal-convection loops were

operated as the first of a series of loops to study
the effect on corrosion of various alkali-metal
fluorides as components of the basic fluoride fuel
mixture MF-ZrF ,.-UF, (50-46-4 mole %), where M
stands for potassium, rubidium, or lithium. The
attack with the lithium-bearing mixture was severe,
but the attack with the potassium and rubidium
mixtures was similar to that normally found with

NaF-ZrF 4-UF“.

Inconel tube-to-header joints brazed with nickel-
chromium-germanium-silicon low-cross-section al-
loy s were tested in sodium, NaK, and NaF-ZrF ,.UF
in seesaw apparatus. Preliminary data indicate
that alloys with high nickel contents are the most
corrosion-resistant to NaK. In a study of the effect
of brazing time on corrosion, the depths of attack
on joints brazed slowly (4 hr) and rapidly (10 min)
were found to be the same.

Cathalloy A31 {4 wt % W-96 wt % Ni) was
found to have good corrosion resistance to static

sodium at 1500°F.

One specimen of Sm203 and two specimens of a
commercial rare-earth oxide mixture (63.8 wt %
Sm,0;-26.9 wt % Gd,0;—balance primarily other
rare-earth oxides) were tested in static sodium in
Inconel containers at 1500°F, One of the com-
mercial specimens was tested for 500 hr, and the
other commercial specimen and the Sm O, speci-
men were tested for 1000 hr. All the specimens
changed color; however, powder x-ray diffraction
comparisons of the untested and tested specimens
did not show any reaction products,

Two type 316 stainless steel thermal-convection
loops were operated with sodium to study the effect
of a diffusion cold trap on the amount of corrosion
and mass transfer observed in such a system.
There was less attack and fewer mass-transferred
crystals in the loop which had the cold trap.

Three type 348 stainless steel-boiling-sodium
loops have been operated for various lengths of
time to study the extent of mass transfer in a
stainless steel—sodium system in which the oxygen
content of the sodium is held to a very low level.
No mass-transfer crystals were detected in the
cold sections of these systems,

Tests have been conducted at 1200°F for 1000 hr
to study the extent of dissimilar-metal mass trans-
fer of beryllium metal to Hastelloy B in contact
with sodium as a function of the distance between
the two materials. The results indicated that to
avoid extensive alloying of beryllium with Hastel-
loy B at 1200°F a minimum separation distance of
20 mils is required,

A solid-phase-bonding screening test of K151A
(70% TiC-10% NbTaTiC,-20% Ni) against K152B
(64% TiC—6% NbTaTiC,-30% Ni) showed that
these cermets solid-phase bond when in line con-
tact at a pressure of approximately 280 Ib per
linear inch in NaF-ZrF ,.UF, (53.5-40-6.5 mole %)
at 1500°F for 1000 hr.

A series of differential-thermal-analysis tests
were performed to determine the solubility of
lithium in NaK. The data obtained showed that
the solubility is very low, 0.25 wt % Li being
soluble at 75°C,

Nine nickel- and iron-base alloys of special
compositions were prepared and were subjected to
static corrosion tests in sodium hydroxide for
100 hr at 1500°F. The most promising alloy of the
group was a 90% Ni-10% Mo alloy, which was
attacked to a depth of less than 0.5 mil. Further
tests are to be conducted on specimens of this
composition,

Thermal-convection loop tests of fused sodium
hydroxide in promising container metals are also
under way. The preliminary tests with Inconel
loops showed some discoloration of the hot zones
and etching in the cold legs. There were no
massive deposits in any of the loops. Preliminary
tests of nickel are under way.

Screening tests of elastomers for possible use
as valve seat materials
initiated,

in NaK circuits were
and several promising materials were
selected for further testing. Three General Electric
Company silicone specimens remained intact after
exposure to NaK at 200°C for three days under a
helium atmosphere,

A search for a moderator that can be cooled by
direct contact with the molten fivoride fuel mix-
ture was initiated, Thus far, beryllium carbide,
beryllium oxide, and boron carbide have been
tested in NaF-ZrF .UF, in Inconel capsules at
800°C for 24 hr. The beryllium carbide specimen
reacted completely with the fuel melt; the beryllium
oxide specimen showed a 2% change in dimensions
and a 6% gain in weight; and the boron carbide
specimen showed a slight loss in weight.

Experiments are under way in which activation
analyses are being used to study the diffusion of
chromium in Inconel. It is believed that succes-
sively milling small amounts of metal from the
inside of an irradiated tube will produce a satis-
factary series of samples from increasing depths,
The amount of the sample, the depth of each milling,
and « measurement of the amount of Cr51 (278-day)
radioactivity in each portion will be used to de-
termine the degree of chromium diffusion as a
result of corrosive attack by molten fluorides.

6. Metaliurgy and Ceramics

Metallographic investigations are under way on
NaK-to-air radiators that failed after various
periods of service at high temperatures. A corre-
lation of the degree of adherence of the braze
material at the tube~to-fin joints and the degree of
oxidation of the copper fins with heat transfer
performance is expected to give a better under-
standing of the relative importance of the fabri-
cation features. An experiment was conducted to
determine the influence of the stresses set up
during welding on the crack susceptibility of back-
brazed tube-to-header joints., Two core halves
were brazed, one in the conventional upright
position and one in the horizontal position, and
dye-penetrant inspection indicated freedom from
cracks both before and after welding.

Tests of transverse shrinkage of welds of the
type to be used in the fabrication of the ART were
made. [t was found that weld shrinkage increased
with plate thickness increases and that the Heliarc
welding process resulted in a larger shrinkage for
a given joint design in a plate of given thickness
than did the metallic-arc welding process.

PERIOD ENDING MARCH 10, 1956

Methods for controlling, during fabrication, the
critical spacing between the two volutes which
constitute the pump housing of the ART NaK pump
are being studied, A combination of welding and
annealing was developed that gave satisfactory
spacing, and a brazing procedure is being in-
vestigated.

An experimental program is under way for de-
veloping a reliable and consistent procedure for
joining cermet valve components to metallic
structural materials, such as Inconel. A highly
promising procedure has been found in which an
interfacial reaction between the cermet and the
nickel transition layer at 1350°C results in a
metallurgical bond. The nickel layer is then
brazed to the Inconel component,

Flow=-point determinations were made on several
low-cross-section brazing alloys being considered
for use in the fabrication of the ART. Four alloys
with satisfactorily low flow points were submitted
for corrosion festing.

Satisfactory brazing of aluminum bronze fins to
Inconel was obtained by wusing a thin
(0.002-in.) electroplate of iron to facilitate welding,
The addition of manganese to nickel-base high-

tubes

temperature alloys was also found to promote
melting,

The creep-rupture testing program for Inconel
was nearly completed, The design curves for as-
received (fine-grained) and annealed (coarse-
grained) Inconel in argon and in the fuel mixture
NaF-ZrF -UF, (50-46-4 mole %) at 1300, 1500,
and 1650°F were completed.

The effect of a biaxial-stress system, such as
is found in pressurized tubing, on the ductility of
Inconel was evaluated by comparing the ductility
obtained in a tube-burst type of test to that ob-
tained for uniaxially-stressed sheet-type speci-
mens. The decrease in ductility shown by the
tube-burst specimen may be attributed to the 2-to-]
hoop-to-axial stress ratio set up in a closed-end
pressurized tube,

Metal-bonded boron bodies are being studied as
possible alternates for hot-pressed B ,C as neutron
shielding material because of uncertainties with
regard to the effects of irradiation on hot-pressed
B,C. The densities obtained have varied from
about 0.30 to 1.5 g of boron per cubic centimeter of
material.  Metal matrices of copper, iron, and
molybdenum were investigated. Rapid deteriora-
tion of the metallic properties was found to occur
ANP PROJECT PROGRESS REPORT

as the boron content was increased. The most
promising cermet was the copper-B,C mixture,

A design scheme for the ART neutron shield was
devised which will permit the radiation damage to
be absorbed in a ductile matrix of copper and, also,
permit the use of B C ceramic tiles to increase
the boron density and, hence, reduce the NaK
activation. This was achieved through the use of a
layer of copper-B4C backed by canned B ,C ceramic
tiles.

The reactions which occur between Inconel and
metal borides, oxides, nitrides, carbides, and
carbonitrides were determined at temperatures be-
tween 1100 and 2000°F. The reactions between
lnconel and boron-containing materials would
prohibit their use in contact with each other above
1500°F,

Fabricability studies were carried out with fair
results on heats of the nominal composition of
Hastelloy B with carbon additions of up to 0.1%.
The carbon additions were effective in reducing
oxide-type inclusions, and the ingots were hot-
rolled to sheet successfully., A similar series of
melts of Cr-Mo-Ni compositions with carbon addi-
tions is being prepared in an effort to improve
fabricability of these alloys. Metallographic ex-
amination of a cracked extrusion of Hastelloy W
revealed eutectic melting at the base of a crack.
Identification of the eutectic was not possible,

Room-temperature tensile tests of pack-rolled
clad-niobium sheet have been completed. The
results compare favorably with those obtained with
wrought stock. Studies were continued of diffusion
barriers for cladding of niobium with Inconel. At
present the duplex barrier of copper—stainless
steel looks most promising. A technique has been
developed by which completely clad creep speci-
mens of niobium will be prepared and tested.

Extensive compatibility tests of UO, and
niobium, both powdered and wrought, were carried
out by heating for 500 hr at 1000°C, All evidence
indicates that the two materials are compatible,
Niobium is being considered for use in solid fuel
elements because of its good high-temperature
strength,

Encouraging results were obtained on the ex-
trusion of simulated seamless tubular fuel ele-
ments with mixtures of Al,O, and stainless steel
as cores. lhree-ply extrusions at ratios of 9:1
and 21:1 showed fairly uniform cladding and core
thicknesses.  Sections have been sent to the
Superior Tube Company for redrawing. A similar

technique is to be used for the extrusion of ART
control rods.

Work is continuing on the development of a high-
density barrier shield plug for the ART pumps.
Fabrication procedures have been established for
several suitable materials. A thermal conductivity
apparatus has been constructed and tested for
conductivity measurements on the possible ma-
terials,

A study of the fabricability of a light alloy con-
taining 20% lithium, which would be useful as
shielding material, was initiated. Several pro-
tective coatings and cladding materials for a
lithium-magnesium alloy are being investigated.

Methods are being studied for preparing compacts
of samarium and gadolinium oxides and cermets of
iron and rare-earth oxides and nickel and rare-
earth oxides for use in reactor control elements.

The cyclograph flow-detection instrument is
being used successfully for the inspection of
small-digmeter tubing. The instrument is essen-
tially a tuned oscillator in which the oscillator
coil encircles the tubing, The variables that cause
flow signals have been studied, and spurious
signals have been defined. The wobble of the
tubing in the coil, which was the worst offender
in producing spurious flow indications, can be
eliminated by a well-designed mechanical feed :
mechanism, Optimum test frequencies have been
determined as functions of tubing diameter and
wall thickness,

The results obtained from immersed ultrasound,
radiography, and fluorescent penetrant inspections
were compared. It appears that no one of the
inspection methods alone is adequate and that if
any one of them were eliminated a few defects
would be overlooked.

7. Heat Transfer and Physical Properties

Preliminary heat-transfer data were obtained for
molten NaF-KF-LiF-UF, (11.2-41-45.3-2,5 mole %)
flowing through Hastelloy B tubes in a pressurized
forced-convection system. The data fall about
30% below the normal correlation for ordinary
fluids. Heat-transfer and isothermal friction char-
acteristics were determined in the transition re-
gion for three degrees of spacer density for the
ART heat exchanger. The limited data obtained
thus far appear to indicate that for a given heat
exchanger pressure drop the heat transfer coeffi-
cient is almost independent of the spacer density.
A 5/22-scale model of the ART sodium coolant
annulus was fabricated and assembled for a study
of the fluid flow distribution. Pressure-drop and
flow-distribution measurements are being obtained
for concentric-annulus conditions, for various
radial displacements of the outer core shell, and
for other conditions, ‘

Yisual studies were made of the flow through the
21-in. ART core model after the installation of
vortex generators supplied by Pratt & Whitney
Aircraft, Yane angles of 50, 55, 60, 65, and 70 deg
were used. Although no large reversed flow regions
existed, steady and unsteady flow were observed
in various regions of the core., QOther means to
obtain better flow distribution are being studied.,

The temperature structures in an idealized ART
core and in the region beyond the beryilium re-
flector were calculated. Upper and lower limits
for the temperature rise of the sodium flow over
the outer surface of the beryllium reflector were
25 and 33°F. The core-wall cooling by the sodium
in the core annuli was found to be approximately
2.5 Mw. The half-scale model of the ART core to
be used for volume-heat-source heat-transfer ex-
periments was completed and was operated with
water. Operating procedutes and schedules are
being established,

A study of the effect of composition on heat
capacity for NaF-ZrF , salt mixtures was investi-
gated; no significant differences were found over
the composition range investigated, 53 to 65
mole % NaF and 35 to 47 mole % ZrF ,. A study
has been initiated to determine the influence of
fission products on the viscosities of fuel mix-
tures. The viscosities of NaF-ZrF , (50-50 mole %),
NaF-LiF-BeF, (49-36-15 mole %) and NaF-.LiF-
BeF ,-UF, (56-21-20-3 mole %) were determined,

Performance comparisons of several of the more
important ANP fuels were made for heat and
momentum transfer in the reactor core and in the
heat exchanger. The alkali-metal base fuels
were found to be superior to the zirconium-base
fuels. The major differences between the fuels
occur in the heat exchangers. The heat exchanger
pressure drops were found to be only about half as
great for the alkali-metal base fuels as for the
zirconium-base fuels, and the radial fuel tempera-
ture differences were from 20 to 40% lower.

8. Radiation Damage

The effects of irradiation on the corrosion of
Inconel exposed to a fluoride fuel mixture and on

PERIOD ENDING MARCH 10, 1956

the physical and chemical stability of the fuel
mixture have been investigated by irradiating
Inconel capsules filled with static fuel in the MTR
and by operating in-pile forced-circulation Inconel
loops in the LITR and the MTR. In the many
capsule tests and in the three in-pile loop tests
made to date, no major changes that can be at-
tributed to irradiation, other than the normal burnup
of the uranium, have occurred in the fuel mixtures,
The metallurgical examinations of the lnconel
capsules and tubing have likewise shown no
change in corrosion that can be the result of radia-
tion damage. The low corrosion results obtained
for the in-pile loops have been confirmed by
chemical analyses for corrosion products in the
fuel mixtures.

The sectioning of MTR in-pile loop No. 3 was
completed at NRTS, and, upon receipt of the sec-
tions at ORNL, examinations of the disassembled
loop were initiated, Metallographic examination of
as-polished Inconel specimens from the nose
coil revealed no evidence of corrosive attack.
Etched specimens appeared to have moderate inter-
granular attack to a depth of about 1 mil, but this
may be an etching effect rather than true corrosive
attack.  Cells are being readied for obtaining
specimens from the pump and the remaining lengths
of fuel tubing. Chemical analyses of both the
fuel and fuel tubing are under way. Provisions
for the disassembly of the next loop are being
made,

Eight Inconel, helium-pressurized, tube-burst,
stress-corrosion specimens were expased to radia-
tion in hole HB-3 of the LITR in a helium atmos-
phere, with a circumferential stress on each speci-
men of 2000 psi, Similar specimens are being
tested out-of-pile to obtain confrol data. A plot of
the time-to-rupture vs the operating temperature
during irradiation showed an unusual degree of
scatter in the data and led to an investigation of
the quality of the tubing. All but four of 28 speci-
mens of tubing stock examined by nondestructive
inspection were rejected because of pinholes and
spongy regions. Metallographic inspection of 25
sections from the 24 rejected tubes, however,
resulted in the location of only one appreciable
defect. Better quality tubing will be used for
future experimentation,

An alternate apparatus for stress-corrosion test-
ing of structural materials in contact with a fuel
mixture has been designed for use in the LITR,
The thermal-neutron-flux data needed for the
ANP PROJECT PROGRESS REPORT

design calculations were obtained by irradiating
a mockup capsule at a position about 2 in, from
the inner end of hole HB-3 of the LITR. Measure-
ments on the mockup indicated that a flux of
9 x 1072 neutrons/cm?:isec will exist at the outer
surface of the fuel in the annular stress-corrosion
apparatus and that the flux at the inner surface of
the annulus will be 5 x 1012 neutrons/cm2sec,
Calculations based on these measurements indicate
an unperturbed thermal-neutron flux at this position
of 2 x 1013 neutrons/cm? sec.

Tubular Inconel creep specimens subjected to a
tensile stress of 1500 psi at 1500°F in a helium
atmosphere and irradiated for 33 days in hole HB-3
of the MTR were found to have fractured. Post-
irradiation examination has suggested that localized
high temperatures or the unknown contaminant that
was responsible for the film and the intergranular
attack on the specimen caused premature fracture,

Elevated-temperature tensile tests on monel
have shown that the existence of a ductility mini-
mum at about 1000°F is strain-rate dependent.
The ductility minimum was found at a strain rate
of 2 in./min, but it was not observable at a strain
rate of 0.002 in./min,

Seven Inconel capsules containing NaF-ZrF
base fuels that were irradiated in the MTR for
periods of six to nine weeks are being sampled.
Facilities have been put into service for irradiating
two or three capsules simultaneously in a hitherto
unused high-flux position in the MTR,

The holdup of radickrypton in charceal traps
saturated with nitrogen or with helium was investi.
gated over the temperature range -110°C to +16°C,
A theoretical expression was developed and fitted
to the results; this expression may be used for
predictions of holdup under other conditions or for
optimum design of the charcoal traps.

A second vertical in-pile logp is being prepared
for insertion in the LITR. The loop is essentially
the same as the one operated previously, but the
tip of the loop will extend more than 4‘/2 in. below
the position of maximum flux in the reactor to
obtain a greater temperature differential than that
obtained previously,

Design conditions for the gamma spectrometer
for the off-gas system of the ART were established.

A flux-monitoring system in which Inconel is
used as the monitor has been developed for use
at high temperatures, The Cr3! activity is fol-
lowed for exposure times of up to one or two

10

months, and the Co%0 in the Inconel is used for
longer exposures.

F lux measurements have been obtained from co-
balt foils that were attached to the nose coil of
MTR in-pile loop No. 3. Fast-flux profile measure-
ments were obtained for hole 19 in the ORNL
Graphite Reactor, and similar measurements were
started in hole HB-3 of the LITR.

Although theoretical considerations predict a
substantial increase in the corrosion of Inconel by
fused fluoride fuels under irradiation, this has
never been observed experimentally, The excess
oxidizing power of the irradiated fuel, which re-
sults from the imbalance of fluorine between UF
and the fission-product fluoride mixture, might be
expected to increase corrosion. The deposits of
the fission-product metals, ruthenium, niobium, and
(presumably) molybdenum appear, however, to
inhibit attack of such fuel on the container surface.

The tritium produced by the Li%(n,¢)He4 reaction
is expected to cause a large increase in the cor-
rosion of metal capsules by fuels in the NaF-KF-
LiF-UF, system. In order to prevent this, any
lithium used in MTR capsule irradiations must
contain less than 0.2% Li%, In in-pile loops, up to
1% Li% can be used without depressing the neuiron
flux unduly and without causing serious increases
in corrosion.

A study of radiation effects on B4C and related
thermal-neutron shielding materials is under way.
The first irradiations were made on uncoated hot-
pressed B,C (Norbide) and slip-cast B,C bonded
with SiC (The Carborundum Company) at low tem-
peratures (~200°C)., The hot-pressed samples
retained their physical dimensions and bulk struc-
ture with no helium evolution. The slip-cast
bonded material showed inadequate radiation sta-
bility, In order to determine whether the hot-
pressed samples will retain helium at elevated
temperatures, samples will be irradiated at tem-
peratures up to 2150°F.

9. Analytical Chemistry of Reactor Materials

Instruments are being developed for detecting
leaks of NaK into the exhaust air stream from the
NaK-to-air radiators of the ART and from radiators
installed in test stands. The instrument to be used
on the ART is capable of detecting an incremental
increase of about 0.01 ppm of alkali metals in the
air while it is crossing the bank of radiators,
whereas the instrument for monitoring the exit
gases from the test stands is to have a sensitivity
of about 10 ppm. The detection method for use
with the test-stand radiators is based on the
measurement of the hydroxyl ions which are formed
when the alkali metal oxides and hydroxides are
absorbed from a sample of the contaminated exhaust
air in an aqueous solution containing the color
indicator bromcresol green. The absorbance of
this indicator is increased by a factor of 2 by the
addition of 1 to 10 ppm of NaK. The instrument
will be designed so that an alarm will be activated
if the coulometric current required to hold the pH
of the absorber solution at a constant value ex-
ceeds a predetermined rate. A more sensitive in-
strument for use with the ART radiators is being
studied that is based on the absorption of light of
the frequency of the sodium doublet resonance
line at 5890 A, Calculations show that 0,01 ppm
of sodium in air would produce a 50% attenuation
in a beam of sodium-doublet radiation which
traversed a 50-cm optical path.

A spectrophotometric method for the determina-
tion of trace amounts of titanium in mixtures of
fluoride salts was developed. The method is
based on the yellow color of the titanium-Tiron
complex, and uranyl, zirconium, and ferric ions
are masked to prevent their interference.

The method for the determination of trace amounts
of tantalum in NaF-KF-LiF mixtures without
uranium was modified and was applied to the de-
termination of tantalum in uranium-bearing mix-
tures. The tantalum was separated from the inter-
fering uranyl ions with cupferron,

Studies were continued on the bromination method
for the determination of combined oxygen in fluo-
ride salts. Attempts to improve the recovery of
oxygen as CO, from the bromination of ZrO, by
the addition of NiF2 were unsuccessful, However,
the solid products resulting from the bromination of
the mixture of Zr02, NiF,, and graphite were large
crystalline masses rather than the finely divided
solids obtained previously from the bromination of
the mixture of ZrO,, FeF,, and graphite. Analyses
of the crystalline product are expected to provide
a basis for a better understanding of the reaction
mechanism,

A method for the determination of small amounts
of boron in fluoride salt mixtures was developed
in which the salts are dissolved, at room tempera-
ture, in a solution of AiCl;+6H,0 and 2 M HCI,

The boron remains in solution and is extracted by

PERIOD ENDING MARCH 10, 1956

ethyl ether and determined by the carminic acid
method. The extraction coefficient was found to
be approximately 0.45 + 0.02. By this method,
boron can be determined in concentrations as low
as 10 ppm.

A spectrophotometric method for the determina-
tion of bismuth as the tetraiodobismuthate(lll)
complex was applied to the determination of bis-
muth in NaF-ZrF -UF,. The effect of uranium on
the absorbance of the complex was negligible for
uranium-to-bismuth weight ratios of less than 30:1.

A method for the determination of dissolved oxy-
gen in lubricating fluids was developed that is
based on the displacement of the dissolved oxygen
by carbon dioxide. The oxygen content of a number
of lubricants ranged from a maximum of 42 ug per
milliliter of fluid to @ minimum of 15 ng/ml,

Modifications were made to the method for the
compleximetric determination of zirconium in
zirconium-base fuels. The zirconium-VYersene com-
plex was formed at pH 6, and the solution was
digested on the steam bath before tifration, Zir-
conium was determined in a total of 21 samples by
the modified procedure. The average difference
between the results and those obtained by the
gravimetric procedure was 0.6%.

A significant decrease in the time required for
the determination of microgram quantities of ele-
ments of the rare-earth group in stainless steels
and in Inconel was obtained by removing the
chromium, before electrolysis, as the volatile
chromyl chloride,

The optimum conditions for the distillation of
sodium metal from Na,O in the distillation method
for the determination of oxygen in metallic sodium
were established. The optimum temperature range
for distillation is 800 to 850°F. Distillation
periods of as much as 4 hr can be tolerated at
these temperatures without loss of Na,0. Dis-
tillations at higher temperatures confirmed earlier

observations that NQZO is volatilized at tempera-
tures above 900°F.

10. Recovery and Reprocessing of Reactor Fuel

The completion of the pilot plant for recovering
fused salt fuels is now scheduled for May 15, 1956.
Tests of the percolator type of contactor planned
for use in the fluorination vessel showed it to be
satisfactory, A freeze valve designed for use in
the pilot plant held satisfactorily against a pressure

of 20 psig in 50 freezing and melting cycles,

T
Continued studies of the fluorination process
for the recovery of uranium from fused salts have
indicated that the presence of salt impurities has
a much greater effect on the rate of UF , volatiliza-
tion than other factors, such as use of nitrogen
with the fluorine or variations of the method of
introducing the fluorine into the molten salt. A
fluorine efficiency of 70% was achieved with
>99% UF , volatilization when the impurity content
in the initial salt was kept to a minimum, The
vapor pressure of UF , over the complex UF ;:3NaF
was determined at various points in the tempera-
ture range 80 to 320°C, The data were success-
fully fitted by the linear equation

log P = 10.88 - (5.09 x 103/1) .

mm Hg

An enthalpy change of +23.2 kcal per mole of UF
was calculated for the reaction

3NaF-UF , —> 3NaF + UF,

Uranium losses in NaF beds under process con-
ditions did not exceed 0.05%, even with repeated
use of the NaF over three cycles,

PART IlIl, SHIELDING RESEARCH

11. Shielding Analysis

The code of a Monte Carlo calculation of energy
penetration and deposition resulting from transport
gamma-ray radigtion in a shield of slab geometry
was used in a parametric study of a two-region
lead-water shieid, The radiation was 1-Mev gamma
rays incident on a slab at 0 deg, 60 deg, 70 deg
32 min, and 75 deg 31 min. The first region of
the slab was composed of water 1.5 mean free
paths (mfp) thick at the initial gamma-ray energy,
and the second region was composed of lead
0.5 mfp thick. The coding was also extended to
calculate energy flux and tissue dose rate, The
results indicate that lead is more effective in
stratified slab shields when it is placed behind
a good scatterer, such as water.

12. Reactor Shield Design

The heating in the lead and alkylbenzene shield
of a 300-Mw circulating-fuel reflector-moderated
was calculated for the following com-

(1) primary gamma rays originating in
or near the reactor core, (2) fission-product gamma

reactor
ponents:

12

rays from the heat exchanger, and (3) thermal
neutron captures in the lead and borated (2% boron)
alkylbenzene. = The third component was sub-
divided to take into account the secondary gamma
lead, hydrogen, and boron capture.
particles resulting from thermal-neutron

rays from
Alpha
captures by boron were also considered.

The gamma-ray and fast-neutron dose rates at
a distance of 50 ft from the ART were calculated
as a function of the thickness of the lead shield.
Contributions to the total gamma-ray dose rate by
primary gamma rays from the core, secondary gamma
rays originating in the shield, and heat exchanger
gamma rays were determined,

13. Lid Tank Shielding Facility

Analysis of the dynamic source tests on mockups
of a reflector-moderated reactor and shield was
completed, In the analysis the experimental dose
rate resulting from fission-product gamma rays
emitted in the heat exchanger mockup was sepa-
rated out and compared with dose rates calculated
by two different methods. In one calculation it
was assumed that all the fission-product gamma
rays were of a single energy (2.7 Mev/photon).
The other calculation was based on the spectrum
of gamma rays from the fuel belt, previously meas-
ured at the LLTSF. The two calculated values
agreed, but they differed by about 30% from the
measured value, |t was found that the difference
could be attributed to the dose buildup factor for
water having been used in the calculation (that is,
the buildup factor was chosen as if the lead were
an equivalent thickness, in mean free paths, of
water), Substitution of a new buildup factor for
the total mean free paths (lead and water), based
on Monte Carlo studies of laminated shields, re-
sulted in agreement between the measured and
calculated dose rates,

14. Bulk Shielding Facility

A mockup experiment at the ORNL Graphite
Reactor thermal column was initiated to measure
the gamma rays streaming through the NaK pipes
of the ART,

The measurements of the decay rates of fission-
product gamma rays as a function of time after
fission were extended to include six energy groups
between 0.28 and 5 Mev for times up to 1600 sec
after fission.
Part |

REACTOR THEORY, COMPONENT DEVELOPMENT,

AND CONSTRUCTION
1. REFLECTOR-MODERATED REACTOR

W. F. Boudreau

A. P. Fraas

H. W. Savage

Aircraft Reactor Engineering Division

A. M, Perry

Electronuclear Research 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 in Building 7503 was
continued. Package 1 construction work on the
building additions, building alterations, and cell
installation was on schedule and at the 50% com-
pletion point. The principal work accomplished
during the quarter included completion of the
building-addition structure, completion of the spec-
trometer room and tunnel, erection of the 30-ton
crane, modification of the 10-ton crane, con-
struction of the blower house and switch house,
placement of reinforced concrete for the stack
base and for the walls of the special-equipment
room, and partial erection of the cell tanks and
surrounding structure, Some of this work can be
seen in Fig. 1.1, which is a view from a location
southwest of the altered 7503 building, The
dark-sided portion of the main building is the
addition which will house the reactor cell, heat-
dump systems, and spectrometer room and tunnel,
In the foreground, from left to right, can be seen
the switch house, blower house, and stack foun-
dation, A closeup view of the blower-house
interior, prior to erection of the exterior walls,
is shown in Fig. 1.2. In the left foreground can
be seen the walls of the ramp access way from
an elevation of 840 ft to the radiator pit at an
elevation of 820 ft in the main building. In the
center foreground are the forms and the reinforcing
steel for the upstream end of the main air duct.
An interior view across the building addition
looking toward the southeast corner of the building
is shown in Fig. 1.3. In the bottom center can
be seen the spectrometer tubes, which extend to
the spectrometer tunnel beneath the low bay and
which will penetrate the cell water tank. At the

bottom right can be seen portions of the special-
equipment-room walls, An interior view across
the building addition [ooking toward the southwest
comer is shown in Fig. 1.4. The lower sections
of the cell tanks can be seen in the foreground,
and the opening into the blower house for the main
heat-dump air duct is shown in the center left,
The duct will enter the building and pass hori-
zontally across the corner to the stack, The main
and auxiliary radiatoers will be mounted in the
duct on a diagonal line approximately as repre-
sented by the left edge of the picture. The bottom
sections of the cell tanks are shown in Fig. 1.5.
The inner tank is the bottom hemispherical head
of the pressure vessel, and the outer tank is the
water tank. The incomplete cylindrical portion of
the pressure vessel is shown in Fig. 1.6. At the
top center can be seen the 4-in.-thick panel which
contains sleeves for the NaK and the off-gas
piping. In the left and right foreground are the
4-in ~thick panels which will receive the junction
panels for auxiliary-services lines, heater leads,
and the instrumentation and control lines. The
large 2-in. panel at the upper left contains the
small nozzles for beam penetrations into the spec-
trometer tubes. In the bottom of the tank, under
the scaffolding, can be seen the structural framing
for the floor, equipment support, and reactor
support,

Bids were received for package 2 construction
work, and a contract for $58,400 was awarded on
January 20, 1956, to the Rentenbach Engineering
Company, Knoxville, Tennessee. This unit of
work consists in the installation of the diese!
generators and facility, the electrical control
centers, and the spectrometer-room electrical and
air-conditioning equipment. At the end of the
quarter, this work was in the material procurement
stage, with no site work having been started.
Approximate completion dates for package 2 are

15
91

 

Fig. 1.1. View of Altered 7503 Building from Southwest.

 

Photograph taken Feb. 17, 1956.

UNCLASSIFIED
PHOTO 16503

 

LA0dIY SSTYO0Yd LDIF0¥d dNY
Ll

Fig. 1.2.

Interior of Blower House Prior to Erection of Exterior Walls, Photograph taken Feb, 10, 1956.

"UMCL ASSIFIED

 

9661 ‘0L HOYVYW ONIAGN3T Q01334
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
PHOTO 16456

 

o
o "
oy

e

 

Fig. 1.3. Interior View Across Building Addition Looking Toward Southeast. Photograph taken Feb.
10, 1956. |

18
PERIOD ENDING MARCH 10, 1956

. UNCLASSIFIED
'PHOTO 16455

 

Fig. 1.4, Interior View Across Building Addition Looking Toward Southwest. Photograph taken Feb.
10, 1956.

19
ANP PROJECT PROGRESS REPORT

 

 

 

 

£$991 0LOHd
AF4ISSYIONN

9661 .m 'qa4 uajp} _._aEmoho_._n_ ‘SHuUD) ||oD) j0 suoldeg wojjog

S

l

nm_m

 

 

 

 

20
 

PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED §
PHOTO 16506

Fig. 1.6. Incomplete Cylindrical Portion of Pressure VYessel, Photograph taken Feb. 17, 1956.

June 24, 1956, and July 15, 1956, respectively,
for installation of contractor-supplied and Govern-
ment-supplied items,

Design was completed and negotiations for a
proposal are in progress with the package 1
contractor, V. L. Nicholson Company, for the
installation of package A work, which consists
in auxiliary-services piping. Two proposals are
being requested, one based on the package 1
completion schedule, which currently is June 8,
1956, and the other based on the contractor’'s
proposed schedule in the event the first schedule
presents unreasonable situations, The proposals
are to be submitted March 7, 1956.

Design work on package 3, which includes the
process equipment, the process piping, etc., is
currently under way, and it is expected to be
complete early in June 1956. Package 3 work will
be performed by ORNL forces,

ART DESIGN
A. P. Fraas

Aircraft Reactor Engineering Division

H. C. Gray
Pratt & Whitney Aircraft

Detailed layouts have been completed on all
major subassemblies of the ART., Detailing of
the north head and the main heat exchangers has
been completed, and the detailing of the control
rod, reflector-moderator, island, and pressure-shell
subassemblies and the reactor support structure
is under way. Typical weld joints are being
fabricated to determine weld shrinkage allowances
(see Sec. 6, ‘‘Metallurgy and Ceramics'') for the
final weld designs. Detailed layouts of the reactor
shield and the piping inside the reactor cell are
being prepared.

21
ANP PROJECT PROGRESS REPORT

Stress Analyses
R. V. Meghreblian

Aircraft Reactor Engineering Division

An extensive series of calculations has been
under way for several months in order to determine
the stress distribution in major structural com-
ponents at the top of the reactor assembly, referred
to here as the ‘‘north head.’”” The design loads
being used in these calculations are based on
full-power {or ‘‘design-point'’) operation of the
reactor. Two components have been studied in
detail: the double-ring structure, which transmits
the up-load from the reflector-moderator to the
north head (and thence to the pressure shell), and
the composite double-deck structure of the north
head. The proposed design of these members is
based on these calculations.

It is believed that these analyses provide a
reasonably accurate picture of the gross features
of the stress distribution and, therefore, are
adequate for determining basic dimensions and
proportions of structural members. However, the
structural configuration of the north head is so
complex that analytical studies must be based on
relatively simple geometric models which cannot
include all the details of the actual design. Thus
the finer details of the stress distribution must
be determined by other techniques.
experimental

A program of
stress analyses was prepared for
this purpose, and the work is presently under way
at the University of Tennessee. The program
includes stress studies of the two north-head
components mentioned above, as well as studies
of the structural integrity of the pump-barrel and
NaK pipe attachments to the pressure shell and
full-scale tests on portions of the NaK piping
outside the pressure cell. The NaK circuit
selected for this experiment is that portion of the
system which serves the NaK pump-and-radiator
assembly on the reactor side of the main air duct
(Fig. 1.7). This circuit is the shortest portion of
the entire piping system, and it is believed to be
the least flexible; thus stresses from thermal
expansion of the piping are expected to be most
severe in this circuit. At the conclusion of the
thermal-stress study, the facility will be used to
determine the vibrational characteristics of the
complete NaK pump-radiator-piping assembly.

The program of detailed stress analysis of the
ART is being supplemented by a parallel program

22

 

ORNL-LR-DWG 13200

aft s¥in,

 

 

 

7 ft Qin,

 

 

      

A

e |
RADIATORS - /,

e

 

 

 

 

 

 

 

 

 

 

Fig. 1.7.
Assembly on Reactor Side of Main Air Duct.

Nak Circuit for Radiator-and-Pump

of material tests on the high-temperature character-
istics of Inconel and beryllium. The purpose of
these tests is to establish a better basis than
that presently available for the design analysis
of structural components that are exposed to
cyclic load conditions. A research contract has
been awarded to the University of Alabama for a
program of tests on the high-temperature strain
and thermal-cycling properties of Inconel. Con-
tracts are being negotiated with the Universities
of Michigan and Syracuse for a series of tests to
determine the relaxation and creep properties of
Inconel in the temperature range 1200 to 1600°F.
The OSyracuse program is to

include a joint
analytical-experimental study of the creep-buckling
of thin hemispherical shells. The data to be
obtained through these contracts are needed for
the design analysis of the core and reflector
shells, which will contain the circulating-fuel
system. An analytical and experimental study of
the reflector shells is being undertaken by the
Lewis Laboratory of the National Advisory Com-
mittee for Aeronautics.

Sodium System
R. I. Gray

Aircraft Reactor Engineering Division

Recent calculations of the effects of the non-
uniform power distribution in the reactor core
indicate that the heat flux through the core shells
of the ART will be higher than previously ex-
pected. The combined effects of heat transfer and
radiation heating yield a total estimated heat load
to the sodium system of 6.2 Mw. The increase
in the heat load required either an increase in
the design sodium flow rate or an increase in the
design temperature rise of the sodium system,
Since an increased sodium flow rate would result
in excessive pressure drops in the system, the
sodium temperature rise was increased from 150
to 200°F, and the sodium-system operating temper-
ature range became 1050 to 1250°F. Minor changes
in the dimensions of the reflector-moderator cooling
circuit were also made to stay within the pressure-
drop limitation of 30 psi and the maximum sodium
temperature limitation of 1250°F. These changes
included the removal of the staple-spacer support
lands of the island and the reflector annuli, the
increase of the island and the reflector-annuli
thicknesses to 0.188 in., the shortening of the
reflector-return-annulus spiral spacers to 3.1 ft
(6, = 45 deg), and the increase of the reflector-
retum-annulus thickness to 0.125 in. The flow
of sodium in the island and reflector cooling holes
is to be adjusted by variation of the entrance
conditions of each tube row., In this manner a
balanced flow condition can be obtained with a
minimum pressure drop.

Detailed relaxation calculations have determined
an island cooling-hole distribution which limits
the maximum beryllium temperature difference to
50°F, while maintaining @ minimum number of
cooling holes. This design includes a total of
120 holes. The previously determined analog
solution for the reflector cooling-hole distribution

PERIOD ENDING MARCH 10, 1956

specified 288 holes, and that solution remains
unchanged.

The increased sodium-system heat load alse
necessitated modification of the auxiliary radi-
ators. The four radiators are to be mounted to
form a V in order to obtain low NaK and air
pressure drops.

Reactor Shield
W. B. Cottrell

Aircraft Reactor Engineering Division

The thickness of the lead portion of the ART
reactor shield was reduced from 7 to 4.3 in, to
obtain space for increased neutron shielding
around the pumps. This decrease in lead thickness
will increase the gamma-ray dose rate by a factor
of 10 (from 7/8 r/hr at 50 ft to ~ 9 r/hr), but the
dose rate will still be substantially less than that
proposed by most of the airframe companies for
nuclear aircraft. Collars of gamma-ray shielding
material are being designed to relieve the problem
of gamma-ray streaming through the various ducts
in the lead shield.

Reactor Study Models
W. L. Scott

Aircraft Reactor Engineering Division

A detailed one-sixth-scale model of the ART is
being fabricated to assist in the layout of piping
and in the study of fabrication and installation
procedures. Tentative mockups of the oil, helium,
electrical, and instrumentation lines are now being
made. All the basic cell structure, the reactor-
and-shield support structure, and the NaK lines
have been installed.

The existing one-half-scale plastic model of the
north head is being modified to reflect the current
design. This model is to be used for the determi-
nation of the locations of the pressure taps
required for flow studies and for layout design of
the shielding of the north head.

ART COMPONENT DEVELOPMENT,
PROCUREMENT, AND TESTING
Water Test of Aluminum Mockup of North Head
| D. R. Ward

Aircraft Reactor Engineering Division

The fabrication and assembly of the fuel-system
components of the aluminum north-head mockups
are nearly completed, and tests of the fuel system

23
ANP PROJECT PROGRESS REPORT

The sodium-te-NaK heat
exchanger mockups and other parts of the sodium
system will be fabricated while the fuel-system
tests are being performed. No unexpected problems

are to begin soon.

have yet been encountered in the fabrication and
assembly of this mockup. Warping of aluminum
parts during welding has somewhat reduced the
precision of the assembly, and similar warping
may be encountered in welding the Inconel reactor
parts.

Fuel-pump performance,
studied under both normal and unbalanced con-
ditions. Several view ports have been provided
for observation of fluid ingassing and degassing.
The interchange of liquid between fuel pumps and
the fuel swirl chamber will be studied thoroughly.
Special instrumentation is being provided for de-
termining the level to which the liquid will rise
around a pump shaft when one pump is stopped
while the other pump is running.

with water, will be

Core Flow Studies

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

Studies of the flow in a full-scale plastic model
of the ART core were continued through the use
of the techniques described previously.! Several
guide-vane configurations that had been previously
tested in the full-scale aluminum core model were
retested in the plastic model, and photegraphic,
conductivity, and pressure-drop data were obtained
for each inlet configuration. Attempts were made
to alter the flow patterns obtained, by adding baffle
plates and turbulators to the trailing edges of
the vanes, but flow-reversal regions still continued
to exist in the core annulus gbove the equator,

An inlet-guide-vane system designed on the
basis of one-quarter-scale-model axial-flow data?
provided the lowest core-entrance pressure drop
to date, but it generated a flow pattern similar to
that obtained with the axial-flow type of header
that had no means of auxiliary flow guidance.
Further modification and testing of this guide-vane
system by means of systematically attaching and
relocating a conical baffle plate on the trailing

 

]G. D. Whitman, W. J. Stelzman, and W. T. Furgerson,
ANP Quar. Prog, Rep. Dec. 10, 1955, ORNL-2012; p 23.

2k, E. Lynch et al.,, ANP Quar. Prog, Rep. Dec. 10,
1955, ORNL-2012, p 171.

24

edges of the guide vanes yielded the arrangement
shown in Fig. 1.8, which generated a core flow
pattern containing no flow reversal throughout the
core annulus and, ot the same time, provided good
surface scrubbing, good midstream mixing, and
fairly high velocities throughout the upper half
of the core. Below the equator the velocities were
high, but the fluid adjacent to the inner and outer
sutfaces tended to hug these surfaces as it
approached the core outlet; thus the movement
of the fluid from these surfaces was slight, and,
in general, the midstream mixing was poor. Further
modifications to this design are being made so
that the flow in the lower half of the core will
be improved.

Engineering Test Unit

M. Bender W. C. Tunnell
G. W, Peach G. D. Whitman

Aircraft Reactor Engineering Division

Flow diagrams and instrumentation lists have
been prepared for the Engineering Test Unit
(ETU), which is a nonnuclear mockup of the ART,
It will differ from the ART only in heat-input and
heat-removal arrangements. For the ETU, two
gas furnaces will supply 1.5 Mw of heat to the
fuel system through two adjacent NaK systems.
The remaining two NaK systems will operate
isothermally., The 1.5 Mw of heat input will be
removed from the fuel through two sodium-to-NaK
reflector-coolant systems. The ART auxiliary
systems, such as the fuel drain tank, the off-gas
system, the fuel recovery system, the fuel
sampling system, the drain-tank cooling system,
and the various gas systems, will not be mocked
up as part of the ETU.

Work has been started on a detailed study of
the test program for the ETU, and a manual
covering all phases of the program is to be issued.
The building modifications required to accommo-
date the ETU are being made. Contracts have
been let for the core shells, the heat exchangers,
and the radigtors. The design of the boron-
containing tiles was modified to conform to lower
temperature limits and new radiation-damage infor-
mation, and therefore procurement of the boron-
containing materials has been delayed.

Present design and shop schedules indicate that
the ETU will be ready for operation in January
1957.
4

 

PERIOD ENDING MARCH 10, 1956

PHOTO 25581

Fig. 1.8, Guide<Vane and Baffle-Plate Arrangement for Distribution of Flow in Plastic Model of ART

Core,

The north head for the ETU has been designed,
and a fabrication sequence for its assembly has
been prepared. A work estimate shows that 27
calendar weeks will be required for its fabrication
and assembly after the materials are received. A
similar study of the moderator assembly is now
being made.

Methods for inspection during assembly are
being studied. It has been established that the
dimensional tolerances specified will require that
the assembly operations be carried out in a
temperature-controlled area. Dimensional in-
spection will be performed with templates and
gages designed especially for this application.

Procurement of Special Reactor Materials
and Components

W. F, Boudreau
Aircraft Reactor Engineering Division

Considerable progress was made during the
quarter in the establishment of contracts and
facilities for the procurement of some major ART
and ETU components and facility items. The
construction of components, however, did not
proceed as rapidly as had been forecast because
of a fack of firm design information. The current
status of some of the major items is presenied
below.

25
ANP PROJECT PROGRESS REFPORT

Beryllivmn. — One beryllium hemisphere pressing
was mgggfluczessfully in the new facility of
The Brush Beryllium Co., and machining is to
proceed as soon as firm design information is
received. The pressing of the beryllium for the
islands was scheduled to be completed by the end
of the quarter.

Shell Fabrication. — Modification of the Hydro-
spin machine to be used for fabrication of the core
shells was delayed during the quarter by a combi-
nation of contract negotiation difficulties and a
fack of firm information as to the size of the
largest shells to be produced. The provision of
facilities was advancing at a satisfactory rate
at the end of the quarter, however.

CX900 Inconel. — The first lot of 5000 ft of
3 5-in.-0D  CX-900 Inconel tubing was received
from the Superior Tube Co. near the end of the
Although
considerable progress was made in processing the
remaining quantity of Inconel to its final form,
this was also delayed by lack of firm design
information,

quarter, and inspection was started.

Main Heat Exchangers and Radiators, — An order
was placed with the York Corp. for the fabrication
of all the radiators and one-half the heat ex-
changers required for the ETU and the ART. An
order was placed with Black, Sivalls & Bryson,
Ine., for the remaining main heat exchangers., The
activities of both companies during the quarter
were limited chiefly to the preparation of facilities,
the training of welders, and general planning,
because of the lack of final design information.

Dump Valve. — The second prototype dump valve
was delivered on February 12, 1956. Reasonably
firm information has now been developed which
will permit the production of additional prototype
valves ot o greatly accelerated pace during the
next two quarters.

Main Blowers and Louver Dampers. — The order
for the four main radiator blowers progressed on
schedule during the quarter. Difficulties have
been encountered in obtaining a vendor for the
louver dampers, primarily because the service
requirements are so unusual, Negotiations were
in progress with the American Foundry & Furnace
Co. at the end of the quarter.

26

Inspection of Materials and Fabricated Parts

A. Taboada
Metallurgy Division

R. L. Heestand
Pratt & Whitney Aircraft

An inspection group has been established for
quality control of reactor and test materials and
components., Weld-joint designs were established
that are to be used as a basis for all criticadl
welding, and a system for identifying and classi-
fying critical welds was formulated. All drawings
are to be reviewed by the inspection group for the
detection of undesirable weldments,

Welding procedures and inspection techniques
were established for all anticipated types of
welds. Since welders who are to work on critical
welds must pass the qualification tests set up
in the weld specifications, training groups were
set up to qualify welders. The techniques es-
tablished for weld inspection include visual, dye-
and radiographic examination of all
critical welds.

penetrant,
Specifications for the inspection
of all critical materials are being prepared, and
amangements were made for all purchase orders
for critical materials to be reviewed by the in-
spection group.

The welds inspected during the quarter totaled
2005, with 14% rejection. A total of 5655 ft of
tubing of various sizes was inspected, and 491 ft,
or 9%, was rejected. Defective areas were found
in some of the 36 pieces of various sizes and
lengths of pipe and the 99 pieces of plate that
were inspected, but the defects were ground or
filed out and the pipes and plates were accepted.
Dye-penetrant and reflectoscope techniques were
used for the inspection of 86 pieces of bar stock,
and five pieces were rejected, Other items that
were inspected included eight Inconel dished
heads, 329 fuel sampling assemblies, York Corpo-
ration radiator No. 3, and ORNL radiator No. 3.

ART INSTRUMENTATION AND CONTROL
R. G. Affel

Instrumentation and Controls Division

J. M. Eastman

Bendix Products Division

Information obtained with the simulator has

shown that in the event _of an ART fuel-pump
failure at design power the maximum fuel core-
outlet temperature can be limited to 1700°F by
inserting the control rod at the rate of ]/8%
(Ak/k)/sec. The magnetic, emergency, 1% rod
drop has therefore been abandoned because of the
thermal-shock conditions that would develop at
the core outlet in the event of a spurious rod drop
at design power, and an emergency rod insert rate
of 1/8% (Ak/k)/sec has accordingly been adopted.

Reactor performance with U233 fuel was also
investigated on the simulator, The lower pro-
portion of delayed neutrons, in comparison with
U235 fyel, made the core-outlet fuel temperature
less sensitive to a fuel-pump failure, |t was
found that the core-outlet fuel temperature would
not exceed 1700°F, even with no corrective
control-rod action. For this test, the over-all
temperature coefficient for the ART with U235
fuel, —2.5 x 10-3 (Ak/k)/°F, was assumed as
the temperature coefficient. The U233 fuel could
be expected to give a somewhat different temper-
ature coefficient, The micromicroammeter amplifier
and the servo amplifier used for the ARE have
been obtained and will be connected to control
the simulator for studies of the ART flux servo.

The heat-exchanger-inlet minimum-NaK-temper-
ature limit is being studied to determine control
and louver-actuator performance requirements.
initial tests indicated that a louver closing time
of 2 sec was probably faster than necessary.
However, the method of transport-lag simulation
used for the tests was considered to be inadequate.
Further work will be done with improved simulation
of this lag, and control compenents will be used
in the hookup.

The louvers are being supplied with pneumatic
actuators to facilitate testing, Current plans call
for two speeds of actuation of the louvers installed
in the ART — one for normal load changes and the
other for fast louver movement to prevent the NaK
going into the reactor from dropping below the fuel
freezing temperature during unusual transients
such as an emergency rod insertion, The exact
mode of control to be used for NaK temperature
limiting and the means of correlating manual and
automatic louver action are to be determined from
the simulator study.

Tests indicate that, if the surface temperature
of the stagnant sodium in the control-rod thimble
is held to below 300°F, sodium deposition on the
cool walls above the surface will be very small,

PERIOD ENDING MARCH 710, 1956

Accordingly the rod drive mechanism is being
designed without a bellows to seal it from the
helium atmosphere cbove the sodium. Two rod
drive motors and drive speeds are being used;
the faster one will provide 1/8% (Ak/k)/sec. The
motors are geared down through worm reductions,
and they drive the rod through differential gearing
and a pinion and rack, A conventional shaft seal
is used on the pinion shaft to retain oil in the
gear casing. The sodium in the upper part of
the rod thimble is to be cooled with a water jacket.
The rod drive mechanism is housed above the
shield and the reactor support structure. The
layout is essentially complete,

The ART control technique has been revised to
accommodate the changed emergency-rod-insertion
method.  The information block diagrams and
elementary wiring diagrams have been further
revised to place more dependence on the operator
and less on instrumentation to prevent dangerous
misoperation of the reactor. The original complex
chains of instrument permissives were considered
to be too sensitive to the reliability of instrument
and control components. It is still planned to
provide automatic corrective action whenever an
emergency requires this action in too short a time
to permit operator deliberation and to warn the
operator by annunciators whenever an emergency
requires his attention and corrective action. Inter-
locks to prevent misoperation will be reduced. I
is expected that when the operator is changing
power level he will have foreknowledge of the
proper operating sequence and the response to be
expected, and he will have time to judge the
validity of instrument information and act ac-
cordingly. It is planned to use the simulator to
familiarize the operator with the reactor and
system response characteristics.

Preliminary instrument specifications for the
ART helium and nitrogen systems were issued to
provide engineering forces with space requirements
for these devices. Preliminary installation flow-
sheets for both the ART and the ETU are now
available, and detailed design work is proceeding.

A number of methods for detecting NaK leaks
into the high-velocity air streams from the ART
radiators have been analyzed, and experimental
work is being done on two of these methods. This
work is reported in Sec. 9, ‘‘Analytical Chemistry
of Reactor Materials.”

Tests of solenocid-valve-seat materials suitable
for use in systems involving liquid-metal vapor

27
ANP PROJECT PROGRESS REPORT

have been made (see Sec. 5, ‘“‘Corrosion Re-
search’), as well as tests of suitable valves to
meet various pressure and leaktightness require-
ments. All-welded solenoid valves, which were
found to be leaktight by a spectrometer check, are
being tested. A survey of a number of commercial
vendors of solencid valves disclosed that the
special requirements dictated by the disaster con-
ditions postulated for the ART impose severe
solencid-valve problems. It is hoped, however,
that a suitable design will be developed. Tests
of a number of commercial solencid valves have
demonstrated that only a soft seat material will
meet leaktightness requirements, Metal seats are
completely unreliable, in that their leakage rates
are not reproducible under repeated cycling.

Approximately 12 pressure transmitters designed
for high-temperature use have been tested, and
performance data have been compiled. Two units
of a new design have also been received and
are being tested. Initial test data indicate that
the newer units have performance characteristics
somewhat better than those of the currently used
devices. A number of other newly designed trans-
mitters are on order and will be tested. Trans-
mitters used on test loop instrumentation have
been analyzed in an effort to determine the point
of failure.
as vet,

An initial test of a high-temperature ORNL-
designed turbine-type flowmeter was attempted.
It was found that expansion of the magnetic core
of the flowmeter and expansion of the circulated
fluid forced one of the bearing inserts out of the
rotor and thus jammed the device. The flowmeter
has been redesigned and will be retested.
Improved methods of measuring the magnetic field
intensity of magnets used for electromagnetic flow-
meters have been investigated, and a new
measuring system was ordered. It is felt that
with an accurate measuring system it will be
possible to increase the accuracy of flow measure-
ments made on liquid metals with the wuse of
electromagnetic flowmeters.

The first of a series of test rigs for measuring
the fluid level in the reactor fuel expansion tank
was placed in operation and then shut down in
approximately 2 hr because of plugging of a z-in.
line with ZrF, vapor deposits. A second fest,
for which @ %-in. line was used, was terminated
in approximately 10 hr by plugging. The units

No conclusive information is available

28

have been redesigned and further testing is under
way. A contract has been placed with the General
Electric Company to supply a trial quantity of
high-temperature liquid-level-sensing elements for
liquid-metal systems of the type used in the SIR
These units may be applicable to the
ART liquid-metal level measurements.

An investigation of the different types of thermo-

system.

couples that might be used to measure reactor
internal temperatures has indicated the swaged,
magnesium-oxide-insulated, Calrod type of thermo-
couple to be the most suitable for ART use,
Quantities of this type of thermocouple have been
ordered for use in an evaluation program, and a
contract has been given to a vendor to supply
these thermocouples for the ETU. Thermocouples
that use both Chromel-Alumel and platinum—
platinum-rhodium wires will be installed in the
ETU in an effort to obtain the highest accuracies
possible. Accuracy requirements on the thermo-
couples have had to be relaxed from the initial
specifications because of the difficulty in cali-
brating the thermocouples in their final locations,
A preinstallation calibration will be performed,
but it is not yet known how the installation will
offect the accuracy of the temperature measure-
ments.

A number of commercially available standard
instrument components have been examined and
tested for various functions, Among these items
are pressure regulators, pressure controllers,
temperature controllers, instrument check valves,
and electrical pressure transmitters. Electrical
pressure transmitters have been explored on a
specification basis, and orders have been placed
for a number of trial units.

REACTOR PHYSICS
A, M. Perry

Electronuclear Research Division

NaK Activation in the Fuel-to-NaK Heat
Exchanger of the ART

H. W. Bertini

Aircraft Reactor Engineering Division

Three separate calculations have been made of
the activity produced in the NaK in the fuel-to-NaK
heat exchangers of the ART by core neutrons. One
of these calculations was made by the ORNL Lid
Tank Shielding Facility (LTSF) group,® one by a
group at the Nuclear Development Corporation of
America,4 and one by the author, The configura-
tions used for these calculations are described in
Table 1.1. An attempt to normalize the three cal-
culations to the present ART configuration is
described here. The configuration used for the
normalization is also described in Table 1.1, It
is expected that the final ART configuration will
be essentially the same as that used for the
normalization,

The results of the normalization of the data are
presented in Table 1.2. No corrections were made
to the calculations to account for differences in
the size of the core. Corrections for differences
in the beryllium thickness were made by taking a
relaxation length of 5.7 cm for the decrease of
NaK activity vs the increase in beryllium reflector
thickness,> Corrections for variations of the boron
content in the first boron curtain were made by
using a plot of sodium activity vs boron layer

 

3J. B. Dee et al.,, ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 209, Table 12,1.

4w, Frank, Nuclear Development Corporation of
America (NDA), private communications to H. Bertini.

51, B. Dee, op. cit., p 204, Fig. 12.1.

PERIOD ENDING MARCH 10, 1956

thickness® and an extrapolation for the thicknesses
expressed as 0.587 and 0.681 g of B'0 per square
centimeter, No corrections were made for the
variations in heat exchanger thicknesses or for the
variations in the second boron curtain thickness.
The activity was assumed to be directly propor-
tional to the mass of the sodium (in the NaK) and
to the design power. An additional correction was
made to the LTSF data to account for the incorrect
assumption that, at the time the LTSF results were
reported, the power of the LTSF source plate was
3.6 w. A recent measurement’ indicates that the
source-plate power was more nearly 2.1 w.

The delayed-neutron activity calculations by
NDA are nof included in Table 1.2, because there
is no simple correction factor that will take into
account the large heat exchanger thickness used.
Only the L TSF-source-power and sodium-mass
corrections were applied to the LLTSF delayed-
neutron results, and only the sodium-mass correc-
tion was applied to the author’s delayed-neutron
results,

 

6G. T. Chapman, J. B. Dee, and H. C. Woodsum, ANP
Quar. Prog. Rep. Jume 10, 1955, ORNL-1896, p 199,
Fig. 12.14

7D. Otis, private communication to H. Bertini.

TABLE 1.1. CONFIGURATIONS USED FOR CALCULATIONS OF THE ACTIVITY OF
THE NoK IN THE ART FUEL-TO-NaK HEAT EXCHANGERS

 

Configuration

 

 

For LTSF For NDA For For
or ! Author’s Normalization
Calculation Calculation .
Calculation Calculation
Core radius, in. 10.5 13.19 10.63 10.5
Beryllium-reflector thickness, in. 1 11.81 11.81 1
First boron curtain, grams of g1l per square 0.325* 0.587 0.681 0.216**
centimeter
Heat exchanger thickness, in. _ 2.1 6.5 1.97 3.25
Second boron curtain, grams of g0 per square 0.325 0.162 0.341 0.216**
centimeter
Sodium content of NaK in heat exchanger, kg 42 200.67 13.4 35
Reactor design power, Mw 60 60 60

 

*A ‘,lf'a-in. layer of pure B'0 has 0.731 g of pl0 per square centimeter.

el 3/8-in. layer with a natural-boron density of 1.3 g/cm3.

29
ANP PROJECT PROGRESS REPORT

TABLE 1,2. RESULTS OF NORMALIZATION TO THE ART CONFIGURATION OF CALCULATIONS OF THE
ACTIVITY OF THE NaK IN THE FUEL.TO-NaK RADIATORS OF CIRCULATING-FUEL REACTORS

 

LTSF Calculation

NDA Calculation Author’s Calculation

 

Correction Factors Used

Reflector thickness

First boron curtain 1.45
Sodium mass 0.833
LTSF source power 1.71

NDA to ART normalization

factor

1.45 1.45
2.5 2.9
2.61

426 x 107 kg of sodium
per core neufron per

curie per activation

Previous and Normalized Values

Previous values of NaK 470 curies
activity

Normalized NaK activity 970 curies
from core neutrons

Normalized NaK activity 87 curies
from delayed neutrons

Total normalized NakK 1057 curies

activity

0.3 x 'Il')_7 activations 55 curies
per core neutron per
kg of sodium (in NaK)
463 curies 600 curies
50 curies
650 curies

 

Activity of Mass-Transferred Material
in the ART Radiators

A, M, Perry

Electronuclear Research Division

A potentially troublesome aspect of the wrap-
around heat exchanger design of the ART, when
applied to an aircraft power plant, is the degree of
radioactivity that will be induced in the secondary
heat-transfer circuit, Activation of the primary
coolant, NaK, has been extensively discussed
(see above). Estimates of Na?4 activity in the
NaK have varied, depending upon the particular
reactor design being studied, from a few hundred
to o few thousand curies. While the activity would
not be sufficient, in general, to affect shield
design appreciably, it would certainly interfere
with routine maintenance of the aircraft engines,
At least two solutions have been proposed for
dealing with this problem. One solution entails
the installation of an intermediate heat-transfer
circuit to isolate the coolant for the fuel circuit
from the engine radiator circuit; the other consists

30

in draining the radioactive NaK from the radiator
circuit and replacing it with fresh NaK before the
aircraft is approached for maintenance. While the
latter solution is much to be preferred from the
standpoint of aircraft weight and performance, there
still remains the possibility that constituents of
Inconel, which have become activated in the heat
exchanger and carried in the fluid stream to the
radiators, will be deposited on the radiator tube
or header walls and will create a residual radiation
field too high to permit unshielded access to the
engines. An estimate of the activity of the mass-
transferred material that will be deposited in the
radiators of the ART has been made, and extrapo-
lations to higher power reactors would ke reasonably
straightforward,

Specific Activity of Mass-Transferred Material, =
It is assumed, for the present, that all materials
in the heat exchanger region are activated primarily
by thermal neutfrons. Activation rates are then
proportional to the thermal-neutron absorption
cross sections of the several materials present,
The Na?4 activity in the NaK is known, and, since
it serves as a normalizing factor, the thermal-
neutron flux need not be known,

{f a material is being activated in a constant
neutron flux, the saturated decay rate is equal to
the rate of activation, Thus the saturated gamma-
ray activity per gram of an element resulting from
activation of an isotope of the element is given by

A
0
1, = - S0l (in gamma rays per gram-second),
where
A, = Avogadro’s number, 6,023 x 1023,

<

A = atomic weight of the element,

S; = abundance of isotope j in the element,

o; = capture cross section of isotope 7,

1’]- = gamma rays per disintegration of the

product isotope,
activating flux,

¢

In particular, the sodium activity is given by

A
0
Ive = BTy (0.49 barns) (2 gamma rays

per disintegration) (¢} ,

and, then, using the sodium activity as the normal-
izing factor,

S.a./.
177
- @)

 

oJ

For the ART the saturated sodium activity is 1000
curies and the sodium mass is 35,000 g, Therefore
I, = 0.029 curies/g, and

Si94 5

 

l,; = 0.66

After operation of the reactor at constant power
for a time ¢, the activity of isotope j is less than
I . by the factor (1 - e"t/Tf'), where 7. is the
mean life of the product isotope. The activity of
isotope j (or, rather, of the product isotope re-
sulting from neutron capture by isotope j) after
operation for a time T is therefore

PERIOD ENDING MARCH 10, 1956

T —-t/T. =(T=t)/T;
- j i
A]. = fo mW Io].('l - e e dt
) =T/T. =T/,
- N . I
= mW, Ioj [Tj (1 —e Y- Te ]
T,
] ~T/7. ~-T/T.
MW L iy - 7
fiMWkIOj. [T (1 ~e e } .
where
m = rate of mass transfer, grams per unit time,
W, = weight fraction of the element in the mass-
transferred material,
M = mT = the total amount of mass-transferred

material.

Spectrographic analyses of Inconel and of mass-
transferred material taken from the cold leg of a
sodium-Inconel forced-circulation loop are shown

in Table 1.3.

The contributions of each of the materials to the
activity in the ART radiators, after 1000 hr of
operation, are shown in Table 1.4, Only those
isotopes are listed that lead to gamma-emitting
isotopes. Isotopes that are stable, that emit beta
particles only, or that decay by electron capture
with no nuclear gamma rays are omitted from the
table. The over-all activity of the mass-transferred
material is about 1.2 x 10™% curies/g. (By con-
trast, the saturated activities of pure sodium, pure
cobalt, and pure manganese in a thermal-neutron
flux of 10" neutrons/cm?.sec are 0.035, 1.0, and
0.4 curies/q, respectively., The saturated fission-
product activity of the fuel in the ART is about
100 curies/g.)

Since chromium, manganese, and cobalt are the
principal contributors to the activity in the mass-
transferred material, it is well to re-examine the
assumption that their activation rates are pro-
portional to their thermal-neutron cross sections,
especially since cobalt and manganese are known
to have pronounced resonances in the same general
energy range as that of sodium. The neutron flux
as a function of energy was obtained from multi-
group calculations performed by the Curtiss-Wright
Corp. Absorption cross sections as a function of
energy were calculated from the Breit-Wigner

31
ANP PROJECT PROGRESS REPORT

TABLE 1.3.

COMPOSITIONS OF INCONEL AND OF MASS-TRANSFERRED DEPOSITS IN AN

INCONEL FORCED-CIRCULATION LOOP THAT CIRCULATED SODIUM

 

Abundance in Mass=Transferred

 

Element Abundance in Inconel (%)* Deposits (%)*
Nickel 72.0 10 79.0 N  to 95
Chromium 14.0 to 17.0 5 te 10
Iron 6.0 to010.0 0.2t0 0.5
Mangoanese 0.25 te 0.50 1 to?2
Carbon 0.02 to 0.07
Copper 0.15 to 0.50 <0.1**
Silicon 0.15 to 0.50 <0.1**
Sulfur 0.007 1o 0.015
Titanium 0.15 to 0.50 <O YR
Cobalt 0.08 to 0.12 0.240 0.5
Aluminum 0.10 to 0.15 <0.1**
Barium <0.02** <0.02**
Beryllium <0.001** <0.001**
Calcium <0.05** ~0.1*
Magnesium ~0.02** ~Q0.02**
Molybdenum <0.02** <0.02**
Vanadium ~0,02** <0.02**
Zinc <0,2** <0.5**
Zirconium <0.1** <0,T**

 

*Ranges are given when the data are known to be consistent; otherwise the analyses are for typical samples.

**The accuracy of these analyses is 1100%.

formula for each resonance level,%and the separate

contributions from each level were added. Com-
parisons of thermal cross sections and of the
integrals

]04 ev

0,1 ev UO(E) ¢(E) dE !

for manganese and cobalt normaiized to the values
for sodium, are shown below:

 

 

Mn55 C059
o.fh
a
26 74
oth {(for Na)
a
f G, @ dE
21 156

fo_ ¢ dE (For Na)

 

8 Resonance parameters from BNL-325, Neutron Cross
Sections, D. J. Hughes and J. A. Harvey (July 1, 1955);
B. T. Feld, Table of Neutron Resonance Constants,
NYO-3078 (Aug. 28, 1953); D. J. Hughes, letter to
W. K. Ergen, March 15, 1955,

32

The cobalt activity based on thermal cross sections
was therefore too low by a factor of 2. A correc-
tion for this activity raises the total activity to
about 2 x 10~4 curies/q.

Amount of Mass-Transferred Material. - The
total weight of material that will be carried to the
radiators of the ART by the mass-transfer process
can be roughly estimated from results of forced-
circulation loop experiments,
lation

Four forced-circu-
sodium-Inconel loops were operated for
1000 hr with a maximum fluid temperature of
1500°F and a temperature differential of 300°F.
The average weight of deposit in the cold leg
(mostly in the economizer) was® 15 g. The average
surface density of the deposit in the cold leg was
8 mg/cmz, which would yield in the ART radiators
a total deposit of 5 kg.
certain,

This figure is quite un-
because information on the effects of
system volume, surface-area-to-volume ratios, and

Reynolds number is not yet available. it may be

 

?J. H. DeVan, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 106.
seen that the estimate is probably pessimistic,
however, by considering its implications relative
to attack on the heat exchanger, The NaK-wetted
area of the heat exchanger is 10° cm?, The
average attack on the heat exchanger would there-

fore be 5 mg/cm?, or about 0.25 mil, in 1000 hr,

Since the maximum intergranular attack found!'®

 

196, M. Adamson and A. Taboada, ANP Quar. Prog.
Rep, Sept. 10, 1955, ORNL-1947, p 97.

PERIOD ENDING MARCH 10, 1956

in sodium-Inconel loops in 1000 hr is about 1 mil,
the estimate of a total transferred deposit of 5 kg
in the ART radiators is believed to be conservative,

Total Transported Activity. — The total activity
of the mass-transferred material to be found in the
ART radiators, based on the above estimates of
specific activity and weight of deposit, will be
about 1 curie.

It is of interest also to estimate the radiation
{evel that will be found in front of the radiators.

TABLE 1.4. ACTIVITY OF MASS-TRANSFERRED MATERIAL IN THE ART RADIATORS

AFTER 1000 hr OF OPERATION AT 60 Mw®

 

 

S x 100,
Gamma fr Gamma c . d
Isotope Energyb Rays per }sotope g W X 100 Activity F, Factor® Half Life
o ) Abundance (barns) (%) {curies/g)
{(Mev) Disintegration
(%)
Cr50 0.32 0.10 4.3 16.3 10 3.8 x 103 1 26 d
M3 0.85 1.0 100 13.2 2 1.2 x 10~ 1 2.6 h
1.8 0.30 100 13.2 2 0.4 x 103
2.1 0.20 100 13.2 2 0.2 x 1077
Co5? 1.17 1.0 100 37.0 0.5 3.0 x 10™3 1 52y
1.33 1.0 100 37.0 0.5 3.0 x 10~3
Znb4 1.11 0.455 48.9 0.5 0.5 2.3 x 10~6 10 250 d
N;64 0.37 0.04 1.2 2.0 95 0.5 x10~7 (100) 2.6 h
1.1 0.13 1.2 2.0 95 1.4 x10~7
1.50 0.26 1.2 2.0 95 3.6 x 107
Fe38 1.1 0.54 0.31 1.0 0.5 3 x10"8 1000 46 d
1.3 0.46 0.31 1.0 0.5 2.6 x10~8
Mo?®  0.726 Strong 23.8 0.13 0.02 5.7 x 10~8 1000 67 h
cu83 1.34 0.0043 69 4.3 0.2 7.4 x10™7 104 12.8 h
cv® 105 0.09 31 2.2 0.2 2.3 %10~ 10 5.14 m
y31 1.45 1.0 99.76 4.5 0.02 1.6 x 107 104 3.76 m
Ti50 0.32 0.955 5.34 0.4 0.2 1.1 x10=10 10° 5.8 m
0.61 0.01 5.34 0.4 0.2 1.2 x 1012
0.93 0.045 5.34 0.4 0.2 5.3 x 10~ 12
Bal3® o5 Strong 0.1 0.024 0.02 1 x10~"! 10° 12 d
Ba'38 0,163 0.26 7.7 0.68 0.02 3.7 x 10~ 10 85 m
1.05 0.006 7.7 0.68 0.02 8.4 x 10~ 12
Catb 1.31 0.83 0.0033  0.25 0.1 2.5 x 10~ 1 108 4.8 d
ca%® 30 1.0 0.19 1.1 0.1 8.8 x 10~ 8.5 m
536 2.7 0.9 0.017 0.14 0.01 7 x 1011 108 5.0 m

 

ZInformation about isotopes taken from:

D. J. Hughes and J. A. Harvey, Neutron Cross Sections, BNL+325 (July

1955); K. Way et al., Nuclear Level Schemes, TID-5300 (June 1955); K. Way et al., Nuclear Data, NBS Circular 499

(Sept. 1950).

b Beta-particle emitters ignored.

“Weight per cent of the element in mass-transferred material; see Table 1.3.

Activity in curies per gram of mass-transferred material,

€Factor by which abundance of element in mass-transferred material would have to be increased to make its activity

significant.

33
ANP PROJECT PROGRESS REPORT

[t is assumed that the active material will be more
or less uniformly distributed throughout the hotter
one-half of each radiator. The frontal area of the
radiators will be about 140 ft2, and the depth of
the hot side will be less than 3 in. The average
density of the radiators will be about 1 g/m?.
Since the absorption cross section for 1-Mev
gamma rays is 0.06 em?/g, the mean free path of
gammg rays in the radiators is about 16 cm, or
6 in. There will be little self-shielding, and all
the activity may be assumed to be at the surface
of the radiators, Since 1 curie of 1-Mev gamma
rays gives about 1 r/hr at 1 meter, the dose at
d =1 ft, or 0.3 meters, from the surface of the
radiators and opposite the center of the frontal
area will be approximately D = W, that is, the
average value of 1/72 measured to points on the
surface of the radiator. The radius of the circle of
140 ft2 area is 6.7 ft, or 2,1 meters, and therefore

L5 /2 2mp dp

 

f

2 In (R/d)

R 2
./; p dp R

Since R = 2.1 meters and d = 0.3 meters, 1//2 =
0.9 r/hr. The dose level 1 ft from the hot surface
of the ART radiators is thus expected to be about
1 v/hr and to be due mostly to Co®9.

Analysis of Estimates. = The actual dose rate
may be less than expected for several reasons.
The total amount of mass-transferred material,
while very uncertain, has probably been over-
estimated. The quoted abundance of cobalt in the
mass-transferred material was by far the highest
observed in several loops. The typical abundance
is much lower, A thicker boron curtain between
the reflector and the heat exchanger, while not
required in the ART, could reduce the rate of
activation by at least a factor of 4. The use of
cobalt-free nickel in the heat exchangers would
give a substantial further reduction of activity.

34

The actual dose rate may be greater than that
calculated for several reasons. The effect of
small discontinuities in the boron curtain, while
believed to be small, has not been thoroughly
evaluated. The activation of materials in the
sodium-to-NaK heat exchanger, located in the
north-head assembly, is as yet an unknown factor,
The importance of an exchange reaction, in which
radioactive atoms in the heat exchanger may ex-
change with stable atoms in the radiator, with no
net transfer of material, is difficult to evaluate,

A considerable amount of experimental work has
been done on the transfer of radioactivity in stain-
less steel-sodium systems in connection with the
SIR program.!! So far as the processes involved
are understood, the following conclusions have
been tentatively reached:

1. The
materials in the isothermal systems studied may
have been limited by diffusion in the metal walls.

2. The rate-limiting process (or processes) is
steeply temperature-dependent.

3. In systems having a temperature differential
around the loop, the mass-transfer process (pre-
sumed to result from different solubilities of wall
materials in the fluid at hot- and cold-leg tempera-
tures) approximately compensates for a decreased
diffusion rate in the metal in the cold zone. The
net transfer of radioactivity in a nonisothermal
system is about equal to that in an isothermal

rate of redistribution of radioactive

system operated at the temperature of the hot zone
of the nonisothermal system.

These conclusions appear to indicate that the
exchange process will be far less important than
the mass-transfer process in the ART secondary
heat-transfer loop., There is still considerable
uncertainty, however; for example, there may be a
substantial rate of exchange to the isothermal hot
leg that links the heat exchanger to the radiators
and in the entrance headers to the radiators. Re-
liable evaluation of all these possibilities must
await results of radicactivity-transport experiments
being performed by the Callery Chemical Company
for Pratt & Whitney Aircraft and, possibly, must
await operation of the ART.

 

W padioactive Accessibility Reports, CTU-1 to CTU-18
(Jan. 1953 to present),
RADIATION HEATING IN THE ART
H. W. Bertini

Aircraft Reactor Engineering Division

C. M. Copenhaver

Reactor Experimental Engineering Division

R. B. Stevenson
Pratt & Whitney Aircraft

Radiation heating in the ART is being studied
by four methods. Two-dimensional multigroup cal-
culations of neutron flux distributions in the ART,
which are being performed by the Curtiss-Wright
Corp., will provide the information necessary to
determine heating by neutron elastic scattering
and to determine the source distributions for all
sources of gamma radiation. Gamma-ray heating
in the island-core-refiector region will be calcu-
lated by a two-dimensional numerical-analytical
method developed by Pratt & Whitney Aircraft and
by a two-dimensional Monte Carlo method being
developed at ORNL (see Sec. 11, “‘Shielding Analy-
sis'’), Heating in the heat-exchanger—pressure-
shell region and in other special areas, such as
the north and south headers, is being studied by
analytical and numerical methods with the use
of desk computers,

Status of Information and Description of Calcu-
lational Methods. — Information regarding all pos-
sible sources of gamma rays and the factors affect-
ing their transmission and absorption is being
collected, and, at this time, the compilation is
nearly complete, except for some data on the gamma
rays from inelastic scattering. The compilation
will include buildup factors; energy absorption
and total absorption gamma-ray cross sections;
prompt, decay, and capture gamma-ray neutron
spectra; inelastic scattering cross sections; and
the gamma rays from inelastic scattering, A brief
summary of the data that have been obtained is
given below,

Energy-absorption buildup factors were calcu-
lated for all materials. The form assumed for the
buildup factor was

B, =A(e™ - 1)+1,
where A and « are constants for a particular ma-
terial at a specified gamma-ray energy, and pur
is the distance in mean free paths from the source
to the point of interest. The constants A and a
were determined for all materials at several en-

PERIOD ENDING MARCH 10, 1956

ergies by interpolation in atomic-number, Z, values
from data given by Goldstein and Wilkins.!'2 The
energy-absorption coefficients and the total ab-
sorption coefficients were calculated by interpo-
iation of values of these quantities given by
Moteff,

The prompt-gamma-ray spectrum was put into an
analytical form that approximates the recently ob-
tained experimental data.'* Superimposed on this
spectrum was an estimate of the UZ35 rqdiative
capture gamma rays, which contribute approximately
14% of the total prompt-gamma-ray energy per
fission. The resulting analytical expression for
the total prompt-gamma-ray spectrum is 8.8¢~ 1-01E
photons/Mev-fission, This expression was inte-
grated between several fixed energy limits to get
the prompt-gamma-ray energy within these energy
limits, The data, modified in this way, can be
used directly, in conjunction with the buildup
factors, in numerically integrating over the spec-
frum for the purpose of calculating the heat depo-
sition rate resulting from these gamma rays in
various parts of the reactor,

The experimental curve of the spectrum of decay
gamma rays'> was also approximated by an ana-
fytical expression., This expression was modified
to include the fraction of the long-lived decay
gamma-ray emitters, which were not measured ex-
perimentally, and to omit the fission-product gases,
which are expected to escape from the reactor.
The resulting spectrum is 10e=*33F photons/Mev-
fission. This expression was integrated between
the same energy limits as those used for the
prompt-spectrum calculation, and the integration
yielded information of the same nature.

Almost all the capture-gamma-ray data were taken
from papers by Kinsey, Bartholomew, and Walker,
who give capture-gamma-ray spectra for gamma-
ray energies above about 3 Mev. The references
to these papers, and some of the data used, are
given by Mittleman and Liedtke.!® Because of

 

12y, Goldsteih and J. E. Wilkins, Jr., Calculations of
the Penetrations of Gamma Rays. Final Report, NYO-
3075 (June 30, 1954).

13,, Motetf, Miscellaneous Data for Shielding Calcu-
lations, APEX-176 (Dec. 1, 1954).

g, Gamble, Prompt Fission Gamma Rays from
Uranium 235, unpublished Ph.D. thesis (June 1955).

]5R. W. Peele et al., ANP Quar. Prog, Rep. Dec. 10,
1955, ORNL-2012, p 226, Fig. 13.22,

]6P. S. Mittelman and R. A. Liedtke, Nucleonics
13(5), 50 (1955).

35
ANP PROJECT PROGRESS REPORT

the lack of information on the spectrum below
3 Mev, a spectrum for these gamma rays was chosen
rather arbitrarily in such a way that the integral
under the total spectrum gave the neutron binding
energy of the compound nucleus. Most of the
capture-gamma-ray energy is given off above 3 Mev;
so errors in the assumptions should not contribute
significantly to the total error. Partial integrations
were performed numerically over the same energy
range as that used for the prompt- and decay-
spectra calculations.

The results of a fairly extensive literature search
on the inelastic scattering cross sections and the
(n,n",y) gamma-ray spectrum of the materials in the
ART are being assimilated. There is insufficient
experimental information available, however, and
it has been necessary to make approximations and
extrapolations over large energy regions. The
calculated heat deposition rate, as a result, will
have to be used rather cautiously., However, pre-
liminary calculations indicate that, in general,
radiation resulting from inelastic scattering is not
the major contributor to the heating, and therefore
the error introduced into the total heat deposition
rate is relatively small,

Preliminary Calculations. — Some preliminary
calculations, as mentioned above, have been made
to justify calculational techniques and to deter-
mine the gamma-ray sources that can be neglected
in calculating the heat deposition rate. Calcu-
lations were made, by using slab geometry, of the
heat deposition rate at a peint in the pressure
shell resulting from decay gamma rays produced
in the heat exchanger. In one case the buildup
factors for the heat exchanger were used, and in
another case the buildup factor for Inconel was
used, but the results were essentially the same.
The negligible differences in the results indicate
that, for gamma rays passing through boundaries
of materials of similar Z, the heat deposition rate
is nearly independent of the choice of buildup
factor, so long as it is restricted to those for
similar Z.

In another case, it was assumed that the decay-
gamma-ray energy was monoenergetic, and the
results were compared with the results obtained
by integrating over the gamma-ray energy spectrum,
Monoenergetic gamma rays of 0.5, 0.75, and 1.0
Mev were used, and differences from the spectrum
calculations of as much as 20% were obtained,
This indicates that it will be necessary to inte-

36

grate over the spectrum of gamma-ray energies in
order to achieve the desired accuracy,

The !{‘-in.-thick Inconel liner on the outside of
the beryllium reflector becomes a fairly hot source
of capture gamma rays, Test calculations of the
heat deposition, in slab geometry, made by assuming
the Inconel liner to be a plane surface source of
gamma rays were compared with calculations made
by assuming a plane distributed source of gamma
rays. For regions of small penetrations (0.05 mean
free paths) there was a difference of 10% between
the results for the two calculations, and for regions
of deeper penetrations (2.5 mean free paths) the
difference became as high as 30%. Therefore care
must be taken before it is assumed in the calcula-
tions that apparently thin sources of gamma rays
may be approximated by plane surface sources.

Test calculations were also made to determine
the relative source strengths of gamma rays re-
sulting from inelastic scattering in the fuel, in
the core shells, in the island, and in the first
5 em of the beryllium reflector, By using the
fluxes and the absorption rates determined by
multigroup calculations in spherical geometry,!?
it was found that the energy of the gamma rays
produced by inelastic scattering in the fuel con-
stitutes about 20% of the total gamma-ray energy
released from the fuel. This is a surprisingly
large contribution to the total heating, and it cannot
be neglected.

The contribution to the total heating by the
gamma rays resulting from inelastic scattering in
the core shells was found to be negligible. How-
ever, the gamma rays arising from inelastic col-
lisions in the island and in the beryllium reflector
near the fuel are significant. They are comparable
to the capture gamma rays in intensity, and, al-
though their source strength per unit volume is
small, their contribution to the heating cannot be
neglected because of the large volumes involved.

Calculations were also made of the source
strength of the capture gamma rays in the pressure
shell. The strength was found to be extremely
small in comparison with the strengths of other
sources,

Calculations of the heating in the heat-exchanger—
pressure-shell region have been nearly completed,
Heat deposition rates for all regions amenable to

 

17H. Reese, Jr., 8. Strauch, and J. T. Mihalczo,
Geometry Study for an ANP Circulating Fuel Reactor,
WAD-1901 (Sept. 1, 1954).
the analyticodexd numerical approach will be
published in a summary report when all the cal-
culations have been completed.

Application of Methods to a Test Case. — The
energy deposition from gamma rays was determined
for a test case by the Monte Carlo method and by
an analytical approach, for which an exponential
form of the buildup factor was used, in order to
make a comparison between the two methods and
also to gain some insight into the validity of the
analytical buildup-factor method for nonhomogene-
ous media. The results of these calculations are
shown in Fig. 1.9.

The source was a uniform plane source of 1-Mev
gamma rays adjacent to the fuel region shown in
Fig. 1.9. The Monte Carlo calculations were per-
formed by using a code developed by Auslender
for slab geometry (see Sec, 11, ‘‘Shielding Analy-
sis''), The following equation was used for the
analytical calculations:

S,ue

 

(1) H = {AE1 [(1 v ) Z (pTX)i]

+ (1 = AE, [ ,- (,,,TX)Z.J} ,

where

H = energy deposition,

S = gamma-ray source strength,

{t, = linear energy absorption coefficient,

gy = linear total absorption coefficient,

{(pyX); = total thickness traveled through each
slab in mean free paths,

A and a are the energy-absorption buildup-factor
constants, which depend on the material and on
the energy of the incident gamma rays, and E, is
the usval exponential integral

E,[)/] = J;m e~ Yy~ Vdu

The energy deposition was determined in each
region by using the buildup factor for each of the
materials present. When it is considered that the
statistics of the Monte Carlo calculation are not
too good because of the relatively small number
of histories computed and that the buildup factors
are derived on the assumption of a homogeneous
medium, the results shown in Fig. 1.9 are guite
encouraging. They seem to indicate that the ana-
Iytical buildup-factor approach will give reason-

PERIOD ENDING MARCH 10, 1956

ably accurate values of the energy deposition for
multilayer regions if the buildup factor appropriate
to each material is used in computing the depo-
sition in that material,

The greatest discrepancies between the two
methods occur near the boundaries of different
materials, where the buildup-factor approach should
be incorrect. However, since the equivalent Z of
the fuel is very close to the Z of the Inconel, the
ratio of the energy deposition on either side of
the first boundary (fuel-Inconel interface) should
be approximately equal to the ratio of the energy
absorption coefficients, which is 0.221/0.094 =
2.35. If the values of the energy deposition, as
given by Eq. 1, are used with the fuel buildup
factor, the ratio is 19.4/8.2 = 2.36, while from
the Monte Carlo calculation the ratio is 13.6/9.7 =
1.40. Therefore it is felt that the results predicted
by the Monte Carlo calculation for the first one-
half of a mean free path in the first Inconel layer
must have a large statistical error and that the
analytical buildup-factor approach probably gives
the more nearly correct energy deposition at this
point.

This reasoning cannot be applied to the Inconel-
sodium boundaries, since the Z's for these ma-
terials are quite different. However, it is felt that
the increase in the energy deposition in the last
one-quarter of a mean free path in sodium (as pre-
dicted by the Monte Carlo calculation) is not o
real effect and is caused by statistical uncertain-
ties in the calculation.

More test cases are being planned in order to
get a better estimate on the validity of the buildup-
factor approach and to improve the statistics of
the Monte Carlo calculation. However, from this
one test case, it seems that the buildup-factor
approach will give reasonably good answers for
the gamma-ray energy deposition in multilayered
regions, even though the buildup factors are derived
for homogeneous media,

Self-Absorption of the Decay Gamma Rays
in the ART Fuel Dump Tank

H. W, Bertini
Aircraft Reactor Engineering Division

The self-absorption of the decay gamma rays
in the ART fuel dump tank has been calculated,'®

 

By, w. Bertini, An Estimate of the Self Absorption of
the Decay Gammas in the ART Fuel Dump Tank, ORNL
CF-55-12-47 (Dec. 9, 1955).

37
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o
ORNL=-LR-DWG 13201
0.20 ’ |
\ o |
org ‘ | —— MONTE CARLO METHOD |
W —— — — FUEL BUILDUP FACTOR USED
‘\‘\ —— - —— INCONEL BUILDUP FACTOR USED |
—— \\ [—=--— SODIUM BUILDUP FACTOR USED
‘ A i
A \\
0.16 \
! \ \\\‘.
\ 1 )
o \ \ ]
= \ \
S o4 \ \
= ' \ v\
5 \ \\\ K
& \ A L\
: \ \
2 \ \\
e 012 \ X
o \ \ :
§ \ \ \ ' sopiumM |
S \‘ ‘ (MULTIPLY SCALE BY 10) | INCONEL
\
@ 10 W N ;
< 3\ X
> \\ P\ \ ;
i ‘ \ W\
W \ W\
> \ W
o \ \\\
- \ \
= 008 \\} \\ -
— :
[T
e | ‘ fi \
=z ! re N
5 | AN
S | NI\
< 0.06 I\ ]
[T \ .
i \\l‘
‘ x\ ‘
| N
0.04 | | | ‘\\
: [ I \ \ NS
FUEL INCONEL \\\\\ \
— by
| SO NI
. < O
002 N N
’ \\\\\.\
0 [
0 i 2 3

DISTANCE (MEAN FREE PATHS)

Fig. 1.9. Gamma-Ray Heating as a Function of Distance from o Plane Source of 1-Mev Gamma Rays
Adjacent to the Fuel Region.

38
The results indicate that 90% of the decay gamma
rays originating in the fuel dump tank will be
absorbed there, and therefore cooling facilities
must be provided in the dump tank to remove almost
all the decay heat in the fuel.

It was assumed for the calculations that the
dump tank contained only fuel and that it was
spherical rather than cylindrical. Spheres of two
different radii were used, so that one sphere had
the same volume as the cylinder and the other had
the same surface-area-to-volume ratio as the
cylinder. The results obtained for the two spheres
differed by only 2%.

The expression for the fraction of gamma rays
escaping from a sphere was taken from the work
of Storm, Hurwitz, and Roe, ! ? for which the straight-
ahead approximation was made; that is, it was
assumed that only those gamma rays that made
first-flight absorption collisions in the sphere
were absorbed there, Since multiple-scattering
effects were neglected, the expression yields an
underestimate of the self-absorption. If the totai
gamma-ray cross section is used, rather than the
energy-absorption cross section, the expression
will yield an overestimate, since every collision
is counted as an absorption collision. The energy
absorbed in each sphere was calculated by using
both the total and the energy-absorption cross
sections, The maximum difference in the results
obtained was about 10%,

A numerical integration was made over the decay-
gamma-ray spectrum?® to check its effects, and
the value obtained was about 3.5% lower than that
obtained by assuming that all the gamma rays were
monoenergetic, The estimate that 90% of the decay
gamma rays originating in the fuel dump tank will
be absorbed there appears therefore to be con-
servative.

Activation of the Sodium in the ART
H. W. Bertini

Aircraft Reactor Engineering Division

The activation of the sodium to be used as the
reflector coolant in the ART was recalculated?’
to correspond to the mid-December design of the

 

]9M. L. Storm, H. Hurwitz, Jr., and G. M. Roe, Gamma
Ray Absorption Distributions for Plane, Spherical, and
Cylindrical Geometries, KAPL-783, p 67, Eq. (87)
(July 24, 1952).

2OR. W. Peele et al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 226, Fig. 13.22.

PERIOD ENDING MARCH 10, 1956

reactor., The starting point for the calculations
was the Curtiss-Wright multigroup calculations on
spherically symmetric systems, specifically their
reactor No. 675,22 Attempts were made to modify
the multigroup results to account for the asym-
metry of the ART and to account for the changed
hole distribution in the beryllium island and
reflector,

The results of this work indicate that the ac-
tivity of the sodium will be 5.3 x 101 d/sec or
1.45 x 10% curies. For a total sodium volume of
2.45 13 this represents a decay-gamma-ray source
strength of 0.51 w/cm®. An increase of 20% in
the volume of the sodium passages in the island
and reflector would increase the source strength
only about 10%.

Shutdown Reactivity of Lithium in the
ART Reflector Coolant

A. M. Perry

Electronuclear Research Division

In the event that it becomes impossible to dump
the fuel out of the ART, it might be necessary
to augment the reactivity effect of the control rod
by some poison in order to shut down the reactor
to room temperature, The change in reactivity in
going from iscothermal operation at 1200°F to room
temperature involves the separate effects of di-
mensional changes, o thermal-base change, and
a fuel-density change,

The slow net temperature coefficient of reactivity
calculated by Curtiss-Wright,?® when corrected for
the fuel expansion coefficient, is

1k ~2.34 x 10~5/°F

k dt
The temperature coefficient measured in the
high-temperature critical experiment was =-2.3 x
10~5/°F. The close agreement of the calculated
and the experimental values was considered to
justify the use of the Curtiss-Wright calculations
of the separate effects involved in going from

 

24, w. Bertini, Activity of the Na Coolant in the
ART, ORNL CF-55-12-78 (Dec. 16, 1955).

22, Reese, Jr., S. Strauch, and J. T. Mihalczo,
Geometry Study for an ANP Circulating Fuel Reactor,
WAD-1901 (Sept. 1, 1954).

2crRE Temperature Coefficients for 1/6 in. and 1/8
in. Thick Inconel Core Shell, WNE 55-6-2 (May 18, 1955).

39
ANP PROJECT PROGRESS REPORT

1200°F to room temperature. Therefore, for the

dimensional changes,

1 4k
— — = =077 x10~5/°F ,
E dt

AR = +0.92% ;

for the thermal-base change,

1 4k
— — = +1.99 x 10~5/°F |,
E dt

Ak = —2.4% .

The results of the high-temperature critical ex-
periment?4 showed the specific mass reactivity
coefficient of the uranium in the fuel,

M dE
YR o

to be about 0.14, and the Curtiss-Wright calcula-
tions2% gave y = 0.18 for the uranium and 0.22 for

 

24A, D. Callihan et al., ANP Quar. Prog. Rep. Sepl.
10, 1955, ORNL-1947, p 60.

25¢, B. Mills and H. Reese, Jr., Desz'gn Study of an

ANP Circulating Fuel Reactor, WAD-1930 (Nov. 30,
1954).

40

the whole fuel salt. By using y = 0.22 and o fuel
density of 3.3 g/cm? at 1200°F and 4.13 g/em3
at room temperature, the change in reactivity as
a result of the fuel-density change is

Ak = +5,0% .

Thus the over-all change in reactivity is +3.5%.

The poisoning effect of the sodium in the ART
moderator (island and reflector) has been esti-
mated to be about 0.8% in reactivity, The required
reactivity effect is therefore about 4.5 times as
great as that from sodium alone. The absorption
cross section of natural lithium is 140 times that
of sodium, and therefore the required shutdown
reactivity could be achieved by adding 4.5/140 =
3.2 at. % of lithium to the sodium, The sodium
volume in the moderator cooling system is 2.45 ft3,
and, since lithium is less dense than sodium (0.47
vs 0.8 g/cm®), the volume of lithium required
would be

2.45 x 0.032 x 0.8/0.47 = 0.13 f+3 ,

Shutdown from 1400°F would involve a reactivity
change of about +4.1% and would require about
0.15 13 of lithium. Even if some credit is taken
for the control rod, which could be as little as 1%
at 1200°F, 0.1 £t3 of lithium would be required.
PERIOD ENDING MARCH 10, 1956

2. EXPERIMENTAL REACTOR ENGINEERING

H. W. Savage
Aircraft Reactor Engineering Division

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

Aircraft Reactor Engineering Division
Loop No. 3
C. W. Cunningham J. A. Conlin

Aircraft Reactor Engineering Division

In-pile loop No. 3, which was operated in the
MTR,! was sectioned in the General Electric
Company's hot cell at the National Reactor Testing
Station (NRTS) for convenience in shipping to
ORNL. The loop was sectioned inte 14-in. lengths
that could be accommodated by the available
carriers. The sectioning sequence proceeded from
the front, or nose, end of the loop to the rear of
the pump drive unit, with one additional cut at the
intermediate bulkhead. It was possible to return
the nose section and the entire pump assembly
intact.

An unsuccessful attempt was made during the
sectioning to determine the cause for plugging
of the fission-gas lines, The lines extending
through the concrete shield were observed to be
open, but it was not feasible to check the passage
into the pump or the lines out of the pump in the
region where the plugging probably occurred.
These lines will be examined in a hot cell at
ORNL (see Sec. 8, ‘*Radiation Damage’’),

After the sectioned loop was received at ORNL,
it was placed in a hot cell, and the fuel was
removed. The loop was then sectioned further,
and samples were prepared for metallographic
examination. This loop, as previously reported,
operated with a maximum fuel temperature of about
1500°F. There was a temperature differential in
the fuel system of 155°F for 103 hr and 100°F for
168 hr. The metallographic examination revealed
corrosion penetration to a maximum depth of 1 mil.
Comparison with extrapolations of results ob-
tained with forced-circulation loops operated
out-of-pile indicates that the 1 mil of corrosion
is about what would be expected out-of-pile under
similar conditions, and hence irradiation appears

 

L. A. Carpenter et al., ANP Quar. Prog., Rep. Dec.
10, 1955, ORNL-2012, p 28.

to have little, or no, effect. However, the oper-
ating time, buildup of fission products, and temper-
ature differential in the fuel were less than are
anticipated in the ART, and the results can only
be treated as preliminary,

The pump has now been disassembled, except
for the impeller housing. Parts in the bearing-
housing region appeared, in general, to be quite
clean, Less oil has been observed than was
expected from the lubrication rate for the seals
and bearings. The excess oil may have been
absorbed in the Fiberglas insulation used on
thermocouple and electrical leads, but this was
not fully evident, No direct cause for the stoppage
of the purge-gas flow through the bearing housing
has been found. The seals and bearings were
found to be in good condition,

Considerable quantities of decomposed or partially
decomposed materials were found forward of the
front seal, The seal bellows contained a thick,
reddish«black, waxlike substance, and there was
a similar deposit on surfaces in this vicinity,
The fuel in the pump sump appeared to be mixed
with a carbonaceous material, which increased
the sump level to above the point for normal
operation, It is possible that the presence of these
materials caused a friction drag on the pump shaft
and produced the erratic-speed behavior observed
during operation. Similar materials were found in
the off-gas line for the sump purge, but it seems
doubtful that these materials could have com-
pletely plugged the lines in the regions where they
have been observed, No ZrF, vapor (**snow”)
deposits were noted.

Loop No, 4
C. C. Bolta R. A. Dreisbach
D. M. Haines
Pratt & Whitney Aircraft

J. A, Conlin C. W. Cunningham

Aircraft Reactor Engineering Division

Assembly of in-pile loop No. 4 was completed
on January 20, and it was shipped to NRTS for
checkout and operation. Loop No, 4 was the same
as the previous loops, except for minor modifi-
cations, which included larger diameter tubing for

4]
ANP PROJECT PROGRESS REPORT

the gas purge system and the substitution of
nitrogen for helium in the nose purge system, along
with a separate cold trap external to the reactor
cubicle, The liquid-nitrogen heat-transfer system
for cooling the fission-gas adsorption traps was
also revised to provide for more automatic control,
and vacuum-jacketed lines were added to minimize
nitrogen losses,

The loop was found to be satisfactory and was
inserted in the MTR HB-3 beam hole on February 6.
Isothermal operation started on February 8, and
the reactor was brought to full power on February 9,
The operating conditions were approximately the
same as those for loop No. 3, that is, a temper-
ature differential of about 150°F and a maximum
fuel temperature of about 1500°F. It is apparent,
however, from several observations, that both
loop No. 4 and loop No. 3 operated at a somewhat
higher power level than was estimated for loop
No. 3. The temperature differential observed from
the thermocouples on loop No., 4 was between
175 and 200°F and that estimated from heat removal
in the air was approximately 200°F, The higher
power level indicated by the temperature difference
has been substantiated by analyses of the activity
in the cobalt-foil flux monitors in loop No. 3.
The flux is now estimated to have been 50% greater
than that originally calculated from operating
conditions,

It was not possible to use the nitrogen purge
system for the nose region, because the cold traps
plugged. The plugging was apparently caused by
moisture in the gas. Extensive checks had been
made at ORNL to determine the reliability of this
purge system, and no trouble had been encountered.
As many as 60 bottles of nitrogen, prepared to the
same specification as that for the nitrogen normally
purchased at the MTR, were passed through a
similar cold trap without noticeable plugging.
Less than two bottles resulted in plugging of
traps at the MTR. Helium was therefore utilized
again as the nose purge gas, but no difficulty
occurred from heater short circuits, such as was
experienced previously,?2

Both the bearing-housing purge system and the
pump-sump purge system plugged after ] I/2 weeks of
operation, and the helium inlet lines to these
systems were clamped to prevent diffusion of
radioactive fission gases to the operating area.

 

2p, A. Gnadt and A, A, Abbatiello, ANP Quar, Prog.
Rep. Dec, 10, 1955, ORNL-2012, p 29.

42

Activity appeared in the nose purge gas at one
time, but it was contained in the nose-purge-gas
traps. This activity apparently came from a leak
in the pump bulkhead and did not indicate trouble
in the main circuit, Operation was terminated
during the scheduled MTR shutdown of February
27, and the loop was removed from hole HB-3 on
February 29, High radiation levels were observed
in the cubicle during removal of the loop and were
apparently due to plated material on the walls of
the gas lines. In general, the activity of the
lines and of the traps appeared to be higher than
had been originally estimated or extrapolated
from experience with foop No, 3. The loop circu-
lated the fuel for approximately 500 hr and operated
with a temperature differential for approximately

390 hr.

Loop No, 5
J. A, Conlin P. A, Gnadt

Aircraft Reactor Engineering Division

Loop No. 5 has been assembled except for
thermocouple attachment, insulation, and instal-
lation of the water jackets., It is leaktight and
pump operation has been checked. Completion of
assembly has been delayed until the operation of
ioop No. 4 can be analyzed sufficiently to indicate
any required changes. The schedule for utilization
of the reactor beam hole provides adequate time for
this delay.

FORCED-CIRCULATION CORROSION AND
MASS5-TRANSFER TESTS

Fused Salts in Inconel

C. P. Coughlen G. E. Mills
P. G. Smith

Aircraft Reactor Engineering Division

R. A. Dreisbach
Pratt & Whitney Aircraft

Ten electrical-resistance-heated and two gas-
heated forced-circulation Inconel loops were
operated with fused salts as the circulated fluids
during this quarter, The operational data for these
loops are summarized in Table 2.1. The results
of metallurgical examinations of the loops are
presented in Sec. 5, ““Corrosion Research.”

A development study of the electrical-resistance-
heated loops has revealed that temperature differ-
ences of up to 270°F exist between opposite sides
e . - — ——

TABLE 2.1. SUMMARY OF OPERATING CONDIT

PERIOD ENDING MARCH 10, 1956

IONS FOR INCONEL FORCED-CIRCULATION LOOPS

THAT CIRCULATED Nc:F-ZrF“-UF4 (50-4 6-4 mole %)

 

Maximum

Maximum

 

. Reynolds Temperaf:.lre Fluid Tube-Wall Operating
Loop No. Differential Time Comments
Number (OF) Temperature Temperature b
(°F) CF) ()
Loops with Electrical-Resistance-Heated Straight Sections
7425-6 10,000 200 1500 1565 1000 Terminated on schedule
<7A 10,000 200 1500 1610 1000 Terminated on schedule
-8 10,000 200 1500 1570 1000 Area of cooled surface increased
_ to five times that of standard
loop; tetrminated on schedule
-9 4,500 200 1500 1670 Life test Accumulated time in test: 1991 hr
. on Feb. 10, 1956
-10 10,000 200 1500 1600 1000 Terminated on schedule
-11 16,000 200 1500 In test Heated volume increased fourfold
-42  Variable 200 1300 to 1500 1700 Variable Development loop for studying
heater design
-43 6,500 200 1500 1700 1000 Part of series to determine effect
of wall temperature; terminated
on schedule
. -44 Variable Variable Variable Variable Variable Development loop for studying
heater design
. =45 6,500 200 1300 1700 320 Terminated because of a failure
in auxiliary equipment
Loops with Gas-MHeated Coiled Sections
<71 Variable Variable Yariable Variable Variable Development loop for studying
thermocouples
4935-6 8,000 200 1500 1575 Life test Accumulated time in test: 6637 hr

on Feb., 10, 1956

 

of the heated tubing, top and bottom, when the
Reynolds number of the circulated fluid is less
than about 5000. A lower limit has therefore been
set for the Reynolds number, and the problem is
being studied further,

Liquid Metals in Inconel and in Stainless Steel

C. P. Coughlen A, G, Smith
. Aircraft Reactor Engineering Division
R. S, Dreisbach
Pratt & Whitney Aircraft
Six forced-circulation loops were operated with
sodium and noneutectic NaK in Inconel and in
stainless steel systems during the quarter, A

summary of the conditions of operation of these
loops is given in Table 2.2,

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

The fluid used in initial tests of ART pump
bearings and seals for lubrication and heat removal

43
ANP PROJECT PROGRESS REPORT

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

 

 

T ; Maximum 0 .
Cold Reynolds %mpera ‘ure Fluid Operating perating
Loop Ne. Differential . Time Remarks
Trap  Number o Temperature Fluid
( F) OF (hr)
("F)
74266~ Yes 15,000 300 1500 Sadium In test Cold trap at 250°F
test (cleaned)
-6 Yes Variable Variable Variable Sodium Variable Development loop for studying
service plug indicators
-7 No 15,000 300 1250 Sodium In test Beryllium insert in hot leg;
massive beryllium in stream
7439-3 Yes 15,000 300 1500 Noneutectic 1000 Cold trap at 100°F; terminated
NaK on schedule
-51 Yes 15,000 Maximum 1500 Noneutectic Life test Started February 13, 1956
possible NaK
-52 Yes 15,000 Maximum 1600 Noneutectic Life test Started February 13, 1956
possible NaK

 

was a light-weight, refined, petroleum-base mineral
oil, having a viscosity of 65 SSU at 100°F, When
it was determined, however, that the temperature
of the pump lubricating fluid would be 200 to
240°F in the operating ART, an investigation of
suitable synthetic [ubricants was begun, since
most low-viscosity petroleum oils have a tendency
to “‘coke’ at temperatures above 180°F. Mechani-
cally, coking would be of no consequence, but it
was feared that the reduction, with time, in the
amount of heat removed by the lubricant might be
hcrm{ul.

The UCON LB series of synthetic lubricants
was tested first, These are water-insoluble
polyalkylene glycol-base fluids that are available
with or without various inhibitors in o wide range
of viscosities. Other properties being equal, they
are superior to petroleum products in that their
specific heat is higher, and, more important, they
do not coke, even when used at temperatures well
above those specified for ART pump operation,
These two advantages are offset, however, by
operational limitations. The loosely held hydroxyl
ions in the UCON fluids have a tendency to pick
up copper from system components and to pla’re it
out elsewhere. Unfortunately, this copper plating
occurs primarily on the mechanical-seal interfaces.
Although this problem could be circumvented

44

by eliminating all copper-brass parts, equally
difficult design problems would be introduced,
It has also been found to be impossible to es-
tablish and maintain good seal performance with
the UCON fluids. Early successful operation with
the UCON fluids could not be reproduced. One out
of ten seals tested with the UCON fluids might
be acceptable, whereas further testing with mineral
oil showed six to eight out of ten seals to be
acceptable, '

The lack of success with UCON fluids does not
preclude the eventual use of a synthetic fiuid.
Testing and evaluation of other possible fluids
ore continuing, However, to permit further pump
studies, all pump rotary assemblies are now being
tested with the originally rejected petroleum-base
mineral oil, but the oil operating temperature is
being limited to 180°F. If no single fluid is found
to be serviceable, it is possible that a bifluid
system will be required, that is, one fluid for
cooling and another fluid for lubrication and seal
contact., As an aid in the solution of the fluid
problem, tests are being conducted to determine
the severity of the effect of petroleum-oil coking
on heat removal,

Since proper control and removal of the lower
seal oil leakage are essential to pump operation,
the seal design has been modified to include a
baffle that extends upward beyond the seal interface
to direct any oil leakage downward into a catch
basin, from which it can easily be removed by
means of a dip tube and continuous gas purging.
The Durametallic seals discussed previously3
cannot be used with the modified design, and it
now appears that the seal will be a bellows type
with a stationary Graphitar 39 nose piece running
against a hardened-steel ring (tool steel or
AISI-8620 hardened to Rockwell C 50-55) clamped
to the rotating shaft. The original leakage-rate
specification has been changed from 2 cm3/day or
less to 4 to 5 cm3/day.

Sodium-Pump Performance Tests with Water

H. C. Young
Pratt & Whitney Aircraft

M. E. Lackey

Aircraft Reactor Engineering Division

Additional tests of the performance of the ART
sodium pump, model MN, with water were performed

 

Sw. L. Snapp and W. K, Stair, ANP Quar. Prog., Rep.
Dec, 10, 1955, ORNL-2012, p 33.

PERIOD ENDING MARCH 10, 1956

to determine the effect of inlet conditions on
performance and cavitation. The volute inlet
conditions for the tests described previously4 did
not mock up the final design, and the additional
tests were therefore made to determine the effect
of the modified volute inlet on the pump character-
istics. The results of these tests are compared
with those of previous tests 5 and 6 in Table 2.3,
and performance curves for tests 5 and 7 are
presented in Figs. 2.1 and 2.2. The impeller used
in test 5 was fabricated according to the final
design, and the volute inlet conditions of test 7
were those to be used in the ART pump. A test
of @ pump that incorporated all the final design
conditions was started, but the inlet shroud vane
separated from the impeller because of the failure
of a soldered joint, and the test could not be
completed. Since all data obtained had shown
good agreement, it was decided to consider the
tests as having been completed,

A typical cavitation curve is presented in
Fig. 2.3. Cavitation tests were conducted with

 

4M. E. Lackey and H. C. Young, ANP Quar. Prog,
Rep, Dec. 10, 1955, ORNL-2012, p 35.

UNCLASSIFIED
CRNL-LR-DWG 13082

 

 

 

 

 
 
 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

140 - ‘ ‘
EFFICIENGY (%) ‘
|( _ DESIGN POINT ‘
120 | |
| TN |
‘ / | |- VOLUTE
/1/0 L ’ © \ \/,/ BALANCE LINE
/ ' |
100 M A o Lo e _ . |
>/ m //‘;, // ’ /J )
- .
2900 e 7 / \
— 80 "7/—/ _'.00 rpm - /’l _% /71 \;i T
+— 2 / / -~
+, / , - ///, .7\/\ //
e 4+ [ L F
it 400 R8T 7~ ' / /
I 60 240 —-60-62—65 ,
T | s 7 / /
& h’\ /
60 Y~
4 o 55 "~
0 A ‘
|
20 e
|
0 . 1
0 50 100 150 200 250 300 350 400 450 500 550 600 650
FLOW (gpm)

Fig. 2.1. Sodium-Pump Performance Test 5.

45
TABLE 2.3,

RESULTS OF WATER PERFORMANCE TESTS OF ART SODIUM PUMP

Design point: 430 gpm, 90-ft head

 

Impeller Clearance

 

, T f Inlet Effici
Test No, Inlet Conditions (in.) ype ob Inle retency Remarks
Shroud (%)
Axial Radial
5* Volute inlet plate 3162 in. G.050 0.0625 Shroud 0.258 in. shorter 64
thick; opening 3.5 in. in than in previous test;
diameter with 3]{32-in. shroud vanes removed
radius at edge; 0.37Q~in.
spacetr between plate
and bottom of impeller
& Same as in test 5 0.050 0.0625 Shroud saome as in test 5; 63
shroud vanes reinstalled
and extended 3/32 in.
7** Volute inlet plate 1/2 in. 0.050 0.0625 Same impeller conditions 63
thick; opening 3.5 in. as fest §
in diameter with 1,/z-in.
rod at edge; no spacer
8 Same as in test 7 0.050 0.0625 Same as in fest 5§ Inlet shroud broke away
from impeller; test
incomplete
Q Same as in test 5 0.050 0.0625 Same as in test 5 Test made to check volute

unbalance

 

*Tests 5 and 6 run during previous qucn'fel‘4 with final impeller configuration.

**Final volute inlet conditions used.

L¥0d3Y $§3¥J0¥d LIO2Ir0dd ANV
PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
ORNL-LR-DWG 13083

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

140 — ‘ ’ I T T ] } T \ T
L [/DES\GN POINT |
120 - { - {
// |
100 | =L /l .
-
_ !
-~ ® -~
-y ‘\\ \ \‘:\ i
| s 17 _ _[ /.
— 80 66 ' | 2
= o6 / \
2 . DA A
Lfi ‘ ‘ > 60 .'\ // *
60 ‘ 55 /X‘Q
| voLuTE 50
BALANCE LINE ;
40 fl ’\ J
%0 | l T | )
0 L ] _
0 50 100 150 200 250 300 350 400 450 500 580 600 650
FLOW {gpm}

Fig. 2.2. Sodium-Pump Performance Test 7.

UNCLASSIFIED
ORNL—LR~DWG 13084
VALUE OF CAVITATION PARAMETER, o

02 03 0.4 05 06

 

 

 

  

CHANGE N MAXIMUM EFFICIENCY (%)

 

 

 

 

 

 

 

 

24 — .2900rpm 90°F, FIXED THROTTLEl

102900 rpm, 92°F, 430-gpm FLOW

[0 2400rpm 90°F 380 gpm FLOW
28 ,
32— .

TOTAL INLET HEAD LESS VAPOR PRESSURE
PUMP HEAD

36 | |

 

Fig. 2.3. Sodium-Pump Cavitation Characteristics.

two different inlet conditions and with and without
vanes on the impeller inlet shroud. The variations
in the data were slight, but the data obtained under
the conditions that most nearly simulated the
design conditions appeared to give the best
cavitation characteristics.

Estimates of performance with sodium, based on
the water test data, show that to prevent cavitation
at a sodium temperature of about 1500°F a total
inlet head of 6.26 psig will be required; at a
sodium temperature of 1285°F a total inlet head
of 0.96 psig will be required. The ART sodium
pumps are to operate at an inlet pressure of
15 psig, and therefore the cavitation character-
istics of the pump appear to be satisfactory,
Cavitation data will also be obtained in high-
temperature tests with sodium.

Four static-pressure taps were equaily spaced
on the pump volute, at the periphery of the impeller,
to obtain measurements of the radial hydraulic
forces on the impeller. It was found, however, that
one of the taps had been improperly located in a
corner of the volute near the tongue, and it
measured part of the velocity head in addition to
the static pressure, The other three pressure taps
indicated that the balance was satisfactory. |In
order to investigate the balance more thoroughly,

47
ANP PROJECT PROGRESS REPORT

the number of taps was then increased to nine.
The data indicated that the unbalance at design
head and flow was somewhat [ess than that
measured with the improperly located tap and
somewhat greater than that measured by the other
three taps. The volute balance lines are shown
in Figs. 2.1 and 2.2. At design point the maximum
unbalance is less than 7 psi, aond therefore the
balance is considered to be satisfactory.

It appears from the water test data that the
sodium pump will produce the required design
head of 90 ft and flow rote of 430 gpm at an
efficiency of 63% at 2860 rpm and that it will
require 15.6 bhp. To obtain the high head required
with a minimum speed, an impeller vane angle of
90 deg is used. As a result, the head-vs-flow
curve is rather flat for parallel pump operation.
The design point lies on the negative slope of the
curve, however, and it is believed that satisfactory,
stable, parallel pump operation will be achieved,

NoK-Pump Performance Tests with Water

H. C. Young
Pratt & Whitney Aircraft

M. E. Lackey
Aircraft Reactor Engineering Division

The test stand used for water tests of the ART
NaK pump consists of a 6-in, closed-pipe loop; a
200-hp wound-rotor motor, with speed control by
variable resistance; a flow-metering orifice; a heat
exchanger to remove the heat added to the water by
the pumping power; a throttling valve; a test
volute; a test impeller; and the necessary instru-
mentation, The test volute was a precision
aluminum casting, and the test impeller was
fabricated of bronze.,  The preliminary tests
indicated that the primary design heod and flow
rate could be obtained at speeds below the design
speed. lLeckage flow around the impeller hub
was excessive, and serious ingassing occurred at
design point, with the pump pot acting as a surge
tank. For the first test the pump was operated
with the pump pot full of water, and a separate
surge tank was connected to the loop. Disassembly
of the pump after the first test disclosed that
the volute tongue had been damaged by local
cavitation and that the impeller had rubbed against
the side of the volute,

For further testing, the volute tongue was filed
to @ smooth contour and re-used. Vanes were
installed on the back of the impeller hub to

48

decrease bypass flow and thus to decrease the
ingassing that resulted from splashing, Also, a
sleeve was installed in the volute to decrease the
cross-sectional area at the point of highest hy-
draulic unbalance.  As the tests progressed,
changes were made to hub vanes and the volute as
indicated by the tests. The performance curve
obtained in test 7 is shown in Fig. 2.4, The curve
shows that the design point can be reached at
considerably below the maximum motor speed of
3550 rpm. The efficiency is approximately 75%
at design speed, and the volute unbalance is
within satisfactory limits,

Additional tests indicated that the impeller will
cavitate at an inlet pressure of 5 psig, which is
below the design inlet pressure of 10 psig under
ART conditions. This condition is considered
to be acceptable, but an alternate design is being
prepared. Further tests are to be conducted to
investigate more thoroughly the degassing character-
istics, to determine the optimum volute tongue
angle, and to check further the performance and
cavitation after minor revisions have been made to
the pump assembly.

High-Temperature Tests of ART Fuel Pumps
S. M, DeCamp, Jr,

Aircraft Reactor Engineering Division

The startup of an ART fuel pump (model MF-2)
on a short-circuit test stand (No. 1) was described
previously,  After the initial difficulties had
been overcome, the pump operated continuously at
¢ maximum fluid (NaK) temperature of 1400°F
for 2056 hr. No further operating difficulties were
encountered until about 4 hr before termination of
the test, when the power trace of the drive motor
increased sharply to about 1Y times normal, or
7 kw. This power increase lasted several minutes,
and then the power level returned te normal,
Approximately 4 hr later the power consumption
again increased to about 'l‘/z times normal and
continved to vary rapidly over a wide range, When
the power consumption failed to return fto normal,
the test was terminated to prevent possible damage
to the pump.

Upon disassembly of the pump it was discovered
that the lower seal oil-leakage removal system had
not been functioning properly and that some oil had
gone down the shaft, The low seal leakage rate

 

35, M. DeCamp, ANP Quar. Prog, Rep. Dec, 10, 1955,
ORNL-2012, p 36.
previously reported® was, of course, invalidated.
The seal was not damaged. Since this test was
run without continuous purge-gas flow through the
system, hot NaK vapor apparently diffused up the
annulus around the shaft, contacted the lower
seal leakage oil, and formed the greasy deposits
shown in Fig. 2.5. Close examination has revealed
no damage to the pump,

The pump is designed so that oil leaking past
the seal will flow across the seal runner to the

PERIOD ENDING MARCH 10, 1956

shaft and down the shaft into the pumped fluid or
will flow down the side of the seal bellows and
into the oil catch basin, Qil is forced from the
catch basin by applying gas pressure to the surface
of the oil and forcing it up an annulus around one
of the tubes that carries coolant to the shield
plug. In this test it was possible, apparently, for
the gas to pass through the annulus without
carrying the oil over with it. Efforts are being
made to improve the catch-basin drainage system

UNCLASSIFIED

 

 

ORNL-LR-DWG 13085

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

450 !
400 |-
o
‘5660 "
‘5600 ©
350 ‘5“00 (et "
3309 o0
‘ m
300 ‘ L 30— o , cne
/ DESIGN POINT — / \
| .
_ \ ’ EFFICIENCY (%)
' | VOLUTE BALANCE LINE /“/ ‘
250 _
£ |
2 |
w i
200 i ]
150
100 -
50 - —
0 l i I
0 200 400 600 800 1000 1200 1400 1600 1800
FLOW (gpm}

Fig. 2.4. NaK-Pump Performance Test 7.

49
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
PHOTC 25418

 

UNCLASSIFIED
PHCTO 17449

 

Fig. 2.5. Disassembled ART Fuel Pump (Model MF-2), Showing Greasy Material That Formed When
Leakage Oil Contacted the Vapors from the NaK Being Circulated in the Pump at 1400°F. This pump
operated continuously for 2056 hr before disassembly,

50
to ensure the removal of all seal oil leakage and
to prevent oil from being slung from the seal onto
the shaft,

Short-circuit test stand No. 2 was ready for use
in November 1955, and another fuel pump was
installed, This system was filled with a fluoride
fuel mixture and brought to temperature for seal
tests, However, the lubrication oil pump was not
functioning properly, and the system was shut
down. An examination indicated that the sparging
impeller located just below the top pump bearing
was causing ingassing of the oil system, Also,
it was found that the UCON oil used in this system
retained the trapped gas to a greater extent than

did the Gulfspin 60 oil used in stand No. 1.

In an attempt to reduce the ingassing, the oil
was changed from UCON LB-300X to UCON LB-65
(the number of the oil indicates the SSU viscosity).
The stand was then again prepared for high-
temperature operation. After three days of oper-
ation, the lower seal developed a gross leak, and
operation was therefore stopped.  Examination
showed that the UCON oil had removed copper
from the system and deposited it on steel and
graphite surfaces, especially in the seal region,
Particles of the copper had entered the space
between the graphite seal and the seal runner and
caused the leak. Studies are now in progress to
find a more satisfactory oil. It appears that
Guifspin 60 oil will be acceptable if its heat-
stability and heat-transfer properties prove to be

PERIOD ENDING MARCH 10, 1956

satisfactory., The ingassing problem is also being
studied further, It may be possible to circumvent
the problem by using a larger oil reservoir or by
removing the sparging impeiler,

The tests conducted to date on the short-circuit
test stands are described in Table 2.4.

High-Temperature Pump-Performance Test Stands

J. J. W, Simon R. Curry
H. C. Young
Pratt & Whitney Aircraft

S. M. DeCamp
Aircraft Reactor Engineering Division

The design layouts and details and the fabrication
of components for two test stands for high-temper-
ature performance and endurance tests of ART fuel
pumps were completed. These loops, which were
described previously,® will be used to obtain
performance data on head, flow rate, cavitation
and vibration characteristics, and the functioning
of the xenon-removal system and to test endurance,

A similar stand for testing sodium pumps is
being fabricated, and stands for testing NaK pumps
and various special pumps have been designed.

 

6R. Curry and H. Young, ANP Quar. Prog. Rep,
Sept- 10, 1955' 0RNL']947' p 44.

TABLE 2.4, ART FUEL PUMP SHORT-CIRCUIT LOOP TESTS

 

 

Test Stand Hours Speed Temperature . . R n for Test
Date Started No. No. Type of Test Run (fpm) P Fluid Type of Gil e'?::minourfio:s Troubles Discovered
8-31-55 1 1 Endurance 168 2600 1400 NaK Gulfspin-60 Bad shield plug Bad heater and bad
O-ring seal
9-13.55 2 1 Endurance 26 2600 1400 NaK  Gulfspin-60 Bad shield plug
9.20.55 3 1 Endurance 2056 2600 1400 NaK Gulfspin-60 Increased power Qil leak to system
consumption
12-1.56 4 2 Seal 16 2600 1200 He  UCON-300 Oil system in-
gassing
12-16-56 5 2 Seal 96 2600 1200 He  UCON-65 Seal failure Copper plated out on
seal face
2.8-56 6 2 Seal 12 2600 1000 He Gulfspin-60 Seal failure

 

51
ANP PROJECT PROGRESS REPORT

HEAT EXCHANGER DEYELOPMENT

E. R. Dytko
Pratt & Whitney Aircraft

R. E. MacPherson J. C. Amos

Aircraft Reactor Engineering Division

Intermediate Heat Exchanger Tests

J. W. Cooke H. C. Hopkins
L. R. Enstice
Pratt & Whitney Aircraft

Test operations were continued on intermediate
heat exchanger test stands A and B. A summary
of the tests conducted is presented in Table 2.5.
York radiator No. 3 and Pratt & Whitney (PWA)
radiator No. 2 failed in stands A and B, re-
spectively, during the quarter. Both radiators
failed on the air-upstream side at or near the base
plate, The side plates of PWA radiators Nos. 1
and 2 were slit prior to initial operation (modi-
fication 1). The York radiator No. 3 side plates
were removed, and the support and base plates
were split (modification 2) prior to operation.
Eighteen thermocouples were installed on the tubes
of York radiater No. 3 to obtain data on individual
tube temperatures. The radiator is shown in
Fig. 2.6, with the thermocouples installed.
Temperature readings provided by these thermo-
couples indicated that, when NaK flow was stopped
and restarted, transient temperature conditions
occurred because of unequal NaK temperatures
throughout the loop. Typical transient temperatures
are indicated in Fig. 2.7. The tube temperatures
measured by thermocouples 1, 2, 5, and 7 varied
much more rapidly and through greater extremes
than did the bulk NaK inlet and outlet temperatures
measured by thermocouples 8 and 9.

York radiator No. 9, which is being installed in
stand A, is the first radiator fabricated according
to the revised design. It has no side plates,
support plates, or base plate., Nickel plates,
10 mils thick, with oversize holes provide top and
bottom air seals. It is hoped that these design
revisions will alleviate thermal-stress concen-
trations in the fin-tube matrix,

Cambridge radiators Nos. 1 and 2, modified to be
identical to York radiator No, 3 except for two
additional slits cut in the base plate parallel to
the air flow (modification 3), were installed in
stand B, and test operations were started. This
test was interrupted, however, by failure of the

52

 

Fig. 2.6. York Radiator No. 3, Showing Thermo-

couple Installations.

intermediate heat exchanger, ORNL-1 (type IHE-3),
which had operated for 1825 hr.

Construction of test stand C was delayed be-
cause of difficulties in procurement of the heat
exchanger, and therefore the Cambridge radiators
Nos. 1 and 2 originally procured for stand C were
installed in stand B. The decision was then made
to convert stand C to a test facility for prototype
ART radiators, Each radiator to be tested will
be one half of a complete ART radiator unit, The
modified test facility will be capable of testing
these radiators at approximately 85% performance,
The thermal stresses which will be encountered
by the ART radiators will be substantially
duplicated in these tests,

Small Heat Exchanger Tests

L. H. Devlin J. G. Turner
Pratt & Whitney Aircraft

Test operations continued on small heat ex-
changer test stand B, and stand C was placed in
£9

TABLE 2.5, SUMMARY OF OPERATION OF INTERMEDIATE HEAT EXCHANGER TEST STAND

 

Test Unit

Hours of Nonisothermal Total Hours of Number of

. R o . .
Operation Operation* Thermal Cycles eason for Termination

 

ORNL heat exchangers Nos. 1 and 2 {type IHE-2)
ORNL radiators Nes, 1 and 2

York radiators Nos, 1 and 2

York radiator No, 2

York radiator No. 3 (medification 2)

Circulating cold trap (modified diffusion)

Circulating cold trap (80-gal system)
NaK screen filter
Plugging indicator (5 hole, 0,030 mil)

Plugging indicator (7 hele, 0.030 mil)

ORML heat exchangers Nos, 1 and 2
PWA radiators Nos. 1 and 2 {modification 1)
Cambridge radiators Nos. 1 and 2 (modification 3)

Circulating cold trap (with precooler)

Circulating cold trap (80-gal system)

Plugging indicator

Test Stand A

15 358 73 Heat exchanger failed
70 621 8 ORNL radiator No. 1 failed
45 150 20 York radiator No. 1 failed
105 437 39 York radiator No. 2 failed
1 361 20 York radiator No, 3 failed
437 Replaced by 80-gal+system
cold trap
361 Test continuing
120 Test continuing
360 Test continuing
648 Test continuing

Test Stand B

993 1825 21 Heat exchanger failed
585 1195 20 PWA radiator No, 2 failed
408 603 1 Test continuing
1195 Replaced by 80-gal-system
cold trap
630 Test continuing
1825 Test continuing

 

*For tests in progress, the total operating time is shown as of February 15, 1956.

9561 ‘0L HOIVYW ONIONI Q0I¥3d
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DOWG 130488

 

 

 

 

 

 

 

 

 

 

 

 

 

1400 T
Nak PUMP RESTARTED
!
g | i
1000 2 \\
800 —— \ -
600
£ V
w
5
|q_: 400 m
r
w
a
Z |
ul |
200 : -

 

THERMOCOUPLE
LOCATIONS

 

 

 

 

 

 

 

O
o
o

TIME (min)

Fig. 2.7. Typical Transient Temperotures That

Resulted from Stopping and Restarting NaK Flow
Through York Radiator No. 3.

initial operation during the quarter. A summary of
operation of these stands, which were described
previously,” is presented in Table 2.6.

ORNL radiator No. 3 failed in stand B after
716 hr of operation. The failure occurred on the
air-upstream face of the radiator at or near the
center support plate., The design of this radiator
was identical to the design of ORNL radiator
Ne. 1 and York radiators Nos. 1 and 2, which
failed in intermediate heat exchanger test stand A,
except that the side plates were not welded
together (modification 1). This radiator is shown
before and after failure in Figs. 2.8 and 2.9.

York radiators Nos. 4 and 5 were modified by
removing the side plates and slitting the support

 

7). C. Amos, L. H. Devlin, and J. G. Turner, ANP
Quar. Prog. Rep, Dec. 10, 1955, ORNL-2012, p 49.

54

UNCLASSIFIED
Y6655

  

SUPPORT
PLATES

    

NOT
WELDED =

Fig. 2.8. ORNL Radiator No. 3 Prior to Installa-
tion in Small Heat Exchanger Test Stand B.

plates and base plate (modification 2) and are now
in operation in test stands B and C, respectively.
The installation of radiator No. 4 is shown in

Fig. 2.10.

Heot-Transfer and Pressure-Drop Correlations

J. C. Amos

Aircraft Reactor Engineering Division

Radiator air pressure-drop and heat-transfer data
obtained on the heat exchanger test stands operated
during the quarter have been added to previous
data® and are presented in Figs. 2.11 and 2.12,
The additional heat exchanger heat-transfer data

 

8R. D. Peak and J. C, Amos, ANP Quar, Prog. Rep.
Dec. 10, 1955, ORNL-2012, p 49.
 

Fig. 2.9.

PERIOD ENDING MARCH 10, 1956

Photo 25131 '

f‘\ ’;P’fl“
‘ UNCL ASSIFIED

ORNL Radiator No. 3 After Failure in Small Heat Exchanger Test Stand B.

55
TABLE 2.6. SUMMARY OF OPERATION OF SMALL HEAT EXCHANGER TEST STAND

 

Hours of

Tota! Hours

Number of

 

i Nonisoth ] R for Terminati
Test Unit oniso .erma of Operation* Thermal Cycles eason tor lermination
Cperation
Test Stand A**

ORNL heat exchanger No. 1 1540 1645 12 Test completed

(type SHE-1)
Test Stand B

ORNL heat exchanger No. 1 816 1710 27 Test continuing
(type SHE-2)

ORNL radiator Na. 3 295 716 5 Radiator failed
(modification 1)

Y otk radiator No. 4 520 1000 22 Test continuing
(modification 2)

Circulating cold trap (80-gal 1710 Test continuing
system)

Plugging indicator 1710 Test continuing

Test Stand C

ORNL heat exchanger No. 2 0 400 0 Test continuing
(type SHE-2)

Yotk radiator No, 5 0 400 0 Test continuing

{modification 2)

Circulating cold trap (80-gal

system)

Plugging indicator

Test continuing

Test continuing

 

*For tests in progress, the total operating time is shown as of February 15, 1956.

**Rebuilt for core-shell stability tests upon completion of one heat exchanger test.

and fluoride fuel pressure-drop data are in good
agreement with the previous data, Radiator NaK
pressure-drop data are presented in Fig. 2.13.
The NaK pressure drops measured during nearly
isothermal operation were observed to increase
when oxide plugging temperatures were above
900°F (370 ppm O,). Subsequent reduction of
contamination (plugging temperatures below 600°F,
100 ppm O,) by cold-trap operation reduced the
NaK pressure drops to the initial levels, In every
instance the NaK pressure drop gradually increased
when the system was operated with a continuous
temperature differential, For example, after
430 hr of nonisothermal operation, the NaK
pressure drop across Cambridge radiators Nos. 1
and 2 increased over 30%, while the pressure drop

56

across the heat exchangers in the same stand
increased only 6%.

All heat exchanger test stands are currently
equipped with 80-gal-system circulating cold
traps of the type previously described.? These
cold traps are capable of maintaining a contami-
nation level of 100 ppm 0,, as represented by a
plugging temperature of approximately 600°F, in
the 80-gal NaK system in the intermediate heat
exchanger test stands. On the small heat ex-
changer test stands, with approximately 20-gal
NaK systems, the cold traps are capable of
maintaining a contamination level that is below

 

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

Fig. 2.10.

PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
Photo 25333

SUPPORT PLATES AND
BASE PLATES SLIT.

 

York Radiator No. 4 as Installed in Small Heat Exchanger Test Stand B.

57
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR--DWG 13092

 

 

 

 

 

 

 

 

 

   

 

Of I CAMBRIDGE—1,-2 ~-
! [

 

AIR PRESSURE DROP PER TUBE ROW (in. OF Hs0
{CORRECTED FROM AVERAGE AIR TEMPERATURE TO 60°F)

 

PWA—1,—2- | |
i 1 | i
‘ | L]
0.05 : J
0.2 04 086 10 2.0 40 60 100

AIR MASS FLOW RATE THROUGH MINIMUM RADIATOR FLOW AREA (Ib/secflz)

Fig. 2.11. Correlation of Radiator Air-Pressure-
Drop Data Obtained from the Operation of Heat
E xchanger Test Stands.

UNCLASSIFIED
ORNL - LR-DWG 13083

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o] T T
’7%— CAMBIRJID(ESEE-]:(Z o l 1
. - ORNL-3 T \/ | - |
i
3
Z
2
3
: B
<o b bl *'l : ——
| L
- . ‘ ] ! ol
| L l "f( | J{ - j
osb | T - |

 

 

 

 

 

150 200 400 60C 80Q 1000 2000
AIR FILM REYNCLDS NUMBER

4000 6000

Fig. 2.12. Correlation of Radiator Heat-Transfer
Data Obtained from the Operation of Heat Ex-
changer Test Stands.

the sensitivity of the plugging indicators; that
is, the plugging temperature is below 350°F and
represents 34 ppm 02.

The plugging temperatures obtained from the
two plugging indicators installed on intermediate
heat exchanger test stand A were in substantial
agreement (average deviation, 5 ppm). These
indicators are identical except that one has «a

58

UNCLASSIFIED
ORNL~LR—DWG 13094

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

010 ! ‘ | —
. R - _ - . ‘ _\; _
S |
E ! HOURS OF !
a ‘ NONISOTHERMAL. | TOTAL HOURS
= CURVE RADIATOR QPERATION OF OPERATION
5 A PWA—1,-2 0 100
5 B " 100 530
& c " 500 1000
- D (WATER) " o 0
CO2 |-E (WATER) " 585 1200
F ORNL—1—2 0 100
G i 15 500
H ORNL—3 0 100
1 " 200 600
J YORK —1—2 0 100
K YORK—2 70 250
C YORK —4& o 160
oo M n 400 900 |
' | T T T
10,000 20,000 50,000 100,000

REYNOLDS NUMBER

Fig, 2.13. Correlation of Data on NaK Pressure
Drop Across Radiators Operated on Heat Exchanger
Test Stands,

plugging disk with five 0.030-in. holes and the
other has a plugging disk with seven 0,030-in.
holes.

STRUCTURAL TESTS
Ovuter-Core-Shell Thermal-Stability Test
G. D. Whitman A. M, Smith

Aircraft Reactor Engineering Division

R. Curry
Pratt & Whitney Aircraft

The core-shell test stand was put in operation
on December 28, 1955, and after 24 hr of operation
the system had to be shut down because of an
unaccountably high pressure drop in the loop
supplying sodium to the outer annulus of the
model, A recheck of the pressure-drop calcu-
lations for the outer-annulus circuit revealed no
significant error; so it was decided to x-ray the
model to check for a possible restriction, The
x rays disclosed that the outer thermal baffle had
moved up about ¥ in, and was restricting the flow
in the outer annulus (Fig. 2.14). The model was
therefore removed from the system and partly
disassembled in order to move the outer thermal
baffie back into place. To prevent a possible
recurrence of the trouble, four Inconel pins were
PERIOD ENDING MARCH 10, 1956

ORNL-LR-DWG {3095

  
    
  

EXIT
INNER ANNULUS
————-

 

ENTRANCE
QUTER ANNULUS
-

 

 

 

 

 

 

 

—INNER ANNULUS
7 5 — TEST SHELL
e " N —— QUTER ANNULUS

 

"THERMOCOUPLE

 

 

  

 

 

THERMOCOUPLE - i
AN,
N
A
S SPACER
A
iy N
A= — ——— ISLAND
\\\;\\
VAN — —— CORE SHELL HOUSING
THERMOCOUPLE ..__ WO
o .
L 1
\.\‘
7 , ~ THERMOCOUPLE

 

 

 

 

EXIT
QUTER ANNULUS
-———

———INCONEL PINS

 

 

 

 

 

 

O R

 

 

 

 

 

 

1-—-0UTER THERMAL BAFFLE

 

 

 

 

 

 

§ _L“; ‘
\L"fi- e ‘ 5
g L

Al

INNER THERMAL BAFFLE f 1

i i

Al A

Al g T

§ L

Al

i LA

ENTRANCE

INNER ANNULUS

Fig. 2.14. Core-Shell Thermal-Stability Test Assembly.

59
ANP PROJECT PROGRESS REPORT

attached to the thermal baffle to restrict its
movement to less than V in.

Repairs to the model were completed, and the
test was again started on February 2, 1956.
Isothermal conditions were established, and after
several days of checkout and operation the first
of the programed 100 thermal cycles 10 was applied
on February 7, 1956.

The cycle time has been increased from 1 hr to
approximately 2]/2 hr, with a 1-hr hold time under
temperature differential and isothermal conditions
and an approximately 15-min transient time to
and from each condition. To date, 55 such cycles
have been completed, Upon completion of the
thermal cycles the core shell will be examined for
dimensional stability,

Inconel Strain-Cycling Tests

J. C. Amos L. P. Carpenter

Aircraft Reactor Engineering Division

J. H. DeVan
Metallurgy Division

C. H. Wells
Pratt & Whitney Aircraft

Testing of Inconel specimens in o helium
atmosphere in the two anvil bending-test as-
semblies, described previously,!! was continued
to provide data for longer test periods. With
increased testing time, pronounced bending
occurred in the specimens at right angles to the
direction of the applied force, To eliminate the
side thrusts that caused this warpage, new anvils
were designed with slots, rather than flat surfaces,
on which the specimens are bent. Specimens
riding in these slots are strained to the same
curvature as in previous tests, but, since the slots
are the same width as the specimens, side move-
ment is constrained, Tests with up to 117 cycles
have been run successfully on the new anvil, with
specimen warpage confined to the small tolerance
between the siot and the specimen,

Automatic actuators are being installed to allow
pneumatic, rather than manual, operation of the
rods which tronsmit bending strains to the speci-
men., Since such actuators will eliminate a

 

10, W, Bell, ANP Quar. Prog. Rep. Sept, 10, 1955,
ORNL-+1947, p 52.

1y, c. Amos, J. H. DeVan, and C. H, Wells, ANP
Quar, Prog. Rep, Dec, 10, 1955, ORNL-2012, p 56.

60

possible source of torque transmittal to the
specimens, it is hoped that this modification will
further correct the problem of side thrusts.

Metallographic results have been received for
several of the specimens tested previously,
Cracks typical of those which initiate at the
outside fibers of the specimen and propagate
inward are shown in Fig., 2,15,

The metallographic examinations have verified an
earlier observation that the number of cracks
produced for a given number of cycles at 1% strain
is higher at 1400°F than at 1600°F, Results of
tests at 1200°F are not yet complete enough to
indicate a trend at this temperature,

The results of recent tests, based on macroscopic
examination, are summarized in Table 2.7. Several
tests which were run for longer times were omitted
from the table because of the definite warpage at
right angles to the induced bending direction,
which created a stress condition quite different
from that intended.

Swagelok Tubing-Connector Tests
P. G. Smith

Aircraft Reactor Engineering Division

Tests have been conducted to determine whether
Swagelok fittings can be used to replace welded
joints in forced-circulation loops operated with
fused salts at high temperature., A loop has been
operated with NaF-ZrF ,-UF, (50-46-4 mole %)
at 1300°F for 168 hr, at 1400°F for 200 hr, and at
1500°F for 100 hr. There have been no leaks in
the four Swagelok test pieces. In further tests the
reliability of the joints under thermal cycling of
the loop will be determined, On the basis of tests
to date, it appears that the commercial fitting can
probably be used in forced-circulation corrosion-
loop fabrication as a replacement for butt-welded
connections,

REACTOR COMPONENT DEVELOPMENT
Dump Valve

L.. P. Carpenter J. W, Kingsley
Aircraft Reactor Engineering Division
J. J. Milich
Pratt & Whitney Aircraft
The first prototype dump valve was tested with
the fuel mixture NaF-ZrF -UF , (50-46-4 mole %).

The initial tests with helium were described
PERIOD ENDING MARCH 10, 1956

 

Fig. 2.15. Inconel S$train-Cycling Specimen Showing Typical Cracks Which Initiate at the Outside and
Propagate Inward. 150X,

previously,'2 The plug material of the valve is
Kennametal 152B (64% TiC-30% Ni-6% NbTaTiC,)
and the seat material is Kennametal 162B (64%
TiC=25% Ni-5% Mo-6% NbTaTiCS). The vaive
was fabricated by Black, Sivalls & Bryson, Inc,
In the recent tests the fuel temperature was
1200°F, the actuator oil pressure was 250 psi,
and the stem thrust was 1220 psi. The estimated
seat contact pressure under these conditions
was 10,000 psi. The tests, to date, show that the
fuel leakage past the valve seat decreases with
time while the valve is closed and increases to a
high value when the valve is opened and reclosed,

The test apparatus consists of a storage reser-
voir and a sump tank connected by a dump line in
which the dump valve is installed. A small-
diameter riser line connects the storage reservoir
with a surge tank, and the fuel level in this riser

 

« 12y, 3, Milich and J. W, Kingsley, ANP Quar. Prog.
Rep., Dec. 10, 1955, ORNL-2012, p 59.

is detected with an adjustable spark-plug probe.
Leakage rates are determined from the fuel-level
fluctuations,

In the initial tests a pressure of 90 psi was
imposed on the fuel above the valve, and leakage
was such that the fuel level moved out of the
range of the adjustable probe in 5 min (leakage of
3000 to 4000 cm3/hr). In the following 64-hr
period, during which the valve remained closed
and only the fluid head was imposed on the valve
seat, leakage averaged 0.59 c¢m3/hr. Then the
initial pressure of 90 psi was again imposed on
the fuel above the valve, and during a subsequent
24-hr period the leakage averaged 0.91 cm3/hr,
The valve was then opened and closed, and the
leakage rate after it was closed was 30 cm3/hr.,

A mechanical noise was heard each time the
valve was opened after having been closed for
1 hr or more. Upon disassembly, evidence of
galling of the valve stem at the guides was found.

61
ANP PROQJECT PROGRESS REPORT

TABLE 2.7, SUMMARY OF RESULTS OF INCONEL STRAIN-CYCLING TESTS

 

 

Specimen Strain Number of Half-Cycle Specimen Temperature ‘

No. (%) ‘Cycles Time (hr) (OF) Macroscopic Observations
1C5 1 60 2 1400 No cracks

1D5 1 50 2 1400 No cracks

1G8 0.6 150 % 1200 No cracks

1H8 0.6 100 Y 1200 No cracks

1Cé 0.6 300 % 1600 No cracks

106 0.6 200 % 1600 No cracks

1G5 1 120 % 1200 No cracks

1H5 1 100 % 1200 No cracks

1K3 1 60 2 1200 Cracks observed
1L3 1 50 2 1200 Cracks observed
1 1 103 % 1400 No ceacks

10 1 117 % 1400 No cracks

 

There was no evidence, however, that the noise
was caused by galling of the valve seat or stem.
Visual examination indicated that the valve plug
had not been resting uniformly on the valve seat.
The plug had a distinct **wear ring,”’ as shown
in Fig, 2.16, where the sharp corner of the valve
seat gouged into the plug. The seat design has
been changed to provide a conically shaped seating
surface.  Metallographic examinations are being
made of the plug and the seat ring, as well as the
valve, to determine the effects of temperature and
cortrosion by the fuel mixture.

Cold Trap and Plugging Indicator

J. J. Milich R. D. Peak
Pratt & Whitney Aircraft

The test assembly for cold-trap and plugging-
indicator evaluation, previously described,!3 was
completed, and shakedown tests were started,
The shakedown period has been prolonged, because
the original design of the stand had many faults,

 

135, 3. Milich, ANP Quar. Prog, Rep. Dec. 10, 1955,
ORNL-2012, p 60.

62

including gas pockets, which made loop filling
difficult, and twe free NaK surfaces, which created
a very difficult operating problem. In efforts to
solve the shakedown troubles, the piping has been
modified, thermocouple and electric-heater wiring

GNCL ASSIFIED
Photo 25209

INCHES

 

 

Fig. 2.16. Dump-VYalve Plug After Testing with
NaF-ZrF UF , (50-46-4 mole %) at 1200°F. (Seewse

with caption)
has been changed, new liquid-level indicators have
been provided for the surge tanks, and a new
helium system and a gas-control panel have been
installed. To further the development, a second
stand has been designed and is now being built
that will be free of almost all the troubles dis-
covered with the first stand and will provide much
greater flexibility in operation and test methods.

Zirconium Fluoride Yapor Trap

J. J. Milich
Pratt & Whitney Aircraft

The ZrF ,-vapor-trap test system, which was
described previously, 4 was operated for 795 hr at
an average sump temperature of 1300°F and an
average exit-line temperature of 1397°F, with an
average helium flow of 3.75 liters/min and a
minimum trap temperature of 115°F. Analyses of
the exhaust gos showed a carry-over of 0.04 g
of zirconium per liter and 10 ug of HF per liter,
This shows that the trap removed ZrF, vapor
under these test conditions, but further tests are
needed to validate the results for application to

ETU and ART design. These test results are

 

145, J. Milich and J. W. Kingsley, ANP Quar. Prog,
Rep. Dec. 10, 1955, ORNL-2012, p 60.

PERIOD ENDING MARCH 10, 1956

clouded somewhat by the uncontrolied conditions
that resulted when failure occurred, The failure
was caused by corrosion of a tube in the trap.
An examination of the dismantled system revealed
fuel constituents in the entrance to the trap. The
fuel was probably entrained because of the high
gas flow and the agitation of the fuel sump at the
time of failure, Some of the tubes at the beginning
of the cooling section were plugged with ZrF4,

Two tests are now in progress to gather funda-
mental data concerning ZrF, vapor traps. in the
first test, helium scturated with ZrF . from
NaF-ZrF ,-UF , (50-46-4 mole %) is passed through
a 4-in, pipe, 4 ft long, which is filled with York
Demister packing. The exhaust gos is metered
and is analyzed for ZrF, and HF. The efficiency
of the trap at various trap temperatures with
constant gas flow will be determined.

The second test unit consists of a 2-ft length
of 1-in. pipe filled with an adsorbent (NaF pellets)
through which the ZrF ,-saturated helium is forced.
The ZrF , is adsorbed on the NaF pellets. Oper-
ating temperatures are limited by the melting point
of the adsorbent. The efficiency of the trap at
constant temperature and various flow rates will
be determined by metering and analyzing the
exhaust gas,

63
ANP PROJECT PROGRESS REPORT

3. CRITICAL EXPERIMENTS

A. D, Callihan

J. J. Lynn
Applied Nuclear Physics Division

D. Scott, Aircraft Reactor Engineering Division

E. Demski
W. J. Fader

E. V. Sandin
S. Snyder

Pratt & Whitney Aircraft

COMPACT-CORE REFLECTOR-MODERATED-
REACTOR CRITICAL EXPERIMENTS

The study of the sodium-cooled reflector-mods=r-
ated reactor with solid fuel proposed by the
Nuclear Development Corporation of America
(NDA)! has been continued. The room-temperature
mockup of the reactor was described previously?
as having a critical mass of 31 kg of U235, The
foading in the experiment has been increased to
33 kg by the addition of 2%-in.-dia by 0.01-in.-
thick disks in one section of the fuel region.
Although the over-all yranium density is uniform,
the local distribution in this one section differs
from that in the other sections, which are loaded
with 0.004-in.-thick sheets., This addition to the
loading gives some excess reactivity for use in
measurements of the assembly.
the neutron-flux distribution and the mass-reactivity
coefficients of various reactor component materials
are symmarized below. The measurements have
been reported in detail and correlated with the
predicted reactor parameters by NDA,?

Measurements of

 

1CCR-Z.' A Compact Core Reactor for Aircraft Pro-
pulsion, NYO-3080 (July 30, 1954).

A, D. Callihan et al., ANP Quar. Prog. Rep., Dec. 10,
1955, ORNL-2012, p 73.

3Q:.xu:lrtejv'ly Progress Report, ANP Reactor Develop-
ment, October 1, 1955 through December 31, 1955,
NDA-20, p 35 (Jan. 23, 1956).

Neutron-Flux Distribution

The neutron-flux distributien in the assembly
was measured with both bare and cadmium-covered
indium foils along one diametric and two longi-
tudinal traverses, The diametric traverse, made
at the mid-plane, was extended 203{1 in. above
and below the axis so that any asymmetry re-
sulting from wvariations in the vranium loading
could be observed. The data, shown in Fig. 3.1,
indicate that no significant differences exist in
the flux patterns adjacent to the two fuel sections,
One longitudinal traverse was made over one half
of the reactor axis; the other was made adjacent
to the core shell in the fuel-region mockup. The
results are plotted in Figs. 3.2 and 3.3.

Mass-Reactivity Coefficients of
Reactor Component Materials

Samples of various materials of interest in the
reactor structure were placed at a number of
locations in the assembly, and their effects on
the reactivity were noted from the critical positions
of calibrated control rods. The samples included
type 316 stainless steel, sodium, boral, uranium,
beryllium, iron, molybdenum, niobium,
Hastelloy B, and nickel. The sizes of the samples,
their locations within the assembly, and their
effects on the reactivity of the critical assembly
are shown in Figs. 3.4, 3.5, and 3.6.

Inconel,
69

RELATIVE ACTIVITY (orbitrary units)

o 15 PO

-y
ORNL~LR—DWG 13133

 

 

 

 

  
     
 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

  

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.0 ‘ : 1
| | | ‘_ |
| | e | |
18 ————+—= o= i i —_t l -- : — ‘
o BARE INDIUM FOIL 1 ; ' i i \
o CADMIUM-COVERED INDIUM FOIL ‘ | ‘//-f-"'—"— 0-03"”-|‘TH'CK ?TA'NLESS STEEL SHELL
16 |-— ; S e 4+t -t e - i i — — 1
| - | 0.06-in-THICK
; ! | —STAINLESS STEEL SHELL
o ‘ _ z 1 o o L
— 1 o
BERYLLIUM REFLECTOR | FUEL REGION FUEL REGION BERYLLIUM REFLECTOR
1.2 — | l i
o | | |
| ~CADMIUM FRACTION CADMIUM FRACTION-—
1.0 | :
"F 7
0.8 - — - - A
0.6 — ‘\ \
0.4 — \\\\\ K
o2 ///’ ] ‘ | ‘ "
o | | BERYLLIUM ISLAND | ~ \
oL et | T 1 ~ 3
e —20 -8 —I8 14 2 -0 -8 -6 -4 —2 0 2 4 18 20 22

Fig. 3.1. Neutron-Flux Distribution Along the Mid-Plane of the Compact-Core Reflector-Moderated-Reactor Critical Assembly,

RADIAL DISTANCE FROM AXIS (in}

 

1.00

0.80

0.60

040

0.20

CADMIUM FRACTION

9661 ‘0L HOY VYW 9NIAN3 gOly¥3d
ANP PROJECT PROGRESS REPORT

Fig. 3.2. Neutron-Flux

2.0

RELATIVE ACTIVITY (arbitrary units)

04

0.2 t

o
w

o
o

 

o~

N

o

 
   
    
 

ORNL—LR—DWG 13134

1

 

—— L.

| © BARE INDIUM FOIL
|« CADMIUM-COVERED INDIUM FOIL
i | -

 

 

CADMIUM FRACTION

 

 

Critical Assembly.

RELATIVE ACTIVITY larbitrary units)

o7

0.6

o
w

o
B

©
N

<
o

oA

  

 

 

 

i

‘
6 8 10 12
DISTANCE FROM MIDPLANE (in.)

Distribution Along the Axis of the Campact-Core Reflector-Moderated-Reactor

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SN
ORNL-LR-DWG 13435
- T
FUEL _}— HEADER—-—-{
|
\
A BARE INDIUM FOIL
® CADMIUM-COVERED INDIUM FOIL 1.0
— CADMIUM FRACTION |
! A
_.._/
V4 0.8
g
‘/ .
i——i* - L3 0.6
A/‘ -—._‘_____¥ . ‘_—/ E
® ra
y / =
. ° jn—/ L —oa 2
° > i o
e J—— =<1
/—-——————‘ °
0.2
‘ 0
0 i 2 3 4 5 6 7 8 9 10 1

DISTANCE FROM MIDPLANE (in.)

Fig. 3.3. Neutron-Flux Distribution Along the Fuel-Reflector Interface of the Compact-Core Reflector-

Moderated-Reactor Critical Assembly.

66
PERIOD ENDING MARCH 10, 1956

ORNL-LR-DWG 13437

 

« \ ,

 

ref————|SLAND ———®="~mt— FUEL REGION —— g -—BERYLLIUM REFLECTOR o

 

 

 

 

 

 

30 - ; - - cem - - e

 

 

 

 

A INCONEL, 140 g, 2% x 275 x 043 in.

O IRON, 103.8 g, 2% x 2% x 040 in.

A MOLYBDENUM, 148.5 g, 27 x 2% x 040 in.
a

0

o

 

 

 

NIOBIUM, 94 g, 2% x 2% x 0.082 in.
HASTELLOY B, 410.6 g, 27 x 2% x 0090 in.
NICKEL, 122 g, 2% x 275 x 0104

 

 

 

 

 

 

 

LOSS IN REACTIVITY PER 100 g CF SAMPLE (cents)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ‘it :
\[1 ‘
| . | o - _
: v ’% ‘ '\j\\o\\
0 ’ ‘ T%—‘D ‘ i \.
0 5 10 15 20 25

SAMPLE POSITION; RADIAL DISTANCE FROM AXIS (in.)

Fig. 3.4. Reactivity Change Effected by the Addition of Yarious Materials in the Compact-Core Reflector-
Moderated-Reactor Critical Assembly.

67
LOSS IN REACTIVITY PER 100 g OF SAMPLE (cents)

30

25

n
o

o

o

ORNL—LR—DWG 13136

 

   
   
 
  
  

 

>

»

i
J-—MULTIPLY SCALE BY {00

FUEL BERYLL!IUM
J_REGION_| :  REFLECTOR

 

 

 

  

| |
TYPE 316 STAINLESS STEEL, 49.06 q,

2% x 1% x 0.086 in.

TYPE 316 STAINLESS STEEL, 525 g,
4%g x 10Y5 x 0.096in.

SODIUM, 12¢73 g,
2T % 2% x 1in.

BORAL (165 mg OF BORON /em?), 4.4 g,
27/8 X '-016 X '/8 in,

i e —

|
|
| 4
|

 

 

 

 

1

 

25

ISLAND _|
O L
O 5
SAMPLE POSITION; RADIAL DISTANGE FROM AXIS (in)
Fig. 3.5, Reactivity Change Effected by the

Addition of Various Materials in the Compact-Core
Reflector-Moderated-Reactor Critical Assembly.

68

ORNL—LR—OWG 13138

 

 

 

 

  

 

 

 

 

 

25 I :
' i |
- | © URANIUM, 27.47 g OF uese,
S 4%¢ x 2% x 0.008in.
£ 20 5 1
o ® BERYLLIUM, 25tq,
a 2 x 2% x lin,
S B }v(I g x1in
< i :
« ! i
5 15 Lo
o BERYLLIUM |
o 'REFLECTOR!
Q T I
o« i
b i
L ;
. 10 ———————— —|
t i
> o i
= . |
2 i‘. |
Ll \ i ‘e \
Sl N
= “-DIVIDE SCALE BY 10 |
<t '\.
© FUEL
| _ISLAND | REGION
0 |
0 5 10 15 20 25
SAMPLE POSITION; RADIAL DISTANCE FROM AXIS (in)
Fig. 3.6. Reactivity Change Effected by the

Addition of Various Materials in the Compact-Core
Reflector-Moderated-Reactor Critical Assembly.
Part lI

MATERIALS RESEARCH
Wiy

4, CHEMISTRY OF REACTOR MATERIALS

W. R. Grimes

Materials Chemistry Division

PHASE EQUILIBRIUM STUDIES!

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

Materials Chemistry Division
H. Insley, Consultant

Phase equilibrium studies of fluoride fuel mix-
tures were continued. A material containing 63.5
mole % NaF, 18.0 mole % ZrF,, and 18.5 mole %
UF, is being investigated that shows some promise
as a substitute for the high-melting-point compound
Na,UF as the fuel concentrate for “‘enrichment™
of reactor systems.

Examinations of the binary systems KF-ZrF, and
RbF-ZrF , are virtually complete, and examinations
of ternary solvents that involve these and other
binary systems are well under way, Extensive low-
melting-point areas in the NaF-RbF-ZrF, triangle
are apparent, and it appears likely that useful fuel
systems may be developed by the addition of UF .

Examinations of relatively simple systems of
alkali-metal and alkaline-earth fluorides are pro-
ceeding rapidly. While it is apparent that mixtures
with genuinely low melting points will not be
achieved with these materials, their possible
at any high temperature seems to
warrant some additional study. Preliminary investi-
gation of the alkali-metal-fluoride—CeF, binary
systems has disclosed no melting points below

about 650°C,
The System UF“-UO2
R. J. Sheil

Materials Chemistry Division

application

Uranium dioxide has frequently been observed as
an impurity in fuel mixtures containing UFA, and,
since there is very little published information on
the UF ,.UQ, system, it appeared to be desirable
to conduct a limited investigation of the UF4-rich
part of the system. As the starting point in this

study, an attempt was made to determine the

 

The petrographic examinations reported here were
performed by D. White, Metallurgy Division, ond
T. N. McVay and H. Insley, Consultants. The x-roy
examinations were performed by R. E. Thoma and
B. A. Soderberg, Materials Chemistry Division.

freezing point of pure UF4. The presently ac-
cepted freezing point of 1036°C was determined by
Ryon and Twichell.?2 They corrected for the effect
of UO, in the UF, used for their freezing-point
determination by measuring the freezing point of
several UF ,-UQ, mixtures and extrapolating a plot
of log mole fraction of UF, vs the reciprocal of
absolute temperature.  The measured freezing
points of two recent UF, preparations averaged
1026°C. One sample was purified by treatment
with  NH,F.HF and the other by bubbling HF
through liquid UF, at 1100°C. A correction of
+2°C should be applied to correct for the UO,
content of the samples (0.18 wt % average, by
chemical analysis) according to the above-mentioned
plot of Ryon and Twichell. The corrected value
of 1034°C agrees quite well with the published
figure,

Cooling curves obtained with UF ,-UO, mixtures
containing approximately 2 and 4 wt % UO, gave
inconsistent results, In contrast to the nearly
pure UF, preparations, which showed little tendency
to undercool, it was difficult to aveid undercooling
these mixtures. A mixture containing 2.2 wt % UO
showed a liquidus temperature of 1019°C, with the
slowest cooling rate, while other cooling curves
obtained with the same mixture indicated liquidus
temperatures of 1006 to 1009°C. Ryon and Twichell
reported a freezing point of 1017°C for a mixture
containing 2.06 wt % UO,. One cooling curve
obtained with a mixture containing 4.1 wt % UO
showed no thermal effect between 1135 and 985°C,
while a second cooling curve (starting at 1095°C)
obtained with the saome mixture showed thermal
effects at 1038, 990, and 974°C. Petrographic and
x-ray diffraction examination of the slowly cooled
melts showed no evidence of solid solution or
compound formation between UF  and UO,, but the
possibility of solid-solution formation at elevated
temperatures followed by ‘‘exsolution”’
temperatures cannot be ruled out by the available
data.

at lower

 

25, D. Ryon and L. P. Twichell, Vapor Pressure and
Related Physical Constants of Uranium Tetrafluoride,
H-5.385.2 (July 25, 1947).

71
ANP PROJECT PROGRESS REPORT

The System NaF.UF

R. E. Moore
Materials Chemistry Division

A compound observed in slowly cooled melts in
the system NaF-LiF-UF, has been found to be the
subsolidus binary compound NaF.:2UF, It was
identified (x-ray and optical examination) by ob-
taining a single phase with 66.7 mole % UF ,-33.3
mole % NaF in quenched samples held for 64 hr at
well below the solidus temperature (680°C). On
heating to 660°C, the compound NaF2UF de-
composes into 7NaF-6UF,. and UF,. The ap-
pearance of NaF-2UF, in the slowly cooled melts
of several indicates that the
compound has a primary phase field in these
ternary systems.

ternary systems

The System KF-ZiF ,
R. P. Metcalf

Materials Chemistry Division

Quenches in the system KF-ZrF, show that the
compound 3KF-2ZrF, melts incongruently ar 417°C
to liquid and 2KF-ZrF,. The liquidus is at 438°C.

Two forms of KF.ZrF, have been observed, one
of which is optically similar to **R-3"’ metastable
NaF.ZrF,. The marked undercooling of slowly
cooled melts has suggested the likelihood of one
of these forms being unstable. Seeding experi-
ments, in which a crystal of R-3 type KF.ZrF, is
dropped into a molten undercooled melt, produced
only R-3 type KF-ZrF, on cooling. Quenches
below 461°C show only R-3 type KF.ZrF,. The
normal KF.ZrF, is probably metastable and has
no region of stability in the phase diagram.

The System RbF-ZrF
H. A, Friedman R. J. Sheil

Materials Chemistry Division

A detailed study of the system RbF-ZrF, by
thermal analysis and petrographic and x-ray dif-
fraction examinations of quenched and slowly
cooled samples was initiated. The system diagram
presented in Fig. 4.1 must be considered to be
tentative, since more information is needed to
clear up questionable areas. A previously published
diagram of this system®
thermal analysis.

was based solely on

 

3L. M. Bratcher, R. E. Traber, Jr., and C. J. Barton,
ANP Quar. Prog. Rep. June 10, 1952, ORNL-1294, p 92,

72

ORNL-LR-DWG 12778

 

   

 

 

 

 

 

 

 

1009 T T T
900 Ny
| ™
P\
~ 800
g b\
Ll ; !
£ 700 i
- R
§ e00 ,!' \ E o
o) wIlF Ay ¥
» ~ ,' 11 LL\N
500 o : \ @ L_t.l:5
O |I Oy OF
|l 4
400 l;! - ‘
f
. | e

 

 

 

 

RbF {0 20 30 40 50 60 70 80 90 ZrF4
Zrfy {mole 7o)

Fig. 4.1. Tentative Phase Dicgram of the System
RbF-Z:F .

The melting point of 3RbF-ZrF , was determined
by thermal analysis. No inversions are evident
from either thermal analysis or gradient quenching.
Thermal analysis and visual observation show
clearly that 2RbF.ZrF, melts incongruently at
about 620°C and that the liquidus is near 750°C.
A very sharp and energetic thermal-analysis halt
occurs at 440 to 450°C. Small samples which were
annealted at temperatures above 460°C before being
quenched contained poorly formed crystals of
2RbF.ZrF ,, while those annealed at temperatures
below 460°C contained well-formed crystals. The
evidence is clear that a very rapid inversion oc-
curs at 460°C. This conclusion has been con-
firmed by visual observation. The compound
3RbF-ZrF4 could not be found in samples of
2RbF.ZrF, which were annealed and quenched
from above the incongruent melting point. Further-
more, a series of quenches of a composition
containing 28 mole % ZrF, over the temperature
range 545 to 679°C contained nothing but ex-
tremely fine-grained material, which was not
identifiable by petrographic examination but which
was shown by x-ray diffraction to be a mixture of
3RbF.-ZrF, and 2RbF.ZrF,. In order to explain
these results, it is necessary to assume a limited
solid-solution region (dotted line) for 3RbF.ZrF,
and to assume further that this solid solution,
which would be the primary phase at 28 and 33.3
mole % ZrF ,, undergoes very rapid *‘exsolution”
on cooling.

Acgording to thermal-analysis results and the
examination of quenched samples at 36 mole %
ZrF,, the 2RbF.ZrF, inversion appears to take
place at a lower temperature (410°C) in composi-
tions containing more than 33.3 mole % ZrF .
Limited solid solutions of a2RbF-ZrF, and
B2RbF-ZrF, must be postulated to explain this
lowering of the inversion temperature. A shift in
the x-ray pattern for an annealed and quenched
sample of a 36 mole % ZrF, composition at 393°C
has confirmed the B2RbF-ZrF, solid solution,

The compound 5RbF.4ZrF, was found by petro-
graphic recognition of almost single-phase material
in slowly cooled preparations containing 44.5 mole
% ZrF,. Petrographic examinations of quenched
samples containing 36 and 40 mole % ZrF, show
that 2RbF.ZrF, ond 5RbF.4ZrF, are the primary
phases, respectively, and that the eutectic between
the compositions occurs at a temperature near
375°C. Cooling curves obtained in this region
show, in general, a hait at about 340°C, which is
possibly the undercooled eutectic halt, but this
point is still in question.

The liquidus temperature and the inversion of
RbF.ZrF, were established by petrographic ex-
amination of quenched samples. Slowly cooled
mixtures containing 50 mole % ZrF , often contain
some OSRbF.4ZrF,, but annealed and quenched
samples contain only single-phase material. The
dotted line at 50 mole % ZrF, indicates a meto-
stable extension of the liquidus curve of 5RbF.
4ZrF,. The eutectic between 5RbF.4ZrF, and
RbF-ZrF , has not yet been located exactly.

Samples of .a composition containing 57 mole %
ZrF , which were annealed and quenched from below
the solidus contained only RbF.ZrF, and ZrF ,
and thus the absence of compounds with more than
50 mole % ZrF , is indicated.

The System NaF-ZrF -UF ,
L. M. Bratcher

Materials Chemistry Division

Thermal-analysis data obtained previously* were
checked for a few compositions in the system
NaF-ZrF ,-UF ,, and the solid phases were examined
in an effort to find a low-melting, nonsegregating
composition with @ high concentration of UF, that
might make a suitable substitute for Na,UF . as
the enriched-uranium carrier for reactor startup

 

4C. J. Barton et al., ANP Quar. Prog., Rep. Sept. 10,
1954, ORNL-1771, p 55. i

PERIOD ENDING MARCH 10, 1956

operations,  The most promising mixture, NaF-
ZrF ~UF, (63.5-18-18.5 mole %), showed only a
single break on the cooling curves. The break was
607°C, which is about 50°C lower than the melting
point of Na,UF,. The crystalline phases indi-
cated, by petrographic and x-ray diffraction studies,
to be present in the slowly cooled melt were
Na,ZrF ., Na,U,F4,, and NayZrF,. The last two
phases showed evidence of a slight amount of
solid solution. It appears from the available data
that this mixture is an invariant point in the system,
and therefore it should be nonsegregating. It has
a calculated uranium content of 1.44 g/cm? at
700°C, as compared with 2.75 g/cm? for I\Ic12UF6

at the same temperature.

The System NaF-KF-ZrF

H. A. Friedman
Materials Chemistry Division

H. Davis
Pratt & Whitney Aircraft

Phase relationships are now being studied in the
system NaF-KF-ZrF, by using small quenched
samples, Data obtained with slowly cooled melts
indicate that there is a congruently melting com-
pound 3NaF.KF.2ZrF,, which has a melting point
of 530°C and a lower inversion temperature, The
data for quenched samples show that, at the 590°C
liquidus, B-1 2NaF.Z:F, is the primary phase.
Two phases, formerly believed to be enantiomorphs
of 3NaF.-KF.2ZrF,, coexist below 500°C and are,
therefore, two distinct incongruently melting
compounds of undetermined composition.

The compound near the composition NaF-KF-
ZrF , (24-33-43 mole %) has not yet been identified
in quenched samples, but it has been observed
below 400°C, which is its incongruent melting
point. The primary phase at the 425°C liquidus
is 3KF.2ZrF . ‘

Quenches have shown that the primary phase
field of 3KF.ZrF, extends over both 3NaF.3KF.
27rF, ond NaF.KF.ZrF,. The liquidus at NaF.
KF.ZrF, is 557°C, and the incongruent melting
point is 550°C.

The System NaF-RbF-ZrF
H. A. Friedman

Materials Chemistry Division

Large mélts of approximately 500 to 700 g are
used, in general, when thermal data are to be taken

73
ANP PROJECT PROGRESS REPORT

from stirred, slowly cooled melts in order to
achieve sensitive detection of small thermal ef-
fects. Large samples are not economically feasible,
however, when RbF is one of the constituents, and
a new opparatus arrangement was devised for use
with samples of about 50 g. Four samples can be
melted simultanecusly in the heating chamber of
the new apparatus. The thermocouples, rather than
being contained in the conventional ¥ -in.-dia
nickel-tubing stirrer (0.065-in. wall thickness), are
contained in Y-in.-dia nickel or Inconel thermo-
couple wells with wall thicknesses of 0.006 to
0,012 in. The thin walls of the thermocouple wells
permit the detection of small thermal effects with
a sensitivity comparable to that obtained by using
large samples.

The system NaF-RbF-ZrF, has several charac-
teristics in common with the system NaF-KF-
ZrF ,; that is, the liquidus surfaces of both systems
have similar profiles and temperatures; there is a
1: 1: 1 ternary compound in each system; and the
lowest melting region of each system contains
approximately 40 mole % ZrF, at high RbF or KF
concentrations,  The NaF-RbF-ZrF, system is
distinguished from the NaF-KF-ZrF, system in
that few ternary compounds in the NaF-RbF-ZrF
system have analogs in the NaF-KF-ZrF , system,
and the eutectics in the NaF-RbF-ZrF , system are
located, in general, closer to the RbF-Z¢F, binary
than the corresponding eutectics in the NaF-KF-
ZrF , system are to the KF-ZtF , binary.

The existence of two ternary compounds in the
system NaF-RbF-ZtF, has been established; their
formulas are NaF.RbF.ZrF, and NaF.RbF.2ZrF,.
The former has an analog in the system NaF-KF-
ZrF ,; the latter does not. Both are congruently
melting compounds. A third ternary compound
exists at or near the composition 3RbF.3NaF.
4ZrF,. It hos a melting point of approximately
450°C, The melting relationships of the compounds
have not yet been established.

The following joins on the phase diagram of the
system are sides of compatibility triangles:

NaF-3RbF.ZrF ,

3NaF.ZrF ,-3RbF.ZrF
3NaF-ZrF ;-RbF-NaF.ZrF
3RbF.ZrF -RbF.NaF.ZrF
2NaF.ZrF ,-RbF-NaF-ZrF
2RbF.ZrF -RbF:NaF.ZrF

74

INaF-6ZrF ,-RbF:NaF.2ZrF
RbF.ZrF ,-RbF:NaF.2ZrF
Of these joins, the quasi binaries are:

Nc:F-?:Rl:aF-ZrF4 [eutectic, NaF-RbF-Zer
(33.5-50-16.5 mole %); mp, 777°C]

3NuF-ZrF4-3R|:)F-Z|'F4 eutectic, NaF-RbF-
ZrF4 (50-25-25 mole %); mp, 738°C]

7NaF.6ZrF -RbF-NaF.2ZF,
RbF-NaF.2ZrF -RbF.ZrF

The primary phase field of RbF.NaF-2ZrF , appears
to be quite small across the latter two quasi
binaries.

The System RbF-BeF,

L. M. Bratcher R. E. Meadows
R. J. Sheil

Materials Chemistry Division

Previous thermal-analysis studies of the system
RbF-BfiF2 gave such scattered data, probably
because of the presence of variable amounts of
oxygen in the melts, that the results were not
published. Data on three points in the BeF rich
part of the system were found in the literature,
and thermal-analysis studies were resumed during
the quarter with mixtures covering the range 5 to
67.5 mole % Ber. Most of the mixtures were
prepared by adding RbF or BeF, to portions of a
purified 50-50 mole % mixture. This procedure
appeared to give reproducible mixtures.

The phase relationships in this system are
quite similar to those in the KF-Bef, system.®
Four compounds have been tentatively identified
on the basis of thermal-analysis and solid-phase
studies. The compound Rb,BeF , melts congruently
at approximately 800°C, and Rb BeF . melts at about
725°C. The compound RbBeF, melts incongruently
at 455°C to give RbBeF, and liquid. Quenching
data will be required to determine the melting
behavior of the compound RbBe,F.. The lowest
melting composition in the system is the RbBeF -
RbBe,F. eutectic, which contains approximately
60 mole % Ber and has a melting point of 375 t
10°C. This mixture is, however, too viscous to be

 

5D, M. Roy, R. Roy, and E. F. Osborn, J. Am. Ceram.
Soc. 33, 85 (1950).

SL. M. Bratcher et al, ANP Quar. Prog. Rep. Dec.
10, 1955, ORNL-2012, p 84.
of practical interest, even though the viscosity of

liquid RbBeF 5 is less than that” of KBeF .

The System NaF-BeF .UF
C. J. Barton

Materials Chemistry Division

Viscosity data obtained previously for the
composition NaF-BeF ,-UF , (70-26-4 mole %) showed
rather high values (10 centipoises at 600°C and
4,1 centipoises at 800°C),% but data obtained
more recently from the same source gave lower
valves? of 6.3 aond 2.6 centipoises at 00 and
800°C, respectively. It the values are
confirmed, the kinematic viscosity of this mixture
is competitive with that of an alkali-metal fluoride
mixture with a comparable uranium concentration.
The calculated uranium content of this mixture is
0.42 g/cm?® at 700°C, which is about the same as
that of an NaF-ZrF,-UF, mixture containing 5.5
mole % UF,. The liquidus temperature was not
reported for this exact composition, but filtration
data® indicate that the liquidus temperature is
probably between 545 and 550°C. |t is possible
that further exploration of the ternary system in
this region may demonstrate the existence of
suitable fuel compositions with
liquidus temperatures.

lower

slightly lower

The System NaF.LiF-BeF,

L. M. Bratcher R. E. Meadows
R. J. Sheil

Materials Chemistry Division

Viscosity data reported in Sec. 7, ‘‘Heat Transfer
and Physical Properties,’”’ for the composition
NflF-LiF-BeF2 (49-36-15 mole %) confirm the
conclusion expressed previously'9 that lowering
the BeF, content of melts in this system to below
23 mole % would produce little improvement in

viscosity. This finding is supported by viscosity

 

7S. 1. Cohen, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL-1947, p 155,

81, F. Eichelberger, Progress Report on the NaF-BeF ,-
UF'4 Ternary Phase Diagram, MLM-CF 55-11-14 (Nov.
7, 1955).

). F. Eichelberger,
Activities Report for December 13,

56-1-46 (Jan. 18, 1956).

10, M. Bratcher et al., ANP Quar. Prog. Rep. Dec.
10, 1955, ORNL-2012, p 80,

Mound Laboratory Technical
1955, MLM-CF

PERIOD ENDING MARCH 10, 1956

data reported by Mound Laboratory®? for NaF-
BeF,, LiF-BeF,, and NaF-BeF,-UF, mixtures,
which show that little improvement in viscosity
results from lowering the BeF. concentration in
the melts to below approximately 30 mole %, where
all the BeF, would be expected to be complexed
as BeF,” ion. A study of viscosity-composition
relationships in the ternary system was started
during the quarter, but the available data do not
appear to be encouraging with regard to finding a
low-melting fuel carrier having a kinematic vis-
cosity as low as that of alkali-metal fluoride
mixtures.

The solid-phase studies have been confined,
to date, to the section of the system having LiF,
NaF, and NaBeF, at the apexes of the triangle.
Quenching studies and examinations of slowly
cooled melts have shown that the join LiF-NaBeF,
is a quasi binory which has a ternary compound
Na,LiBe,F, that melts congruently at 355°C. The
NaBeF ,-Na,LiBe,F, eutectic appears to be near
the composition NaF-LiF-BeF, (44,5-11-44.5 mole
%), with a melting point of 332°C. The LiF-
Na,LiBe,F, eutectic has not been located, but
thermal-analysis data indicate that it must be
close to the ternary compound.

As previously reported,!! it has not been pos-
sible to confirm the existence of the compound
Na,LiBe,F, reported by Jahn,'? but considerable
effort was exerted in an attempt to identify a
ternary compound that has appeared in both quenched
and slowly cooled melts in the NaF-rich part of
the Na,BeF ,-Li,BeF, join as an isotropic or very
slightly birefringent crystalline phase. A slowly
cooled mixture having the composition NaF-LiF-
BeF, (57-9.7-33.3 mole %) contained the unknown
compound, as the major phase, and a small amount
of Na,LiBe,F,. Although this isotropic phase
has not been conclusively identified, its composi-
tion must be near that of the 57-9,7-33.3 mole %
mixture, and the formula LiF.7NaF.4BeF, has
been tentatively assigned.

Gradient quenches with the mixture NaF-LiF-
BeF, (50-12.5-37.5 mole %) indicated a liquidus
temperature of 418°C. The compound LiF-7NaF.
4BeF, was found to be the primary phase at this
composition, and Na,LiBe,F, appeared as a

 

1L, M. Bratcher et al.,, ANP Quar. Prog. Rep. Dec.
10, 1955, ORNL-2012, p 83.

12y, Jahn, Z. anorg, u. allgem. Chem. 277, 274 (1954).

75
ANP PROJECT PROGRESS REPORT

secondary phase at 340°C. This indicates that
these two ternary compounds form a quasi-binary
system. One question concerning this part of the
ternary system, which will require further research,
is whether a cubic crystalline phase that has
appeared in some melts is a modification of one of
the known compounds in the system or a new
compound,

The System NaF-LiF-BeF .UF

L. M. Bratcher R. E. Meadows
R. J. Sheil

Materials Chemistry Division

Investigation of the four-component system NaF-
LiF-BeF,-UF , was previously limited to thermal-
analysis and filtration studies of mixtures containing
2 to 3 mole % UF,, which is approximately the
concentration range of interest for circulating-fuel
reactors. 1here haove been indications from filtra-
tion data that higher uranium concentrations might
give lower liquidus temperatures for mixtures with
low BeF, content. Consequently, thermal-analysis
data were obtained with o number of mixtures
formed by adding UF, to purified NaF-LiF-BeF,
compositions in 2 mole % increments up to 10
mole %, The data obtained are summarized in
Table 4,1,

Each of the liquidus temperatures shown in
Table 4.1 represents the ‘“‘best figure’'’ obtained
from two to five cooling curves for each composi-
tion, In general, the solvent mixtures with higher
BeF, content showed the poorest reproducibility
of cooling curves. |t seems likely that a part of
the lack of reproducibility of thermal data is due
to segregation of the high-density uranium phase,

although in these experiments the mixtures were
stirred until they solidified. The data in Table 4.1
show that the etfect of UF , on the liquidus temper-
atures varies considerably with the composition
of the solvent mixture, as might be expected, but
it seems clear that moderately high UF, concen-
trations can be obtained with low-BeF ,-content
mixtures in this system with acceptable liquidus
temperatures. However, the viscosity data reported
in Sec, 7 for the mixture, NaF-LiF-Ber-UF4 (56-
21-20-3 mole %) do not encourage the hope that in
this system a suitable mixture will be found that
has a kinematic viscosity competitive with that of
the best ZrF ,-base fuel compositions.

Filtration data were obtained with two mixtures
in this system by using a new filtration technique
which mokes it possible to take a series of seven
filtered samples from a mixture (usually about 250
g of material) without having to remove the filter
sticks from the apparatus until the experiment is
completed; the samples are cooled to room tempera-
ture in a helium atmosphere. Chemical analysis of
the filtrates obtained from the mixture NaF-LiF-
BeF ,-UF, (51.4-23.3-22.3-3.0 mole %) did not give
a clear-cut indication of the liquidus temperature
for this mixture, which previously reported thermal-
analysis data!?® had indicated to be 530°C. Petro-
graphic and x-ray diffraction studies of the filtered
samples gave inconclusive results.
Uranium determinations on filtered samples of the
mixture NaF-LiF-BeF -UF, (62.6-20.6-11,8-5 mole
%) indicated that the liquidus temperature for this
composition probably lies between 500 and 480°C.

similarly

 

13| . M. Bratcher e: al., ANP Quar. Prog. Rep. Dec.
10, 1955, ORNL-2012, p 83.

TABLE 4.1. LIQUIDUS TEMPERATURES OF NaF-LiF-BeF »-UF ; MIXTURES

 

Composition of Solvent

 

Liquidus Temperature {(°C)

 

 

{mole %}
NoF LiE BeF, No UF, 2 nLujoFIe % 4 nL1JoFIe % 6 nLquFIe % 8 rlr;o;e % 10 J:Ie %
4 4 4 4 4
49 36 15 597 594 590 583 570 545
53 24 23 535 510 510 500 485 (4257)
56 16 28 478 475 455 443 465 475
63.5 7.5 31 535 532 510 483 505 493

 

76
Petrographic and x-ray diffraction examination of
these samples, as well as of others containing 10

PERIOD ENDING MARCH 10, 1956

The System NaF-LiF-CaF,
L. M. Bratcher

mole % UF4 and about 20 to 30 mole % Ber,
showed that the uranium is present as Na,UF  and
that the beryllium-bearing phase cannot be identified
in such mixtures by the techniques presently
available.

Recent experience with the filtration technique
has shown that the filtered samples taken below
liquidus temperatures are frequently small (less
than 1 g). This indicates a need for chemical-
analysis methods more sensitive and reliable than
those presently available in order to follow changes
in liquid composition in this system between
liquidus and solidus temperatures.

Materials Chemistry Division

Thermal-analysis data obtained with 29 composi-
tions in the NaF-LiF-CaF, system are shown in
Fig. 4.2, which also indicates approximate drainage
paths and temperature contours derived from these
data and published data on the three binary systems
involved. The composition and melting point of
the ternary eutectic in this system were reported
previously.'#  Petrographic and x-ray diffraction

 

ML. M. Bratcher et al.,

ANP Quar. Prog. Rep. Dec,
10, 1955, ORNL-2012, p 84,

UNCLASSIFIED
ORNL - LR-DWG 12779

TEMPERATURES ARE IN ©C

 

 

/\\ 1150
/ SN/ \

1100 T
A

\ 1050
7

 

 

   
 
    
  
 
 
    

 

 

958—1000 ”/
\ 610 /
\ *
\ N
. 950
— - o
N 588_——900—"_’_’865

 

DAL

B _XWL_sso

A AN N
612
m 800 =792 725 e
6157 72 __.4-680-613"608
750-——-745._664 61?/645 615 ®
< 693 ?’5 .
/730

612\ 70 /615. 663,620
615 ’ / ,

69
620
8
aoo 660 615 700 / 756 \
\\
LiF //
844

Fig. 4.2. The System NaF-LiF-CaF,.

818

 

     

 

825

77
ANP PROJECT PROGRESS REPORT

studies of the fused melts showed only the pure
components, and thus the absence of compounds

or solid solutions in the system is indicated. In
the course of this investigation thermal-analysis
data were also obtained with a few mixtures in the
NaF-CaF, system. The results, which were es-
sentially in agreement with the published diagram, '3
indicated a slightly convex liquidus curve in the
NaF primary-phase field, in contrast to the straight-
line liquidus shown in the published diagram.

The System NaF-MgF -CaF,
L. M. Bratcher

Material s Chemistry Division

Phase relationships in the system NaF-MgF -
CaF, are more complex than those in the NaF-
LiF-CaF, system because of the existence of the
stable compound'é NaMgF,. The study of this
system is incomplete, but the data obtained thus
far indicate that NaMgF, and CaF, form a quasi-
binary system, as shown in Fig. 4.3, with a eutectic
at about 28 mole % C0F2, which melts at 900
t 5°C. The significance of the thermal effect
noted on the cooling curves at 730°C is not quite
clear, but it probably is due to a small amount of
the ternary eutectic between NaF, NaMgF ., and
CaF

o+ This eutectic was located at the approximate

 

15p, P, Fedotieff and W. P. Hjinsky, Z. arorg. w
allgem. Chem, 129, 101 (1923).

164, G. Bergman and E. P. Dergunov, Compt, rend
acad, sci. U. R. 5. §. 31,755 (1941)}.

UNCLASSIFIED
QRNL—LR—DWG 12780

 

 

 

 

 

 

 

 

TEMPERATURE {°C)

 

 

 

 

 

 

 

 

 

 

 

 

800 i —_—
NaMgF; 10 20 30 40 50 60 70 80

CaF; (mole %)

 

 

90 CGF2

Fig. 4.3. The System NaMgF,-CaF,,.

78

composition NaF-MgF -CaF, (65-12-23 mole %),
and it melts at 745 £ 5°C., Work on this system
will be concluded in the near future.

The Systems NaF-CeF, and RbF-CeF3

L. M. Bratcher
Materials Chemistry Division

The Group !l metal fluorides have received
comparatively little consideration as possible
components of fused-salt fuels. Published partial
phase diagrams for alkali-metal-fluoride~AlF,,
~-YF,, and —LaF, systems make these systems
appear to be unattractive as fuel materials. The
only published data on an alkali-metal-fluoride—
CeF, system appears to be that of Puschin and
Baskow.'”  Their data showed that there was a
eutectic in the KF-CeF, system which contained
25 mole % CeF ; and had a melting point of 650°C.
Since cerium has a cross section only slightly
higher than that of sodium and since CeF, has a
very high free energy of formation, it seemed to be
desirable to conduct exploratory phase studies of
alkali-metal-fluoride—CeF; systems. A study of
the NaF-CeF, and RbF-CeF, systems was there-
fore initiated.  Preliminary data indicate that
compounds are formed in both systems, There
appears to be a eutectic in the RbF-CeF, system
with 15 mole % CeF, and a melting point of about
655°C.  In the NaF-CeF, system the cooling
curves indicate the existence of a eutectic that
melts at approximately 720°C. The composition
of this eutectic has not yet been determined.

The System LiF-NiF )

L. M. Bratcher
Materials Chemistry Division

A study of the LiF-NiF, system was started
recently to obtain information of interest in con-
nection with electrochemical studies involving
structural-metal fluorides. Thermal-analysis data
obtained with mixtures containing 10, 20, and 30
mole % NiF, show only a gradual lowering of
liquidus temperatures from 844°C for pure LiF to
805°C for the 30 mole % NiF, mixture, Other
small thermal effects were observed on the cooling
curves, and these effects increased in temperature
with increasing Nin content of the melts from

 

17N, Puschin and A. Baskow, Z. anorg. u. allgem.
Chem. 81, 359 (1913).
PERIOD ENDING MARCH 10, 1956

545°C at 10 mole % NiF2 to 705°C at 30 mole %. SRbF.4Z:F
Petrographic and x-ray diffraction studies of the
slowly cooled melts indicated a slight amount of Optical data:
solid solution of NiF, in LiF and, also, the Biaxial negative;
formation of @ compound of unknown composition. 2V very large
Optical Properties and X-Ray Patterns for Ny = 1.442
Compounds in Fluoride Systems N, = 1.452
R. E. Thoma X _
) ] -ray data:
Materials Chemistry Division i
G. D. White d(A) 11
Metallurgy Division 5.57 10
H. Insley T. N. McVay 551 10
Consultants
. g . L 5.04 9
The identifying characteristics of some new
compounds encountered in phase studies are 4.77 16
listed below. The symbol d(A) denotes the distance 4.50 12
between reflecting planes measured in angstroms. 4.29 65
The term I/I, refers to the relative intensity as
compared with an arbitrary value of 100 for the 3.98 25
strongest line. Under optical properties, N _ and 3.82 30
N, refer to the lowest and highest indices of 3.78 16
refraction, respectively, for biaxial crystals, and 3 75 ' 6

2V refers to the acute angle between the optic
axes and biaxial crystals; O and E refer to the 3.63 32
ordinary and extraordinary indices of refraction

 

. . 3.59 95
of uniaxial crystals.

The optical data on the compounds 2RbF.ZrF, 3.52 8
and 3RbF.ZrF, differ slightly from the original 3.44 100
data reported previously.'®  The samples of 3.42 100
2RbF.ZrF, and 3RbF.ZrF, available for the recent

.4 L 3.36 100
measyrements are believed to be freer of impurities
than were the samples used before. The optical 3.29 100
data presented here therefore supercede the data 3.05 12
eported iously.
reported previously 2 40 .
3R|:>F-ZrF4 2 508 10
Optical data: 2.386 12
Cubic, N = 1,420 2.350 75
2,281 1n
X-ray data:
o 2.275 12
Simple cubic, a = 3.288 A
2,260 12
0
d(A) /1, 2.53 12
3.290 100 2.245 8
2.326 60
1.900 100 18T, N. McVay and G. D. White, The Optical Properties
of Some Inorganic Fluoride and Chloride Compounds,
1.646 20 ORNL-1712 (May 5, 1954).

79
ANP PROJECT PROGRESS REPORT

80

O
d(A)

2.226
2.152
2.067
2.027
1.959
1.932
1.897
1.868
1.833
1.770
1.763
1.679

d(;‘\)

5.350
4.835
3.590
3.080
2.414
2.338
2,199
1.861
1.789
1.569
1.422

Vi

12
45
12
12
30
30
15
20
20

30
12

2RbF'ZrF4

Optical dota:
Uniaxial negative
O =1.440
E =1.426

X-ray data:
I/[‘

10

100
34

40
21
17
17

B-RbF-ZrF ,

Optical data:

d(A)
7.37
6.66
6.32
5.53
517
5.01
4.67
4.54
4.46
4,35
3.99
3.88
3.77
3.72
3.69
3.61
3.42
3.34
3.325
3.090
3.060
2.960
2.900
2.822
2.788
2.767
2.678

Biaxial negative;

2V ="75 deg
N, = 1.490 to 1,502

Ny = 1,502 to 1.516

X-ray data:

I/1

12
42

12
12

30
36
42
42
100
100
100

14

38
45
PERIOD ENDING MARCH 10, 1956

2.614 6 1716 17
2.564 6 1,709 17
2.514 8 1.699 10
2.501 8 1.694 10
2.482 6 1.679 18
2.455 6 1.662 6
2.433 6 1.654 18
2.423 6
2,398 . RbF-NaF-ZrF
2.380 12 Optical data:
2.355 12 Biaxial negative;
2,344 12 2V = 60 deg
2.309 6 Ng =71.385
2.292 8 N, ="~1.395
2.253 6 Often fibrous or platy; polysynthetic and
rectangular twinning common
2.240 12
2.211 14 X-ray data:
2,206 15 d(A) 1,
2.191 32 5.21 10
2.116 42 482 s
2.074 10 468 38
2.034 6 4.21 45
1.997 30 375 5
1.989 48 2,65 5
1.935 48 3.59 10
1.929 48 339 65
1.926 55 3.34 10
1.903 30 299 20
1.894 6 3.24 35
1.872 12 2,571 10
1.857 10 2.514 8
1.836 10 2.350 15
1.773 6 2.332 28
1.757 6 2.321 28
1.751 6 2.152 100
1.742 6 2.094 100
1.731 15 1.886 10
1.728 14 1.763 18
1.719 24 1.614 8

81
ANP PROJECT PROGRESS REPORT

82

RbF-NaF-ZZrF4

Optical data:

Uniaxial negative

0 =1.446
E =1,435

Often shows platy habit

7.43°
6.71
6.54
5.91
4.55
4.41
4.13
3.90
3.71
3.52
3.40
3.35
3.29
3.23
3.13
2.826
2,481
2,430
2,392
2.380
2.344
2.286
2.237
2.194
2.159
2,142
2.083
2.060
1.983
1.963

X-ray data:

/1

40
22

50

60
45

100
12

M—l
NN

o o W n h " n

— -— N e
o o O O

12

1.943
1.935
1.920
1.854
1.836
1.757
1.754
1.694
1.676
1.670

d(/?\)
7.80
6.61
6.19
6.03
5.61
5.47
5.21
4.98
4.92
4.80
4.65
4.33
4.09
4.04
3.91
3.80
3.69
3.60
3.51

14
25

40
45

38
30

NaF.2U F4

Optical data:

Biaxial negative;

2V =20 to 30 deg
N, =1.520
N), =1.584

X-ray data:

/1,

21
21
100
PERIOD ENDING MARCH 10, 1956

3.43 21 1.840 7
3.42 21 1.817 27
3.41 21 1.792 7
3.30 43 1.790 7
3.26 70 1.754 11
3.15 21 : 1.742 11
3.12 7 1.730 21
3.08 65 1.720 14
2.98 7 1.643 14
2.90 7 1.638 1
2.87 9 1.586 14
2.813 45 1.584 n
2.747 7 1.576 13
2.731 7 1.572 13
2.675 7 1.554 18
2.622 7
2.600 7 Na,UF o
2.449 8

Biaxial negative;
2.411 7 2V = 60 deg
2.355 1 N, = 1.520
2.290 1 N, =1.600
2.250 7
2,211 7 KU,Fy
2.184 7 Biaxial negative;
2,152 7 2V = 10 deg
2,144 7 N, = 1.520
2.137 7 N,, =1.584
2.128 24
2.090 7 “aFeCly
2.065 28 Uniaxial positive
2.034 14 0 =1.600
2.010 9 E =1.636
1.983 13

I(Fe:CI3

1.947 13
1,939 21 Biaxial positive;
1.931 32 2V =20 deg
1.912 35 Ny =1.700
1.865 10 ' Ny =1.740
ANP PROJECT PROGRESS REPORT

ZnF2

Uniaxial positive

0 =1.501
E =1,526
szBeF4
Biaxial;
2V =90 deg

Refractive indices near 1.394

Low birefringence
RbgBer

Uniaxial positive

Refractive indices near 1,402
RbBeF3

Biaxial positive
Refractive indices near 1.390

Low birefringence

CHEMICAL REACTIONS IN MOLTEN SALTS

L. G. Overholser F. F. Blankenship
G. M. Watson

Matertals Chemistry Division

R. F. Newton

Director's Division

A diversified program of chemical studies of
molten fluorides at high temperature was continued
during the quarter. Evaluations of activities of
materials such as FeF, or NiF, in molten NaZrF,
are being made by measuring emf’s and by de-
termining equilibrium conditions for reactions
such as

H2 + FeF2 = Fe® + 2HF

In addition, the reversible electrodes M%/MF , were
shown to be valuable for estimation of solubilities
of the metal fluorides in selected fluoride solvents.
Measurements of electrode potentials were aiso
shown to be applicable to the estimation of the
activity of chromium in nickel-chromium alloys.
Preliminary studies of the kinetics of oxidation
of UF, by KF were conducted. It is not possible

84

to separate the oxidation reaction from the simul-
taneous disproportionation of UF,, but, by using
some reasonable assumptions, a number of con-
clusions regarding the Redox process can be
drawn,

Determinations of equilibrium concentrations for
the reaction of UF, with pure metals in various
solvents of reactor interest were continued. The
data reveal that Nb° is quite resistant to UF, in
NoaZrF . mixtures but that neither W° nor Ta® is of
value in NaF-KF-LiF solvents,

The list of experimental preparations of fVlI:x-ZrF4
compounds was extended to include Zan-Zer
and FeF,.ZrF,. Solutions of alkali halides in
molten alkali metals, with the sole exception of
lithium halides
exhibit intense, brilliant green-to-blue colors in
the temperature range 450 to 800°C,

in molten Li° were shown to

Equilibrium Reduction of FeF, by
H, in NaF-ZrF

C. M. Blood G. M. Watson

Materials Chemistry Division

The study of the equilibrium constants for the
reaction

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

as the solvent, previously described,'? was brought
to a successful conclusion. Additional measure-
ments were made at 700 and at 600°C. In order to
determine whether the extreme length of the period
of operation at high temperatures had changed the
nature of the solvent, the study was concluded
with a series of measurements at 800°C under
conditions that duplicated those used during the
very early stages of the experiment. Within ex-
perimental variations, essential duplication of the
previous results was obtained after the four-month
period of continuous operation. The experimental
results obtained during this quarter are summarized
in Table 4.2 and presented graphically in Fig. 4.4.
It may be observed that, within the precision of
the experiment, no changes occurred in the values
of the equilibrium constants over relatively large
ranges of concentration of iron.

After the experiment was concluded, the remaining
portion of melt was sampled and analyzed. The

 

¢, M. Blood, ANP Quar. Prog. Rep, Dec. 10, 1955,
ORNL-2012, p 85.
PERIOD ENDING MARCH 10, 1956

TABLE 4,2, APPARENT EQUILIBRIUM CONSTANTS FOR REDUCTION OF Fer
BY H2 IN NuF-ZrF4 (53-47 mole %)

Reaction: Fe:l'-'2 (d) + 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 FeF,
Container wall: mild steel (98% Fe)

 

Partial Pressure of
HF in Effluent Gas K *
{atm X ]02)

Fe in Melt Sampling Time
{ppm) (days at temperature)

 

Measurements Made at 700°C

760 38 2.81 0.62
725 39 2.67 0.58
725 40 2.76 0.62
775 41 2.84 0.61
675 42 2.72 0.65

Average of 15 measurements reported previously  0.64

Av 0.63 £0.04

Measurements Made at 600°C

1715 6 1.60 0.087
1525 7 1.64 0.107
1290 10 1.39 0.087
1315 11 1.52 0.102
1320 12 1.54 0.104
1245 13 1.40 0.091
1165 14 1.33 0.089
150 21 0.516 0.099
350 28 0.779 0.100
260 33 0.634 0.090
385 34 0.839 0.106
475 35 0.890 0.097

Av 0.097 £ 0,007

Measurements Made at 800°C

790 4 6.18 3.0

715 5 6.00 3.1

810 6 6.17 2.9

880 7 6.25 2.7

940 10 6.30 2,6

915 11 6.39 2.7

900 12 6.14 2.6

1010 13 6.34 2.4
Average of 13 measurements reported previously 2,5

Av 2.06 £0.2

 

*Kx = Plz-iF/XFeFZPHz’ where x is mole fraction and P is pressure in atmospheres.

85
ANP PROJECT PROGRESS REPORT

UMCLASSIFIED
ORNL-LR-DWG 12784

 

 

3.0 T T & O T . T
e AR e e e e
Kxao %4% v ! f /fil L~ [:/ "‘flI"

e

 

 

 

 

 

 

 

 

300 400 500 600 700 8O0 900 1000 1400 4200 4300
0.80
K 060
0.40
600 800 1000 1200 1400 {600 4800 2000 2200 2400 2600 2800 3000 3200 3400
02
600°C J
K, 010 [Tty zLL = '!”"”“Trr’ =
] ‘
0 200 400 600 800 {000 {200 1400 {600 1800
Fe IN MELT {ppm)
Fig. 4.4. Apporent Equilibrium Constants for the Reduction of FeF, by H, in NaF-Z:F,

(53-47 mole %).

results of the analyses of the initial and final
mixtures are summarized in Table 4,3. The analy-
ses indicate that the composition of the solvent
substantially constant throughout the
experiment.

In order to determine the effect of the solvent on
the equilibrium, it is pertinent to compare the
experimental values with those that may be calcy-

remained

lated from the values of the free energies of
formation of seolid FeF, and gaseous HF that are
available in the literature.2? Since the FeF, used
for the experiment was not in solid form but, rather,
was dissolved in molten NaF-ZrF ,, another useful
basis of comparison is the corresponding equilibrium
constant that would be obtained by assuming the
FeF2 to be a supercooled liquid. A comparison of
the temperature effect on the calculated and ex-
perimental constants is also of interest. From
these calculations, activity coefficients for the
dissolved FeF_ were obtained easily., Theactivity
coefficients are useful for describing, in a concise
form, the behavior of the equilibrium in solution,
with reference to a given standard state.

The equations of interest in the calculations are:

FeF, (s) + H, (g) = Fe(s) + 2HF (g)

il

and

FeF, (1) + H, (g) = Fe (s) + 2HF (g) .

 

20L. Brewer et al.,, pp 65, 110 in The Chemistry and
Metallurgy of Miscellaneous Materials, Thermodynamics,
NNES IV-19B, ed. by L. L. Quill, McGraw-Hill, New
York, 1950,

86

TABLE 4.3, NOMINAL NaF-ZrF4 (53-47 mole %)
COMPOSITION BEFORE AND AFTER
EQUILIBRATION MEASUREMENTS

 

Chemical Analyses

 

 

 

Days of Major Constituents  Minor Constituents
Operation (wt %) (ppm)
Na Zr F Ni Cr
1 12,0 42.4 453 55 110
121 12.2 42,8 45.1 30 90
Theoretical  12.1 42.6 45.3

 

The values of free energies of formation needed
are those of FeF, (s), FeF, (1}, and HF (g). These
values?? are listed in Table 4.4. |t is realized
that the number of significant figures used in the
tabulation is greater than is warranted by the
precision indicated for the published values, but,
since the calculations involve differences, all the
significant figures were employed, except for the
end results, Thus the significance of future com-
parisons with other systems will be limited by
experimental accuracy and not by the inherent
variations introduced by employing small differences
of rounded numbers.

The values of the free energies of formation of
liquid FeF, were obtained by calculation of the
free energies of fusion. A heat of fusion, inde-
pendent of temperature, of 8000 cal/mole was
TABLE 4.4, FREE ENERGIES OF FORMATION OF
FeF, (s), FeF, (1), AND HF (g)

 

Free Energy of Formation {kcal/mole)

 

 

Temperature
0 FeF, (s) FeF,(l) HF (g) FeF,(d)
800 —-130.45 -128.69 —65.84 -—129.64
700 ~133.95 ~131.61 ~65.76 -—132.41
600 ~-137.45 -134.53 =65.67 ~135.39

 

used,20 and integration of the equation
A(AF/T) AH

oT T2

between the temperature in question and the melting
point of FeF, (1375°K) gave the free energies of
fusion employed.

The values listed in the last column of Table
4.4 are the free energies of formation of FeF, (<)
obtained from the experimental data by using the
free-energy changes listed in Table 4,5, The
calculated free energies of formation should pre-
sumably incorporate the free-energy changes
resulting from solution, dilution, and solvation.
At any rate, the calculated values of AF° for
FeF, (d) will precisely describe the behavior of
the equilibrium over the experimental range studied,
subject to the precision of the experimental measure-
ments, if the experimental K_ is treated as a true
thermodynamic equilibrium constant.

TABLE 4,5.

PERIOD ENDING MARCH 10, 1956

The free-energy changes of the pertinent reac-
tions, as well as the corresponding thermodynamic
equilibrium constants, are given in Table 4.5. For
purposes of comparison, the experimentally de-
termined equilibrium constants are also given,
along with the calculated free-energy change for

reaction 3 of Table 4.5.

The activities of HF at equilibrium with certain
ranges of activities of FeF, (s) and FeF, (/)
have been calculated and are presented graphically
in Fig. 4.5, along with the experimentally de-
termined values. The activity of FeF, was taken
to be equal to its mole fraction. The mole fractions
of FeF , were calculated as described previously. 19
The activities of HF are expressed in atmospheres.
It may be observed that the experimental points
are, in every case, between the calculated curves
corresponding to the two standard states used.

The temperature dependences of the calculated
and experimental equilibrium constants are com-
pared graphically in Fig. 4.6. An inspection
shows that the slope of the experimental curve is
the same as the slope of the curve calculated by
taking supercooled liquid FeF, as the standard
state. The slope of the curve calculated by taking
solid FeF, as the standord state is steeper. The
heats of reaction indicated by the slopes and the
entropy changes of the reactions are summarized

in Table 4.6.

The activity coefficients of FeF
were also calculated relative to the
standard states that have been considered.

in solution
different
The

FREE-ENERGY CHANGES AND EQUILIBRIUM CONSTANTS FOR THE

REDUCTION OF Fer BY HYDROGEN

Reactions: FeF2 (s) + H2 (2)

FeF, (1) + H, (g) = Fe(s) + 2HF (g)

FeF, (d) + Hy (g)

= Fe (s) + 2HF (g) (1)
(2)

= Fe (s) + 2HF (g) (3)

 

Free-Energy Changes, AF°

Equilibrium Constants

 

 

 

Temperature (kcal/mole of Fer)
(°C)
Reaction 1 Reaction 2 Reaction 3 Ka for Reaction 1 Ka for Reaction 2 Kx for Reaction 3
800 -1.23 ~2.99 —-2.04 £0.16 1.78 4,06 2.6 £0.2
700 +2.43 +0.09 +0.89 £0.12 0.285 0.955 0.63 £0.04
600 +6.11 +3.19 +4,05 £0.12 0.0294 0.159 0.097 £ 0.007

 

87
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR—DWG 12782

10~

ACTIVITY, OR PARTIAL PRESSURE, OF HF {atm)

 

3
o= 2 5 103 2 5 o2
ACTOATY, OR MOLE FRACTION, OF FeFy \N NoF—Zrfy (53—47 mole %)

10

Fig. 4.5. Comparison of the Activities of HF
at Equilibrium with a Range of Activities of Liquid
and Solid FeF, in NaF-Z¢F, (53-47 mole %).

UNCLASSIFIED
ORNL-LR-DWG '2783

TEMPERATURE (°C)

 

EQUILIBRIUM CONSTANT, A

 

 

12.00 (x107%)

9.00 "~ 10.00 .00
Y (2K)

Fig. 4.6. Temperature Dependence of Calculated
and Experimental Equilibrium Constonts for the
Reaction FeF, + H,==Fe + 2HF.

88

TABLE 4.6, HEATS OF REACTION AND
ENTROPY CHANGES FOR THE
REDUCTION OF FeF, BY HYDROGEN

 

 

Heats of Entropy .
Fer Standard State Reaction, Chonges,
o
AH® (cal} AS® (eu)
Solid {calculated) 37,500 36.0
Liquid (calculated) 29,600 30.3
Dissolved (experimental) 29,600 29.4

 

equilibrium constants for the reduction reactions
are given by the expression:

2
IHF “%Fe (s) .
o
a, a
H2 FeF2
If the activity of solid iron is taken to be unity
and the activities are converted to partial pressures

or mole fractions, the expression becomes

2 2
Phe YHF

 

Yy
H2VFeF2
2
YHE
x
YH, YFeF,

High-temperature extrapolation of the available
data?1=23 on association factors of HF (g) indi-
cates that, at the temperatures used, HF is monom-
eric, If ideal gas behavior is assumed for both HF
and H,, it appears quite reasonable to believe that
YHE = YH, = 1.0. Thus the activity coefficient for
FeF,, as defined by the previous equation, is
given by .

K

X

VFer = E;

The activity coefficients of FeF, for the different

 

2lp, L, Jarry and W. Davis, Jr., J. Pbys. Chem. 57, .
600 (1953).

22R, W, Long, J. H. Hildebrand, and W. E. Morrell,
J. Am. Chem. Soc. 65, 182 (1943).

23, Simons and J. H. Hildebrand, J. Am. Chem. Soc.
46, 2183 (1924).
standard states, calculated with this equation by
using the experimental equilibrium constants, are
summarized in Table 4.7,

Within the precision of the experimental measure-
ments, the activity coefficients listed in Table
4,7 are independent of the FeF2 concentration
over the ranges investigated. |t is apparent that,
as far as this investigation is concerned, the most
convenient choice of standard state is FeF, (d).

 

M{ MFQ(Q]):

 

However, since it is believed to be unlikely that
the activity coefficient is independent of the
concentration for the general case involving other
solvents and solutes, it would appear to be more
desirable to adopt the hypothetical pure super-
cooled liquid as the standard state.

TABLE 4.7. ACTIVITY COEFFICIENTS OF
FeF, IN MOLTEN NaF-ZrF , (53-47 mole %)

 

FeF2 Activity Coefficient

 

Standard State Ay 800°C At 700°C At 600°C

 

Sofid 1.46 2.20 3.28
Liquid 0.64 0.66 0.66
Dissolved 1.00 1.00 1.00

 

It is hoped that the equilibrium-reduction reactions
of other metallic fluorides will prove to be amenable
to investigation by this method and that, by com-
parison of similar thermodynamic correlations,
information may be obtained on the condition of
the solute in the molten fluoride solvent. A study
of the reduction of FeF, by H, in other solvents
in the absence or presence of UF, is plonned in
the hope of obtaining a better understanding of the
effect of composition on corrosion.

EMF Measurements in Fused Salts

L. E. Topol
Materials Chemistry Division

Determinations of activities and activity co-
efficients of metal fluorides from the emf’'s of

PERIOD ENDING MARCH 10, 1956

concentration cells in molten NaF-ZrF, (53-47
mole %) were continued.?4 The cells were measured
in the temperature range 550 to 700°C in a helium
atmosphere,  The half cells were contained in
recrystallized alumina (morganite} crucibles, and
electrical contact between them was achieved with
a porous ZrO, bridge impregnated with the NaF-
ZrF , solvent.
The cells measured were of the type

NaF(a,), ZrF ,(a,) [} NaF(aj), ZrF (aj), MF,(a]){M ,

where M = Fe, Cr, or Ni and (a) denotes the activity
of a species. If it is assumed that Na* and F~
are the sole carriers of current, the cell reaction is

MF,(af) + 2iNaF(a,) = MF,(a;) + 2:NaF(a}),

where ¢ is the transference number of the Na*. If
the amounts of added MF, are small, the activities
of NaF are approximately equal in both half cells,
and the cell potential should be given by

 

 

RT 4 RT 1%y

E = M—E II’I = —2_1—'«‘ n ’
n ’ r
4 V1%,

where a,a’, x,x", and y,y” are the activities, mole
fractions, and activity coefficients, respectively.

Duplicate cells were measured, each for a period
of two days, by using iron electrodes and FeF
electrolytes. Three different concentration cells
were used in the study. Agreement between dupli-
cate cells was, in general, better than 13% of the
emf observed. The values obtained on the second
day for each cell were lower by 2 to 8% than those
obtained on the first day. Typical data for the
first day are presented in Table 4.8. From these
data it is obvious that, for all the unsaturated
cells, the change in activity-coefficient ratio with
temperature is very small. A comparisen of the
activity-coefficient ratios of the various concen-
trations studied with the ratio of the most concen-
trated mixture studied (x = 0.015) is shown in
Table 4.9. The data show that the activity
coefficient increases with increased dilution.

It should be noted that the decrease in activity
coefficient with concentration of FeF, is greatest

 

24y | E, Topol, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 97.

89
ANP PROJECT PROGRESS REPORT

TABLE 4.8. ACTIVITY RATIOS FOR FeF, IN NaF-ZrF , (53-47 mole %) FROM THE CELL
Fe| FeF(x,), NaF(x,), ZrF ,(x;) || ZrF (x3), NaF(xJ), FeF ,(x7)| Fe

 

 

 

Emf (v)
Temp;erufure a]/a{ V]/V;

Q) Measured ldeal

For x, = 0,000607; x| = 0.002309*
700 0.0505 0.0560 0.300 1.14
650 0.0472 0.0531 0.305 1.16
600 0.0450 0.0502 0.302 1.15
550 0.0429 0.0473 0.298 1.13

For Xy = 0.000607; x; =0,00714*
700 0.0830 0.1034 0.138 1.62
650 0.0790 0.0980 0.137 1.61
600 0.0750 0.0926 0.135 1.60
550 0.0502 0.0874 0.242** 2.85

For x, =0.00776; x; =0,01484*

700 0.0252 0.0268 0.548 1.05
650 0.0245 0.0254 0.540 1.03
600 0.0155 0.0240 0.662** 1.26
550 0.0008 0.0226 * &

 

*|In mole fractions,

**Solubility of FeF, in at least one half cell was less than the added concentration of FeF,,

TABLE 4.9. VARIATION OF ACTIVITY COEFFICIENTS
OF FeF, IN NaF-ZrF , (53-47 mole %) AT
650°C WITH CONCENTRATION

 

 

xq xy Y17, Y1/Y0.015
0.00778 0.01485 1.05 1.05
0.00270 0.00744 1.17 1.24
0.000607 0.00714 1.63 1.72
0.000607 0.00231 1.18 1.46

 

in the dilute melts. Unfortunately, it is in the
very dilute cells that any slight error in concen-
tration (from weighing or from oxidation of the
FeF,, etc.) leads to a fairly large error in the
emf reading. Further experiments along this line
are planned.

90

In the course of these measurements it was
noted that a sharp change in emf occurred when
the solubility of FeF, was exceeded. Thus, by
adding different amounts of the same metal fluoride
to the NaF-ZrF, solvent in two half cells and
plotting the measured emf as a function of tempera-
ture, the temperature at which the solute dissolves
completely on heating {or precipitates out on
cooling) can be detected by a break or discontinuity
in the plot of emf against temperature. In the
actual experiments, less than equimolar amounts
of metal fluoride and ZrF , were added to NaF-ZrF ,
(53-47 mole %), since it is believed that the
saturating phase in this solvent has the general
formula MF,-ZrF,. Thus the solubilities were
determined along a quasi binary between 7NaF-
6ZrF, and MF ,-ZrF , so that the composition of
the saturated solution would be independent of the
total amount of MF, present. The solubilities of

-
NiF,, FeF,, and CrF, contained in vessels of Ni,
Fe, and Pt, respectively, were measured from 550
to 750°C. In all runs the melts were first heated to
the temperature at which complete solubility in
both half cells was attained. Emf readings were
then taken at decreasing temperatures, with ap-
proximately 30 min being allowed at each point for
equilibrium to be reached. (Constant emf’s at each
temperature were usually approached in 15 min.)
Each cell was run in this way several times, and
both heating and cooling curves were plotted.
With both cells saturated, the emf should be zero,
and it should be constant during heating until
dissolution of the MF,-ZrF, in the dilute melt is
complete. At this point the change in emf with
temperature is abrupt and continues to be rapid
until all the MF_.ZrF, in the concentrated melt
dissolves. Subsequent increases of temperature
result in fairly small changes in emf.

The ZrF, concentrations in the half cells become
increasingly different between the solubility
temperature of the dilute melt and that of the
concentrated one. A typical emf-vs-temperature
plot for an FeF. cell is shown in Fig. 4.7. In all
cases the data obtained at the lower temperature
could be reproduced much better than the data
obtained at the higher temperatures. The average
deviation of the lower temperature, T, was about
3°C, whereas the higher temperature, T,, varied
by about 20°C. Therefore, in practically all
cases, the saturation temperature of the more
concentrated half cell in each run (T,) was re-
determined by measuring this half cell against
another still more concentrated half cell, The

UNCLASSIFIED
CORNL—LR—-DWG 12784

 

   
 
    

 

 

30
o ON COOLING )
o ON HEATING |
20 - e ]
E i
- ____..-——.-——-.—'_
’ |
10 —— li,,,./ , —_— —
,___-——-#/ ¢, = 6.07 mole %
! T¢)= 1103 male %
0 ! '
650 700 750 800

TEMPERATURE (°C)

Fig, 4,7. Temperatute Dependence of the EMF of
the Cell Fe|FeF,, ZrF (c,) || FeF,, ZrF (c,"}|Fe
in the Solvent NaF-ZrF , (53-47 mole %),

PERIOD ENDING MARCH 10, 1956

solubilities given in Fig. 4.8 were found in this
manner,

The log of the solubility in mole % MF -ZeF, is
plotted against the reciprocal of the absolute
temperature in Fig. 4.9. As may be seen, the
solubilities for the dilute melts obey a straight-
line relation with respect to 1/T, but in the more

UNCLASSIFIED
ORNL-LR-DWG 12785

 

800 - — ‘ —

 

TEMPERATURE (°C)

 

 

 

 

0 4 8 {2 16 20
MF-ZrF, (mole 7o)

Fig. 4.8. Temperature Dependence of Solubility
of MF:_,-ZrF4 in NaF-Z(F .

UNCLASSIFIED
ORNL-LR—-DWG 12786

NiFp ZrFs-

 

3

0.90 1.00 140 1.20 130 {(x10”

Y1 (oK)

)

Fig. 4.9. Solubility of MF,.ZrF, in NaF.Z(F,

vs Reciprocal of Absolute Temperature.
ANP PROJECT PROGRESS REPORT

concentrated mixtures there is a deviation from
linearity, This deviation is in the direction op-
posite to that expected. It appears likely that at
the higher temperatures the soaturating phase is
no longer MFQ-ZrF‘; it may well prove to be the
simple MF, compound, The slopes of the lines at
lower temperatures and concentrations in Fig. 4.9
yield heats of solution of 37.5, 35.7, and 23.3
kcal/mole MF ,.ZrF, for the CrF,, FeF,, and
Nil:2 compounds, respectively. By extrapolating
these lines to unit mole fraction of MF,.Z¢F ,
“ideal’’ melting points of 775, 857, and 1155°C
are found for CrF,-ZrF,, FeF,.ZrF,, and NiF,.
ZrF ,. These ideal melting points have no physical
meaning, but they may serve as a rough measure
of the relative stability of the three compounds.
The unexpected difference in both the ideal melt-
ing point and the heat of solution of the NiF,
complex from that of either the CrF, or the FeF,
leads to the supposition that different types ot
species may be involved. |t may be that NiF .
ZrF , breaks down at temperatures lower than any
used for these tests.

Activity of Chromium in Alioys
M. B, Panish

Materials Chemistry Division

The activity of chromium in alloys of chromium
and nickel is being examined because of the great
importance of such alloys as container materials
for molten salts, Grube and Flad?25 studied the
reaction

Cr2O3 + 3H,=— 3H20 + 2Cr

and tabulated the pressures of H,0 and H, in
equilibrium with pure chromium and nickel-chromium
alloys at 1100 and 1200°C, An activity diagram,
not previously published, which was calculated
from these data is presented in Fig. 4.10. This
diagram must be considered to be surprising, since
considerable negative deviation from Raoult’s
law is exhibited, in spite of a region of immiscibility
on the chromium-rich portion of the diagram, [t has
been suggested that the experimental work was in

error. 26

 

25G. Grube and M. Flad, Z. Elektrochem. 48, 377
(1942).

26K, w, Wagner, Thermodynamics of Alloys, Addison-
Wesley Press, Cambridge, 1952.

92

UNCLASSIFIED
ORNL-LR—DWG 12787

.

 

100 ——

N

i .
| IDEAL— / '

 

i
‘ y I
0.75 |——— - : b ;-
\ 3 /

 

i
I
i
c | / !
= !
= | ¢
5 f - / : ;
< J ; a
w 050 |——- o / — | /
2 ' | :
2 X ‘
< | REGION OF
& = IMMISCIBILITY—-
s |
oasl I S N

 

 

e

|

|
50 75 i
Cr IN Ni {mole %)

 

Fig. 4.10.  Apparent Activities of Nickel-
Chromium Alloys at 1100°C.

The activity of chromium in nickel-chromium
alioys is being investigated here by emf measure-
ments of electrolytic cells of the type

Cr®|NaCl, RbCI, CrCl, | Cr(Ni) alloy

A schematic diagram of the cell, which utilizes a
central reservoir and four electrode compartments,
is shown in Fig. 4.11. This assembly is of fused
quartz with standard glass taper joints in the lower
temperature regions. The nickel-chromium elec-
trodes were prepared for this study by using powder
metallurgical techniques and high-purity chromijum
and nickel. The electrode connections are nickel
wires that are joined to tungsten wires sealed
through the glass. The alkali halide mixtures are
prepared from clear crystals selected from slowly
cooled melts of reagent-grade materials.

Curves plotted from emf measurements at various
temperatures for two alloys containing different
proportions of chromium {22.1 and 27.6 mole %)
are shown in Fig. 4.12, For these runs the
electrolyte was the NaCl-RbCl eutectic containing
0.1 to 0.5 wt % CrCl,. An activity coefficient of
about 0.9 at 800°C was determined from these
curves, in good agreement with the 1100°C data of
Grube and Flad.23 Attempts to utilize electrodes
containing 10 mole % or less of chromium showed
that the emf obtained was quite unreproducible,

-

fl
«
,
UNCLASSIFIED
ORNL-LR-DWG 12788

  

TO VACUUM

 

 

 

/"/
| | | | TUNGSTEN
41" i e
| I /k/—NlCKEL WIRE
Ul I
0 I

aUARTZ I ’

 

 

6““' ‘ 1

 

 

 

 

 

  
  
   

 

Il LIQUID LEVEL

 

 

 

 

 

 

 

 

 

L ~ELECTRODE Y%, TO 1-in. LONG

 

 

 

 

 

 

 

LIQUID LEVEL

o
/-J Ya-in. DIA {-mm CAPILLARY

-~

Fig. 4.11. Cell Vessel with Central Reservoir
and Compartments for Four Electrodes.

PERIOD ENDING MARCH 10, 1956

The voltages obtained for several cells in which
the electrolyte was the LiCl-KCl| eutectic with
1 wt % CrCl, and in which the electrodes were
chromium cmd2 Inconel are shown in Fig. 4.13.
The line drawn through the points of Fig. 4.13
was calculated for an ideal chromium-nickel alloy
containing 15 wt % chromium. Although Inconel
nominally contains 5 wt % iron and the points
scatter rather badly, the data seem to be represented
reasonably well by the ideal case.

The electromotive forces obtained should be
reproducible after raising and lowering of the
temperature and after brief shorting or electrolysis.
Electrodes containing less than 10% chromium and
the Inconel electrodes, however, responded poorly
to the temperature test, All the electrodes re-
sponded well to the shorting test. This indicates
that a diffusive equilibrium between chromium on
the surface and chromium within the body of the
electrode is established at the temperatures at
which these tests are performed. !t was found that
mechanical agitation of the electrodes within the
cell had no effect upon the electromotive force,
This indicates that there is no sharp concentration
gradient in the electrolyte in the immediate
vicinity of the electrodes,

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

Materials Chemistry Division

Data obtained from filtration studies of the
reduction of UF, by Cr® or Fe® with NaF-ZrF,
(50-50 mole %),27 NaF-LiF-KF (11.5-46.5-42
mole %),28 NaF-ZrF , (53-47 mole %),2% or NaF-
LiF-ZrF , (22-55-23 mole %)30 as reaction mediums
were given in previous reports. More recently,
studies have been made on the reduction of UF

by Fe® or Cr® with molten NaF-ZrF , (59-41 mole %)

 

27), D. Redman and C. F. Weaver, ANP Quar. Prog.
Rep. June 10, 1954, ORNL-1729, p 50; ANP Quar. Prog.
Rep. Sept. 10, 1954, ORNL-1771, p 60.

28) D. Redman and C. F. Weaver, ANP Quar. Prog.
Rep. March 10, 1955, ORNL-1864, p 56.

295, D. Redman and C. F. Weaver, ANP Quar. Prog.
Rep. June 10, 1955, ORNL-1896, p 60.

30), D. Redman and C. F. Weaver, ANP Quar. Prog.
REP. Septo 10, 1955' ORNL"]947, p 74-
ANP PROJECT PROGRESS REPORT

as the solvent. The results obtained in these
studies show that Cr® is not stable with respect to
UF, dissolved in these different fluoride mixtures,
This observed instability of chromium, either as
pure metal or in alloys, has led to studies of the
stability of UI:4 toward other structural metals
which might serve as substitutes for chromium in

alloys. Earlier work31 showed that Mo® was more
stable toward UF4 than Cr° was in both NaF-ZrF
(50-50 mole %) and NaF-LiF-KF {11.5-46.5-42
mole %). The systems UF -V and UF ,-Nb® also
have been studied with the NaF-LiF-KF mixture as

 

315, D. Redman and C. F. Weaver, ANP Quar. Prog.
Rep. March 10, 1955, ORNL-1864, p 61.

UNCLASSIFIED
ORNL—~LR—DWG 12789

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.10 r .
0.08
A AT T
(o)
(o) 0 g . o ®
0.06 * veT t ¥ ©
) O
»
3
Z
W
Q.04
L J
o } ALLOY ELECTRODE CONTAINING 27.6 mole % Cr
o
o0z A ALLOY ELECTRODE CONTAINING 22.1 mole % Cr
(®) DOUBTFUL VALUE
O !
600 700 800 200 1000
TEMPERATURE (°C)
Fig. 4.12. EMF Measurements Obtained with a Concentration Cell at 600 to 1000°C for

Chromium-Nickel Alloys vs Pure Chromium in Molten NaCl-RbCl Eutectic Containing CiCl,.

UNCILASSIFIED
ORNL—-LR—DWG 12790

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

012 4
»
0.08 ¥ / —t—
:_u? ¢ IDEAL *
2
Wy
0.04
o L
640 660 680 TO0 720 740 760 780 80C 820 840

TEMPERATURE (°C)

Fig. 4.13.

EMF Measurements Obtained with a Concentration Cell at 600 to 1000°C for

Inconel vs Pure Chromium in Molten LiCI-KC| Eutectic Containing CCl,.

94
the reaction medium. Data for these systems,
reported previously,32 show that Nb° is unstable
toward UF, in the NoF-LiF-KF solvent. During
the quarter, studies have been made on the systems
UF ,-Ta® and UF ,-W® with the NaF-LiF-KF eutectic
as the reaction medium. Studies also have been
carried out on the UF ,-V® and UF ,-Nb°® systems
with NoF-ZrF4 (50-50 mole %) and NaF-LiF-Zer
(22-55-23 mole %) as the reaction mediums,

The results of the studies of the reduction of

UF, by Cr°® at 600 and 800°C with NaF-ZrF,

 

32 p, Redman, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 86.

PERIOD ENDING MARCH 10, 1956

(59-41 mole %) as the reaction medium are given in
Table 4.10. In these runs approximately 2 g of
Cr® was reacted with UF  (11.4 wt %, 3.7 mole %)
dissolved in opproximcte1y 40 g of the Nc:F-ZrF’1
mixture contained in nickel. The data given in
Table 4.10 show considerable scatter; however, it
appears that the chromium concentrations are
essentially the same at the two temperatures
studied. Data in Table 4,11 present a comparison
of equilibrium Cr** values for the several mixtures
studied. As may be seen, the concentration of
chromium fluoride at equilibrium becomes pro-
gressively smaller as the ratio of alkali-metal
fluoride to ZrF4 is increased, This is in general

TABLE 4.10. DATA FOR THE REACTION OF UF, WITH C:° IN MOLTEN NaF-ZrF, (59-41 mole %)
AT 600 AND 800°C

 

Conditions of Equilibration

 

Present in Filtrate

 

 

Temperature Time Total U Total Ce* Total Ni
(°C) (hr) (wt %) (Ppm) (pPm)
600 3 8.9 710 65

3 8.7 1010 95
5 9.4 1060 70
5 8.8 970 95
S 8.6 1030 35
800 3 9.9 1000 20
3 8.7 950 45
5 9.6 1190 60
5 10.4 1080 70

 

*Blank of 270 ppm of Cr at 800°C.

TABLE 4.11. EQUILIBRIUM CONCENTRATIONS OF Cr** FOR THE REACTION
Ccr® + 2UF, -r-_—‘ZUF3 + CrF, IN VARIOUS SOLVENTS

 

Sclvent Composition

Equilibrium Concentration of cett (ppm)

 

 

(mole %) At 600°C At 800°C
NaF-ZrF ,; 50-50 2250 2550
NaF-Z¢F ; 53-47 1700 2100
NaF-ZrF ,; 59-41 975 1050
NaF-LiF-Z¢F ;; 22-55-23 500 650

 

95
ANP PROJECT PROGRESS REPORT

agreement with the concept that UF, is stabilized
by the formation of a UF_= or UF,~= complex
ion which must compete with ZrF , for the free
fluoride ion,

As the data in Table 4.12 indicate, however, the
concentrations of Fe*t in equilibrium when pure
Fe®is used as the reducing agent are approximately
the same as those found when either NaF-ZrF
(53-47 mole %) or NaF-ZrF, (50-50 mole %) is
used as the reaction medium., Temperature ap-
pears to have little effect on the reaction, although
a slight decrease in iron concentration with
increased temperature is indicated. A similar
temperature effect occurs when the other two
NaF-Zer mixtures are used as solvents,

Data for the reaction of UF , with Ta® and with
WS at 600 and 800°C with NaF-LiF-KF (11.5-46.5-
42 mole %) as the reaction medium are given in
Table 4.13. Approximately 2 g of the metal was
equilibrated with UF, (15 wt %, 2.3 mole %)
dissolved in approximately 20 g of the NaF-LiF-KF
mixture contained in nickel.

The data presented in Table 4.13 indicate that
neither Ta® nor W° is stable with respect to UF
under the conditions studied, and therefore they
afford little or no advantage over Cr° The data
collected for the UF ,-W® system show considerable

TABLE 4,12, DATA FOR THE REACT
NUF'Z'F4 (59+41 mole %)

scatter, the cause of which is unknown. Since
there was some indication that equilibrium was not
being attained during the short heating periods, a
few experiments were run with 12-hr equilibration
periods. The results obtained for these runs were
not precise, but they seem to demonstrate that no
significant increase in concentration of tungsten
occurs after 5 hr of equilibration.

Data are given in Table 4.14 for the reaction of
UF, with Nb® at 600 and 800°C with NaF-ZrF,
(50-50 mole %) and with Nc;:F-LiF--ZrF:4 (22-55-23
mole %) as the reaction mediums, The runs were
made in nickel equipment with 11.4 wt % UF
dissolved in approximately 40 g of the fluoride
mixtures and 2 g of Nb°, These data show that no
significant interaction between Nb®and UF, occurs
in either of the ZrF ~containing fluoride mixtures
studied. This behavior is in marked contrast to
that observed when NaF-LiF-KF (11.5-46.5-42
mole %) is used as the reaction medium; with this
medium, niobium concentrations of 700 and 1500
ppm were found at 600 and 800°C, respectively,
The markedly different niobium concentrations
found for the solvents are surprising and must be
associated with the differences in free alkali-metal
fluoride concentrations present and the resulting

fON OF UF, WITH Fe® IN MOLTEN
AT 600 AND 800°C

 

Conditions of Equilibration

 

Present in Filirate

 

 

Temperature Time Total U Total Fe* Total Ni
(°C) (hr) (wt %) (ppm) (ppm)
600 3 8.5 510 35

3 8.6 €30 40
5 8.6 370 60
5 8.5 600 70
800 3 8.9 680 70
3 8.6 290 65
3 9.2 460 70
3 9.2 440 75
5 9.3 465 100
5 8.8 280 35

 

*Blank of 140 ppm of Fe at 800°C.

96
PERIOD ENDING MARCH 10, 1956

TABLE 4.13. DATA FOR THE REACTION OF UF, WITH To® AND WITH W° IN MOLTEN
NaF-LiF-KF (11.5-46.5-42 mole %) AT 600 AND 800°C

 

Conditions of Equilibration

 

Present in Filtrate

 

 

Temperature Time Total U Total Ta* Total Wx=* Total Ni
(°C) (hr) (wt %) (ppm) (ppm) (Ppm)
For Reaction with Ta®
600 3 11.2 1320 140
3 1.6 1220 20
5 12.1 1150 110
5 11.8 1150 240
800 3 11.2 3220 170
5 11.9 2800 80
5 11.8 3020 180
For Reaction with W°
600 3 11.2 950 215
3 11.2 920 435
5 11.6 1410 250
5 11.2 1350 230
5 10.5 900
5 11.0 1125
800 3 11.3 980 260
3 11.2 920 205
5 11.2 1480 65
5 10.8 2150 250
5 12.0 580 155
5 11.8 1260 165
5 11.6 1210
5 11.8 1600
12 10.5 1880 115
12 11.8 1400
12 11.3 1430

 

*Blank of approximately 50 ppm of Ta ot 800°C. Several attempts were made to reduce the size of the blank, but

all were unsuccessful,

**Blank of 350 ppm of W at 800°C.

changes in activities of the uranium and niobium
fluoride complexes.

A previous study of the reaction of UF , with V°
in the NaF-KF-LiF eutectic showed vanadium
concentrations in the melt to be about 25 ppm.
However, repeated checking of this rather surprising
result has served to show that the chemical
analyses were grossly in error. |t now appears
that significant interaction of UF, with vanadium

in  either NaF-KF-LiF or NaF-ZrF4.

oCccurs

However, it seems advisable to withhold the data
until the inconsistencies in analyses have been
clarified.

Reaction of UF3 with NaF-KF-LiF Evutectic

B. H. Clampitt
Materials Chemistry Division

Alkali-metal vapors are evoived from a heated
solution of UF3 in the molten NaF-KF-LiF eutectic,

97
ANP PROJECT PROGRESS REPORT

TABLE 4.14, DATA FOR THE REACTION OF UF4 WITH Nb° AT 600 AND 800°C IN MOLTEN
NuF-ZrF4 (50-50 mole %) AND NaF-LiF-ZrF, (22-55-23 mole %)

 

Conditions of Equilibration

 

Present in Filtrate

 

 

Temperature Time Total U Total Nb* Total Ni

(°c) (hr) (wt %) (pPm) (ppm)
For Reaction Medium Nc:F-ZrF4 (5050 mole %)

600 3 8.2 25 55
3 7.9 20 60
5 8.9 25 55
5 8.7 25 80

800 3 8.8 20 40
3 8.8 20 70
5 8.8 20 80
5 8.5 20 85

For Reaction Medium NuF—LiF-ZrF4 (22-55-23 mole %)

600 3 8.2 1 100
3 8.8 1 120
5 8.4 2 120
5 8.9 1 45

800 3 8.5 1 110
3 8.6 1 110
5 8.9 5 15
5 8.3 4 40

 

*Blank of 25 ppm of Nb at 800°C in NaF-ZrF ,; approximately 15 ppm in NaF-LiF-ZrF ,.

despite the standard free-energy changes for the
reaction

KF + UF, —>UF, + K°

being unfaverable by about 30 kcal per gram atom
of potassium. The vapor is about 95% K° with the
baflance almost entirely Na® but, for purposes of
the following discussion, it is assumed that KF is
the only component of the melt reduced by UF ..
That this reaction should proceed perceptibly is
additional evidence that UF, is strongly complexed
and stabilized in the molten fluoride solution; this
stabilization, which in general makes UF |, solutions
less corrosive than they would otherwise be,

98

contributes, unfortunately, to instability of UF,
in such systems, '

The kinetics of conversion of UF3 to UF , in this
solvent are complicated by the disproporticnation
reaction

4UF ,—>3UF, + U

which takes place simultaneously with the redox
reaction and produces an alloy of uranium metal
with the container wall, A rough study of the
kinetics of attrition of UF, in this solvent has
been made by analysis for UF, of samples taken
after 2 hr of heating of standard mixtures in open
crucibles of copper under helium atmospheres at
varying pressure,
A previous report33 showed that at 650°C, under
these conditions, the rate of reaction (measured by
the UF ; remaining after 2 hr) was independent of
helium pressure at pressures above 50 mm Hg; at
pressures below 50 mm Hg the amount of UF,
remaining varied nearly linearly with pressure,
The results obtained at 750°C, shown in Fig. 4.14,
are quite similar; at this temperature the reaction
rate is insensitive to pressures above 130 mm Hg.
For samples which were held at pressures above
this “‘critical’’ pressure and for which 21.4 wt % of
UF ; was initially present, 58% of the UF; remained
after 2 hr at 650°C and 47% remained after a
similar period at 750°C.

 

A
ORNL-LR-DWG 12791
|

10 s : J
./.’ !

 

® a0

 

 

 

 

 

 

 

 

 

UF3 REMAINING AFTER 2 HOURS (wf %)

 

 

 

 

 

 

0 . b
0 100 200 300 400 50C 600 700 800

HELIUM PRESSURE (mm Hg}

 

Fig. 4.14. Oxidation of UF; at 750°C in NaF-
LiF-KF (11.5-46.5-42 mole %).

This behavior suggests strongly that the sharp
break in the reaction rate as a function of pressure
occurs at the ‘‘bubble-point’’ pressure for the
solution at a given temperature, At pressures
above this level- the evolution of potassium ap-
pears to reach a near-equilibrium steady state and
to proceed at a rate controlled by the rate of
vaporization of potassium from its solution in the
salt. If the values 50 and 130 mm Hg are taken
as the bubble pressure of K° from the solution at
650 and 750°C, respectively, then, since the
vapor pressures of pure K° at those temperatures
are 240 and 750 mm Hg, the activity of the

 

33g, H, Clampitt and C. J. Barton, ANP Quar. Prog.
Rep. Dec. 10, 1955, ORNL-2012, p 90.

PERIOD ENDING MARCH 10, 1956

potassium is approximately 0.2 at each temper-
ature, (The absence of a brilliant blue-green
phase seems to be sufficient assurance that the
solution is not saturated.)

The effect of temperature on the two competing
reactions cannot be specified with certainty,
The disproportionation reaction should be rapid,
initially, ot all temperatures, but it would be
expected to slow to a rate controlled by the slow
diffusion of U° into the copper wall, In each 2-hr
test, therefore, an almost constant amount of UF
might be expected to form, regardless of temper-
ature,

If the reaction to produce K° vapor is assumed
to be more rapid, by a ratic of 130/50, at the higher
temperature and if the disproportionation rate is
assumed to be nearly independent of temperature,
then, by using the experimental data for reactions
above the ‘‘critical'’ pressure, a set of three
simultaneous equations in three unknowns can be
formulated and solved, The calculations show
that, if these assumptions are correct, about 7 and
18% of the original UF, was oxidized by KF at
650 and 750°C, respectively, and 35% dispro-

portionated at each temperature,

The 35% lost by disproportionation is con-
siderably higher than the 5 to 10% shown experi-
mentally for similar conditions in the NaF-LiF
eutectic, This discrepancy may well be associated
with the fact that a deposited mirror of uranium
alloy is always noticed when KF is present but not
when NaF-LiF alone is used; potassium seems to
‘““catalyze’’ the disproportionation in some fashion,

For the steady-state condition, the apparent heat
of activation for the evolution of potassium is
18 kcal if the rates are considered to be proportional
to the steady-state potassium pressure at the two
temperatures. This is in good agreement with the
20-kcal heat of vaporization of potassium, particu-
larly since the 18-kcal figure applies to solutions
in which the activity of potassium is only 0.2,

It is interesting to note that, since the UF -to-UI:4
ratio does not change markedly with temperature
and the near-equilibrium steady-state activity of
potassium does not change from the value 0.2,
the equilibrium constant for the reaction is nearly
independent of temperature. This means that the
heat of reaction is very small, in spite of the
large positive standard free energy for the reaction,
Evidently, a large negative heat of complexing is

99
ANP PROJECT PROGRESS REPORT

involved, and the reaction is better represented by
the equation:

(X + DKF + UF,—> K UF,, + K° .

On the assumption that the activity coefficients
of KF and UF, are of the order of unity, the
relation
ayE Ao
o 4
K = o=AF°/RT _
a a
UF, “KF
()’XUF )(0-2)
-4 _qp-1s
(0.02)(0.4)

yields about 10-7 for the activity coefficient of
UF, (the concentration X is approximately 0.02)
in the NaF-KF-LiF eutectic at 1000°K. There
are as yet no data which permit an independent
estimate of the activity coefficient of UF, in this
system, This value seems to be surprisingly
small, since in the NaF-ZrF  solutions (in which
UF, must compete with ZrF, for the complexing
fluoride ions) the activity coefficient of UF,
from the reaction

2UF4 + Fec’:Fer + 2UF3

is about 10-1,

On the basis of these results, the NaF-LiF-KF
type of melts, for which the beneficial effects of
UF, are most needed, are the melts with which
UF, is least compatible, This is a consequence
of the extremely small activity coefficient for
UF, in the presence of the large excess of fluoride
ions and the correspondingly high activity of U°
required by the disproportionation equilibration to
balance the low UF , activity,

Experimental Preparation of Fluorides

B. J. Sturm
Materials Chemistry Division

The preparation of several structural-metal
fluorides was accomplished during the quarter.
Additional Crl':3 and FeF, were prepared by
hydrofluorination of anhydrous CrC[3 and FeCI2
at 600°C. Substantial quantities of CeF"3 and
Lc:l:3 were synthesized, and additional work was
done on complexes of ZrF
fluorides,

with several metal
The identities and the purities of all
the preparations were established through chemical,
x-ray, ond petrogrophic examinations,

100

The preparation of CeF . was accomplished by
treating aqueous solutions of either CeCl3 or
C:e(NC):,’)3 with aqueous HF, washing the precipi-
tated CeF , thoroughly with water by centrifugation,
and drying at 150°C. The compound LaF, was
prepared in a similar manner by using a solution
of LdCl3 obtained by dissolving La,0,. A small
quantity of CUF2 was prepared by dehydration of
CuF ,-2H,0 with HF at 600°C. The compounds
NaCrF, and NaNiF, were prepared by fusing the
proper quantities of NaF and structural-metal
fluorides in sealed nickel capsules,

Data were presented previously34 for the
compounds CrFQ-ZrF4, FeF,-ZrF ,, and NiF ,-ZrF ,
which weee.prepared by heating Zr 4 With equimolar
quantities of the structural-metal fluorides ot
950°C. The instability of CrF,.ZrF, and
FeF,-ZrF, was found to be due to atmospheric
contamination; if stored properly, these compounds
are, apparently, stable indefinitely. A compound
ZnF _.ZrF , with x-ray properties very similar to
those of Nin-ZrF4 was prepared by heating
equimolar quantities of ZnF2 and ZrF4 at 950°C
in sealed nickel capsules.

An equimolar mixture of AIF_ and ZiF , was
heated to 950°C in a sealed nickel capsule in an
attempt to prepare A|F3-ZrF4. Examination of
the product revealed that no reaction occurred;
only A|F3 and ZrF , were observed.

An attempt to prepare I'_'ei::,’-ZrF4 by heating
equimolar quantities of Fer and ZrF , in a nickel
capsule lined with silver gave material which
x-ray examination revealed to be FeF _.ZrF, and
an unknown phase, Chemical analysis szhoweé that
this material contained equivalent quantities of
Ag* and Fe** and, thus, that Fe3* had been
reduced by Ag® The replacement of the silver
liner by a gold liner gave a material that contained
small amounts of FeF_ and ZrF, but which
consisted primarily of a phase having essentially
the FeF,.ZrF, structure but a smaller cell size.
It is believed that this phase is FeF .Z¢F . It is
interesting to note that decomposition of FeF -ZrF
on exposure to the atmosphere yields a ma’reriof
with a cell size intermediate between that of
FeF ,-ZrF , and the presumed FeF,.ZrF,.

 

348. J. Sturm, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 91.
Color Studies of Alkali Metals in Equilibrium
with Alkali Halides at High Temperatures

B. H. Clampitt
Materials Chemistry Division

Davy335 and Bunsen and Kirshhoff3¢ observed
long ago that when a molten alkali halide was
subjected to electrolysis a highly colored phase
appeared near the cathode. This color has been
explained as being due to colloidal metal fogs37.38
or to subhalide formation. Bredig and his co-
workers39:40 and Cubicciotti (reference 41, for
example), who have made extensive studies of
metal salt systems, have investigated metallic
systems in which no colors were produced.
Corbett and von Winbush4? have studied the
solubility of metals such as cadmium and antimony
in their chlorides and have found relatively slight
color changes in the salts as the metals were
dissolved, In general, it appears that color changes
in molten salts have been often observed but that
colors in discrete metal phases have not been
described, Recent experiments in this laboratory
have shown that, in general, alkali-metal phases
in contact with alkali-metal salts are highly and
strikingly colored,

The experiments were conducted in a dry box
which could be evacuated and subsequently filled
with an inert atmosphere. A 2-in. steel pipe
welded through the bottom of the box and sealed
at the projecting end served to provide, when fitted
with a surrounding tube furnace, a heated zone
intregal with the box, Crucibles of nickel were
used for the containing vessels., The salts and
metals were of reagent grade, Fluorides and
chlorides were fused prior to use; bromides were

 

3SH. Davy, Trans. Roy. Soc. {Londonr) 97, 1 (18C7).

36R. Bunsen and G. Kirshhoff, Pogg. Ann. 113, 364
(1861).

37R. Lorenz and W. Eite!, Pyrosole, Lupzig, 1926.

38, Mollwe, Akad, Wissensch. Gottingen Nachrichten
Math. Physikal. Klasse., Fachgruppe 1l, Neue Folge 1,
Neo, 18, 203-207 (1935).

M. A. Bredig, J. W. Johnson, and Wm. T. Smith, Jr.,
J. Am. Chem. Soc. 77, 307 (1955).

40M. A. Bredig, H. R. Bronstein, and Wm. T. Smith, Jr.,
J. Am. Chem, Soc. 77, 1454 (1955).

41p, Cubicciotti, J. Am. Chem. Soc. 74, 1198 (1952).

42y D, Corbett and $. von Winbush, J. Am. Chem.
Soc., 77, 3964 (1955).

PERIOD ENDING MARCH 10, 1956

given no previous treatment., The alkali metals
were washed in the dry box with cyclohexane to
remove the protecting mineral oil, and the cyclo-
hexane was removed by evaporation,

The colors observed when a small amount of
metal is added to a molten salt mixture are listed
in Table 4.15, and the colors observed when a
small quantity of salt is added to a relatively
large quantity of liquid metal are listed in Table
4.16. Colors have been observed with every
combination tested, except lithium metal with a
lithium salt, Repeated tests have shown that the
colors are much more dependent on temperature
than on composition of the system, At 400°C the
mercury-like sheen of the alkali metal begins to
develop a green tinge, At 500°C, the green
changes to turquoise, Above 600°C the color
becomes increasingly deep blue and eventually
changes to violet, The color produced is inde-
pendent of relative quantities of metal and salt,
The color is produced whether or not the temper-
ature is sufficiently high to melt the salt; molten
metal is colored by solid salt,

Since the metal phase is sometimes on the top
and sometimes on the bottom, because of differ-
ences in interfacial tensions, it is difficult to
determine with certainty whether the salt or the
metal is colored. There are several reasons,
however, for the belief that the color is in the
metal phase. A definite quantity of salt, too small
to measure conveniently, can be added to a large
quantity of molten sodium without color production;
but the addition of one more grain may bring the
color in full intensity. The color production is
independent of whether or not the salt that can be
added without producing the color is so small
that it cannot be detected with certainty. If the
NaF-KF-LiF eutectic, for example, is contained at
600°C in a crucible with a partition that extends
to within ¥ in. of the bottom and sodium metal is
added to the salt on one side, the material on that
side turns blue immediately; the material on the
other side becomes dark but shows no blue color,
When NaCl is added to one compartment of a
similar crucible containing sodium metal at 600°C,
the contents of both compartments turn blue. Since
NaC!l sinks in molten sodium, however, this last
observation is not necessarily evidence for the
color being in the metal phase.

Bredig's3? results show a decrease in solubility
of the salt in the metal as the cation size becomes

101
ANP PROJECT PROGRESS REPORT

TABLE 4,15. COLORS OBSERVED WHEN SMALL AMOUNTS OF METAL ARE ADDED TO
LARGE AMOUNTS OF MOLTEN SALT

 

 

Composition of Salt Phase Experimental Temperature* Position of Metal
Alkali Metal o Color
(mole %) (TC) Phase

NaF-LiF-KF; 11.5-46.5-42 Li 500 Top Turquoise

Na 500 Top Turquoise

K 500 Top Turquoise
NaF-LiF; 40-60 Na 670 Top Blue

Li 670 Top Blue
LiF Li 845 Top * %
L.iCl Li 650 Bottom Metallic
LiCl-NaCl; 59-41 Na 600 Bottom Blue

Li 600 Bottom Blue
LiCl-LiF; 80-20 Li 600 Bottom Metallic
LiCl-LiBr; 25-75 Li 600 Bottom Metallic

 

*Determined by vapor pressure of alkali metal and melting point of the salt.

**Color not defined because furnace was at red heat,

TABLE 4.16. COLORS OBSERVED WHEN SMALL AMOUNTS OF SOLID SALT ARE ADDED TO
LARGE QUANTITIES OF METAL

 

Experimental Temperature

 

Alkali Metal Salt Phase (°C) Color
Li LiF 600 Metallic
LiCl 600 Metallic
LiBr 600 Metallic
NaF 600 Blue
NaCl 600 Blue
KF 600 Blue
KClI 600 Blue
Na LiF 600 Blue
NaF 600 Blue
NaCl 600 Blue
K LiF 500 Green
LiCl 500 Green
NaF 500 Green
KF 500 Green
KCI 500 Green

 

102
smaller. It is possible, therefore, that lithium
halides are too insoluble in metallic lithium at
700°C or below for the colors to be observed in
such systems,

The salt phases do not remain transparently
clear in these experiments. The salt phase darkens
in all cases but does not, apparently, show true
colors, Darkening under these conditions might
reasonably be related to F-center production in
the molten alkali halides. Bredig called attention
to a study of this problem by Mollwo,43 who
measured absorption spectra of molten alkali
halides which had been exposed to alkali-metal
vapor. Mollwo found the abserption spectra to
consist of broad bands, without structure, and to
have maximums at, for example, 790 my for sodium
salts. No relationship with the band maximums of
the F-centers was found, and, for a given metal,
there was no dependence on the anion used.

It appears to be very likely that the dark melts
observed in the present work resemble those which
gave the broad absorption spectra reported by
Mollwo and that they represent ““loose’’ F-centers,
that is, loosely bound electrons which can absorb
a broad range of energies, This in in contrast to
the lattice-locked F-centers in solid alkali halides,
for which the frequency of the absorption band is
related to the lattice constant, d, by the equation

ved? = 5.02 x 10=3 m2/sec

Also, the blue-green ‘‘metal fogs'' or ‘pyrosols,’’
mentioned by Mollwo and others as observable
during electrolyses of alkali halides, would now be
interpreted, according to the results obtained in
the present work, as being due to a colored metal
phase,

PHYSICAL PROPERTIES OF MOLTEN
MATERIALS

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

Vapor Pressures in the System KF-ZrF

S. Cantor
Materials Chemistry Division

The vapor pressures of various compositions
in the system KF-ZrF , are being measured by the

 

43g, Mollwo, Akad. Wissensch, Gottingen Nachrichten
Math. Physikal. Klasse., Fachgruppe II, Neue Folge 1,
No. 18, 203-207 (1935).

PERIOD ENDING MARCH 10, 1956

Rodebush-Dixon method44 previously used in these
laboratories. 4>  The method depends on the
determination of the pressure of inert gas in the
limbs of the vapor-pressure cell at which the
passage of the gas through the cell is blocked by
the vapor. This pressure is determined by the
change of levels in a differential manometer
connected to the two limbs of the cell, Calcu-
lations that involve the interdiffusion of the vapor
and the gas at the temperatures of interest indicate
that 5 to 10 mm Hg represents the lower limit of
accurate determination of vapor pressure,

Calibrated platinum—platinum-rhodium thermo-
couples were used in conjunction with a Leeds &
Northrup portable precision potentiometer (model
8662) to measure temperatures, The apparatus
was tested by determining the vapor pressure of
ZrF4 and comparing the value obtained with that
obtained by Moore46 with a similar apparatus and
with values obtained at Battelle Memorial Institute
(BMI} by the transpiration method.47 The vapor-
pressure equations obtained from these three
investigations are shown in Table 4.17.

[t was assumed that the pressure measured was
due to ZrF, only for the region studied in the
KF-ZrF , system, This assumption appeared to be
justified, since the equilibrium vapor pressure of
pure KF is quite small at the temperatures studied
and petrographic and x-ray analysis indicated
essentially pure ZrF4 in the sublimates obtained.
The results of these measurements are compiled
in Table 4.18 in terms of the constants A and B
of the vapor-pressure equation,

B

fog P (mm) = A - TCK)

 

The constants were obtained by a least-squares
treatment of the data., The heat of vaperization,
which was obtained from the quantity B, appears
to decrease with increases in the KF fraction of
the mixture, while the low volatility of ZrF4 from
the mixtures suggests that the heat of vaporization

 

44w, H. Rodebush and A. L. Dixon, Phys. Reuv. 26,
851 (1925).

45R. E. Moore and C. J. Barton, ANP Quar. Prog, Rep.
Sept, 10, 1951, ORNL-1154, p 136,

46p, E. Moore, ANP Quar. Prog. Rep. Sept. 10, 1952,
ORNL-1375, p 147.

47, A, Sense, M. J. Snyder, and R. B. Filbert, Jr.,
J. Phbys. Chem. 58, 995 (1954).

103
ANP PROJECT PROGRESS REPORT

TABLE 4.17. YAPOR-PRESSURE EQUATIONS FOR ZrF4 OBTAINED BY VARIOUS INVESTIGATORS

 

Calculated Pressure (mm Hg)

 

 

 

Investigator Vapor-Pressure Equation
o= P = At 1000°K At 1100°K

11,320

Moore log P = 12.46 — 13.8 145
12,376

BMI log P = 1340 — — 10.5 141
11,506

This study log P = 12,70 — — 15.6 174

 

TABLE 4,18. SUMMARY OF CONSTANTS OBTAINED FROM VAPOR-PRESSURE

MEASUREMENTS OF THE SYSTEM KF-ZrF4

 

Cemposition

 

 

 

 

(mole %) Constants Heat of Vaporization, H
(keal)

Zrl""4 KF A B

84.50 15.50 11.88 10,820 49.5

69.97 30.03 11.63 10,720 49.0

65.87 3413 9.754 8,881 40.6 .

60.07 39.93 7.635 6,806 31.1

49.94 50.06 8.211 7,925 36.2 -
could be expected to increase, A further investi- oy UNCLASSIFIED
gation is being carried out to ascertain whether the 150 — i ‘ | .
data are subject to some source of systematic ' L
error which might be responsible for the apparent - ':LHO'(?R?SU[;ZTA o NGFlz;; [
discrepancy., For this purpose an independent IE Vo o BMI DATA ON Nof -Zrf, 4 . i
static method of measuring vapor pressures will be = A iy
used, % %0 i LT 7

a RAOULT’S LAW FOR |
The B800°C isotherms of vapor pressure vs £ 70 |- UNDISSOCIATED ""QU--'-Q—S“‘\}A;'----NGF*--Z‘& . -
|
composition for the NoF-ZrF‘11 and KF-ZrF, 2 o L . //l W /|
systems are shown in Fig. 4.15. The vapor pressure 2 | s Law //.\\_K}_ZrFq
for pure liquid ZrF was calculated from the 20 '__._.;__fifgifpopzrw 215 34 /]
sublimation pressure and the heat of fusion /r,//! | /fi“/
‘ |

(9 kcal/mole) as determined from the loweringof 10 s . M/fif ‘
the freezing point of ZrF , by the addition of NaCl. NeF 10 20 30 40 50 60 70 80 90 ZrF,
The results indicate clearly that lower vapor kF Z¢F, (mole %) .
pressures can be obtained in the KF- ZrF system
than in the NaF- ZrF system, presumably because Fig. 4.15. Changes in Vapor Pressure with

KF furnishes a greater activity of fluoride ion as
a result of the greater radius of the potassium ion,

104

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

at 800°C

Composition for the Systems NaF- ZeF,
ZiF,

and KF-
Vapor Pressure of FeCl,

C. C. Beusman
Oak Ridge Institute of Nuclear Studies

Measurements of the activity of ferrous chloride
in melts of the alkali-metal chlorides have been
started with the use of vapor-pressure methods,
A transportation apparatus has been assembled
and will be used in the temperature ranges where
both components are volatile, For lower temper-
atures, the Rodebush-Dixon method44 will be used.

For calibration purposes, the vapor pressure of
pure ferrous chloride has been measured in the
Rodebush-Dixon apparatus. The data are in fair
agreement with literature values and are well
represented by the equation

8,985
log P (atm) = 23,080 — TR

Surface Tensions of Molten Salts

 

— 5.234 log T (°K) .

S. Langer
Materials Chemistry Division

Surface tensions of molten salts have been
examined by a number of investigators48~33 who
have documented their findings in the open liter-
ature; a determination of surface tension of a
NaF-ZrF -UF, fuel mixture has been made by
Cohen and Jones®4 of ORNL. However, no
systematic study of surface tension as a function

 

48y, k. Boardman, A. R. Palmer, and E. Heymann,
Trans, Faraday Soc. 51, 277 (1955).

49y, |. Semenchenko and L. P. Shekhobalova, J. Phys.
Chem. (USSR) 21, 613, 707, 1387 (1947). AEC trans.

S0F, M. Jaeger, Optical Activity and High Temperature
Measurements, p 235-307, McGraw-Hill, New York, 1930.

51g. s. Ellefson and N. W, Taylor, J. Am. Ceram. Soc.
21, 193, 205 (1938).

52¢. F, Baes, Jr., and H. H. Kellogg, J. Metals 5,
643 (1953).

53y, D. Kingery et al., Development of Metal-Ceramic
Compositions Suitable for Service at Elevated Tempera-
tures, NEPA-1170 (Sept. 1949); NEPA-1234 (Nov. 1949);
NEPA-1848 (April 30, 1951); and F. H. Norton et al.,
Study of Metal-Ceramic Interactions at Elevated Temper
atures, NY0-3136 (Oct. 1, 1951); NYO-3137 (Jan. 1,
1952); NYO-3139 (April 1, 1952); NYO-3142 (Jan. 1,
1953); NYO-3143 (Feb. 1, 1953); NYO-3144 (Feb. 1,
1953); NYO-3145 (April 1, 1953); NY0-6290 (Oct. 1,
1953); NYO-6291 (July 1, 1953); NYO-6294 (April 1,
1954).

545, |. Cohen and T. N. Jones, Preliminary Surface
Tension Measurements of the ARE Fuel (Fluoride
Mixture No. 30), ORNL CF-53-3-259 (March 27, 1953},

PERIOD ENDING MARCH 10, 1956

of composition, purity, etc., has yet been made for
materials of interest to the ANP program. Since
surface tension has been shown to be important,
in certain cases, to heat-transfer>> and corrosion>6
processes, a study of this property is under way.

The method chosen for this study is the sessile-
drop technique, which has been successfully
employed by Ellefson and Taylor,”7 Baes and
Kellogg,52 and Kingery et al.33 This method was
selected in preference to the maximum-bubble-
pressure technique because of the uncertainty of
the maximum-bubble-pressure method with liquids
having contact angles greater than 90 deg (measured
through the liquid). A profile drawing of a sessile
drop is shown in Fig, 4.16. A photograph is taken
of the sessile drop on a supporting plaque, and the
dimensions 2x, the maximum diameter of the drop,
and z, the height from the maximum diameter to the
apex, are measured. These dimensions, along with
the density of the molten salt, are sufficient to
determine the surface tension, The contact angle
8, which is difficult to measure directly, can be
calculated if the additional quantities x“ and z°
are known,

UNCLASSIFIED
ORNL—LR--DWG 127933

 

 

 

 

 

Fig. 4.16. Profile of a Sessile Drop.

The equipment required for this study has been
assembled and is in working order, Preliminary
experiments are being carried out with NaF-ZrF
(53-47 mole %), and nickel, Inconel, platinum, an
graphite are being used as supporting plaques,
The condition of the surface of the supporting
plaque is known to be extremely important in

 

55w, k. Stromquist and R, M. Boarts, Effect of Wetting
on Heat Transfer Characteristics of Liquid Metals,
ORO-60 (Jan. 31, 1952).

36, W. Taylor, J. Nuclear Energy 2, 15 (1955).

57B. 5. Ellefson and N. W. Taylor, J. Am. Ceram. Soc.
21, 193, 205 (1938).

105
ANP PROJECT PROGRESS REPORT

determining whether the molten drop will wet the
plaque, and the effects of various atmospheres on
the wetting properties are also being studied. For
the sessile-drop method to be successful, the
molten drop must not wet the supporting surface,

The preliminary experiments show that nickel
and Inconel surfaces are not immediately wetted by
the sessile drop when it melts, even if the surfaces
have previously been hydrogen-fired, The contact
angle is only slightly greater than 90 deg, and, on
standing, the contact angle slowly decreases until
the surface is wetted by the salt, whether the
atmosphere is helium or a vacuum of about 2 4.
A clean, polished, degreased platinum surface is
wetted by the drop immediately upon melting in
both vacuum and helium, However, if a nickel
plaque that has been previously hydrogen-fired is
used as the supporting surface and a hydrogen
atmosphere is maintained in the furnace tube so
that any thin film of oxide which has formed since
the hydrogen-firing process will be reduced, the
sessile drop will wet the surface as soon as it
starts to melt, The droplet will spread over the
nickel surface by the time it is entirely molten,
It oppears from these preliminary experiments that
truly clean hydrogen-fired nickel surfaces are
completely wetted by molten Nc:ZrF5 and that, if
thin oxide films are aliowed to form on the metal
surface after hydrogen-firing, the films will prevent
wetting until they are dissolved by the molten salt,
These observations are difficult to reconcile with
the lack of evidence of wetting in nickel purifi-
cation vessels in which intensive hydrogen
treatments of molten fluorides are carried out for
hours at 800°C, and therefore a more thorough
study of the behavior of the sessile drop is being
made.

Ellefson and Taylor57 measured the surface
tension of LiF by the sessile-drop technique with
‘*electrode’’ graphite as the supporting surface,
Their results are in good agreement with those of
Jaeger,?0 which were obtained by the maximum-
bubble-pressure technique, Thus it can be assumed
that at least some fluoride melts will not wet
graphite, Experiments with the use of graphite
surfaces are under way,

 

58G. M. Watson and F. W. Miles, ANP Quar. Prog.
Rep. Dec. 10, 1955, ORNL-2012, p 96.

5%, D. Harkins, Physical Methods of Organic Chem-

istry, vol. |, Part 1ll, 2d ed., chap. IX, p 355 (ed. by
A. Weissberger), Interscience, New York, 1949.

106

Density and Surface Tension of Molten

KCI ot 800°C
F. W. Miles

Materials Chemistry Division

Apparatus 38 added to the conventional equipment
for fuel purification is being used for the study of
variations of density and surface tension of molten
salts with changes in chemical composition,
Calibration experiments with molten-salt standards
are under way, and measurements were made of the
surface tension and the density of molten KCl to
determine the accuracy and precision of the
methods,

Both the density and the surface tension of a
molten salt can be determined from the same set of
measurements of the maximum pressures required
to blow helium bubbles into the melt through a
capillary orifice at various accurately known
depths. This is the standard method for surface-
tension measurement and has been frequently
described in the literature, The procedure being
used at present was obtained from various
sources,”?-61 and the design of the capillary
orifices was substantially guided by Porter's
analysis,82 For the calculations Sugden’s method
of correction®9:63 for bubble distortion, as modified
by Tripp and Young®4 for single bubbles, was used.

The density and the surface tension of molten
KCl at 800°C were determined with three different
orifices (two with the same radius) for purposes
of calibration. Measurements were made with each
orifice at five different depths. The depths used
were approximately ‘4, ]/2, Y, ]?32, and 2 in, The
results are summarized in TaABIe 4.19.

The second determination with orifice No., 1 was
purposely performed without recalibration of the
tip aofter exposure to the atmosphere following

immersion in KCl, The nickel tube carrying

 

60g, Sugden, The Parachor and Valancy, p 209,
Routledge, LLondon, 1930.

61, w. Taylor, Solid Metal-Liquid Metal Interaction
Studies. Part I: The Surface Tension of Sodium, AJE.R.E,
Report M/R 1247 (Sept. 1953).

$2A. W. Porter, Phil. Mag. 9, 7th Series, 1065 (1930).
633, Sugden, J. Chem. Soc. (London) 121, 858 (1922).

64y, Pp. Tripp with T. F. Young, Maximum Bubble
Pressure Method for Measurement of Surface Tension,
Ph.D. thesis, University of Chicago, 1934. Many help-
ful suggestions regarding the experimental assembly
grnd procedure were received from Professors Young and
ripp.
PERIOD ENDING MARCH 10, 1956

TABLE 4.19., DENSITY AND SURFACE TENSION OF KCI AT 800°C

 

Orifice Radius*

Observed Density

Observed Surface Tension

 

Oritice No. (em) (g/ml) (dynes/cm)
1 0.0729 1.516 95.2
1 0.0729 1.523 92.9
2 0.0729 1.501 91.9
3 0.0430 1.510 93.8

Av 1.513 * 0.007

Av 93.5 * 1.1

 

*Orifice radius obtained by calibration with water for which the surface tension was simultaneocusly determined with

a ring tensiometer,

orifice No. 2 was found to be slightly bowed after
the experiment; this may account for the lower
surface-tension values found with this orifice,

Previously published values for these properties
of KCl are 1.509 g/cm3 and 95.8 dynes/cm by
Jaeger65 and 1.5102 g/em3 by Van Artsdalen and
Jaffe. 66 Although the precision and accuracy
indicated above are perhaps acceptable, it is
hoped that further improvement can be obtained
without undue difficulty, Additional tests with
KC| and with KCl-NaCl mixtures will be completed
before measurements on fluoride mixtures are
attempted,

PHYSICAL CHEMISTRY OF FUSED SALTS®?

E. R. Van Artsdalen
Chemistry Division

It appears to be established that, in the solid,
Cc:tCl2 forms 1:1 complexes with KCl, RbCIi, and
CsCl. It has been claimed6® that viscosity
measurements of the liquid mixtures indicate that
these complexes exist in the molten state.
Freezing-point depression studies were made in
fused sodium nitrate to test whether any such
complexes are stable in the dilute range, Mixtures
of CsCl and CaCl, depressed the freezing point in

 

65F. M. Jaeger, Z. anorg. u. allgem. Chem 101, 185
(1917).

66E. R. Van Artsdalen and I. S. Yaffe, J. Phys. Chem.
59, 118 (1955).

67 Details of this work will be published in separate
reports and articles from the ORNL Chemistry Division.

68y, L. Strelets, V, N. Zhludneva, and |. L. Reznikoyv,
Zhur. Priklad, Khim, 28, 643 (1955).

a strictly additive, ideal manner, Therefore the
evidence indicates that no CsCl-CaCl, complexes
exist in the dilute range when dissolved in sodium
nitrate at about 300°C.

Density and preliminary conductance data have
been obtained for several molten rare-earth halides.
These liquid salts tend to attack quartz rather
severely, and, as a result, the conductance
measurements show poorer precision than that
previously obtained for alkali halides. Some data
for LuuC[3 and LaBr, are given in Table 4.20.

Density and conductance of several fused
alkaline-earth halides and their mixtures with
some alkali halides are being investigated, It
is hoped that information obtained from this study
will throw light on the question of the existence
of complexes in these liquid-salt mixtures and,
therefore, be complementary to the freezing-point
studies of dilute mixtures,

PRODUCTION OF PURIFIED FUEL MIXTURES
G. J. Nessle G. M. Watson

Materials Chemistry Division

Use of Copper<Lined Equipment for Fuel
Purification

F. L. Daley

Materials Chemistry Division

Reaction vessels consisting of copper liners
in stainless steel jackets have been proposed as
substitutes for the nickel reactor cans that have
failed frequently in recent months as a result of
sulfur contamination in the raw materials vused
for large-scale fuel production. Three smail-scale
experiments (2.5 kg) were performed to determine
the effects that might be encountered with the

107
ANP PROJECT PROGRESS REPORT

change from nickel to copper-lined stainless steel
reactors, In the first two experiments, mixtures
of NaF-ZrF -UF, (50-46-4 mole %) were purified
by the standord procedure, The same copper liner
was used in both preparations, and no corrosion
was apparent, The copper liner was contained in
a nickel reactor in one experiment and in a stainless
steel reactor in the other. Both experiments were
successful in that the impurities in the product
were at a normal level, averaging 70 ppm iron and
40 ppm chromium, The copper contamination was
less than 20 ppm in each case,

On the basis of these two experiments, it was
decided to attempt purification of large batches in
copper-lined stainless steel reactors. It appeared
possible, however, that the introduction of stainless

steel reactors would change the existing corre-
lations between the concentration of impurities in
the melts and the concentration of HF in the effluent
hydrogen during the stripping process, Since this
correlation is used for control of the processing
operation, it was necessary to establish that its
utility was not impaired by the change to the
copper—stainless steel system,

Accordingly, a 2.5-kg batch of prepurified NaF-
7_rF4-UF4 mixture was deliberately contaminated
with 2000 ppm iron (as FeF,) and treated with
hydrogen at a flow rate of 300 cm3®/min. Filtered
samples of the melt were drawn at intervals and
submitted for chemical analysis, and the HF con-
centration of the exit hydrogen was monitored in
the usual way. The data obtained are shown in

Table 4.21.

TABLE 4.20. DENSITY DATA FOR LaCI3 AND LaBrqy

 

Density, p = a - bt

 

 

o Experim;:fol Melting Temperature (a/em) Standard Deviation
int Range (g/cm‘?')
(°C) (°C) a b x 103
La C13 857.5 873 to 973 3.8773 0.774¢ 0.0001,
La Br3 779.6 796 to 912 5.0089 1.0960 0.0004

 

TABLE 4.21. CONCENTRATION OF IRON IN A PREPURIFIED NuF-ZrF4-UF4 MELT CONTAMINATED WITH
2000 ppm Fe (AS Fer) IN RELATION TO HF CONCENTRATION IN EFFLUENT GAS

 

Hydrogen Passed Through Melt

HfF Concentration in Effluent

Contaminant Concentration (ppm)

 

 

(liters) (moles /liter) Fe Cr Ni Cu

7 3.3 x 10-3 1675 90 55

41 1.3 x 10-3 1640 100 100

88 1.0 x 10-3 1070 100 110
167 8.3 x 10~4 330 100 90 5
492 2.7 x 10~4 75 90 20 5
581 2.0 x 10~4 75 90 20 5
663 1.5 x 104 55 100 25 5
987 1.1 x 104 55 90 60 5

 

108
From these results, it appears that the melt is
adequately purified at effluent-gas concentrations
of 2 x 10~4 moles/liter, This is similar to the
situation in unlined nickel equipment, As antici-
pated, no perceptible changes were observed in the
small chromium and nickel contents,

To test corrosion of the copper liner under
extreme conditions, a purified melt was treated
with HF for 30 hr at 800°C; no hydrogen stripping
was used, Six samples of melt obtained at 4- to
8-hr intervals were submitted for chemical analysis,
Less than 5 ppm of copper was found in any
sample. The nickel concentration rose from about
50 ppm to about 300 ppm during this test. The
nickel undoubtedly came from the nickel dip tube
through which the HF was introduced; whether
anodic dissolution of the nickel and resultant
protection of the copper occurred is not clear.
In any event, the quantity of nickel introduced
could easily be removed by hydrogenation and
should not prove to be serious from the standpoint
of corrosion of the purification equipment,

Visual examination of the liner and the assembly
after these tests showed no perceptible deteri-
oration. |t appears that the use of copper-lined
stainless steel reactors on the 250-1b or larger
scale equipment should be quite satisfactory with
ZrF ,-bearing mixtures, Tests with NaF-KF-LiF-
UF, and BeF ,-bearing materials will be made in
the near future,

Preparation of ZrF, by Gas-Phase
Hydrofluorination of ZCl,

F. L. Daley
Materials Chemistry Division

The equipment for conversion of ZrCl 4 to ZrF, by
direct hydrofluorination of the solid%? is being re-
built to overcome certain mechanical difficulties,
and an alternate process for the conversion is
being studied on a small scale. Results of calcu-
lations by Gully indicate that the over-all rate
of reaction may be controlled by the rate of gaseous
diffusion into and out of the solid phase rather
than the interfacial reaction rate. Also, one of the
unsolved problems in this process is the elimination
of small percentages of oxide or water from the

product. It is apparent that an alternate process

 

69A. J. Gully, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL-1947, p 83.

PERIOD ENDING MARCH 10, 1956

that would eliminate the solid phase and exclude
possibilities of atmospheric contamination might
effect substantial in the rate of
reaction, the yield, and the quality of product.

improvements

In the alternate process being studied, a stream
of hydrogen is passed at a constant flow rate
through a bed of ZrCl, held at a constant temper-
ature, in the neighborhood of 300°C. The effluent
gas, a mixture of hydrogen and ZrC|4 vapor of a
definite composition, governed by the temperature
and gas flow rate, is then introduced into a mixing
chamber, at 325°C, into which HF is also intro-
duced at a constant rate. The reaction chamber
is baffled, and the gases are introduced tan-
gentially to allow for better separation of the
solid ZrF,. A vent to a KOH trap and a hydrogen
burner allows continuous disposal of gases.

A small experimental assembly was constructed,
and twa preliminary experiments were made, These
experiments were encouraging in that they yielded
a high-quality product; however, it was found that
plugging tended to occur at the entrance to the
mixing chamber,  Some indicated changes in
design are being made in an attempt to avoid
plugging.

Laboratory-Scale Purification Operations

F. L. Daley

Materials Chemistry Division

The standard hydrofluorination-hydrogenation
process was used, with appropriate modifications,
to prepare a number of especially pure materials
requested for various purposes. Of these materials,
only BaF, and RbF were materials not previously
prepared in the routine processing schedule,

Pilot-Scale Purification Operations

C. R. Croft J. P, Blakely
J. Truitt W. T. Ward

Materials Chemistry Division

The 5- and 50-1b facilities processed 70 batches,
totaling about 700 Ib, of various compositions for
use in small-scale corrosion testing, phase-
equilibrium studies, and physical-property de-
terminations., A very considerable increase in
requirements for nonstandard compositions has
been noted in recent months, and, at present, a
backlog of about 1100 |b of special orders is on
hand, with NaF-KF-LiF-UF, mixtures comprising
the bulk of this backlog. Since the stockpile of
standard Zer-bearing fuels is adequate, at present,

109
ANP PROJECT PROGRESS REPORT

the 250-1b scale equipment has been shut down,
and the personnel are being used to operate the
small-scale equipment on a 24-hr, three-shift
basis, It is anticipated that the 250-Ib scale
equipment can be used for NaF-KF-LiF-UF
mixtures if the demand becomes sufficient and i?
the copper-lined stainless steel reactor cans prove
to be compatible with these materials,

In the processing of compositions containing
Bef,, sulfur corrosion of the smali-scale nickel
equipment has become more troublesome. Since
the available BeF, is known to have a high sulfur
content relative to the other salts used, such
corrosion was, to some extent, anticipated, H is
apparent that [arge-scale production of BeF
compositions will necessitate more rigid specifi-
cations for BeF, and probably a better container
material than nickel. A few small-scale copper-
lined stainless steel reactors are being fabricated.
it is possible that reactors of this type may prove
to be useful and versatile in investigations of new
fluoride compositions,

Production-Scale Operations

J. E. Eorgan J. P. Blakely
Materials Chemistry Division

During the past quarter, 23 batches, totaling
approximately 5725 |b, of fluoride compositions
were processed in the 250-ib facility, The arrival
of 2,500 |b of a 10,000-Ib order of hafnium-bearing

ZrF4 from an outside vendor has eliminated the

shortage of usable ZrF, for production purposes.
The remaining 7500 1b will be delivered within
the next three months, and, if specifications are
met, sufficient ZrF  will be on hand to meet all
demands for zirconium-base mixtures the
remainder of the fiscal year. The analyses of
contents of the eight drums in which the first

for

2500 b of ZrF, was shipped are given in
Table 4.22.
The incidence of reactor-can failures in use

in the production-scale facility has continued to
increase, During this quorter five large reactor
cans were damaged beyond repair because of
leaks, A more thorough investigation of the causes
of these failures was initiated in conjunction with
the |International Nickel Company, Inc, Chemical
and metallographic examination revealed that
sulfur embrittlement of the nickel was causing
the cans to rupture. Analyses of the nickel in the
reactors at the points of failure revealed as much
as 0.035% suifur, which is well above the critical
content for failure,

As a result of these findings and in view of the
difficulty suppliers are having in
meeting the present sulfur specifications, studies
of other container materials were made, as
described above, Since the laboratory studies
indicate that copper-lined stainless steel reactors
will prove to be suitable, two 250-Ib reactors have
been ordered,

The production facility was shut down in
January. The duration of the shutdown will depend

commercial

TABLE 4,22, CHEMICAL ANALYSES OF PURCHASED .?_'rF4

 

Major Constituents (wt %)

Minor Constituents (ppm)

 

 

 

Batch Ne. Zr F Na K c Ni Cr Fe S
112-1 54.6 44.7 0.59 0.1 0.09 195 60 500 35
-2 54.4 45.5 0.46 0.1 0.10 2 55 575 40
.3 54.2 43.5 0.59 0.1 0.26 65 35 375 45
1130-1 53.2 44.3 0.35 0.1 0.28 90 35 705 30
-2 55.3 44.2 0.22 0.1 0.30 90 35 760 90
-3 53.8 44.3 0.28 0.1 0.23 70 35 580 100
-4 53.5 44.1 0.59 0.1 0.20 70 30 485 100
-5 53.7 44.5 0.86 0.1 0.21 105 30 340 110

54.5

Theoretical

45.5

 

110
upon the length of time needed for fabrication of
the new reactors, the consumption of the present
stockpile of fuel materials, and the status of an
accelerated program for testing new fluoride
compositions,

The facility for the conversion of hafnium-free
ZrC|4 to ZrF, was operated three times during the
quarter, Mechanical failures caused by binding of
the agitator shaft in the packing gland terminated
each attempt to use the facility. The design of the
reaction chamber has been reviewed, and tolerance
changes have been made. New parts are being
fabricated for assembly; the unit should be
assembled and tested again in the near future,

Quality Control of Raw Materials and Products

F. L. Daley
Materials Chemistry Division

Chemical analyses and petrographic examinations
of fuel raw materials are routinely made. However,
the most satisfactory specification check appears
to be actual preparation of small batches of
purified fuel raw material,

mixtures from the

Accordingly, such checks were made on samples
of various zirconium compounds supplied by
commercial vendors, These tests showed that one
otherwise-suitable material contained sufficient
water (or hydrolysis product) to yield an oxyfluoride
component in the fuel, Tests of this type will be
continued as additional materials are supplied.
It is intended to extend these tests to include
specimens of purchased materials, The information

TABLE 4,23,

PERIOD ENDING MARCH 10, 1956

obtained will be a guide to necessary modifications
of the production procedure,

In addition to the HF-monitoring procedure and
to the chemical analyses of the product, which are
standard quality-control features of the 250-1b
production-scale operation, examination of the
samples by x-ray diffraction and petrographic
techniques has been added as a routine procedure,
Of some 5000 Ib produced during the quarter, only
one 250-Ib batch was found to contain oxides or
oxyfluorides and thus to require reprocessing.
Contamination of this batch probably resuvited
during a repair on the equipment while the batch
was in preparation,

Batching and Dispensing Operations
F. A. Doss J. P. Blakely

Materials Chemistry Division

During this quarter, 111 batches, totaling
approximately 5900 Ib, were dispensed to various
users in batch sizes ranging from 1 to 250 ib.
The total amount dispensed represents an increase
of 2000 Ib in the fluoride~mixture consumption, in
comparison with the amount consumed during the
previous quarter, but consumption is still less
than that expected. As a result, there was a small
gain in the stock inventory. A material balance
for the quarter is given in Table 4.23.

The consumption of the NaF-ZrF, and the
NaF-ZrF ,-UF, mixtures is not expected to in-
crease during the next quarter, and therefore it is
anticipated that requests for these compositions
can be filled from the stockpile until the new
reactors for the production facility are available.

MATERIAL BALANCE

 

 

 

 

 

Material (Ib)

NoF-ZrF4-UF4 chF-ZrF‘1 Special - Total

(50-46-4 mole %) (50-50 mole %)
On hand at beginning of quarter 5,715 2475 661 g8.851
Produced during quarter 5,728 694 6,422
Total 11,443 2475 1355 15,273
Dispensed during quarter 5,208 600 764 6,572
On hand at end of quarter 6,235 1875 591 8,701

 

111
ANP PROJECT PROGRESS REPORT

Filling, Draining, and Sampling Operations

N. V. Smith F. A, Doss
J. P. Blakely W. T. Ward

Materials Chemistry Division

During this quarter, the filling, draining, and
sampling operations were considerably accelerated
because of the large number of engineering tests
that were initiated. |f the present schedules
prevail, this accelerated pace should continue for
the remainder of the fiscal year,

Approximately 4000 lb of processed fluorides
and an equivalent amount of liquid metals were
used in batch sizes ranging from 5 to 500 Ib to
charge tests rigs during this quarter, About 95%
of these operations dealt with engineering tests
requiring charges in the 50-lb range.

Liquid-metal salvage operations are assuming
major proportions, along with the accelerated
operational schedule, The disposal of bulk NaK
is well provided for, but bulk sodium disposal
is becoming a greater problem. A proposed
facility for the disposal of sodium under water, by
a procedure similar to that used for NaK, is being
studied, However, the necessity for providing
heat to melt the sodium for underwater injection
at a remote location makes this approach appear
to be uneconomical.

In-pile loop No. 4 was filled with the enriched-
uranium mixture NoF-ZrFA-UF" {53.5-40-6.5
mole %) during the quarter, A new batch of the
same composition has been processed for filling
loop No. 5.

In the filling of previous in-pile loops, the
danger of blowing the fuel into the pump bowl had
become a major problem, Since gas in the small
sump tank could be vented through the pump, loop
No. 4 was filled by gravity flow. The batch was
transferred to a special storage can which was
equipped with a dump line from the bottom. After
the can had been connected to the sump, heat was
applied and the batch was transferred successfully,
This procedure has therefore been adopted as
standard for the filling of loops of this type,

112

Calibration of ART Enricher
F. A. Doss J. P. Blakely

Materials Chemistry Division

The final portion of the uranium required to bring
the ART to criticality is to be added as molten
NGZUF6 from a semiautomatic, remotely operated
enricher.,  This mechanism utilizes a piston,
driven by an accurately machined screw, which
displaces the molten salt from a sump and causes
it to flow over a weir and then by gravity into the
reactor sump,

A prototype of this equipment was tested during
the high-temperature critical experiment, While
the operation was successful, in general, it was
observed that the machine behaved erratically
when attempts to add small amounts of material
were made, After the experiment, the enricher was
preserved intact and was remounted and carefully
leveled in Building 9212. An inert-atmosphere box
containing a glass window was then fitted around
the orifice for further tests. The amount of material
delivered by a specified number of turns of the
screw was ascertained by weighing a graphite
receiver which caught each discharge; in addition,
visual observation sufficed to disclose the number
of turns required to start the flow. Attempts were
made to force various increments of molten salt
from the equipment with several levels of melt in
the enricher sump. A condensation of the data
obtained is presented in Table 4.24.

In each attempt, flow began only after the piston
had been depressed at least 0.4 turn, and usually
more than 0.6 turn was required. Apparently, the
surface tension of the material is so high that the
liquid ‘‘bulges’’ above the weir before flow begins,
Accordingly, additions smaller than about 125 g
often cannot be made at all, while the percentage
uncertainty in additions of 250 g (1 turn) is of the
of +10%. Reproducibility becomes con-
siderably better, especially on a percentage basis,
when larger additions are made. If this enricher is
used only to add increments greater than about
500 g, it will be satisfactory without change. |f
smaller additions are necessary, the enricher must
be redesigned to give smaller holdup in the weir
(and hence a higher static head per unit of piston
travel) or to change the shape of the dam.

order
PERIOD ENDING MARCH 10, 1956

TABLE 4.24. REPRODUCIBILITY OF INCREMENT DRAWN FROM ENRICHER AS A
FUNCTION OF TURNS OF THE SCREW

 

‘ Mean Deviation from
Increment Drawn (g) ©

 

 

Number of Increment Attempted Mean Increment Mean Increment
Attempts (number of turns) Minimum Maximum (g/turn) (g/turn)
5 0.6 145 320 320 41
5 0.8 207 241 271 13
2 0.9 211 233 246 12
13 1.0 242 286 257 9
6 2.0 497 516 254 4
10 5.0 1238 1282 252 2

 

113
ANP PROJECT PROGRESS REPORT

5. CORROSION RESEARCH

W. D, Manly
Metallurgy Division

FORCED-CIRCULATION STUDIES
J. H. DeVan

E. A, Kovacevich R. S, Crouse
Metallurgy Division

Fluoride Fuel Mixtures in Inconel

Three Inconel forced-circulation loops were op-
erated to study the effect on corrosion by fluoride
fuel mixtures of increases in the ratio of loop
volume to loop wall surface area. The fluoride
fuel mixture circulated in these loops was NaF-
ZeF UF, (53.5-40-6.5 mole %). The following op-

erating conditions were utilized for all three loops:

Operating time, hr 1000
Maximum fluoridesmixture temperature, °F 1500
Temperature gradient, °F 200
Maximum tube wall temperature, °F 1600
Reynolds number 10,000
Yelocity, fps 6.5
Length of heated section, ft 17
Total length of loop, ft a1

The system variables and the corrosion results for
the three loops are given in Table 5.1.

[t may be seen that the quotients of heated sur-
face area and total loop volume were varied in the
ratio 1:2:4. The volume increases in loops 7425-6
and -7A were obtained by introducing reservoirs at
the hairpin turns connecting the first and second
heaters, as shown in Fig. 5.1. Since heat is not
normally introduced at this hairpin turn, Calrod
heaters were placed around the reservoir cans to

maintain the fluid temperature established by the
tirst heated section. The heating of the reservoir
cans of the fluid did not increase the temperature.
The walls of the cans were therefore at the same
temperature as the wall of the tube comprising the
hairpin section in standard loop 7425-1A, and the
temperature distributions around the heated sec-
tions, as well as the rest of the loop, were main-
tained the same for ail three loops. The fluid
temperature at the reservoirs was approximately
75°F lower than the maximum fluid temperature,
which was attained immediately past the second
heated section in each loop.

The variations in depth of attack in the three
loops show no particular correlation with loop
volume, as may be seen in Table 5,1. As previ-
ously reported, however, the attack to a depth of
9 mils in loop 7425-1A is higher than appears to
be normal for standard loops; ! it was expected to
be in the range of 5 to 7 mils., This expected
value provides a more reasonable comparison with
the attack to a depth of 6 mils observed in loop
7425-6, in which the system surface-to-volume
ratio was a factor of 2 less than that in loop
7425-1A. No measurable increase in attack as a
result of the volume increase was found,

The attack to a depth of ¢ mils in loop 7425-7A,
in which the system surface-to-volume ratio was a
factor of 4 less than that in loop 7425.1A, was
accompanied by a very thin, continvous gold-
colored layer that covered substantial portions of

 

1. H. DeVan, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 105.

TABLE 5.1. EFFECT ON CORROSION OF YARYING THE HEATED-SURFACE-TO-VOLUME RATIO

 

 

Ratio of Heated Maximum
Loop Sys‘l‘ern golume Surface Area to Depth of
(in.%) Volume Attack (mils)
7425-1A 80.8 2.10 9
(standard loop)
-6 251 0.98 é
<7 A 518 0.5 9

114
  
  

 

COCLED
SECTION

TERMINALS

@

 
 

PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
ORNL-LR-DWG 12582

   

RESERVOIR

 

 

 

 

Fig. 5.1. Forced-Circulation Loop Showing Position of Reservoir Added to Increase Volume of System.

the cold leg, as well as parts of the pump thatwere
in contact with the fluid. A chemical analysis of
the deposit showed predominantly titanium and,
possibly, some zirconium. A careful search was
made of all parts of the loop for materials with
high titanium content, but none were found. At-
tempts to identify the layer as a titanium compound
by x-ray diffraction have been unsuccessful.
Therefore, although the attack in loop 7425-7A
was approximately 3 mils deeper than the attack
in loop 7425-6 and the attack which would have

been normal in loop 7425-1A, the significance of
the increase in depth of attack cannot be judged
until the source of the deposits is found. How-
ever, even if the increase in depth of attack can
be attributed to the increase in volume, it appears
that varying the system volume in therange studied
does not have a significant effect on corrosive

attack over a 1000-hr period.

Two other forced-circulation Inconel loops (7425-
2 and 7425-4A) were examined. These loops,

115
ANP PROJECT PROGRESS REPORT

which had circulated the fluoride fuel mixture NaF-
KF-LiF«(UF, + UF;) (11.2-41-45.3-2.5 mole %),
clearly demonstrated that the presence of U3'
in such a mixture is very effective in reducing
hot-leg attack. However, disproportionation of UF
occurred in both loops, and therefore continuous
layers of metallic uranium were formed along the
loop walls in the cooled sections. Chemical
analyses of samples (Table 5.2) taken before
filling of the loops showed 1 wt % of the fluoride
mixture to be uranium in the trivalent form. Both
loops operated with a hot-leg temperature of
1500°F, a temperature drop of 200°F, and a
Reynolds number of 10,000.

A pump-motor failure during operation resulted in
the unscheduled termination of loop 7425-2 after
550 hr. The maximum attack during this period
was found metallographically to be less than
1.5 mils. Loop 7425-4A, which operated success-
fully for 1000 hr, showed only slightly greoter
attack, which reached to a depth of 2 mils, as
shown in Fig. 5.2. Analyses of samples of the
fuel mixture taken before filling and after draining
of the loop, as reported in Table 5.2, showed little
change in the chromium content, and thus they
substantiate the low attack. The low attack in
these loops is in contrast with the attack to a
depth of 35 mils reported previously? for loop
46953, which was operated under similar condi-
tions but circuloted an alkali-metal fluoride mix-
ture with the uranium present only as UF .

The continuous metal layers formed on the cold
legs of loops 7425-2 and -4A were well bonded and
almost invisible, When etched, however, the
layers became blackened ond were heavily at-

 

2Gr. M. Adamson and R. S. Crouse, ANP Quar. Prog.
Rep. June 10, 1955, ORNL-1896, p 85.

tacked, and thus the presence of uranium metal
formed from the reaction

4UF ;== 3UF, + U°

was indicated.

Metal crystals were found in the cold legs, in
addition to the continuous deposits. The crystals
were composed predominantly of chromium and can
therefore be attributed to mass transfer, Thus,
even though the presence of trivalent uranium
substantially reduces the attack, the mass transfer
of chromium is apparently not entirely eliminated,
and the disproportionation of UF, to form metallic
uranium is serious.

 

 

Maximum Attack in Inconel Forced-

Fig. 5.2.
Circulation Loop 7425-4A After Circulating NaF-
KF-LiF-(UF, + UF.) (11.2-41-45,3-2.5 mole %) for
1000 hr at a Maximum Fluid Temperature of 1500°F
and a Reynolds Number of 10,000. 250X. Reduced
31%. (owewpt with caption)

TABLE 5.2. URANIUM AND IMPURITY ANALYSES OF SAMPLES OF Nc::F-KF-LiF-(UF4 + UF3)
TAKEN BEFORE FILLING AND AFTER DRAINING OF LOOPS

 

Uranium Content (wt %)

Impurities Found {(ppm)

 

 

 

Loop No. Sample Taken 3+
: Total U U Ni Cr Fe
74252 Before filling 1.9 1.07 135 35 105
After draining 10.7 0 275 25 150
-4 A Before filling 10.8 0.81 115 40 165
After draining 10.8 0.70 390 155 120

 

116
Liquid Metals in Inconel and Stainless Steel

The first Inconel forced-circulation loop (7439-1)
in which NaK was circulated completed 1000 hr of
operation with a temperature gradient of 300°F and
a maximum fluid temperature of 1500°F. The {oop
included a bypass cold trap for removing oxides,
As in the case of Inconel-sodium loops,® metal
deposits were found in the economizer and, to a
lesser extent, in the cooled sections of the loop,
as shown in Fig. 5.3. The dendritic crystals com-
prising the deposits were somewhat finer than the
crystals observed in sodium loops, and they pro-
jected from a thin metallic layer that was integrally
bonded to the base metal, as shown in Fig. 5.4,
Analyses showed the deposits to be 90 to 92%
nickel and 8 to 10% chromium and thus to be
identical to the deposits found in Inconel-sodium
loops.

A second Inconel loop (7439-2) was operated
with NaK for 1000 hr under conditions simjlar to
those for operation of loop 7439-1, but there was
no cold trap in the system, Visual and metallo-

 

31bid., p 86.

UNCL ASSIFIED
T-8758

7

 

Fig. 5.3. Three Sections from the Economizer
and One Section from the Cold Leg of Inconel
F orced-Circulation Loop 7439-1 Which Circulated
NaK for 1000 hr at a Maximum Fluid Temperature of
1500°F and a Temperature Gradient of 300°F.
This loop included a bypass cold trap.

PERIOD ENDING MARCH 10, 1956

e

B8
£.902.)
.603
0
m‘

L.

x
<9
=

T
1

 

 

 

Fig. 5.4.
Inconel Forced-Circulation Loop Which Circulated

NaK for 1000 hr. 250X. Reduced 32%.

Layer Deposited in Economizer of

graphic examinations did not reveal a significant
difference in the amount or in the appearance of
the deposits formed in this loop in comparison
with the deposits in the first loop., Thus the
presence of a bypass cold trap in the first loop
was apparently of little benefit in limiting mass
transfer, as has also been true of Inconel-sodium
loops.4

A comparison of the amount of mass-transferred
material in these NaK (56% Na-44% K) systems
relative to comparable sodium systems has not
been possible because of difficulty in removing
the deposits from the NaK loops for weight de-
terminations, The maximum deposit thicknesses
were approximately the same in both the NaK
systems and the sodium systems; however, such
thickness measurements are not necessarily effec-
tive measures of mass transfer, because the crys-
tals are formed in discontinuous patches. [t ap-
pears that, in general, the restriction to flow in
both systems as a result of mass transfer would be
similar.

A type 316 stainless steel loop (7426-5) also
completed 1000 hr of operation with sodium as the
circulated fluid. The maximum fluid temperature
was 1500°F, and there was a temperature gradient
of 300°F. A bypass cold trap was included in the
system. As in a type 316 stainless steel loop
operated previously? for 476 hr, only a very slight

 

4G. M. Adamson and A. Taboada, ANP Quar. Prog.
Rep. Seps. 10, 1955, ORNL-1947, p 99.

117
ANP PROJECT PROGRESS REPORT

deposit was found in the economizer and in the
cold leg (Fig. 5.5). The amount of deposit which
could be scraped out was sufficient only for
specirographic examination, which indicated it to
be predominantly nickel and chromium,

UNCLASSIFIED
T.9429

 

9
n
Fig. 5.5. Three Sections from the Economizer

and One Section from the Cold Leg of Type 316
Stainless Steel Loop 7426-5 Which Circulated
Sodium for 1000 hr,

THERMAL-CONVECTION STUDIES
J. H. DeVan

E. A. Kovacevich R. S. Crouse
Metallurgy Division

Alkali-Metal Fluoride Fuel Mixtures in Hastelloy B

Three thermal-convection loops {176, 814, and
816) constructed of }’2-in. sched-40 Hastelloy B
pipe (28% Mo-67% Ni-5% Fe) were operated at
1500°F  with NaF-KF-LiF-UF, (11.2-41.45.3-2.5
mole %). Loops 176 and 814, which ran for 500 hr,
showed a maximum hot-leg attack of 2 mils. The
attack appeared as heavy surface pitting, as shown
in Fig. 5.6. Loop 816 was operated for 2000 hr to
determine the effect of increasing the operating
period. As may be seen in Fig. 5.7, the depth of

118

UNCL ASSIFIED
T-5894

 

 

 

 

Fig. 5.6. Maximum Hot-Leg Attack Found in
Hastelloy B Thermal-ConvectionL oop 814 Operated
with NaF-KF.LiF-UF, (11.2-41-45.3-2.5 mole %)
as the Circulated Fluid. 250X. Reduced 32%.
(@wop® \ith caption)

UNCLASSIFIED
T-9217

EEE]

  

Y
!HG;(ES

FEFERREE

e
>

 

 

 

Maximum Hot-Leg Attack Found in

Fig. 5.7.
Hastelloy B Thermal-ConvectionLoop 816 Operated
with NoF-KF.LiF-UF, (11.2.41-45.3-2,5 mole %)

as the Circulated Fluid. 250X. Reduced 32%.

ooy with caption)

attack in loop 816 was about the same as in the
other two loops, but the pits were more concen-
trated. The cold legs of these loops were attacked
to the same degree as the hot legs, and they ap-
peared to be completely free of deposits. Chemical
analyses of the fuel mixtures used in these loops
are presented in Table 5.3.
Effect of Condition of Inner Sutface of Hastelloy B
Tubing on Depth of Attack

Six thermal-convection loops were constructed of
Hastelloy B tubing that had been reamed to ensure
uniformity of the inner surface composition and to
eliminate surface roughness. Three of the loops
were then operated for 500, 1000, and 1500 hr,
respectively, with NaF-ZrF ,-UF , (50-46-4 mole %)
at a maximum temperature of 1500°F, and three
were operated with sodium under the same condi-

PERIOD ENDING MARCH 10, 1956

tions. The test results are presented in Table 5.4.
The loops operated with the fluoride mixture
showed only slight variations in attack as a func-
tion of operating time. In comparison with previ-
ously operated Hastelloy B loops constructed of
as-received or unreamed tubing, the depths of
attack were similar. However, the metal crystals
found in some of the earlier loops were completely
absent in these later tests,

The three loops operated with sodium also
showed no substantial increase in depth of attack

TABLE 5,3, CHEMICAL ANALYSES OF FUEL MIXTURES CIRCULATED IN
HASTELLOY B LOOPS 176, 814, AND 816

 

Total Uranium Content

Impurities Found (ppm)

 

 

Loop No. Sample Taken

(wt %) Ni Cr* Fe Mo

176 Before filling 11.7 70 20 75

After draining 11.5 85 245 360

814 Before filling 11.9 130 80 120
After draining 11.9 20 95 120 75

816 Before filling 12.1 150 95 120
After draining 12.2 145 175 125 140

 

*Chromium is present as an impurity both in the Hastelloy B and in the fuel mixture, which is produced in Inconel

 

 

 

pots.
TABLE 5.4, RESULTS OF TESTS OF THERMAL-CONVECTION LOOPS CONSTRUCTED
OF REAMED HASTELLOY B TUBING
Maximum fluid temperature: 1500°F
ing Ti i Attack Metallurgical Results
Loop Circulated Fluid Operof(ll:lg; 'me Mamnzun.'] ) fac
T mits Hot-l.eg Appearance Cold-l_eg Appearance
766 Sodium 500 0.5 Light surface attack No deposits, some
attack
767 Sodium 1000 0.5 Light surface attack No deposits or attack
768 Sodium 1500 2.5 Intergranular attack with  No deposits or attack
surface pits
769 NoF-Zer-UF4* 500 1.5 Surface pits No deposits or attack
770 NoF-ZrF4-UF4* 1000 3 Intergranular attack with  No deposits or attack
surface pits
771 NuF-ZrF4-UF4* 1500 2 Surface pits Moderate intergranular

attack to o depth of

2 mils

 

*Composition: 50-46-4 mole %.

119
ANP PROJECT PROGRESS REPORT

with increased operating time., There were no
metallic deposits, and there was very little attack
in the cold legs of the loops. However, some
metallic deposits were observed macroscopically
in the traps below the cold legs of all the loops.

Screening Tests of Special Fuel Mixtures

A series of standard Inconel thermal-convection
loops were operated for 500 hr at a maximum fluid
temperature of 1500°F to obtain preliminary data
for evaluating several special fuel mixtures. The
preparation of the special fuel mixtures is de-
scribed in Sec. 4, ''Chemistry of Reactor Ma-

TABLE 5.5.

Since the quantities of the fuel mixtures
available for filling the loops were limited, the
usual precleaning with the mixture being investi-
gated was omitted,

Four loops were operated as the first of a series
for establishing the level of ZrF, below which
mixtures of NaF-LiF-ZrF ,.UF ; exhibit corrosion
behavior typical of alkali-metal fluorides rather
than zirconium-base fluorides such as NaF-ZrF -
UF  (50-46-4 mole %). The results of metallo-
graphic examination of the loops and chemical
analyses of the fuel circulated are presented in

Table 5.5. The attack by NaF-LiF-ZrF ,-UF,

terials,"

RESULTS OF INCONEL THERMAL-CONVECTION LOOP TESTS OF SPECIAL FUEL MIXTURES

Operating period: 500 hr

Maximum fluid temperature:

1500°F

 

Fuel-Mixture Maximum Depth

Impurities in Fuel

Fuel Sample Total Uranium

 

Loep No. {ppm)
Composition (meole %) of Attack (mils) Taken Content (wt %)

Ni Cr Fe
841 Na F-LiF-ZrF4-UF4; 14 Before filling 13.2 100 70 265
20-55-21-4 After draining 13.0 30 220 100
842 Na F-LiF-Zer-UF“; 12.5 Before filling 12.8 405 195 170
20-55-21-4 After draining 13.0 30 370 90
843 NoF-LiF-ZrF4-UF4; 10 Before filling 16.2 335 105 130
53-35-8-4 After draining 16.5 80 500 95
844 NaF-LiF-ZrF4-UF4; 15 Before filling 16.5 300 105 115
53-35-8-4 After draining 16.5 85 300 90
839 RbF-ZrFA-UFA; 9 Before filling 6.90 380 240 235
50-46-4 After draining 6.10 9 280 70
840 RbF-ZrF ,-UF ,; 9 Before filling 6.87 370 235 275
50-46-4 After draining 6.40 30 145 70
845 KF-ZrF4-UF4,' 8 Before filling 8.21 110 110 315
50-46-4 After draining 8.36 20 270 30
846 KF-ZrF4-UF4; 8 Before filling 8.18 120 115 100
5Q0-46-4 After draining 8.06 10 315 40
847 LiF-Z¢F 4-U F4i 17.5 Before filling 9.39 290 155 260
50-46-4 After draining 9.69 10 1750 85
848 LiF-ZrF4-UF4; 19 Before filling 9.21 860 275 825
50-46-4 After draining 9.46 25 1850 80
860 NaF-ZrF ,-UF ; 10.5 Before filling 8.00 <2 80 95
{std) 50-46-4 After draining 8.50 15 1060 80

 

120
(53-35-8-4 mole %) in loops 843 and 844 resulted in
heavy intergranular subsurface void formation, as
shown in Fig. 5.8, The cold legs of loops 843 and
844 revealed light surface roughening and the for-
mation of metallic crystals, Loops 841 and 842,
which operated under similar conditions with NaF-
LiF-ZrF ,-UF , (20-55-21-4 mole %), had moderate-
to-heavy intergranular void formation in the hot leg,
as shown in Fig. 5.9. Although some metal crys-
tals were found in the cold legs of these loops,
the quantities were smaller than those observed in
loops 843 and 844, as would be expected because
of the difference in ZrF , content.

Six other standard
loops were operated as the first of a series of
loops to study the effect on corrosion of various
alkali-metal fluorides as components of the MF-
ZrF 4-UF4 (50-46+4 mole %) fuel mixture, where M
stands for potassium, rubidium, or lithium. The
resuits of metallographic examinations of these
loops and chemical analyses of the fuel circulated
are presented in Table 5.5. The maximum attack,
which was to a depth of 19 mils, occurred in a
loop (848) which circulated LiF-Z¢F ,-UF , (Fig.
5.10), and, as may be noted in Table 5.5, the
chromium contents of the lithium-containing fuels
circulated in both loops 847 and 848 were quite
high. Loops 839 and 840, which circulated RbF-
ZiF -UF ,, under similar conditions, showed hot-

Inconel thermal-convection

EEE]

T
lNC?ES

T
1

EEERESR

2
=

 

 

 

Maximum Hot-Leg Attack Found in

Fig. 5.8.
Inconel Thermal-Convection Loop 843 Which Circu-
{ated NaF-LiF-ZrF UF, (53.35-8-4 mole %) for
500 hr ot a Hot-Leg Temperature of 1500°F. 250X,
Reduced 32%, (Ssmwmet with caption)

PERIOD ENDING MARCH 10, 1956

leg attack to a depth of 9 mils (Fig, 5.11). The
systems in which KF-ZrF ,-UF, was circulated,
floops 845 and 846, showed hot-leg attack to a
depth of 8 mils (Fig. 5.12). The cold legs of all
these loops showed slight surface roughening and
very thin deposits.

E __f]

1

w5
w
T
(&)
z

 

 

 

 

Fig. 5.9. Maximum Hot-Leg Attack Found in
Inconel Thermal-Convection Loop 841 Which Circu-
lated NaF-LiF.ZrF .UF, (20-55-21-4 mole %) for
500 hr at a Hot-Leg Temperature of 1500°F. 250X,
Reduced 32%. (dmwww®t with caption)

 

Maximum Hot-Leg Attack Found in
Inconel Thermal-Convection Loop 848 Which Circu-
lated LiF-ZrF ,<UF , (50-46-4 mole %) for 500 hr at
a Hot-Leg Temperature of 1500°F. 250X. Reduced
32%. (st with caption)

Fig. 5.10.

121
ANP PROJECT PROGRESS REPORT

 

 

 

 

Fig. 5.11. Maximum Hot-Leg Attack Found in
Inconel Thermal-Convection Loop 840 Which Circu-
lated RbF-ZrF .UF, (50-46-4 mole %) for 500 hr at
a Hot-Leg Temperature of 1500°F, 250X. Reduced
32%. (Swewmt with caption)

S

Fig. 5.12. Maximum Hot-Leg Attack Found in
Inconel Thermal-Convection Loop 846 Which Circu-
fated KF-ZrF ,.UF , (50-46-4 mole %) for 500 hr at a
Hot-Leg Temperature of 1500°F, 250X. Reduced
32%. (gt with caption)

Results of examination of a standard Inconel
thermal-convection loop, 860, which operated with
NaF-ZrF ,-UF , (50-46-4 mole %), are included in
Table 5.5 for comparison with the results for the
special fuels, The hot-leg attack in loop 860
appeared as intergranular attack to a depth of
10.5 mils. e -

122

 

GENERAL CORROSION STUDIES
E. E. Hoffman

W. H. Cook D. H. Jansen
Metallurgy Division

R. Carlander
Pratt & Whitney Aircraft

Low-Cross-Section Erazing Alloys in Liquid Metals
and in Fluoride Fuel Mixtures

D. H. Jansen

Inconel tube-to-header joints brazed with nickel-
chromium-germanium-silicon low-cross-sectionbraz-
ing alloys have been corrosion tested in sodium,
NaK, and NaF-ZrF ,-UF , in seesaw apparatus. A
cross section of the type of joint tested is shown
in Fig. 5.13. The advantage of this type of joint
in comparison with the T-joints used in previous
tests is thot both the high- and the low-melting
constituents of the brazing alloy can be exposed to
the test medium at the same time, The specimens
tested to date were given a fast braze, that is,
5 min to reach brazing temperature and 10 min at
brazing temperature. The results of the seesaw
tests of the joints are presented in Table 5.6.

As may be seen, these preliminary data indicate
that the brazing alloy with the highest nickel con-
tent is the most corrosion resistant to NaK. Tube-
to-header joints brazed with the same alloys but
brought to brazing temperature over a 4-hr period
will be tested in the same medium. The 70%
Ni—11% Cr-13% Ge—6% Si alloy after the tests in
NaK and in the fluoride fuel mixture is shown in
Figs. 5.14 and 5.15.

Two tube-to-header joints brazed with Coast
Metals alloy No. 52 (89% Ni-5% Si—4% B-2% Fe)
were also tested in NaK and in NaoF-ZrF -UF
(53.5.40-6.5 mole %) for 100 hr in seesaw apparatus
with a hot-zone temperature of 1500°F. One joint
was brazed rapidly (10 min) and the other was
brazed slowly (4 hr) in order to study the effect of
brazing time on the corrosion rate. Neither joint
was attacked by NaK, and both showed nenuniform
attack to a depth of 0.5 mil in the fluoride fuel
mixture. The brazing time did not appear to affect
the corrosion resistance,

Cathalloy A-31 in Sodium

R. Carlander

A 6-in, capsule of Cathalloy A-31 (4% W-96%
Ni) was half filled with scdium under a helium
PERIOD ENDING MARCH 10, 1956

 

Fig. 5.13. Cross Section of an As-Brazed Inconel Tube.to-Header Joint Showing Expanded Tube at the
Bottom of the Header Plate; Brazed with Coast Metals Alloy No. 52 (89% Ni-5% Si-4% B~2% Fe). As
polished. 10X,

TABLE 5.6, RESULTS OF SEESAW TESTS OF BRAZING ALLOYS ON INCONEL
TUBE-TO-HEADER JOINTS
Test period:” 100 hr
Hotezone temperature: 1500°F
Cold-zone temperature: 1100°F

 

Attack (mils)

 

Brazing~Alloy Composition

 

(wt %) " Sodiom In NaK In NaF.ZcF UF,
(56-44 wt %) (53.5-40-6.5 mole %)

70 Ni=11 Cr=13 Ge-6 Si 5 3 3.5

65 Ni=16 Cr=13 Ge=-6 Si 6.5 7.5 3

62 Ni—19 Cr—-13 Ge—6 Si 7.5 6.5 3

59 Ni=19 Cr~16 Ge=6 Si 8 5 2.5

 

123
ANP PROJECT PROGRESS REPORT

 

 

 

Fig. 5.14. A 70% Ni=13% Ge~11% Cr—6% Si Brazing Alloy After Exposure to NaK (56% Na-44% K) for
100 hr in Seesaw Apparatus at a Hot-Zone Temperature of 1500°F., Note nonuniform attack to a depth of
3 mils. As polished, 100X,

atmosphere and tested statically at 1500°F for
100 hr., No attack occurred in either the vapor or
the bath zones of the capsule.

Rare-Earth Oxides in Sodium
W. H. Cook
A specimen of Sm,0, (5.88 g/cm3, 25.4% ap-

parent porosity) was tested in an Inconel container
in static sodium at 1500°F for 1000 hr, and two
pieces from a body (6.58 g/cm?, apparent porosity
not determined) fabricated from a commercial mix-
ture of rare-earth oxides, 63.8% 5m,0,-26.3%
Gd,0;-bolance primarily other rare-earth oxides, >
were similarly tested, one for 500 hr and the other
for 1000 hr. The purpose of these three tests was
to evaluate the corrosion resistance of the test
specimens in sodium and the effects, if any, on
the Inconel containers.

The weight and dimensional changes of these

124

three specimens were positive, as was to be ex-
pected with porous bodies, and they were less than
0.5%. The only macroscopic change was that the
bufflike color of the specimens was altered to a
gray-black by the tests.

The untested and tested specimens are shown in
Figs. 5.16 and 5.17. Powder x-ray diffraction
comparisons of the untested and tested specimens
did not reveal reaction products. The walls of the
Inconel test containers were tinted slightly yellow
in the liquid-sodium regions.

Metallographic examinations of the specimens
and the Inconel containers, chemical analyses of
the sodium test baths, and microspark spectro-
graphic anolyses of the inner surfaces of the
Inconel containers will complete the examination
of these tests,

 

5C. E. Curtis et al., ANP Quar, Prog. Rep. Sept. 10,
1955, ORNL.-1947, p 140.
 

PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
Y-1731

INCHES

0.02

 

- {0.038

 

 

 

Fig. 5.15. A 70% Ni-13% Ge-11% Cr—6% Si Brazing Alloy After Exposure to NaF.ZrF ;.UF, (53,5-40-
6.5 mole %) for 100 hr ot o Hot<Zone Temperature of 1500°F. Note attack to a depth of 3.5 mils. As

polished. 100X, (W™t with caption)

Sodium in Type 316 Stainless Steel
Thermal-Convection Loops

E. E. Hoffman

Two type 316 stainless steel thermal-convection
loops filled with sodium were operated for 1000 hr
with hot-leg temperatures of approximately 1630°F.
The minimum cold-leg temperature was 950°F for
one loop and 1050°F for the other. The loops were
loaded with sodium from the same fill pot, and the
test conditions were similar, except that a dif-
fusion type of cold trap was placed at the bottom
of the cold leg of one loop. Blasts of air were
impinged on one section of the cold leg of each
loop to induce a very sharp temperature gradient,.
The purpose of the tests was to determine the
effect of a diffusion cold trap on the amount of
mass transfer observed in a stainless steel-
sodium system. Chemical analyses showed that
the sodium used for filling the loops contained

UNCL ASSIFIED UNCLASSIFIED
Y-16544 Y. 17363

 

(o)

ONE INCH

Fig. 5.16. Samarium Oxide (a) Before and (h)
After Being Exposed to Static Sodium in an Inconel
Container for 1000 hr at 1500°F.

20 ppm oxygen. After the tests the sodium from
loop 26, which did not have a cold trap, contained
25 ppm oxygen, while a sample from loop 27, which
had a diffusion cold trap, contained 35 ppm oxygen.

125
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED UNCLASSIFIED

Y-16546 Y-17362

 

ONE INCH

Fig. 5.17. Specimens from a Commercial Mixture
of Rare-Earth Oxides (63.8 wt % Sm203-26.3 wt %
Gdz(la-bolance primarily other rare-earth oxides)
Before (at left) and After Exposure to Static
Sodium in Inconel Capsules at 1500°F. (a) Speci-
mens used in 500-hr test., (b) Specimens used in

1000-hr test,

The difference is attributed to contamination dur-
ing the sampling operation; however, by present
standards, the oxygen content of the sodium from
both loops was quite low. Small amounts of
metallic crystals were found; by visual inspection,
in the coldest sections removed from each loop.
Neither loop yielded sufficient metallic crystals
for chemical analysis, but there was definitely a
greater quantity of crystals present on the surface
of the cold leg of loop 26, which did not have a
cold trap, than on the cold leg of loop 27.

The results of metallographic examinations of
loops 26 and 27 are presented in Table 5.7. The
attack found in the hottest section of loop 26 was
deeper than that detected in loop 27, as shown in
Figs. 5.18 and 5.19. Nickel was preferentially
leached from the hot-zone walls by the sodium.
Deep intergranular cracking was found in the
coldest section of the loop with no cold trap, but
no cracks were observed in the cold leg of the
loop which had a diffusion cold trap. In a similar
test of an Inconel thermal-convection loop filled
with sodium, in which the hot leg reached a maxi-

TABLE 5.7. RESULTS OF METALLOGRAPHIC EXAMINATION OF SECTIONS FROM
TYPE 316 STAINLESS STEEL THERMAL-CONVECTION LOOPS WHICH
CIRCULATED SODIUM FOR 1000 hr

 

Metallographic Notes = -

 

Section Examined

Loop 26

Loop 27

 

Top of hot leg; operating tem-
perature, 1600°F

Middle of hot leg; operating
temperature, 1630°F

Middle of cold leg; operating
temperature, 1330°F

Bottom of cold leg; operating
temperature, 1050°F in loop
26 and 950°F in loop 27; air-

blast cooled

Surface irregular; attack 0.5 to

0.75 mil deep

Surface irregular; attack 3 to
3.5 mils deep; zone 0.25 10
0.5 mil deep transformed from
austenite to ferrite by prefer-

ential leaching of nickel

Attack less than 0.5 mil deep

Heavy precipitation of an un-
identified phase in grains and
in grain boundaries; many
intergranular cracks originat-
ing at exposed surface and
extending to 10 to 20 mils
deep

Surface irreguiar; attack 0.5 to

0.75 mil deep

Attack less than 1 mil deep;
austeniterto-ferrite transfor-

mation as in loop 26

Intergranular attack to a depth
of 0.5 to 1 mil

Heavy precipitation in grain
boundaries and in grains along
surface exposed to sodium; no

intergranular cracks found

 
 

 

 

 

PERIOD ENDING MARCH 10, 1956

 

 

 

 

 

Fig. 5.18. Sections from Type 316 Stainless Steel Thermal-Convection Loop 27 Which Contained o
Diffusion Cold Trap and Circulated Sodium at a Maximum Hot-Leg Temperature of 1630°F and a Minimum
Cold-Leg Temperature of 590°F. (a) Hot leg. (b) Cold leg. Note austenite-to-ferrite transformation on -
surface of hot leg as a result of preferential leaching of nickel., Etchant: aqua regia, 500X. Reduced 8%.

mum of 1500°F, that is, 130°F less than the
maximum temperature of these stainless steel
loops, there was approximately ten times the
amount of mass transfer found in either of these
tests,®

Boiling Sodium in Type 348 Stainless Steel Loops
E. E. Hoffman

Three type 348 stainless steel loops have been
operated with boiling sodium for various lengths
of time. Type 348 stainless steel is similar in
composition to type 347 stainless steel, except
that the carbon content of type 348 is higher. Each
test was terminated before the end of the scheduled
1000-hr period because of leaks that developed at
various locations in the systems. The periods of
operation increased as fabrication techniques were
improved; the first loop was operated for 108 hr,
the second for 316 hr, and the third for 740 hr.

8E, E. Hoffman, W, H. Cook, and C. F. Leitten, Jr.,
ANP Quar. Prog. Rep. June 10, 1955, ORNL-1896,
p 101, Fig. 5.12.

 

These loops were operated to study mass transfer
by sodium in a stainless steel system in which the
oxygen content of the sodium was very low and to
obtain data for comparison with results obtained
from operation of a similar Inconel system.”

The original design of the boiling-sodium loop,
which was shown in a previous report,” has been
modified, as shown in Fig. 5.20. An overflow
reservoir has been incorporated into the receiver,
and a sampling port has been provided so that
samples of the freshly condensed sodium may be
taken to check the oxygen concentration periodi-
cally during a test run. To obtain a sample, the
nickel sample bucket is dropped down through the
open valve, and o sample of sodium is taken from
the reservoir. The sample bucket is then with-
drawn through the sampling port, under a purified
helium atmosphere, and transferred to a pyrex tube,
where it is sealed off under vacuum. The oxygen

 

7E. E. Hoffman, ANP Quar. Prog. Rep. Seps. 10, 1955,
ORNL-1947, p 109, and ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL.2012, p 122.

127
ANP PROJECT PROGRESS REPORT

 

 

 

 

Fig. 5.19. Sections from Type 316 Stainless Steel Thermal-Convection Loop 26 Which Circulated
Sodium at o Maximum Hot.-Leg Temperature of 1630°F and a Minimum Cold-Leg Temperature of 1050°F.

(a) Hot leg. (b) Cold leg.
leg. Etchant: aqua regia. 500X. Reduced 9%.

concentration of the limited number of samples
taken thus far has been found to be approximately
20 ppm; a portion of this oxygen may be attributed
to contamination during sampling.

The condenser sections of the three stainless
steel loops operated in this series of tests were
examined and found to be free of mass-transferred
crystals. This result was expected, since iron-
base alloys have been found to be less susceptible
to mass transfer by sodium than nickel-base alloys
in both thermal-convection and forced-circulation
loop systems.

Compatibility of Hastelloy B and
Berylliuvm in Sodium

E. E. Hoffman

Tests have been performed to study the compati-
bility of Hastelloy B and Leryllium in sodium as a
function of their distance of separation. The
spacing distances employed were 0, 5, 20, 50, and
100 mils, and the specimens were exposed in

static sodium for 1000 hr at 1200°F. The extent

128

Note irregularity of hot-leg surface and deep intergranular cracking in cold

of alloying was quite similar to that experienced
with Inconel, as might be expected, since Hastelloy
B and Inconel are both nickel-base alloys. Drill-
ings were taken from the surfaces of the various
Hastelloy B specimens. The beryllium concentra-
tions found on the surfaces were:

Berytlium Concentration

Spacer Distance {mils) (10~6 g/cmz)

0 8800
3 6.3
20 1.06
50 0.63
100 0.74

The interaction between the beryllium and the
Hastelloy B specimens that were in contact during
testing is shown in Fig. 5.21. As may be seen,
direct contact resulted in the formation of various
extremely hard and brittle phases on the surface
of the Hastelloy B specimen. These phases have
not all been identified, but several of them are
/ SODIUM VAPQR BYPASS \
/

 
 

CONDENSER LEG

COOLING COILS

HOLD-UP TRAPS

 
    
 

PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
ORNL-LR-DWG 11660

GLASS SAMPLING
APPARATUS —-

 

 

4

i

GROUND METAL JOINT —_

B s
[y

e

 

 

   

YACUUM-TIGHT
AR COOLED VALVE ~—__

   

SAMPLING PORT

NICKEL
SAMPLE BUCKET

 
 

RETURN LEG

DIFFUSION COLD TRAP

— 85 °C
1 o1 2 3 4 5
oo
INCH

Fig. 5.20. Type 348 Stainless Steel-Boiling Sodium Loop.

thought to be nickel-beryllium intermetallic com-
pounds similar to those found in tests of Inconel in
contact with beryllium.2 The extent of the transfer
of beryllium to the surface of the Hastelloy B
specimen when the two materials were separated
by a 20-mil sodium gap is shown in Fig. 5.22. The
very fine precipitate along the surface is thought
to be a nickel-beryllium compound. It appears from
these tests that beryllium and Hastelloy B sur-
faces must be separated by more than 20 mils if
extensive alloying is to be avoided during long
periods of exposure to sodium at 1200°F,

 

8E. E. Hoffman, W. H. Cook, and C. F. Leitten, Jr.,
ANP Quar. Prog. Rep. March 10, 1955, ORNL-1864,
p 85, Fig. 6.20.

Solid-Phase Bonding of Cermets
W. H. Cook

L ong-term solid-phase bonding tests of the most
promising cermets are under way. In the first
of these tests, cermets K152B (64% TiC-6%
NbTaTiC,-30% Ni) ond K151A (70% TiC-10%
NbTaTiC,—-20% Ni) were tested at a calculated
contact pressure of 20,000 psi in NaF-ZrF -UF ,
(53.5-40-6.5 mole %) at 1500°F for 1000 hr. These
cermets were in the form of actual valve parts,
with K152B as a disk and K151A as a seat. The
seating surfaces hod 45-deg bevels, 0.0100 +
0.0002-in.-wide contact regions, and surface rough-
nesses of 5 to 10 pin. Solid-phase bonding oc-
curred around approximately five-sixths of the

129
REPORT

ANP PROJECT PROGRESS

 
     

 

 

" UNCLASSIFIED
Y.17710 Y-17708 @
u
l- 1
. o U
. 3 =z
¥ : o i600 ?»w :
: E Li-,“ :
'5:4 w. M
M? ¥ %mffi '
@, 1600 .-c %’sfi AN ;
@ 1095 A u
. i l"v
© 1540
. . . | 004)
° (105 : ) .
. .
n . :
. « - L et s T, in
. e ) ; . v o —~ .
e HASTELLOY B * 9 5 b o ' HASTELLOY B ~ # 2 ”‘.’; x
: : o 4 i - . b | / ! 2 o....
i e e I —en ¥
,. o E . _ . ,, o \\qi ) 61‘ i
L . ; R ' o T . ’1!
!‘Q 540 ;. J‘- 4 : * & - 1 008
A . ! i - : 3 * /
(o) - | (5) v W,
| 8 | “ o L

 

 
 

: UNCLASSIFIED

    
   
   

  

 

 

 

 

VlCKERS HARDNESS

NO.

Fig. 5.21.

(50 g)

Hastelloy B Specimen After Being in Contact with Beryllium in Sodium for 1000 hr at

1200°F. (a) Unetched. (b) Cathodic etch, Note the very hard intermetallic compounds which formed on the

surface of the Hastelloy B. 500X. Reduced 19%.

linear seating distance, but inspection of the
area indicated that seating had been along a line
rather than on the 0.0100-in.-wide surface., Micro-
scopic pieces had been-pulled from both the KI51A
and the K152B cermets. In the area in which no
solid-phase bonding occurred, there was apparently
uniform seating on the 0,0100-in.-wide surface.
Because of the nonuniform seating, it cannot be
determined whether the sclid-phase bonding oc-
curred as a result of (1) the high stress along the
line contact, (2) the solid-phase bonding being
time-dependent (all previous tests were of 100-hr
duration), or (3) a combination of these., The
reason for the nonuniform seating will be deter-
mined, and the test will be repeated.

Solubility of Lithium in NaK
R. Carlander

A series of differential thermal-analysis tests
have been performed in order to determine the
solubility of lithium in NaK (56 wt % Na-44 wt

% K) as a function of temperature. Two capsules,

130

one filled with NaK and the other with NaK plus
lithium, were used for each test. The two capsules
were heated to 200°C and allowed to cool. The
temperature of the NaK-plus-lithium bath was
plotted on a time-temperature recorder, and the
temperature difference between the two bodies was
plotted a temperature-differential recorder.
When solidification begins, the amount of heat that
is evolved is indicated by a break in the cooling
This break was noted on the

on

curve of the alloy.
temperature-differential recorder
with the temperature on the time-temperature re-
corder to obtain the solubilities. The solubilities
determined were:

and correlated

Temperature (°C) Solubility (wt % Li)

75 0.25
104 1
125
164
166
170
175

onoh AN

&
 

PERIOD ENDING MARCH 10, 1956

INCHES

 

I500X

 

 

003

Fig. 5.22. Hastelloy B Specimen That Was Separated from o Beryllium Specimen by a 20-mil Gap
During Exposure to Static Sodium at 1200°F for 1000 hr. The very fine precipitate along the surface is
believed to be a nickel-beryllium intermetallic compound. Unetched. 1500X.

MASS TRANSFER AND CORROSION IN
SODIUM HYDROXIDE

Static Capsule Tests

E. E. Hoffman
Metallurgy Division

Recent work has shown that A-nicke! does not
exhibit mass tronsfer when exposed to sodium
hydroxide in a nonisothermal system if the maxi-
mum temperature in the system does not exceed
approximately 1100°F. Pure nickel, unfortunately,
has relatively poor high-temperature strength, and
therefore it would be desirable to use a nickel-
base alloy if possible as a container material for
high-temperature sodium hydroxide. Five special
nickel-base alloys were tested in static sodium
hydroxide at 1500°F to increase the severity of
the attack and thus make less difficult the selec-
tion of an alloy for future study in a dynamic sys-.
tem at a lower temperature. Four special iron-
base alloys were also tested for comparison with

similar commercial alloys. The compositions of
the special alloys are listed in Table 5.8,

The nine special alloys were cast and then
extruded into l-in.-dia rods. Two capsules, the
closure caps, and the specimens were machined
from the extruded rod. Duplicate tests were run
on all the alloys, except the 60% Ni-20% Mo-20%
Fe alloy. The specimens were weighed before
and after the tests to obtain weight-change data.
The sodium hydroxide used was dehydrated by
heating at 300°C under vacuum (less than 10 p) for
6 hr and then at 360°C for 20 hr, The individual
corrosion test capsules were sealed in protective
capsules and placed in a furnace where they were

held at 1500°F for 100 hr.

The test results are presented in Table 5.9. The
90% Ni-10% Mo alloy, which was the most cor-
rosion resistant of the alloys tested, is shown in
Fig. 5.23. The tests indicated that the iron-base
alloys, in general, have poor resistance to cor-
rosion by sodium hydroxide at 1500°F and that the

131
ANP PROJECT PROGRESS REPORT

TABLE 5.8, COMPOSITION OF SPECIAL ALLOYS TESTED IN SODIUM HYDROXIDE

 

Composition by Chemical Analysis (wt %)

 

 

Nominal |
Composition {wt %) Ni Fe Mo Cr c 8 Mn 5
85 Ni—15 Mo 86,07 15.09 0.021 0.030 0.002 0.0001
90 Ni-10 Mo 89.60 10.23 0.014 0.030 0.002 0.0002
80 Ni-10 Mo—10 Fe 78.48 10.14 9.60 0.010 0.040 0.002 0.0006
60 Ni—20 Mo—20 Fe 59.44 20.38 19.53 0.018 0.040 0,002 0.0003
70 Ni-15 Fe—15 Mo 71.29 13.83 14.31 0.013 0.023 0.017 0.0001
80 Fe—20 Cr 80.91 19.14 0.017 0.060 0.002 0.021
80 Fe~10 Cr—10 Ni 10.40 79.60 9.93 0.018 0.030 0.002 0.012
74 Fe—18 Cr—8 Ni 8.11 73.5 18.70 0.022 0.030 0.002 0.019
60 Fe—20 Cr—20 Ni 19.88 61.43 19.66 0.005 0.040 0,002 0.013

 

UNCLASSIFIED
Y-17473

 

 

 

 

 

 

 

 

 

Fig. 5.23. A 90% Ni-10% Mo Alloy After Exposure to Static Sodium Hydroxide for 100 hr ot 1500°F.,
Etchant: aqua regia. 500X,

132
PERIOD ENDING MARCH 10, 1956

TABLE 5.9. RESULTS OF CORROSION TESTS ON SPECIAL ALLOYS EXPOSED TO STATIC
SODIUM HYDROXIDE FOR 100 hr AT 1500°F

 

Nominal Allcy Composition Weight Change

 

(wt %) (g/cmz) Metallographic Notes
Nickel-Base Alloys
85 Ni-~15 Mo -0.011 Scattered subsurface voids to a depth of 1.5 to 2.5 mils
-0.010 Scattered subsurface voids to a depth of 2 to 3 mils
90 Ni=10 Mo -0.005 Attack to a depth of less than 0.5 mil
-0.005 Same as above
80 Ni—=10 Mo--10 Fe ~0.006 Few scattered subsurface voids to a depth of 1 mil
-0.005 Same as above except that one corner of specimen
attacked heavily to a depth of 3 mils
60 Ni—-20 Mo--20 Fe -0.003 Very small, scattered subsurface voids to a depth of
4 to 5 mils
70 Ni=15 Fe~15 Mo -0.021 Small subsurface voids to a depth of 3 to 4 mils
-0.02¢ Same as chove
Iron-Base Alloys
80 Fe-20 Cr 0.796 Specimen very heavily attacked; thick oxidation type of
corrosion product with thin metallic layers at surface;
thickness of unattacked material decreosed from 250
to 216 mils
0.813 Same as above
80 Fe—10 Cr—10 Ni 0.045 Heavy oxidation type of attack to a depth of 5 to 6 mils
0.042 Same as above
74 Fe~18 Cr—8 Ni 0.104 Heavy oxidation type of attack to a depth of 11 mils
0.112 Heavy attack to a depth of 12 to 13 mils
60 Fe—-20 Cr—20 Ni 0.047 Uniform attack to a depth of 5 mils
0.032 Uniform attack to a depth of 5.5 mils

 

resistance decreases with increases in the iron
content. The additions of chromium also tended to
decrease the corrosion resistance of the iron-base
alloys, whereas additions of nickel were beneficial.
The nickel-base alloys have good resistance if
the nickel content is above 80%. For example the
90% Ni-10% Mo alloy was found to be attacked
less than the 85% Ni—15% Mo alloy. Further tests
are to be made on alloys similar to the 90% Ni—10%
Mo alloy.

Thermal.Convection Loop Tests

G. P. Smith M. E. Steidlitz
Metallurgy Division

Thermal-convection loop tests of fused sodium
hydroxide in promising container metals are under

way. The preliminary work is being done with
Inconel loops of the type used in the fuel testing
program’ modified to permit the bubbling of hy-
drogen through a small portion of the hydroxide
without entrainment of gas in the circulating fluid,
The hydroxide used is the same reagent-grade
material as that used in the cold-finger tests,'?
Dehydration is accomplished by heating under
vacuum at 300°C for 8 hr and then heating at 375°C
in an Inconel pot for 16 hr, The material is trans-
ferred from the pot to the loop by helium pressure,

Four loops have been run and sectioned. Of

 

9G. M. Adamson, ANP Quar. Prog. Rep. March 10,
1954, ORNL-1692, p 67.

IOM. E. Steidlitz, ANP Quar. Prog. Rep. June 10,
1955, ORNL-1896, p 110.

133
ANP PRQJECT PROGRESS REPORT

TABLE 5.10. RESULTS OF TEST OF INCONEL THERMAL-CONVECTION LOOPS OPERATED
WITH SODIUM HYDROXIDE

 

Temperature {°F)

 

 

L Operating

cop

No Time Maximum _ Cold Status Comments

(hl’) WOII Hydroxlde Leg

863 100 1180 1120 1000 Examination completed Spotty corrosion te a depth of
2 mils in hot zone; no mass
transfer

864 500 1180 1120 1000 Examination completed Hot-zone corrosion to a depth
of 4 mils; mass-transfer de-
posit on cold leg up to 1 mil
in depth

869 1000 1165 1120 1000 Sectioned but not examined Appearance similar to that of

metallographically loop 864, with increased

darkening of hot zone and
increased ferromagnetism

871 500 1340 1220 1055 Sectioned but not examined  Same as loop 869, except

metallographically

that ferromagnetism con-

siderably greater

 

these, two have been examined metallographically
and two have had only visual inspection. All the
loops appear to be clean and relatively bright.
There is some discoloration of the hot zones, and
etching is apparent in the cold legs. Thehot zones
show ferromagnetism, which increases with operat-
ing temperature and time and is the normal result
of corrosion of Inconel.!! There are no massive
deposits observable in any of the loops. The
available data are summarized in Table 5,10.

The corrosion observed in loops 863 and 864, as
shown in Figs. 5.24 and 5.25, is similar to that
seen in the cold-finger testing of Inconel.!! The
mass-transferred deposits on the cold leg of loop
864, Fig. 5.26, however, are not similar to the
plating or the dendritic deposits commonly found
on the cold fingers, but rather to an intermediate
type of aggregate. The lack of mass transfer in
loop 863 is shown in Fig. 5.27, a typical cold-leg
specimen, Further testing of Inconel is planned,
and preliminary tests of nickel loops are under
way,

 

”M. E. Steidlitz and C. P. Smith, ANP Quar. Prog.
Rep. Dec. 10, 1955, ORNL-2012, p 133.

1 .
2The DuPont samples were received toc late for
inclusion in the present series of tests.

134

CHEMICAL STUDIES OF CORROSION

F. Kertesz
Materials Chemistry Division

Physical Properties of Elastomers Exposed to
Attack by Liquid Metals

Screening tests of elastomers for possible use
as valve seat materials in NaK circuits were
initiated, and several promising materials were
selected for further testing. A number of samples
of the potentially useful materials were collected
from Wright Air Development Center, Dow Corning
Corp., General Electric Co., and E. |. DuPont de
Nemours & Co.'2 |dentical test strips were cut
and placed under tension and under compression.
The compression force used on the samples was
260 g, which corresponds to a pressure of 4380
g/em?, and the tension force used was 297 g,
which resulted in a tension of 3400 g/cm?,

The testing apparatus with the samples mounted
on it was placed in a vacuum dry box. The dry
box, which was rebuilt from a scrapped box, is
covered by a single l%-in. sheet of Plexiglas and
has an air lock sealed by an O-ring, The glove
ports are closed by 8-in. pipe flanges andNeoprene
gaskets. A separate vacuum line is provided to
PERIOD ENDING MARCH 10, 1956

" UNCLASSIFIED _
v Y-17757  |.007

 

.Q08

 

209

 

010§

 

P

 

 

pig

 

013

 

 

 

 

 

 

 

 

! Fig. 5.24. Hot Zone of Inconel Thermal-Convection Loop 863 Which Circulated Sodium Hydroxide for
100 hr. The wall temperature near this point was 1180°F; the fluid temperature was 1120°F; the cover
i gas was hydrogen. 250X,

 

 

: Fig. 5.25. Hot Zone of Inconel Thermal-Convection Loop 864 Which Circulated Sodium Hydroxide for
' 500 hr. The wall temperature near this point was 1180°F; the fluid temperature was 1120°F; the cover
gas was hydrogen. 250X.

 

135
ANP PROJECT PROGRESS REPORT

 

Fig. 5.26. Cold Zone of Inconel Thermal-Convection Loop 864 Which Circulated Sodium Hydroxide for
500 hr. Wall and fluid temperatures near this point were 1000°F; the cover gas was hydrogen, 250X,

 

" UNCLASSIFIED. [
V773 |

 

 

 

 

Fig. 5.27. Cold Zone of Inconel Thermal-Convection Loop 863 Which Circulated Sodium Hydroxide for
100 hr. Wall and fluid temperatures near this point were 1000°F; the cover gas was hydrogen. Note
slight intergranular attack and lack of mass transfer, 250X,

136
remove the gas behind the gloves at a faster rote
than it is removed from the box to keep the gloves
pulled well toward the flanges and out of the box.
Separate vacuum lines are also provided for the
air lock.

The sample-testing box, made of cold-rolled
steel, has two NaK-filled compartments. In the
larger of these, one end of each test strip was
clamped to the bottom and the other end was at-
tached to a pivoted arm which provided the tension,
All strips were adjusted to the same tension. The
compartment was filled with NaK to about the
center of the test strip.

In the smaller compartment, round disks of the
rubber samples were tested under compression by
forcing small pistons to the center of the speci-
mens under identical loads. This compartment was
filled with pure NaK to a level above the samples,
The whole apparatus was heated by a thermostati-
cally controlied electric heater,

While being heated to 200°C, about one-half the
samples began to react vigorously, with bubbling,
and so much energy was released that the tempera-
ture rose rapidly to about 250°C. By this time a
number of the samples under tension had failed.
Within 30 min the temperature was back to 200°C.
By that time all the specimens, except three
silicone samples, had failed. The three silicone
samples were still intact, however, after three
days of continved exposure at 200°C, The results

PERIOD ENDING MARCH 10, 1956

of the compression tests confirmed these results.
Silicone rubber compound No, 81578 was practi-
cally unaffected; No. 81577 was next in its re-
sistance to attack; and No. 81576 was third but
still fairly satisfactory.'® The heavy surface
attack on some of the samples that failed was
indicative of the probable harmful effects of
oxides, Other samples appeared to be attacked
uniformly, but the attack was slightly greater at
the bottom, where the NaK was hottest. The re-
sults of these experiments, in which the samples
were exposed to NaK at 200°C for three days
under a helium atmosphere, are summarized in

Table 5.11,

Resistance of Possible Moderators to
Fused Fluoride Mixtures

G. F. Schenck
Pratt & Whitney Aircraft

H. J. Buttram
Materials Chemistry Division

Considerable advantage would be realized if a
moderator material could be found which could be
cooled by direct contact with the molten fuel mix-
ture. Although prospects for such a system do not

 

]3These materials are described in General Electric

Company, Silicones Preliminary Product Data, issued
July 1, 1955,

TABLE 5.11. RESISTANCE TO NoK OF ELASTOMERS UNDER TENSION

 

 

Designation

 

Source of Samples Type Results
Wright Air Development Center 453-9A Polyacrylic Failed; showed heat effects even
before addition of NaK
Dow Corning Corp. Silastic 58 Silicone Failed
Sitastic 80 Silicone Failed
Siflastic 152 Silicone Failed
General Electric Co, 81576 Silicone Intact at end of test
81577 Silicone Intact at end of test
81578 Silicone Intact at end of test
SE450 Silicone Failed
SE550 Silicone Failed
SE750A Silicone Failed

 

137
ANP PROJECT PROGRESS REPORT

look hopeful, a simple program of short-term iso-
thermal tests of the relatively few possible modera-
tor materials has been initiated.

Materials such as beryllium metals, ZrH , YH,
and CaH_ are far too strongly reducing to be stable
to the fluoride mixtures and were therefore not
tested, Beryllium carbide, beryllium oxide, and
boron carbide specimens 0.25 x 0.25 x 0.5 in. were
tested in NaF-ZrF -UF, (53.5-40-6.5 mole % in
Inconel capsules at 800°C for 24 hr,

The beryllium carbide specimen appeared to have
reacted completely with the melt, Petrographic
examination of the melt after the test revealed UF 4
and carbon at the surface of the skeleton of the
original specimen. No identifiable bery!lium com-
pound was observed in the original solid; the
solidified melt had been strongly reduced.

Beryllium oxide showed only about 2% change
in the original dimensions and a 6% gain inweight.
The surface of the specimen was covered with a
brown deposit, which was almost certainly uranium
oxide. The reactions

UF4 + 2Be0—> 2BeF, + UQ,
ZrF4 + 2Be0 — ZBeF2 + Zro2

would be expected to take place. It is possible
that BeO, specimens protected by ZrO, might be
stable, but it is not likely.

The B ,C specimen suffered only a slight loss in
weight in this simple test. Exposure of the speci-
men to air after the test resulted in the liberation
of fumes, which may have been BF .. Further tests
of this material seem to be justified.

Ciffusion of Chromium in Alloys

G. F. Schenck
Pratt & Whitney Aircraft

M. G. Leddicotte
Analytical Chemistry Division

Interface reaction of molten fluorides with the
chromium component of Inconel results in the re-
moval of the chromium from the alloy. A chromium
concentration gradient is thus established through-
out the metal wall, which causes migration of the
chromium from the higher concentration area toward
the chromium-depleted surface and results in the
formation of subsurface voids. The thickness of
the layer impoverished in chromium is a function
of the depth of the corrosive attack.

138

Experiments are under way in which activation
anclyses are being used to study the diffusion of
A measure of the diffusion
velocity may be obtained by activation-analysis
measurements over a period of time,

chromium in Inconel.

A clean section of Inconel was activated in the
ORNL Graphite Reactor for use as a standard on
Three hours
after removal of the specimen from the reactor the
radiation rate from a 12-g section was 7 r/hr at
5 in. from the surface. This specimen was allowed
to cool for several days to diminish the activity.
At the end of this time most of the activity was
that of activated chromium, and an autoradiograph
showed the activity to be homogeneously dis-
tributed.

In the current work an effort is being made to
establish whether diffusion studies are sensitive
enough to follow the diffusion of chromium in the
wall of a )-in.-OD tube.
currently being investigated.
graphic plate is exposed to the irradiated sample,
and the optical density of the plate is then de-
termined as a function of distance from the pe-
riphery. It is believed from an evaluation of the
data obtained that the results are inconclusive
because of the limited resolution of the densitom-
eter. In the second approach, the radioactivity in
various portions of each sample is measured. It is
believed that successively milling small amounts
of metal from the inside of the tube will produce a
satisfactory series of samples from increasing
depths, as measured from the inner surface. The
amount of the sample, the depth of each milling,
and a measurement of the amount of Cr3' (27.8-
day) radioactivity in each portion will be used to
determine the degree of chromium diffusion,

Preliminary work has consisted in determining
the radioactivity of the samples used for radio-
avtograph analyses by means of a gamma-ray
spectrometer. The gamma spectral analysis dis-
played a continuous record of the relationship be-
tween the counting rate and the gamma-ray energy,
and it was thus possible to obtain the relative
intensities of all the gamma-emitting radioelements
in the sample. The results indicate that it should
be possible to measure the Cr®! intensity in each
portion milled from the sample., Current work by
the Analytical Chemistry Division includes the
design of suitable equipment for obtaining and
retaining each layer milled from the specimen,

which to base future measurements,

Two approaches are
In one, a photo-
PERIOD ENDING MARCH 10, 1956

6. METALLURGY AND CERAMICS

Metallurgy Division

WELDING AND BRAZING STUDIES

P. Patriarca
A. E. Goldman G. M. Slaughter
Metallurgy Division

Examination of NaK-to-Air Radiators After Service
at High Temperatures

Metallographic investigations are under way on
NaK-to-air radiotors that failed after various
periods of service at high temperatures. A large
number of sections from radiators designated ORNL
Nos. 1 and 3 and York No. 1 have been examined
to determine the degree of adherence of the braze
material at the tube-to-fin joints and the degree of
oxidation of the copper fins. A correlation of these
factors with heat-transfer performance is expected
to give a better understanding of the relative
importance of the fabrication features,

ORNL radiators Nos. 1 and 3 failed, because of
leaks, aofter 608 and 716 hr, respectively, of
primarily isothermal service in the temperature
range 1000 to 1600°F. York radiator No. 1 failed,
as a result of a leak, after it had been in service
in the range 1000 to 1600°F for 152 hr,

Many metallographic specimens were taken from
each radiator, and each specimen contained a
portion of three tubes joined to 15 to 20 fins. The

tubes were cross sectioned to show two opposite
areas of each joint, Counts were made of the
tube-to-fin joints which exhibited 0 to 24, 25 to
49, 50 to 74, and 75 to 100% adherence, and
estimates were made of the degree of oxidation of
the copper fins according to the categories non-
oxidized, slightly oxidized, and heavily oxidized.
The results of these investigations on the three
radiators are presented in Table 6.1,

Also, ORNL radiater No. 3 is being examined
for possible causes of the failure. As may be
seen in Fig. 6.1, the fire that occurred because
of the leak destroyed the area of the leck, and
therefore the examination is being made of the
numbered sections. Similar sections are being
cut from Pratt & Whitney radiator No. 2, which
also failed because of a leak.

Crack Susceptibility of Back-Brazed
Tube-to-Header Joints

One of the problems associated with the fabri-
cation of the ftull-scale ART NaK-to-air radiators,
designated air-cooled radiator 7503, is the brazing
of the core halves and the welding of these into a
single unit.  The specifications suggest the
joining of two core halves into a single unit by
welding after the brazing operation. An expetiment

TABLE 6.1. RESULTS OF EXAMINATIONS OF NaK-TO-AIR RADIATORS FOR
TUBE-TO-FIN ADHERENCE AND FIN OXIDATION

 

Radiator Designation

 

 

ORNL No. 1 ORNL No. 3 York No. 1

Period of operation in the range 1000 to 1600°F, hr 608 76 152

Number of tube-to-fin joints examined 4150 2282 3847

Joints with 75 to 100% adherence, % 91.8 87.7 67.4
Joints with 50 to 74% adherence, % 4.2 3.5 13.0
Joints with 25 to 49% adherence, % 1.4 1.4 7.3
Joints with 0 to 24% adherence, % 2.6 7.4 12.3
Nonoxidized copper fins, % 59.3 12.1 75.4
Slightly oxidized copper fins, % 20.1 2.5 22,0
Heavily oxidized copper fins, % 20.6 85.4 2.6

 

139
ANP PROJECT PROGRESS REFORT

UNCL ASSIFLED
Y17425

UNCL ASSIFIED
Y-17426

 

Fig. 6.1. ORNL Radiator No. 3 Showing Area Destroyed by Fire and Sections Cut for Metallographic

E xomination.

was therefore conducted to determine the influence
of the stresses set up, during welding, on the crack
susceptibility of the back-brazed tube-to-header
joints.

Two core halves of the type shown in Fig. 6.2
were assembled for brazing. Qne core half was
brazed in the conventional upright position, fin
collars down, while the other was brazed in the
horizontal position. Brazing in the horizontal
position eliminates the need for sump plates to
accommodate excess brazing alloy and thus also
eliminates the restraining effects of the sump
plates on the tubes during nonisothermal service.
All welding was done in the down-hand position in
accordance with established welding procedures.
The completed assembly is shown in Fig. 6.3,

Dye-penetrant inspection of the tube-tosheader
back brazes indicated freedom from cracks both
before and after welding, Visual examination of
the tube-to-fin joints indicated that good flow and
fin-coliar protection had been achieved in both the

140

vertically and horizontally brazed core halves.
The feasibility of brazing in the horizontal position
will be investigated further in experiments with
farger fin banks, 23/4 x 8 x 16 in., and with the
brazing alloy preplaced as sintered rings and as
dry powder.

Measurement of Weld Shrinkage

An experimental program was carried out to
obtain information on the transverse weld shrinkage
that will result from the various welding operations
involved in the construction of the ART, The data
will be of value in design and fabrication of the
various components of the reactor,

Butt welds were made by both the Heliarc and
metallic-arc processes in %/16" 3/8-, 12-, and ‘z-in.-
thick Inconel plates with several joint designs.
A  summary of weld shrinkage measured with
micrometers and appropriate dial gages is pre-
sented in Table 6.2, As may be seen from the
data, weld shrinkage increases with plate thick-
PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
Y-17175

 

Fig. 6.2. Component Parts of Crack Susceptibility Test Specimen.

UNCLASSIFIED
Y-17297

 

Fig. 6.3. Completed Crack Susceptibility Test Specimen.

141
ANP PROJECT PROGRESS REPORT

TABLE 6,2. SUMMARY OF SHRINKAGE OF BUTT WELDS ON INCONEL PLATES

Joint Designs

1. J-type bevel with 60-deg included
angle; welded in accordonce with
procedure specification PS-12

2. VY-type bevel with 100-deg included
angle; welded in accordance with
pracedure specification PS-1

3. V-type bevel with 75-deg included
angle; welded in accordance with
operator’s qualification test speci-

fication QST-12

 

 

Plate Weld
Joint Welding Process Thickness Shrinkage
Design (in.) (in.)
Yariables: Joint Design and Plate Thickness
1 Heliarc and metallic arc 3/8 0.091
1/2 0.95
3/4 0.129
2 Heliarc 3/16 0.113
3/8 0.140
1/2 0.192
Variables: Joint Design and Welding Process
1 Heliarc and metallic arc 3/4 0.129
Metallic arc 3/4 0.119
Heliarc 3/4 0.189
3 Metallic arc 1/2 0.090
Heliare 1/2 0.129
Variable: Weld Volume
3 Heliarc 1/2 0.129
2 Heliarc 1/2 0.192

 

ness, and the Heliarc welding process results in
larger shrinkage for a given joint design in a
plate of given thickness than does the metallic-arc
welding process., An analysis of the cross-
sectional areas for the different joint designs in
]/z-in. plate indicates that other parameters, such
as the mean weld width, are also important, Ex-
periments on buttewelded low-carbon-steel test
plates indicate that the shrinkage is slightly
larger than that for Inconel, as would be expected
from a comparison of the coefficients of thermal
expansion: 6.4 x 10~% in./in.:°F for Inconel and
7.3 x 10=6in./in.+°F for low-carbon steel.

Dimensional Control During Fabrication
of Pump Yolutes

One of the problems associated with the fabri-
cation of an Inconel pump of the type designed

142

for pumping NaK in the ART is the maintenance of
the critical spacing between the two volutes which
constitute the pump housing, Two test pieces that
are representative of the volutes were machined
from 2-in. Inconel plate for use in a study of
methods for controlling the spacing during fabri-
cation, Four spacers were also machined from
Inconel plate to act as rigid supports duting the
welding of the wolutes.
of the experimental specimen are shown in Fig.
6.4, The spacers were subjected to an aluminizing
treatment prior to assembly of the components,
to prevent self-welding. The assembly was welded
inaccordance with procedure specifications PS-12.
Micrometer measurements were made at four radial
positions prior to and after each welding operation,

The welded assembly was annealed in a helium
atmosphere at 1850°F for a period of 2 hr, with

These component parts
PERIOD ENDING MARCH 10, 1956

UNCLASSIF{ED
Y.17332

SPACER

 

BOTTOM

TOP

PK-2 WELD TEST

Fig. 6.4. Component Parts for Experimental Fabrication of NaK Pump Casing.

the heating and cooling rates being maintained at
600°F/hr. After the welding and annealing had
been completed, the spacers were removed by
machining. An effort to remove the spacers intact
was abandoned when it became evident that sig-
nificant force would be required and that, as a
result, the volute surfaces might be scored. The
finished pump casing is shown in Fig. 6.5.

Micrometer measurements of the inside surfaces
of the volutes indicated that the maximum dimen-
sional change was 0.015 in. The maximum shrink-
age on the diameter was 0.056 in. It was also
found after annealing and after removal of the
spacers that the dimensions were exactly the
same oas they were after welding. [t is evident
therefore that the annealing cycle completely
removed the residual stresses. A simijlar experi-
ment is now under way to determine the degree of
dimensional control which can be obtained when
a brazing procedure is used in the fabrication of
the pump casing.

Cermet-to-Metal Joints

The usefulness of cermet valve components for
ANP valve applications depends to a large extent
upon the possibility of successfully joining them
to metallic structural materials, such as Inconel,

The fabrication procedure described previously,!
which utilized thin films of Electroless-plated
nickel-phosphorus brazing ailoy on the cermet,
has proved to be too sensitive to the plating
variables, An experimental program is under way
for developing a more reliable and consistent
joining procedure,

Since high-temperature brazing alloys of the
Ni-Si-B type (such as Coast Metals No. 50;
brazing temperature, 1120°C) will bond to cermets
such as K-152B (64 wt % TiC—6 wt % NbTaLiC,~
30 wt % Ni), a series of tests were conducted with
these materials. The cermets are hard to wet, and
it is very difficult to obtain an even layer of
brazing alloy over the entire faying surface,
Therefore a nickel screen was placed in intimate
contact with the cermet to facilitate wetting, The
“‘tinned’’ surface was then ground flat and copper-
brazed to the nickel-transition layer and to the
Inconel, as in the previous procedure.! A brazed
joint that is typical of the joints obtained by this
procedure is shown in Fig. 6.6. A valve disk
fabricated by this new technique was found by
dye-penetrant inspection to be free of flaws.

 

p, Patriarca, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL.-1947, p 131.

143
ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED
5

Y-1743

   

PK-2 WELD TEST

Fig. 6.5.
Casing.

Finished Experimental NaK Pump

 

Fig. 6.6. Cermet-to-Nickel Joint “*Wetted'’ with
Coast Metals No. 50 Alloy and Brazedwith Copper.
100X, Reduced 32%.

Another highly promising procedure now being
investigated eliminates the need for brazing the
cermet to the nickel-transition layer. It has been
found that at a temperature of approximately 1350°C
an interfacial reaction occurs, between the cermet
and the nickel-transition layer, that results in a
metallurgical bond. A joint made by this tech-
nique is shown in Fig. 6.7. At a higher tempera-
ture, approximately 1370°C, the extreme nonuniform

144

 

UNCLASSIFIED *
¥-17611

 

 

Fig. 6.7.
Bonding at Approximately 1350°C.
duced 32%.

Cermet-to-Nickel Joint Formed by
100X, Re-

UNCLASSIFIED
¥1T64S

Fig. 6.8.
treme Nonuniform Interfacial Reaction Resulting
from Heating to Approximately 1370°C.

Cermet-to-Nickel Joint Showing Ex-

interfacial reaction shown in Fig, 6.8 occurred.
Further tests will be made to establish the optimum
conditions.

Brazing Alloy Development

The low-cross-section Ni-Ge-Cr-Si brazing alloy
system has been found to have sufficient corrosion
resistance to fuel mixtures and to sodium to be of
use in circulating-fuel reactor fabrication. There-
fore several sample melts of alloys in the high-
nickel-content range of the system were made for
flow-point determinations. The results of the tests
are presented in Table 6.3, Four of the alloys
with low flow points were submitted for corrosion
testing in sodium, in NeK, and in fused fluoride
salts, and the results are reported in Sec. 5,
**Corrosion Research,’’

TABLE 6.3. RESULTS OF FLOW-POINT
MEASUREMENTS ON THE Ni-Ge-Cr-5i
BRAZING ALLOY SYSTEM

 

 

 

Brazing Alloy Composition (wt %) Flow

Point
Ni Ge Cr Si c)
75 8 ) 6 1120
75 13 6 6 1120
73 13 1 3 1160
70 13 1 6 1100
68 13 13 6 1100
68 10 16 6 1100
67 13 1 9 1120
65 13 " 1 1140
65 13 16 6 1080
62 13 19 6 1080
62 16 16 6 1100
62 13 16 9 1100
59 16 19 6 1080

 

High-temperature oxidation and corrosion tests
on Coast Metals alloy No. 52 (Ni-Si-B) have indi«
cated that removal of a constituent or of con-
stituents occurred during testing. Since this alloy
has been used extensively in the fabrication of
NaK-te-air radiators and fuel-to-NaK heat ex-
changers, a study was made to determine the type
and extent of removal that may be expected to
occur during the intended service time.

Some results of oxidation tests of cast alloy
buttons are available, The structure of the alloy
is shown in Fig. 6.9 as cast and after 100 and
500 hr at 1500°F, It may be seen that the depth
of constituent removal increased with time and
was approximately 0.006 in. after the 500-hr oxi-
dation test. Microspark spectrographic and metal-
lographic examinations of the sample indicated

PERIOD ENDING MARCH 10, 1956

 

UNCL ASSIFIED
CYe174926

 

Fig. 6.9. Results of High-Temperature Oxida-
tion Tests of Cast Buttons of Coast Metals Braz-
ing Alloy No. 52. (a) As cast. (b) Exposed for
100 hr at 1500°F, (c) Exposed for 500 hr at
1500°F. As polished. 100X. Reduced 4%.

145
ANP PROJECT PROGRESS REPORT

that both boron and silicon were removed, Micro-
hardness measurements on the interior of the as-
cast specimen showed a hardness of 700 VHN,
whereas the hardness of the depleted surface was
140 VHN,

Similar tests conducted in a vacuum and in
helium showed no removal of alloy constituents,
Results of tests in NaK and in fused salts will
be reported later,

The brazing of aluminum bronze fins to Inconel
tubes by conventional dry-hydrogen techniques
has been unsatisfactory because of the formation
of a thin film of aluminum oxide on the fin surface,
Thin electroplates (<0.001 in.) of nickel and iron
were inadequate diffusion barriers when the con-
ventional heating time of 4 hr was used; however,
heavier platings (0.002 in.) facilitated wetting.
An Inconel T-joint brazed to a 6% aluminum bronze
fin with Coast Metals alloy No. 52 is shown in
Fig. 6.10; the 0.002-in, electroplate of iron used

UNCLASSIFIED
Y-174%

Fig. 6.10. Inconel T.Joint Brazed to a 6% Alu-
minum Bronze Fin with Coast Metals Brazing
Alloy No. 52 Showing 0.002.in. Electroplate of
Iron on Fin. 100X, Reduced 10%.

146

 

to facilitate wetting may be seen,

The addition of manganese to nickel-base high-
temperature brazing alloys has also been found to
promote wetting. Additions of 30% manganese to
the Coast Metals alloy No. 53 have permitted the
direct wetting of the bronze, while additions of
only 10% promoted flow on material plated with

only 0.005 in. of nickel,

MECHANICAL PROPERTIES OF INCONEL

D. A. Douglas J. R. Weir
Metallurgy Division

C. R. Kennedy
Pratt & Whitney Aircraft

Creep-Rupture Design Data

The creep testing of Inconel in argon and in
NaF-ZrF -UF, (50-46-4 mole %) at 1300, 1500,
and 1650°F is nearly complete, The data obtained
recently have shown that a few minor modifications
of previously published data are necessary, The
data presented here, in Figs. 6,11 through .15,
supersede the data presented in a previous report
in this series.?2 The curves of Figs. 6.11, 6.12,
6.14, and 6.15 include data obtained at higher
and lower stresses than were used previously,
The curves of Fig. 6.13 reflect a better under-
standing of the scatter of test resvits in the 3000-
to 4500-psi stress range. Metallographic examina-
tion of the test specimens showed that strain
aging or strain-induced precipitation had occurred
in some specimens in this stress range and had
caused an increase in rupture time. Since this
phenomenon did not occur in subsequent tests, the
design curve was redrawn to include only the more
reproducible data. Additional design curves for
Inconel tested in argon and in fused salts are
presented in Figs. 6.16 and 6.17,

Low-Stress Creep Data

Information on the creep of Incone! at very low
stresses at 1300°F was obtained because it was
needed in the design of components that must have
dimensional stability at low stresses at the tem-
perature of interest, The test results obtained to
date indicate a total creep strain of 0.03% in 1000
hr for a stress of 1000 psi at 1300°F in air.

 

2R. B. Oliver et al., ANP Quar. Prog. Rep. March 10,
1955, ORNL-1864, Figs. 7.22, 7.24, 7.26, 7.28, and
7.29, p 122-126.
PERIOD ENDING MARCH 10, 1956

ORNL-LR-DWG 12593

 

20,000
1% 5% |10% RUPTURE
10,000
."_;.
2 5000
w
w
e}
or
-
w
2000
1000 R
i 10 100 1000 10,000
TIME {hr}

Fig. 6.11. Stress-Rupture Characteristics of As-Received Inconel Sheet (Heat B) Tested in NaF-ZrF -
UF, (50-46-4 mole %) at 1300°F.

UNCLASSIFIED
ORNL-LR—DWG 12594

20,000

5000

STRESS (psi)

2000

 

1000
1 s 100 1000 10,000

TIME (hr)

Fig. 6.12. Stress-Rupture Characteristics of As-Received Inconel Sheet (Heat B) Tested in Argon at
. 1500°F.

147
\

ANP PROJECT PROGRESS REPORT

20,000

10,000

5000

STRESS (psi )

2000

1000

UNCLASSIFIED

oy L ORNL—LR —DWG 12595

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 5 10 20 50 100 200 500 1000 2000 5000 13,000
TIME (hr)

Fig. 6.13. Stress-Rupture Characteristics of Annealed (2050°F) Inconel Sheet (Heat B) Tested in Argon

ot 1500°F.

20,000

10,000

5000

STRESS (psi)

2000

1000

 

UNCLASSIFIED
ORNL-LR-DWG 12596

 

10% RUPTURE

2 5 10 20 50 100 200 500 1000 2000 5000 10,000
TIME (hr)

Fig. 6.14. Stress-Rupture Characteristics of As-Received Inconel Sheet (Heat B) Tested in Argon

1650°F.

148
PERIOD ENDING MARCH 10, 1956

 

aikinpiini
QRNL-LR~DWG 12597
20,000
10,000
4 5000
w
o
wl
© 1% RUPTURE
w
2000
1000
1 10 100 1000 10,000
TIME (hr)

Fig. 6.15. Stress-Rupture Characteristics of As-Received Inconel Sheet (Heat B) Tested in NaF-ZrF .-

UF , (50-46-4 mole %) at 1650°F.

ORNL-LR-DWG §2598

20,000

RUPTURE

10,000

5000

STRESS (psi)

2000

10,000

 

1000

1000
1 100
TIME [hr)

Fig. 6.16. Stress-Rupture Characteristics of Annealed (2050° F) Inconel Sheet (Heat B) Tested in NaF.

ZrF -UF, (50-46-4 mole %) ot 1300°F.
149
ANP PROJECT PROGRESS REPORT

e Ay g 5
UNCLASSIFIED
ORNL-LR—DWG {2589
20,000

10,000

5C00

2
RUPTURE

STRESS (psi)

2000

 

1000
i 2 5 10 20 20 100 200 500 1000 2000 5000 10,000

TIME (nr)

Fig. 6.17. Stress-Rupture Characterisfics of Annedled (2050°F) Inconel Sheet (Heat B) Tested in Argon
at 1650°F,

B el

TABLE 6.4. COMPARISON OF THE DUCTILITY AT RUPTURE OF 0.060-in.-THICK INCONEL SHEET WITH
THAT OF 0.060-in.-WALL INCONEL TUBING IN ARGON AND IN i~-lc|F-ZrF4-UF4 (50-46-4 mole %)

 

Elongation (%)

 

Temperature Stress

 

 

 

CF) (o) In Argon In NaF-ZrF4-U F4
Tubing Sheet Tubing Sheet
1300 15,000 2.8 50
12,000 8.8 70
10,000 8.8 40
8,000 13.3 25
1500 5,000 9.2 28 10.2 17
4,000 5.2 20 5.5 13
3,000 1.7 12 3.7 8
2,000 1.3 12
1650 3,000 6.7 30 6.8 19
2,500 3.6 30 5.3 16
2,000 3.8 30 3.3 15
1,500 3.1 12 2.8 8

 

150
Effect of Biaxial Stress on Creep Ductility
at High Temperatures

The stress-rupture strength of an internally
pressurized 0.060-in,-wall Inconel tube is very
similar to that of a 0,060-in,-thick Inconel sheet
specimen, but there is a significant difference in
the ductility of the two specimens at rupture, The
relative values for the elongation in the direction
of the maximum stress at several temperatures and
stresses are shown in Table 6.4 for the two types
of specimens tested in argon and in NaF-ZrF ,.UF
(50-46-4 mole %). The decrease in ductility shown
by the tube-burst specimen may be attributed to
the 2-to-1 hoop-to-axial stress ratio set up in a
closed-end pressurized tube. Metals deformed at
room temperature under various stress systems
exhibit increasing or decreasing ductility in com-
parison with their uniaxial tensile ductility,
depending on whether the stress system tends to
increase or decrease the maximum shear stress,3
In the case of a biaxial tensile-stress system,
such as is found in the tube-burst test, the maxi-
mum shear stress is decreased by the action of
the smaller axial stress, and slip is restricted,
This results in lower ductility at rupture, aithough
the time-to-rupture depends entirely on the larger
hoop stress, The data presented here were taken
at elevated temperatures, where the deformation
mechanism is slightly different from that at room
temperature, but the theories applicable to room-
temperature ductility appear to apply.

SPECIAL MATERIALS STUDIES

J. H. Coobs
H. Inouye J. P. Page T. K. Roche
L. M, Doney
J. A, Griffin A. J. Taylor

Metallurgy Division

M. R, D’Amore R. E. McDonald
Pratt & Whitney Aircraft

Y. M, Kolba, Glenn L. Martin Co.

Neutron Shield Material for High- Tempetature Use
H. Inouye

A recent report states that irradiation causes
extensive physical damage in hot-pressed B,C,
Experiments showed that, in a neutron flux of

 

3\, Gensamer, Strength of Metals Under Combined
Stresses, American Society for Metals, 1941,

PERIOD ENDING MARCH 10, 1956

1013 to 104 nv, the B,C bodies began to crack
at about 3% burnup of the B'0 atoms and that
complete granulation occurred at 26% burnup of
the B! atoms., Helium evolution from the irra-
diated bodies was determined at various tempera-
tures up to 1500°F, In general, the quantity of
helium which was released increased with the
burnup and the temperature to which the specimen
was heated, being a maximum of 20% of theoretical
at 36% burnup of the B! gtoms at 1500°F,

The effects of irradiation on molded and sintered
B,C are being determined by the Solid State
Division (see Sec. 8, “Radiation Damage’') and
will be compared with the effects on a hot-pressed
B,C specimen which was irradiated under similar
conditions. lrradiation of the specimens has been
completed, and a cursory examination indicates
that no cracking or crumbling occurred. Since the
hot-pressed body did not show physical damage,
the B'0 burnup is assumed to have been less
than 3%.

It was estimated previously that o B,C layer
with a density of 2.0 g/em® would attain tempera-
tures between 1500 and 2000°F at a flux density
of 1013 nv in the ART. A further evaluation indi-
cates, however, that the maximum temperature of
the B,C layer in the ART during full power opera-
tion will be 1800°F, if the helium gap between
the B4C and the sodium-cooled lnconel shell is
less, tham 0.020 in.

Metal-bonded boron bodies are being studied as
possible alternates for hot-pressed B,C as neutron
shielding material because of the uncertainties
with regard to the effects of irradiation on hot-
pressed B,C. It was thought that, if boron could
be incorporated in a ductile matrix, irradiation of
the bodies probably would not cause fragmentation,
The thermal expansion of a metal-bonded body
would approach that of Inconel, and the thermal
conductivity would be better than thaot of hot-
pressed B,C, Also, the metal-bonded material
would have better resistance to thermal and
mechanical shock and could be fabricated in
shapes which would fit a curved surface with
gaps of less than 0.020 in.

Since a decision as to the method by which a
high-boron-content layer could be incorporated in
an Inconel annulus in the ART was required before
all the fabrication studies could be completed, a

 

4w. p. Valovage, Effect of Irradiation on Hot-Pressed
Boron Carbide, KAPL-1403 {Nov. 15, 1955).

151
ANP PROJECT PROGRESS REPORT

scheme was devised invelving a double layer of
boron to circumvent the unknown effects of irradi-
ation and to relieve the concern over the fit that
could be achieved with ceramic tiles, the tem-
perature and consequently the quantity of helium
released being a function of the helium gap. The
final configuration, which consists of two layers
of boron-containing material, will result in a
boron density of about 1 g/em?, whereas a mini-
mum of 1.2 g/em? is desired. The details of the
annular space containing the boron layer are shown
in Fig. 6,18, The clad Cu-B,C layer will be
nearest to the neutron source.

The sections of the two layers will be staggered
to eliminate neutron windows between adjacent
pieces. Neutron pinholes will exist at line inter-
sections, but they will be insignificant if a fit of
better than 0.020 in. can be attained. In spherical
areas the clad Cu-B,C sheet, about 8 x 8 in, and
0.100-in. thick, will be attached to the Inconel
shell with one centrally located spot weld, Several
canned B, C tiles will be spot-welded to each
Cu-B,C sheet to anchor them in place.

[n the areas that are not spherical the shield will
be built up of layers of CU-B4C. These layers
will not be vented for helium, since their low
boron densities will minimize the amount of
helivm formed.

ORNL—-LR-DWG " 563;\

Q.010~in, STAINLESS STEEL CLADDING
AND COPPER FOIL—=;
\1 -, \\

",

    
 

0.08C-in. COPPER
B,C CERMET——__ |-

———=——=———-0.005~in.
STAINLESS STEEL CLADDING

REFLECTOR REGION \\ HEAT EXCHANGER REGION

~~0.27-in. B4C TILE
AND
COPPER COATING

7\o.osz~;n. INCONEL

0.25-in. INCONEL—"

m\‘\
SPOT WELDS/ b

 

 

 

 

~
- p //
‘Q,;\\: N

 

 

———{0395 in

Fig. 6.18. The Neutron Shield for the ART.

152

Fabrication of Boron-Containing Materials
H. Inouye

Copper-B,C, - Since B,C is chemically inert to
copper, an investigation of the variables associated
with the fabrication of bodies containing from 20
to 74 vol % B,C in copper was initiated. Thus
far almost all the effort has been devoted to the
fabrication of bodies containing 60 and 74 vol %
54C in order to determine the maximum concen-
tration of boron which can be incorporated into a
strong matrix. The theoretical properties of several
Cu-B,C mixtures are listed in Table 6.5,

TABLE 6.5. THEORETICAL PROPERTIES OF
Cu-B4C MIXTURES

 

54C Content

 

 

Density of Boron
Volume Weight Mixture Density*
Per Cent Per Cent (g/cm3) (Q/Cmg)
10 3.03 8.32 0.206
20 6.57 7.67 0.413
30 10.75 7.03 0.620
40 15.80 6.38 0.820
50 22.00 5.74 1.03
60 29.75 5.10 1.24
70 39,60 4.45 1.44
80 52.90 3.81 1.65

 

*Boron in B4C assumed to be 80%.

The 74 vol % B,C-Cu mixture was crumbly and
relatively weak after being sintered at 1650°F and
coined at 33 tsi. A significant increase in both
the strength and the density resulted when the
sintered bodies were coined at 60 tsi. The boron
densities of the best bodies fabricated were
between 1.24 and 1.31 g of boron per cubic centi-
meter of mixture, or between 83 and 88% of theo-
retical. |t was observed that a sintering weight
loss of 1% occurred when the compacted powders
were sintered in hydrogen at 1650°F for 1 hr and
that the bodies expanded, The dimensional in-
crease did not vary significantly with the com-
pacting pressure between 10 and 40 tsi, An addi-
tional weight loss of 0.1 to 0.2% occurred when
the bodies were resintered, at 1800°F, but there
was no improvement in density,

The sintering weight loss could arise from the
volatilization of contaminants of the boron or from
the reduction of copper oxide, A weight loss due
to the volatilization of the boron would be sig-
nificant if the boron content were low or if an
accurate amount of boron were required in the body.
The expansion which occurs does not permit
coining without the removal of material from the
edges, and consequently the thickness of the
body is affected.

The sintering variables were studied in detail
on mixtures of 60 vol % B,C in copper. Sintering
at 1600°F in helium or hydrogen atmospheres
caused linear expansions as great as 3.0%, when
either commercial- or high-grade B,C was used,
The weight losses of bodies sintered in hydrogen
were about 0.75%, and those of bodies sintered in
helium were negligible. Prefiring the B,C powder
resulted in bodies which shrank instead of ex-
panding during sintering, with no weight changes.

The boron densities of the best bodies of the
.60 vol % B,C composition ranged between 0,977
and 0.998 g;cm3, as coined. A :’/lé-in.-thick body
will support itself in a 2-in, span at 1950°F and
will support a T-ib weight at 1650°F, with no
measurable deflection. Further improvements of
the properties can be derived by hot-rolling the
coined body.

Copper-B,C bodies containing 20, 30, and 40
val % B,C have been sintered and then coined to
a 21-in, spherical radius at room temperature, The
40 vol % B,C.Cy composition is strong and shows
some ductility at room temperature. The forming
of this composition is facilitated when the mixture
is hot-roll clad with stainless steel. Strips of the
clad composite can be formed to a 2-in. radius
without cracking, whereas the unclad material
cracks and breaks upon moderate bending.

Copper-AlIB,,. — Bodies containing 70 vol %
AlB,, in copper were made and evaluated, |t was
hoped that sintering this mixture would result in
alloying and also the precipitation of finely
divided boron. Moreover, if alloying occurred,
the resulting composition (90% Cu-10% Al) would
be ductile and oxidation-resistant,

Boron densities of 1.15 g/cm® (78% theoretical)
were attained with this mixture when it was
sintered for 1 hr in hydrogen at 1650°F and
coined at 60 tsi. When the body was resintered at
1850°F, it melted and o bronze was formed, as
was expected. Increasing the AIB,, content to
78 vol % resulted in a body with a boron density
of 1,17 g/em?3,

PERIOD ENDING MARCH 10, 1956

lron-B,C and Molybdenum-B,C. — Cold-pressed
and sintered bodies of Fe-B,C and Mo-B,C mix-
tures were fabricated, Mixtures containing 80 wt %
B,C were cold-pressed at 60 tsi and sintered at
1150, 1700, and 1950°C, In general, the densities
of the bodies and the strengths increased with
the sintering temperature. The maximum boron
densities attained were above 1.50 g/cm®, The
bodies showed tendencies toward lamination and
brittleness.

Ceramic-Boron Bodies., — Several samples of
molded and sintered ceramic-B,C tiles have been
submitted by the Norton Company and the Carbo-
rundum Company for evaluation. Both companies
have made bodies with boron densities of at least
1.2 g/cm® that have rabbeted edges. Thus far
neither company has submitted spherical segments
which could be measured for dimensional tolerances
to obtain an indication of the helium gap that would
exist between them and a curved Inconel surface,

The Carborundum tiles, 3 x 3 x 0.315 in., were
bonded with silicon and had densities of about
60% of theoretical, Radiographic examination of
the tiles showed macroporosity, with voids as
large as ]/8 in. in diameter, Thickness variations
of as much as 0.006 in. were found for both the
rabbeted edges and the total thickness., Numerous
samples coated with ceramic powders and with
flame-sprayed copper and iron have also been
received for examination.

The Norton Company has also submitted tiles of
B,C bonded with carbon, The specimens with
rabbeted edges were machined. Thickness meas-
urements could not be made, because the surfaces
were rough and the material was crumbly, Samples
flame-sprayed with Rokide A (Al,0;) have also
been received for evaluation,

Boron Steels, — Boron steels have been studied
in order to evaluate the feasibility of using them
for a ring to transmit a compression load from the
beryllium reflector to the support strut ring. A
low-boron-content alloy containing a minimum of
0.65% B0 and a maximum of 1,30% would be re-
quired to support a compression load of 500 psi
at 1300°F,

Small arc melts of iron-boron alloys containing
0.5, 1,0, 1.5, 2.0, 2.5, and 3.0% boron were hot-
rolled in air at 1900°F, Edge cracking occurred in
alloys containing 2.5 and 3,0% boron, The hard-
ness of annealed alloys increased with the boron
content, being in the range of 43 to 97 Rockweli-B.

153
ANP PROJECT PROGRESS REPORT

The 1.5% boron alloy was cold-rolled 40% in thick-
ness before cracking occurred, while compositions
with less than 1.5% boron were rolled 65% before
cracks developed. A 30-lb ingot containing 1%
boron has been cast and will be extruded to
1 ]/z-in.-diu rod.

The pressure ring, 37-in, OD, 0.500 in, thick,
and ]1/2 in. wide, will be difficult to form and will
probably have to be in two separate pieces. In
order to determine the feasibility of casting the
ring in one piece, iron-boron alloys containing
0.75, 1.25, 2.00, and 3.00% boron were cast into
l-in.~dia ingots. All the clloy compositions had
hardnesses of about 55 Rockwell-C and showed no
response to heat treatments up to 1500°F, A
larger ingot had a hardness of 77 Rockwell-B,
which indicated that the cooling rate affected the
structure and hence its nardness.

In future tests examinations will be made of
these alloys to determine the effects of irradiation,
the yield strength of castings and wrought plate at
1300°F, and the extent of diffusion between the
various alloys and Inconel. Tensile bars of the
castings have been made; samples of wrought
alloys have been submitted for irradiation; and
diffusion couples of iron-boron alloys and Inconel
have been service-tested for 500 hr at 1300°F,

Bonding of Boron-Containing Materials to
Inconel. — Experiments are in progress at the
Carter Research Laboratories and the Vitro Lab-
oratories to develop methods for thermally bonding
a boron layer to Inconel. The experiments at the
Carter Research Laboratories are feasibility
studies which involve the codeposition of amor-
phous boron and electroplates of copper and nickel.
Thus far 10 vol % boron in copper has been
deposited, and the deposit has good mechanical
properties. The investigation at the Vitro Lab-
oratories involves the deposition of uniform metal
coatings on ceramic-B,C tiles and the bonding of
Cu-B,C and Fe-B,C layers to Inconel by electro-
phoresis.

Conversion of Boron to B,C, — Another attempt
was made to convert amorphous boron to B, C,
The earlier experiment was repeated, and it was
found that the conversion could be accomplished
by firing of a mixture of boron and graphite at
1750°C, A 20% weight loss that occurred during
firing can be attributed to loss of boron, inasmuch
as x-ray patterns of the product showed an excess
of carbon, The method appears to be unsuitable

154

for the conversion of expensive B0 unless the
weight loss can be controlled and the product can
be made crystalline.

Inconel-Boron Compatibility
M. R, D'Amore

The compatibility of Inconel and beron-containing
materials in intimate contact at temperatures in
the range 1100 and 2000°F was investigated. Also,
nine boron-free compounds were tested to determine
their suitability as barrier coatings for Inconel in
Inconel-boron assemblies.

H. Inouye

Diffusion couples of Inconel-B,C and Inconel-
(B,C-Cu) were tested for periods up to 1000 hr at
temperatures between 1100 and 1650°F. The test
specimens were prepared by hot roll-cladding B,C
powder and a 50 vol % B,C~50 vol % Cu mixture
between Inconel cover plates. In addition the
diffusion between Inconel and bodies of B ,C-Cy,
B,C (Norbide), B,C-SiC, and boron powder was
determined under a contact pressure of 50 psi.
The extent of diffusion was evaluated metallo-
graphically, and the results are presented in
Table 6.6.

Under the conditions of these tests, brittle re-
action products were formed at the B,C-Inconel
interface, and boron and carbon penetrated the
Inconel along the grain boundaries, as shown in
Fig. 6.19, which is a photomicrograph of a speci-
men tested at 1500°F with no contact pressure,
As may be seen from the test data, the diffusion
is low at 1100°F, but it increases rapidly with
temperature and with even moderate pressure.
Intergranular penetration was found up to 28 mils
in depth, but it appeared to have little hardening
effect on the Inconel matrix, as illustrated in the
microhardness traverse in Fig. 6.20.

Additional tests were made to determine the
compatibility of Inconel with metal borides, car-
bides, carbonitrides, oxides, and nitrides. The
test specimens, which were prepared by cold-
pressing powder mixtures of Inconel and the com-
pound powder, were heated at 1500 and 2000°F
for 250 hr in evacuated quartz capsules or in
helium, The reactions which occurred between the
Inconel and the compounds were determined visually
and by x-ray diffraction. The extent of the re-
action was estimated by the appearance of new

phases and the disappearance of the starting
compounds. The data obtained are presented in
Table 4.7.
TABLE 6.6,

PERIOD ENDING MARCH 10, 1956

RESULTS OF COMPATIBILITY TESTS OF INCONEL AND BORON-CONTAINING

MATERIALS IN CONTACT AT HIGH TEMPERATURES

 

Test Conditions

 

Compatibility

Depth of Boron or
Carbon Diffusion (mils)

 

 

) Contact Atmosphere
Specimen Time  Temperature -
hr) CF) Pressure Reaction Intergranular
(psi) Layer Penetration
|nconel-B4C 500 1100 None Evacuated Inconel capsule 1.0 1.0
1000 1100 None Evacuated lnconel capsule 1.0 1,0
500 1300 None Evacuated Inconel capsule 2 to3 4108
500 1500 None Evacuated Inconel capsule 1.5t0 2 15
1000 1500 None Evacuated lnconel capsule 5 to8 15t0 18
lnconel-(Cu-BAC) 1000 1100 None Evacuated Inconel capsule None 1.0
500 1300 None Evacuated Inconel capsule 2 to3 4108
1000 1300 None Evacuated Inconel capsule 5to 6
500 1500 None Evacvated Inconel capsule 1210 13
1000 1500 None Evacuated Inconel capsule 16 to 28
100 1500 50 Helium 1 tob 6 to 10
lnconel-54C 100 1500 50 Helium 2 to3 6 to 10
{Norbide)
Inconel-boron 100 1500 50 Helium 1.0 6to 9
{powder)
Inconel-boren 100 1500 50 Helium 1 to2 10 to 13
(hotpressed)
|ncone|-(B4C-SiC) 96 1600 50 Helium 2 1510 20
Inconel-(BAC-SiC) 26 1600 50 Helium No apparent reaction
(A|203 coated)
|ncone|-B4C 96 14600 50 Heliom 2 15t0 25

(Norbide)

 

Of the nine boron-free compounds which might
be suitable as barrier coatings, the oxides ZrO,,
Al,O,, and MgO were found to be inert to Inconel
at temperatures up to 2000°F. The carbonitride
(ZrCN), the SiC, and the SijN, reacted to some
extent with Inconel, as evidenced by a shift in
the diffraction lines of the Inconel and by the
instability of the compound, which was indicated
by the appearance of oxides. The absence of a
reaction between Inconel and BN was unexpected,
and, as yet, there is no explanation. It was ex-
pected that nickel borides would be formed. This

test is being repeated. The reactions between
Inconel and the metal borides make them unsuit-
able for use above 1500°F, presumably, because
of the formation of nickel borides, which melt

below 1900°F,

Nickel-Molybdenum-Base Alloys

T. K. Roche H. Inouye

Fabricability Studies. — The forgeability of an
alloy with the nominal composition of Hastelloy B

(68% Ni—28% Mo—4% Fe) has been studied as a

155
ANP PROJECT PROGRESS REPORT

 

| UNCLASSIFIED
CUYarse

 

 

Fig. 6.19. Inconel-B C Compatibility Specimen Tested ot 1500°F for 1000 hr with No Contact Pres-

sure. 250X,

 

 

 

Fig. 6.20. Microhardness Traverse of Inconel-
(B ,C-Cu) Compatibility Specimen Tested at 1500°F
for 1000 hr with No Contact Pressure. 100X,
Reduced 35%.

156

function of carbon content in an attempt to ascer-
tain the cause for the difficulties encountered in
the fabrication of Hastelloy B, The poor tfabri-
cability of Hastelloy B is believed to be attribut-
able to a stable high-temperature phase that has
been tentatively identified as (Ni Mo,},C. Arc
melts of the alloy were prepared in an argon atmos-
phere with nominal carbon additions of 0.01, 0.02,
0.04, 0.06, and 0.10% in 100-g samples. The
alloys were cast into rectangular bars approxi-
mately 3 in. long, %’ in. wide, and '/2 in. thick,
and then hot-rolled at 1150°C. Reductions in
thickness of 5% were used for the first few passes,
and then 10% reductions per pass were used until
a final thickness of 0.160 in, was attained, Visual
examination of the hot-rolled strips showed slight
edge cracking of the alloys with 0.01, 0.02, and
0.04% carbon additions, The alloys with 0.06
and 0.10% carbon additions were satisfactory. All
the strips were then cold-rolled without difficulty
to 0.100 in. at a rate of 0.01Q in. per pass,

A study of the microstructures of these alloys
revealed that the amounts of second-phase material,
TABLE 6.7.

PERIOD ENDING MARCH 10, 1956

AT 1500 AND 2000°F FOR 250 hr

COMPATIBILITY OF INCONEL AND YAR!OUS COMPOUNDS HEATED

 

Material

Tested at 1500°F

Tested at 2000°F

 

X+Ray Examination

Yisual Examination

X-Ray Examination

Visual Examination

 

Incone|-5i3N4
Inconel-BN

Incone |-(SiC-Zr02)

Inconel-(Ni-MgO)
Inconel-(ZrCN)

|nconel-B4C

Inncc:onel-":"_rO2
!nconel-A|2O3
Inconel-(A|203-Si3N4)
lnconel-(B4C-SiC)

|ncone|-(Zr02-M90)

Inconel-SiC
|n<:on<-:l-('§r82

Innt:or'ael-VB2
lncone |-TiB2
|nconel-Csz

Inconel-WB
|m::ont=,-|--TuB2
Inconel-MoB,
Inconel-ZrB,
InconeI-A|B‘2

Slight reaction

No reaction
Ne reaction

Slight reaction

ZrO2

No Inconel present;

‘no B4C present

No reaction
No reaction
Neo reaction

Slight reaction

No reaction

Slight reaction
Slight reaction

Slight reaction
Slight reaction

Slight reaction

Slight reaction
Slight reaction
Slight reaction

Slight reaction

Dense, strong body

Cracked and expanded
body

Cracked and expanded
body

No visible reaction

Porous, friable body

Porous, friable body

No visible reaction
No visible reaction
No visible reaction

Porous, friable body

No visible reaction

Expanded, porous
body

No apparent reaction

No apparent reaction
No apparent reaction

No apparent reaction

No apparent reaction
No apparent reaction
No apparent reaction

No apparent reaction

Moderate reaction

No reaction
No reaction

No Ni-MgO present
ZrO2 and ZrC formed

No Inconel present;

no BAC present

No reaction
No reaction
Stight reaction

SiC present; no

B4C present
No reaction

Moderate reaction
Moderate reaction

Moderate reaction
Moderate reaction

Complete reaction

Moderate reaction
Moderate reaction
Moderate reaction
Moderate reaction

Complete reaction; no
AIB]2 or lnconel

present

Porous, streng body

Porous, crumbly

body

Cracked and ex-
panded body

No visible reaction

(Sample con-

taminated)

Metallie globules in
a friable gray

powder
Hard, dense body
No visible reaction
Gray, friable body
Gray, friable body

Porous, friable

body

Dense, strong

metallic body
Hard, dense body
Porous, hard body

Melting occurred;
molten phase

mefallic
Hard, dense body
Porous, hard body
Dense, hard body
Porous, hard body

Melting occurred

 

presumably a carbide, in the alloys increased with
increased carbon content. On the other hand, the
carbon additions were effective in reducing the
fine oxide type of impurity carried over from the
melting stock and observed in the control alloy

(0.01% C). The microstructures are illustrated in
Figs. 6.21 and 6.22.

A correlation of the hot forgeability of these
alloys with their microstructures indicates that
the oxide precipitates are more detrimental to

157
ANP PROJECT PROGRESS REPORT

 

Fig. 6.21. Microstructure of 68% Ni—28% Mo-4% Fe Alloy Containing 0.01% Carbon. Oxide precipitates
may be seen in matrix. 500X,

 

Fig. 6.22. Microstructure of 8% Ni~28% Mo~4% Fe Alloy Containing 0,.10% Carbon. Carbide precipi-
tates present, but matrix cleaned of oxide precipitates. 500X,

158
fabricability than is the presence of the carbide,
up to 0.10% total carbon. Thus, close control of
the melting practice will be required in order to
improve the fabricability of nickel-molybdenum
alloys.

As was reported previously, attempts to extrude
Hastelloy W billets to tube blanks for the pro-
duction of seamless tubing have failed, A section
was cuf through a fractured area of one of the
extruded billets and examined metallographically.
Evidence of melting was found adjacent to some
of the fractures, as shown by the presence of the
eutectic in Fig. 6.23. X-ray analysis of the frac-
tured area did not reveal the nature of the eutectic,
and, as a result, no definite information on the
cause of the hot-shortness of the alloy was ob-
tained, It is hoped that the use of a lower billet
preheat temperature (1950°F rather than 2050°F)
and modifications of the present billet lubrication
methods will improve the extrudability of the
Hastelloy-type alloys. '

 

 

 

 

 

 

 

 

 

 

 

PERIOD ENDING MARCH 10, 1956

Effect of Melting Practice. — A program has been
initiated in which the melting practice will be
closely controlled and will be correlated with
mechanical properties and the fabricability of the
Hastelloy B and Hastelloy W alloys. Arrangements
have been made with Battelle Memorial Institute
for the production of arc-melted ingots with the
following compositions:  Nickel, Hastelloy B,
Hastelloy W, 76% Ni-17% Mo-7% Cr, and 83%
Ni~17% Mo. The melts will be made by the
consumable-electrode process to take advantage
of the high arc temperatures for vaporizing ‘‘tramp’’
elements. Extrusion billets will be prepared from
the ingots for the fabrication of suitable test
specimens,

Effect of Chromium Additions. —~ Previous work
showed that chromium additions to a basic 20%
Mo—80% Ni alloy resulted in poor forgeability of
the alloy, which was attributed, in general, to the
high oxygen content of the chromium. Since carbon
additions improved the fabricability of nominal
Hastelloy B, they were made to 5-Ib vacuum melts

UNCLASSIFIED]

Fig. 6.23. Section Through a Fracture Area of a Commercial Hastelloy W Billet Extruded at 2000°F,
Eutectic structure may be seen adjacent to the fracture. 2000X.

159
ANP PROJECT PROGRESS REPORT

of the basic 20% Mo—80% Ni alloy with 7 and 10%
Cr added. If these alloys can be hot-rolled, larger
heats containing 3, 5, 7, and 10% Cr will be pre-
pared for the fabrication of seamless tubing to
be evaluated in corrosion tests,

Special Alloys. — Five special alloys prepared
in 40-1b heats have been received from the Inter-
national Nickel Company feor corrosion testing and
strength evaluation, The compositions of these
alloys are as follows (in each alley, nickel con-
stitutes the balance):

Component Amount (wt %)
Mo 15 17 15 15 15
Cr 5
W 3 3 3
Nb 3 3 3
Al 0.5 0.5 0.5 1.0 0.5
Ti 1.5
C 0.25

Extrusion billets are being machined from these
ingots for fabrication of seamless tubing and test
specimens,

Niobium Fabrication

J. P. Page H. Inouye

V. M. Kolba

Arc-Melted Niobium. — Four fabricated ingots of
pure arc-melted niobium prepared by Battelle Me-
morial Institute have been received. These ingots
were melted in crucibles lined with niobium foil,
and getters were used. The dimensions and im-
purity analyses of the ingots are presented in
Table 6.8.

A program has been outlined for determining some
of the physical and mechanical properties of this

TABLE 6.8, DIMENSIONS AND IMPURITY ANALYSES
OF ARC-MELTED NIOBIUM INGOTS PREPARED
BY BATTELLE MEMORIAL INSTITUTE

 

Approximate Impurity Analysis

Weight

 

I:l?:f Dimensions (ka) (ppm)
(in.) C H N ©
10 0,5x1.5x75 1.07 100 6 440 717
13 0.5x 1.5x 4.5 0.56 100 4 330 448
14 0.5x2.0x 8.0 1.10 100 4 240 270
15 0.5x2.0x9.5 1.21 100 4 200 200

 

material with similar properties of the more common
wrought material prepared by powder metallurgy
techniques will be of particular interest., The
wrought material has a severe limitation as an
engineering material
are available,
Pack-Rolled Niobium. — Attempts to overcome
the limitation of the size of the wrought niobium
pieces have been made by pack-rolling several
laminae

in that only small pieces

in an evacuated capsule. Preliminary
tests indicate that the room-temperature mechanical
properties of the laminated and nonlaminated ma-
terials are quite similar, Room-temperature tensile
properties for 0.065-in.-thick Inconel-clad speci-
mens are presented in Table 6.9, The clad-core-
clad thickness ratio of these specimens was ap-
proximately 1:3:1. The clad-to-core bonds te-
mained intact until fracture, The scatter in results
can be attributed to small differences in the clad-
core-clad ratios and, to some extent, to minor
variations in the rolling technique.

Creep Test Specimens. — The extreme sensitivity
of niobium at elevated temperatures to even trace
amounts of nitrogen and oxygen has made the va-

 

 

material. A comparison of the properties of this lidity of available inert-atmosphere creep data
TABLE 6.9. ROOM-TEMPERATURE TENSILE PROPERTIES OF INCONEL-CLAD NIOBIUM
Number of Range of Range of
o . Number
Niobivm Laminae Diffusion Barrier of Tests Elongation Tensile Strength

in Core (2% in 2 in.) (psi x 1073)
1 Tantalum 3 6.3 to 10.0 83.7 to 87.4
2 Tantalum 2 8.8te 9.0 78.5 to 78.9
5 Tantalum 3 7.5 t0 10.0 82.0 to 83.7
1 Copper~stainless steel 2 10.5 72,8 to 78.5

 

160
questionable. Attempts are therefore being made
to fabricate Inconel-clad niobium creep specimens.
Two methods of protecting the edge of the lami-
nated sheet are being examined, In one method
a clad-niobium panel is machined to slightly over-
size creep-bar dimensions, the niobium is undercut
by preferential chemical attack, an Inconel wire
is inserted in the ‘‘groove,’’ and the assembly is
welded shut. A small test section has been suc-
cessfully sealed by this technique.

in the other method a niobium core and its
matching Inconel frame are machined to creep-bar
dimensions in one direction and to one-fifth the
final dimensions in the other direction; the core
is then clad by hot-rolling in the latter direction
to creep-bar size. A test section, in which the
niobium was simulated by stainless steel, has
been prepared by this method, This fabrication
technique vyields specimens that are completely
edge-protected. For examinations, the core could
be located on radiographs and the excess Inconel
could be cut away,

Creep data should indicate whether the Inconel
cladding supplies appreciable strength to the sheet,
According to published data,® the creep strength
of Inconel is so far below that of niobium that the
effect of the cladding will be negligible.

Fabrication of Large Panels. ~ Large panels
of clad niobium must be fabricated by welding,
and therefore the minimum core-to-core distance
(across the weld) is of considerable interest.
Metallographic examination of an Inconel-clad
niobium weldment has established a minimum
core-to-core distance of approximately ]{1 in, for
',é-in. sheet, with both copper—stainless steel
and tantalum-foil barriers,

Diffusion Barriers. — Further comparison of tanta-
lum and copper—stainless steel diffusion barriers
has yielded some anomalous data. It was reported
previously® that the cold-rolling properties of the
tantalum-barrier material were superior to those
of copper—stainless steel barrier material, These
sheets had been hot-rolied at 950°C prior to cold-
rolling. Similar cold-rolling tests of material that
had been hot-rolled at 1050°C indicate the reverse
to be true. A program has been outlined for de-

 

5General Electric Co., Aircraft Nuclear Propulsion
Department Engineering Progress Report No. 10, APEX-
10 (March 1954§’

6). H. Coobs et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 144.

PERIOD ENDING MARCH 10, 1956

termining the effect of rolling temperature on these

and other properties of roll-clad niobium sheet,

Niobium-UQ, Compatibility. — Niobium is being
considered for use in solid fuel elements because
of its good high-temperature strength. In order
to evaluate the usefulness of niobium in this type
of service, a program was initiated for investi-
gating the compatibility of niobium and UO,. In
a previous examination of swaged and rolled ni-
obium powder-UO, encapsulated compacts, a
phase, other than the UO, particles, was found
in the niobium matrix. An x-ray diffraction pattern
of the compact after it was aged for 500 hr at
1000°C indicated essentially niobium and UO,,
but lines of Nb,O, were also found.

Metallographic samples of as-received and as-
swaged niobium powder also showed the third
phase found in the initial niobium-UO, compacts.
A Bergsmann hardness check on the as-received
powder particles revealed that the hardness of
the niobium matrix was about 210 VHN, whereas
that of the second phase was about 2400 VHN,
as shown in Fig. .24,

Wrought niobium plate obtained from Fansteel
Metallurgical Corp. was metallographically ex-
amined and found to have a few small inclusions
of the compound in an essentially clean matrix
of niobium. Uranium oxide powder was placed
between two pieces of the Fansteel plate, encap-
sulated, and rolled, and a portion of the rolled
plate was etched and then examined at 1000X,
No reaction between the UO, particles and the
niobium matrix was found, A portion of this rolled
specimen has been aged for 500 hr at 1000°C and
is being examined metallographically,

The results obtained thus far indicate that ni-
obium powder of higher purity must be obtained,
A high-purity grade of powder is being prepared
by converting wrought material to the brittle hy-
dride, grinding, and, finally, vacuum annealing to
remove hydrogen. A further test of the compat-
ibility of UO, and niobium will be conducted
with this high-purity powder. A high-temperature
(~2000°C) vacuum furnace will be constructed
to accomplish the sintering of the niobium powder
compacts.,

Seamless Tubular Fuel Elements
J. H. Coobs M. R. D'Amore

The work on simulated seamless tubular fuel
elements was continued with the preparation and

161
ANP PROJECT PROGRESS REPORT

AVERAGE HARDNESS OF COMPOUND, 2390 VHN
(4 IMPRESSIONS)

(3

IMPRESSIONS)

 

UNCLASSIFIED
Y-17489

 

Fig. 6.24. Hardness Impressions on a Polished Section of As-Received Niobium Powder. 500X, Re-

duced 2%.

extrusion of four three-ply extrusion billets, The
core materials were ~100 mesh type 302 stainless
steel powder with 30 vol % Al,O, to simulate U02.
The billet can material was type 316 stainless
steel. Two billets were filled with loose powder
tamped to a density of about 55% of theoretical.
The other two billets had hot-pressed cores with
densities of 80 and 89% of theoretical. The ex-
trusion data for the billets are presented in
Table 6.10.

Sections 18 in. long were cut from the center
of extrusions 3 and 4 and shipped to the Superior
Tube Company for redrawing to small-diameter
tubing. Extrusions 1 and 2 and the remaining por-
tions of extrusions 3 and 4 were then split longi-
tudinally, and the flow of material was examined
visually. It was found that slight misalignment
of billets 2 and 3 in the container of the extrusion
had caused uneven material flow during
The cores in these two extrusions were

press
extrusion,

162

offset in the direction of extrusion. The core cross
sections of the extruded tubes appeared to be
concentric except in extrusion 3. The eccentricity
of this core was attributed to nonuniform tamping
of the core powder during billet preparation.

Segments have been cut from extrusions 1, 2,
and 4 at various intervals, and the layer thick-
nesses are being measured. Thus for the measure-
ments have been completed only on extrusion 4,
A section approximately 33 in. long, which con-
tained about 70% of the original core material,
was found to be uniform, The layers were cal-
culated to have extruded in a Y:%:¥% thickness
ratio, with the actual thicknesses being 0.048 in,
for the outside cladding, 0.041 in. for the core,
and 0.037 in, for the inside cladding.

The work on seamless tubular fuel elements is
currently being directed toward a study of the re-
drawing properties of the material. Efforts are
also being made to increase the length of the
PERIOD ENDING MARCH 10, 1956

TABLE 6,10, EXTRUSION DATA FOR THREE-PLY SIMULATED FUEL ELEMENTS

Billet soaking temperature: 2100°F

 

Extrusion No, Type of Core

A|203 Particle Mesh Size

Extrusion Ratio

 

] Tamped powder
2 Hot pressed
3 Tamped powder
4 Hot pressed

=325+ 5:1
-140 +325 9:1
-140 4325 21:1

-325* 21:1

 

*Grade 38-500.

uniform core section of the extrusion, The Alle-
gheny Ludlum Steel Corp. is aiding ORNL in the
study of three-ply extrusions,

Control Rod Fabrication

J. H. Coobs R. E. McDonald
M. R, D'Amore

An investigation has been initiated of the feas-
ibility of extruding control rods to close dimensional
tolerances, The control material of interest is a
30 wt % Lindsay oxide—~70 wt % Ni mixture, which
is to be formed into a 2.5-in.-0OD, 2.0-in.-ID, 24-
in.-long cylinder ¢lad externally and internally
with 0.020-in.-thick Hastelloy X. Similar control
materials containing iron rather than nickel and
Inconel rather than Hastelloy X are also of interest,
Small rods of these materials will be used for the
initial extrusion studies because the extrusion
press at ORNL does not have the capacity to ex-
trude a full-sized control rod.

Several methods have been investigated in an
effort to determine the optimum means of fabri-
cating cores of the Lindsay oxide—metal mixture
for 3-in.-dia extrusion billets. Small compacts
containing 30 wt % Lindsay oxide and 70 wt % Ni
have been fabricated successfully both by hot-
pressing and by cold-pressing plus sintering., Hot-
pressing of the powder mixtures at 1200°C produced
a sound compact with a density of 86.4% of theo-
retical, Cold-pressing plus sintering resulted in
a compact with a density of 83% of theoretical,
In preparing the compact the powder was pressed
at 80 tsi, sintered at 1900°F for ]/2 hr, coined at
50 tsi, sintered again at 2100°F, and coined again
at 50 tsi. The cold-pressed body cracked in one
area during sintering at 2100°F. The method used
gives satisfactory densities in small bodies, but
it is not amenable to fabrication of cores for 3-in.-

dia extrusion billets because of the excessive
pressure requirements,

Shield Plug for ART Pumps
J. P. Page J. H. Coobs

The shield plugs surrounding the shafts of the
ART fuel and sodium pumps must fulfill three
primary functions; they must serve as neutron,
gamma-ray, and thermal shielding. In their thermal-
shielding capacity they must prevent freezing of
the fuel below them during zero power operation.
Since no known material will satisfy all these
conditions adequately, it has been proposed that
the following layers of materials be stacked in a
vented Inconel can: (1) a ]/B-in. disk of a B,C-Cu
mixture, for neutron shielding; (2) a 5/é°in. disk of
zirconia, for thermal shielding; (3) a 4- to 5-in,
slug of high-density (>12 g/cm3) low-conductivity
(<0.12 cal/em?/sec) material, for gamma-ray shield-
ing. The top surface of the gamma-ray shielding
material is to be brazed to the Inconel can, which
will be cooled on its outer surface.

Tantalum-constantan and tungsten carbide-con-
stantan compacts are being investigated for use
as the gamma-ray shielding material. A successful
diffusion bond of tantalum-constantan to nickel
has been made by heating the parts at 1100°C in
a hydrogen atmosphere for 20 min and then cooling
in @ helium atmosphere. The specimen was cooled
in helium in order to minimize the formation of
tantalum hydride, which probably caused the crack-
ing of several earlier specimens during cooling.

Lithium-Magnesium Alloys
R. E. McDonald

A study of the fabricability of a light alloy con-
taining 20% Li has been initiated. Such an alloy
would be useful as shielding material. Tests were

163
ANP PROJECT PROGRESS REPORT

started on a 20% Li-80% Mg alloy, but it was
found to be highly reactive in both air and water.
A method of surface protection is necessary before
mechanical and physical data can be acquired.
Chemical coatings, flame spraying, and roll cladding
were investigated without success. Thus far the
Dow No., 17 anodizing treatment, which can be
supplemented by a Unichrome plastic coating, has
formed the only successful surface protection.

A roll-claddable alloy would be more suitable
for shielding, and a less reactive surface shouid

be more amenable to roll cladding. Therefore an

effort is being made to produce a less active
surface by adding small amounts of aluminum to
the lithium-magnesium alloy.  The lithium-to-

magnesium ratio is being held constant at 1:4,
and ternary alloys with 0.5, 1, 1.5, and 2% Al are
being investigated.

Fabrication of Ceramic Materials
L. M. Doney
J. A, Griffin A. J. Taylor

L ow:Density Rare-Earth-Oxide Compacts. — Meth-
ods are being studied for preparing compacts of
samarium and gadolinium oxides (Lindsay Code
920) for use in ART control rods. The compacts
are to be 1.275 in, OD, 0.775 in. ID, and 1.000 in.
long, and they are to have densities of between
3.95 and 4.4 g/cm3. Experimental compacts have
been produced with a density of 3.53 g/em?® and
a porosity of 53.5%. These compacts have been
exposed to molten sodium at 1300°F for 100 hr
and have been found to have a porosity to sodium
of 53%. The compacts were not damaged by the
immersion in sodium. The compacts required for
the actual ART control rods will probably be pre-
pared by a method very similar to that used in the
fabrication of these experimental compacts.

Cermets of Iron and Rare-Earth Oxide. — Cermets
of 50 vol % iron and 50 vol % rare-earth oxide
have been requested for use in the extrusion of
Inconel control rods with cermet cores. The fabri-
cation method devised for these cermets involves
mixing 25 vol % iron metal, 25 vol % iron as Fe,0,,
and 50 vol % samarium and gadolinium oxide
(Lindsay Code 920) and pressing the mixture into
a hollow, cylindrical shape at o very low (10,000
psi) pressure in a steel die. The cylinder is then
isostatically pressed at 80,000 psi in a rubber
bag immersed in an oil-filled pressure vessel,
The isostatic pressing assures uniform density,

164

After pressing, the cylinder is sintered in hydrogen
at 1475°C for l/2 hr. Cermets fabricated by this
method have a density of 4.6 g/cm® and a porosity
of 38.8%. Pieces of the size required for extrusion
will be produced by this method.

Cermets of Nickel and Rare-Earth Oxide. —
Cermet compacts of 70 vol % nickel and 30 vol %
rare-earth oxide have also been requested for ex-
trusion studies of Inconel-rare-earth-oxide control
elements. These compacts were fabricated by the
same method as that used for the production of the
iron—rare-earth-oxide cermets, except that the
compacts were sintered in hydrogen at 1400°C for
20 min, The compacts had a sintered density of
6.1 g/em? and a porosity of 28.4%.

NONDESTRUCTIVE TESTING
R. B, Oliver

J. W, Allen K. Reber
Metallurgy Division

Success in the inspection of small-diameter Inconel
and Hastelloy B tubing with the Cyclograph, de-
scribed previously,” prompted further study of the
instrument, It is primarily a tuned oscillator in
which the oscillator coil encircles the tubing, The
instrument measures the amplitude of the oscil-
lations, the amplitude being a measure of the
resistive component of the coil’s impedance. Theo-
retically, since only the resistive component is
measured, all changes in the tubing cause the
same type of variation,® and good tubing could
possibly be rejected because of an inconsequential
variation. However, experience has proved that
the wobble of the tubing in the coil, which is the
worst offender in producing spurious flaw indi-
cations, can be eliminated by a well-designed
mechanical feed mechanism, such as that shown
in Fig. 6,25, The instrument is sensitive to slight
changes in outside tubing diameter and wall thick-
ness, but these changes occur very slowly over
the length of the tube and can be easily separated
from flaws, which give abrupt signals,

An important feature of the Cyclograph is that its
operating frequency is determined by its sensing
coil and, hence, may be altered by merely changing
the coil. The proper frequency for a particular
tube is selected according to its outside diameter,

 

7R. B. Oliver et al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 164.

8R. B. Oliver et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 139.
wall thickness, permeability, and conductivity.
A graph of the Cyclograph reading vs tubing wall
thickness, expressed as a percentage of the out-
side diameter, is shown in Fig., 6.26. The pa-
rameter for the graph is frequency-normalized to

a valve /_, which is determined by the outside

OUNTING HOLES

i

Yg~in MICARTA
BOARD

   

i

-

  

 

 

 

L aiFiLar WINDING/

=
SECTION A-A

iy oA
et

PERIOD ENDING MARCH 10, 71956

diameter, the permecbility, and the conductivity
according to the relationship given on the graph,
The operating frequency that gives the best results
is the one which locates o point near a peak on
the graph of Fig. 6.26, The dashed lines indicate
the nonlinear characteristic of the instrument near

UNCLASSIFIED
ORNL-LR-DWG 11015

   

 
 

    

1
T
1’,*

0
1

 

 

DETAIL OF CYCLOGRAPH COIL FOR TUBING

     
     
 

FLOATING MOUNTING

FIXED MOUNTING WITH
SYNCHRONOUS MOTOR DRIVE

WEIGHTED ROLL /

Fig. 6.25. Cyclograph Coil Assembly and

SYNCHRONOUS MOTOR DRIVE

FIXED MOUNTING WITH ™.

SUPPORT PLATE

 

2

TUBING UNDER INSPECTION

T T,

0
v
"
S
Q
%

COIL ASSEMBLY

-.
o T i,

Feed Mechanism for Tubing Inspection.

165
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORML—=LR—DOWG 11105

 

   
  
 
  

2
Try.yDE

fo =

jes)
o

1 =PERMEABILITY (henries/meter)
— ¥ =CONDUCTIVITY (mhos/meter) -—-
D =0UTSIDE DIAMETER (meters)

D
O

 

o
o

CYCLOGRAPH READING
(% OF FULL SCALE)
N
O

 

O 5 10 15 20 25 30 35 40 45 50

TUBE WALL THICKNESS (% OF D)

Fig. 6.26. Cyclograph Reading vs Tube Wall
Thickness with the Operating Frequency of the
Cyclograph as a Pargmeter.

the '‘quench’ or zero-reading axis, which allows
extreme sensitivity to small flaws in the tubing.
A plot of the frequency which will locate a peak
on the graph of Fig. 6.26 for each value of wall
thickness is given in Fig. 6.27. By consistent
use of operating frequencies chosen from these
plots, flaw types will always be detected in the
same relative magnitudes. For instance, an inside-
diameter crack extending through 10% of the wall
thickness would cause the same indication from
a 0.25-in.-0D, 0.025-in.-wall tube as from a 0.5-
in.-OD, 0.049-in.-wall tube. Although it is not
possible to obtain a calibration of flaw size vs
instrument indication, it is possible to obtain an
approximate determination of flaw size in terms
of percentage reduction in wall thickness averaged
over the length of the coil. Typical Cyclograch
indications of outside-diameter cracks in Hastel-
loy B tubing and inside-diameter intergranular
attack in Inconel tubing are shown in Fig. 6.28,
and photomicrographs of two of the flaws thus

detected are shown in Figs. 6.29 and 6.30.

The "B scan presentations of the eddy-current
probe coil and of immersed ultrasonic indications
are also photographed, in an effort to identify the
several characteristic types of defects found in
small-diameter tubing., Photomicrographs are also
prepared of appropriate sections of the tubing, but
difficulty has been encountered in the preparation
of representative metallographic sections because
of the minute dimensions of the defects and in-
ability to sufficiently pinpoint the defects for good
metallographic presentation,

166

UNCLASSIFIED
ORNL-LR—-DWG 11106
100

2

5 ;2:

Tr;..cyDz

= PERMEABILITY (henries /meter)
7 = CONDUCTIVITY (mhos /meter)
D= DIAMETER {meters)

|

 

1 2 5 10 20 50 100
TUBE WALL THICKNESS (% OF QD)

Fig. 6.27. Frequency vs Tube Wall Thickness

for Maximum Resistive Coil Impedance.

A type of defect that has been prevalent in small-
diameter Hastelloy B tubing is a longitudinal
crack which may penetrate as much as 80% of the
tube wall from the outside surface. A typical
example is illustrated in Fig. 6.29. The ultrasonic
“B'" scan, as well as the Cyclograph, indicated
this crack, as shown in Fig. 6.31, A crack of this
size can be detected by ultrasound through almost
300 deg of tubing rotation, as is demonstrated by
the diagonal indication that extends almost com-
pletely across the screen, The eddy current probe
coil “B"" scan of this crack is shown in Fig. 6.32.

A characteristic defect in Inconel is the inter-
granular type of attack illustrated in Fig. 6.30.
Attempts to detect this defect ultrasonically gave
inconclusive indications, probably because the
penetration was shallow and the orientation was
not parallel to the axis of the tubing, cumulative
factors which might preclude valid ultrasonic in-
spection. Mechanical difficulties caused by the
excessive wobble of the tubing prevented an ade-
quate probe-coil inspection.

The small l/lé-in.-long gouge shown in Fig. 6.33
on the surface of an Inconel tube was readily
detected ultrasonically, A representative ‘‘B"
scan picture for the ultrasonic detection of this
defect is shown in Fig. 6.34.
PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
ORNL—LR-—DWG 11104

 

 

. \ \\ Ty L L LY \ Y CYCLOGRAPH TRACE AT 160 ke OF REJECTED
W Lo L 0.25-in.-0D x 0.049-in.~WALL HASTELLOY B
\ \\ \ \ \L \\ , \\ Ro8EA I ‘\ \\ TUBING. OUTSIDE DIAMETER RADIAL CRACK
~ \ ! L

 

j | | -3y | 0.027 in. DEEP AT POSITION R586A-1{—-1 IS
| Lfll .l——v—k’t JM\J ) i ‘ SHOWN IN FIG. 6.29.
T T T

/
/ ,.-' / | /

; : — ——
? / / // [ /) // [ /

 

3 CYCLOGRAPH TRACE AT 200 kc
\\ OF REJECTED 0.25-in.-0D x
W 0.025-in.-WALL INCONEL TUB-

 

ING. INSIDE DIAMETER INTER-
GRANULAR ATTACK AT POSITION
/L 34RA TO A MAXIMUM PENETRA-

TION OF 0,002 in. IS SHOWN IN
FIG. 6.30.

 

Fig. 6.28. Cyclograph Traces of Defective Tubing.

 

. Fig. 6.29. Radial Crack 0.027 in. Deep on Outside Surface of a 0.25¢in.-0OD, 0,049-in.-Wall Hastelloy B
Tube Detected by Indication Shown in Fig. 6.28. 100X, Reduced 2%.

167
ANP PROJECT PROGRESS REPORT

 

Fig. 6.30. Intergranular Attack to a Maximum Depth of 0.002 in. on a 0.25-in.-0OD, 0.025-in.-Wall Inconel
Tube Detected by Indication Shown in Fig. 6.28. 150X,

UNCLASSIFIED
PHOTO 15158

 

Fig. 6.31. Ultrasonic ‘‘B’’ Scaon of Crack Shown in Fig. 6.29.

168
 

PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
PHGTO 16155

 

Fig. 6.32. Eddy Current ‘*B’’ Scan of Crack Shown in Fig. 6.29.

 

 

Fig. 6.33. Gouge 0.004 in. Deep on the Outer Surface of a 0.25-in.-0D, 0.025-in.-Wall Inconel Tube.
100X, '

169

 
ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED
Y-16454

 

Fig. 6.34. Ultrasonic ‘'B’’ Scan Indication of
the Gouge Shown in Fig. 6.33.

In an effort to compare the results obtained from
immersed ultrasound, radiography, and fluorescent
penetrant inspections, six pieces of 0.1875-in.-0D,
0.025-in.-wall, 7-ft-long CX-900 Inconel tubing
were inspected by the respective methods. Mere
detection was considered to be sufficient to indi-
cate a defect, and no attempt was made to estimate
the defect size. The total number of such defect
indications and the correlation obtained between
the various methods are presented in Table 6.11,
The sampling is insufficient for final decisions
to be made concerning inspection methods, but
definite trends are indicated. |t appears that no
one of the methods alone would be adequate, and
if any one of them was eliminated a few defects
would be overlooked.

TABLE 6.11. CORRELATION OF RESULTS OF INSPECTION OF CX-900 INCONEL TUBING BY
X-RAY, ULTRASOUND, AND FLUORESCENT PENETRANT METHODS

 

Number of Defects Correlated by

 

 

 

Number of Defects Found by Number of Defects Which Did Not Toro Morbodde Number of
“Tube Each Method Correlate With Other Methods Defects
" Krey Ultmosound TSI e Flerescomt ey Tey Ultosond - Correlared by
Penetrant Ultrasound Penetrant
CxX-1 7 13 14 2 5 6 3 3 6 i
CxX2 2 4 16 2 0 12 0 0 4 0
CX3 1 25 13 0 13 1 1 1 12 1
CX<4 6 5 12 6 2 9 0 0 3 0
CX5 3 4 14 1 1 11 1 1 2 0
CX-6 2 Z 13 0 3 Z 2 0 4 o
Total 21 58 82 1 24 48 7 5 30 2

 

170
PERIOD ENDING MARCH 10, 1956

o 7. HEAT TRANSFER AND PHYSICAL PROPERTIES
H. F. Poppendiek

Reactor Experimental Engineering Division

FUSED.SALT HEAT TRANSFER
H. W, Hoffman

Reactor Experimental Engineering Division

D. P, Gregory
Pratt & Whitney Aircraft

Preliminary heat-transfer studies were carried
out on a fresh batch of the fuel mixture NaK-KF-
LiF-UF, (11.241-45.3-2.5 mole %) in a newly
constructed system containing a Hastelloy B test
section. The results, given in terms of the Colburn
function (Fig. 7.1), lie above those obtained with
an earlier batch of the same salt but are still
below the general heat-transfer correlation. The
Reynolds modulus range of the data will be ex-
tended in further experiments. It is believed that
the difference between the two sets of data may
be the result of suspected variations in the two
salt batches. The new set of heat-transfer data
falls about 30% below the correlation for ordinary
fluids, and this difference may be attributed, in
part, to the inaccuracy of the estimated value of

0.010

Pr

2/
. 3
/st N

o

o

S

w

0.002

COLBURN FUNCTION, /

0.001

1000 2000 5000

 

thermal conductivity used in the correlation of
the salt data.

Operation of a heat-transfer system that includes
a pump to circulate the fused salts was initiated.
Operational experience with NaNO,-NaNO,-KNO,
(40-7-53 wt %) shows that the system performs
satisfactorily at temperatures up to 700°F. Heat-
transfer experiments with water (Fig. 7.2) indicate
that no systemic error exists. The pump system
will be used to study the salt mixture NaF-KF-
LiF-UF, (11.2-41-45.3-2.5 mole %) rather than
NaF-ZrF ,-UF, (50-46-4 mole %), as reported
earlier, However, the heat-transfer characteristics
of the latter salt will be examined in o pressurized
system.

At present, fluid flow rates are obtained in the
pressurized systems with volume probes. Although
extreme precautions are taken, small quantities of
oxygen and water vapor possibly enter the system
when probe changes are necessary. In addition,
it is difficult to maintain the calibration of the
volume probes. Therefore a more positive method

ORNL-LR-DWG 12600

O PREVIOUS DATA
® NEW DATA

. : |

GENERAL CORRELATION FOR ORDINARY
FLUIDS: /= 0.023 N;EOE

10,000 20,000 50,000 100,000

REYNOLDS MODULUS, Vg,

Fig., 7.}, Comparison of Heat Transter Measurements on Two Batches of NaF-KF-LiF-UF , (11.2-41.
45,3-2.5 mole %) with the General Correlation for Ordinary Fluids.

171
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 12601

500

200

100

C.
NNu/NPr 4

50

20

 

5x 0%
REYNOLDS MODULUS NRe

Fig. 7.2. Results of Heat Transfer Measurements
with Water in the Test System that Includes a
Pump.

of obtaining the fluid flow rate has been investi-
gated. The fluid weight is obtained by measuring
the deflection of a cantilever beam which supports
one of the tanks, Direct calibration of the system
is accomplished by loading it with known weights,
The beam movement can be measured by using
either an iron core microformer or strain gages.
The deflection signal, when amplified, can also
be used to operate the system cycling mechanism.
Thus, the necessity of opening the system during
the experiment will no longer exist.

ART FUEL-TQ«NaK HEAT EXCHANGER
J. L. Wantland

Reactor Experimental Engineering Division

Heat-transfer and isothermal friction character-
istics of the ART fuel-to-NaK heat exchanger
have been determined for three degrees of spacer
density: 9 sets of spacers at 71/2-in. intervals,
13 sets of spacers at 5-in, intervals, and 25 sets
of spacers at 2Y-in. intervals, The data, a portion
of which have been extrapolated to zero spacers,
are shown in Figs., 7.3 and 7.4. For the normal
spacer density (13 spacers at 5-in, intervals) the
heat-transfer and isothermal friction characteristics
in the Reynolds modulus range of 1,500 to 10,000
can be expressed by NNU/NPI'O-4 = 0.0155 NRGO'B

172

 

 

 

 

 

 

 

 

 

   

 

——

ORNL-LR-DWG {2602

100 | }
1. | L
L | - .

e 25 SPACERS AT 2%-in. INTERVALS | |
>0 4 13 SPACERS AT 5-in. INTERVALS (NORMAL)
w9 SPACERS AT 7',-in. INTERVALS || | |
— GENERAL CORRELATION ‘

 

—

. [

iy s
. |

 

 

! i /.':’L / . I
| "({l
.‘, 'I} . }
/‘A /| . P
o s~=—EXTRAPOLATED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b
S, .»/x‘é}}g, TO ZERO SPACERS
T 10 S A LT
L ) B
¥ | /‘{":;’ | ]T -
Pab- s el ]
R ]
oA |
o2 a0 > 04 _ ", 08
AL Ny Mo, = 0.0155 N,
5 4,
A . .
. [ SN
IR
]
- — .
-
. . e
L
& - | |
10° 2 5 104
REYNOLDS MODULUS, Mg,
Fig. 7.3. Heat Transfer Characteristics of an

Experimental ART Fuel-to-NaK Heat Exchanger
for Three Spacer Densities.

and f = 0.175/Ny _0-39. The limited data presented
in Figs. 7.3 and 7.4 appear to indicate that for
a given heat exchanger pressure drop the heat-
transfer coefficient is almost independent of the
spacer density. For example, as the spacer density
increases, the Reynolds modulus decreases to
such an extent that the Nusselt modulus does not
change significantly,

The available heat-transfer data for flow parallel
to a square array of cylindrical tubes and for flow
through square cross-sectional ducts' are summa-
rized in Fig. 7.5. It is thought that a square duct

 

1J. L. Wantland, Thermal Characteristics of the ART
Fuel-to-NaK Heat Exchanger, QRNL CF-55-12-120 (Dec.
22, 1955).
o>

ORNL-LR-DWG 12603

05

0 25 SPACERS AT 2% —in. INTERVALS

A 13 SPACERS AT 5-in. INTERVALS (NORMAL!}
® 9 SPACERS AT 7% —in. INTERVALS

-—— EXTRAPQOLATED TO ZERC SPACERS
-—— GENERAL CORRELATION

0.2

FRICTION FACTOR, £
o

0.05

0.02

 

0.01
103 2 5 104

REYNOLDS MODULUS, NRe

Fig. 7.4. lsothermal Friction Characteristics of
an Experimental ART Fuel-to-NaK Heat Exchanger
for Three Spacer Densities.

approximates the flow configuration of the ART
heat exchanger better than any other geometry for
which heat-transfer data are known., At high
Reynolds moduli the data are in agreement with
the relation NNu/NPro'A = 0.023 NRGO'S; in the
transitional flow regime all experimental data fall
below this equation. The data for curve 2 were
obtained with 20- and 100-tube heat exchangers
circulating NaK and fuel.

An analysis has been made in an effort to recon-
cile the disagreement between the data presented

PERIOD ENDING MARCH 10, 1956

 

 

 

 

 

 

 

 

 

 

W
5 ORNL—LR—-DWG 12604
10 T ! T T =
1 WATER FLOW PARALLEL TC TUBES IN A 17T
5 | SQUARE ARRAY, S = 1.{§ J
| 2 ART FUEL FLOW PARALLEL TO TUBES
IN A SQUARE ARRAY, § = {11
2 — }
3 AIR FLOW IN A SQUARE DUCT, £/0 =100 E
16* |~ 4 WATER FLOW IN A SQUARE DUCT, L/ = 130 |

 

 

 

 

 

 

— 5 WATER FLOW PARALLEL TO TUBES IN A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ST SQUARE ARRAY, § =112
& WATER FLOW PARALLEL TO TUBES IN A |7
2 |- SQUARE ARRAY, § =1.2 e s
0.4 _ 08
e ~——GENERAL CORRELATION W, /W °* = 0.023 a,
i I L2, 1
i T
g i 5 e = o oo b . st o]
& i P et -
= ; Lo i 1 G 15
S N fi/ i
EZ 2 : // .
102 - "/ il
- e -
i -1
5 — >
TN
| ‘;’/
2 : ,;Z;’
/o
10 pooT < : /1
= - A i
2 & - {
1 | I |
10 2 5 10t 2 5 10° 2 5 108

REYNOLDS MODULUS, Ay,

Fig. 7.5. Heat Transfer Characteristics for Flow
Through Square Ducts or Parallel to Cylindrical
Tubes in a Square Array.

in Fig, 7.4 on the effect of spacers on the fluid
friction characteristics in an ART heat exchanger
tube bundle and the data reported by Cooper.?
This analysis indicated that the differences in
the spacings between the tube bundle and the
container wall for the two experimental configura-
tions were not great enough to explain the differ-
ence between the two sets of friction data,
Actually, this disagreement is not serious, because
the spacer friction is only a small part of the
total friction,

 

M. H. Cooper, Pressure Drop of Heat Exchanger Tube
Spacers, ORNL CF-55-11-180 (Nov. 28, 1955).

173
ANP PROJECT PROGRESS REPORT

. . ART CORE HYDRODYNAMICS
C. M. Copenhaver F. E. Lynch

Reactor Experimental Engineering Division

G. L. Muller
Pratt & Whitney Aircraft

A 5/22-sca|e model of the ART sodium coolant
annulus, which was designed for a study of the
fluid flow distribution, was fabricated and as-
sembled without the spacers and lands, Although
the model does not incorporate the actual reflector-
coolant hole distribution, provisions have been
made for tapping off the flow (representative of
reflector-coolant holes) at various sectors of the
inlet header,

The preliminary experimental results obtained
were based on 38.6% of the reflector-coolant flow
going to the annulus and the remaining 61.4% being
distributed uniformly to the coolant holes. The
limitations of the experimental system made it
impossible to attain the average Reynolds modulus
of the actual coolant annulus, approximately 10%;
however, Reynolds moduli up to 10* were obtained,
and the results were extrapolated to give an esti-
mate of the flow distribution, For a concentric
annufus (width, 0,135 in.) the average flow through
the portions of the annulus farthest away from
the inlet appears to be about 10% greater than that
through the portions nearest the inlet. Spacers
have been inserted in the annulus according to
the ART design, and flow-distribution and pressure-
drop measurements are being obtained for concentric-
annulus conditions, for various radial displace-
ments of the outer core shell, and for other
conditions.

Yisual studies were made of the flow through
the 21-in, ART core model after the installation
of vortex generators supplied by Pratt & Whitney
Aircraft, Vane angles of 50, 55, 60, 65, and 70 deg
were used. The 50- and 55-deg vortex generators
improved the flow in comparison with that obtained
with the 45-deg generators, but the generators with
larger angles gave no further improvement, In
all cases there was a persistent region of fluctu-
ating flow at the equator on the island wall, and
flow through the core was characterized by macro-
scopic turbulence. Although no large reverse-fiow
regions existed, steady and unsteady flow were
observed in various regions of the core. A faired-
in entrance has been built to use with the vortex

174

generad®eemin order to get better flow distribution
going into the vanes,

TEMPERATURE STRUCTURE IN REGION
BEYOND THE ART REFLECTOR

H. W. Hoffman J. L. Wantland
C. M. Copenhaver

Reactor Experimental Engineering Division

One of the several interrelated problems arising
in the ART design is the determination of the
temperature increase of the reflector-cooling sodium
as it returns outside the reflector to the sodium
pumps. Since the outlet temperature of the sodium
is fixed by the allowable sodium-beryllium inter-
tace temperature, the sodium temperature rise on
the return side is important in establishing the
axial temperature gradient available for cooling
both the reflector and the outer core shell. |In
order to evaluate this temperature rise, it was
necessary to determine the heat sources and their
locations. The idealized slab system for which
the temperature analysis was made is indicated
in Fig. 7.6. The actual system, in which the boron
carbide tiles contact the walls at a finite number
of points, has been simplified in terms of two
In system A it is assumed that
no thermal resistance exists between the boron
carbide and the Inconel shells containing it, while
in system B a 10-mil-thick helium gap isolates
the boron carbide from the Inconel shells. Systems
A and B are considered to be lower and upper
limits of the problem. The heat-deposition rates
used in these calculations are given in Table 7.1,

limiting systems.

The temperature profile in the region between:
the sodium and the fuel return circuits was cal-
culated by the method of superposition for a slab
geometry. The temperature profiles obtained at
the outlet end for systems A and B are compared
in Fig. 7.6.

The temperature of the sodium was evaluated by
using the heat-flux distributions on each side of
the sodium return passage. On the reflector side,
the heat flux at any point was the heat generated
in the 2 in, of beryllium immediately adjacent to
the sodium as corrected for the heat flow in the
beryllium resulting from conduction. On the out-
side, the flux was determined from the temperature
profile obtained by iteration. The resultant tem-
perature rise of the sodium was 33°F in system A
and 25°F in system B,
PERIOD ENDING MARCH 10, 1956

S
ORNL~LR —DWG 12605

 

1450

SYSTEM A |

 

1400

—-1—=S50DIUM

 

1350

INCONEL BORON CARBIDE INCONEL FUEL

 

1300

TEMPERATURE (°F}

 

1250
/

1200
1450

 

 

 

 

 

 

 

 

 

 

 

SYSTEM B

1400

SODIUM HELIUM HELIUM

o
1350

iINCONEL BORON CARBIDE INCONEL

TEMPERATURE
o
C
C

1250

 

1200
60 50 40 30 20 10 o —10

DISTANCE FROM INCONFL—FUEL INTERFACE ( x1073in.}

Fig. 7.6. Comparison of Temperature Profiles in Region Beyond the ART Reflector.

TABLE 7.1, HEAT-DEPOSITION RATES IN ART FOR REGION OUTSIDE THE REFLECTOR

 

 

Region Heat Deposition Rate
Beryllium
2 in. from Be-Na interface 4.639 x 104 Btu/hr -ft3
1 in. from Be-Na interface 6.282 x ]04 Btu/hr -h3
At Be-Na interface 9.568 x 104 Bru/hr « it
. 4 3
Sodium 3.479 x 10" Btu/hrft
Inner Incone! shell 5.5467 x 105 Btu/hr -fts
Boron carbide
Volyme 1.894 x 107 Btushr - 5
Flux at reflector-side surface 4.154 x ]04 Btu/hr -f12
Outer Inconel shell 4,842 x 105 Btu/hr -ft3

 

175
ANP PROJECT PROGRESS REPORT

JEMPERATURE STRUCTURE IN AN
IDEALIZED ART CORE

H. F. Poppendiek L. D. Palmer

Reactor Experimental Engineering Division

An analysis was-made of the temperature struc-
ture in an idealized ART core. From the activity
data obtained in the hot-critical experiment,® it
was possible to obtain a simplified radial volume-
heat-source distribution. This source distribution
was substituted into the heat-transter equations,
which were then evaluated for the uncooled-wall
case. The resulting uncooled-wall temperature
difference above the fluid temperature was more
than twice as great as the corresponding tempera-
ture difference in a uniform volume-heat-source
system. This radial temperature information was
then used in a general cooling analysis of the
idealized ART core and cooling annuli such as
that outlined previously.?  The resulting axial
sodium, core wall, and fuel temperature distribu-
tions are shown in Fig. 7.7; the wall cooling by
the sodium coolant in the annuli was approxi-
mately 2.5 Mw,

ART CORE HEAT-TRANSFER EXPERIMENT

N. D, Greene G. W. Greene
H. F. Poppendiek

Reactor Experimental Engineering Division

G. L. Muller
Pratt & Whitney Aircraft

Fabrication of the half-scale model of the ART
core, together with its associated components,
was completed. The assembly, which will be used
for volume-heat-source heat-transfer experiments,
has been operated with water (Fig. 7.8). Prelimi-
nary measurements indicate that a mean Reynolds

number of 100,000 will be achievable with the

existing pumps.

All recording instrumentation for the transient
and steady-state thermocouples, which are located
near the wall-fluid interfaces, as well as below
the wall surfaces of the core and island, has been
developed and installed, Operating procedures and
schedules are now being established, and explora-

 

3A. D. Callihan et al., ANP Duar, Prog. Rep. Dec. 10,
1955, ORNL.-2012, p 71.

4H. F. Poppendiek and L. D. Palmer, Application of
Temperature Solutions for Forced Convection Systems
with Volume Heat Sources to General Convection

Problems, ORNL-1933 (Sept. 29, 1955).

176

——

ORNL~LR-DWG 12606

 

Re=94,000
Pr=273

1600

 

1500

 

1400

TEMPERATURE (°F)

 

1300

1200

 

 

14G0

 

 

 

 

 

1000

 

DISTANCE FROM CORE INLET {ft)

Figo 7-7-
AJ!T-(:Oreo

Temperature Structure in an ldealized

tory, low-power runs will be made within a few
weeks.,

HEAT CAPACITY
W. D, Powers

Reactor Experimental Engineering Division

The heat capacities and enthalpies of four spe-
cific NaF-ZrF, mixtures were measured in the
liquid and solid states in order to determine
whether there was an appreciable variation in the
heat capacity with composition. No major change
in the heat capacity or in the product of the heat
capacity and density was found over the compo-
sition range studied: 53 to 65 mole % NaF and
35 to 47 mole % ZrF,. The following equations
represent the data obtained.

NaF-ZrF , (65-35 mole %)
Solid (120 to 465°C)
PERIOD ENDING MARCH 10, 1956

o
o
vy
o™~
O
-
o
X
0.

 

 

1on,

-

ART Volume-Heat-Source Test Sect

ig. 7.8.

F

177
ANP PROJECT PROGRESS REPORT

H, — Hygoo =—4.1 +0.1835T + (6.12 x 10=3)T2
-, b
c, =0.1835 +(12.23 x 10=3)T
Liquid (605 to 890°C)
Hop = Hygoo ==1.7 +0.3322T ~ (3.87 x 10~%)72

<, =0.3322 — (7.73 x 10~3)T

NaF-ZrF , (61-39 mole %)
Solid (120 to 470°C)
Hyp = Hysoc ==2.7 +0.1648T +(8.30 x 10~3)7172
c, =0.1648 + (16.61 x 10-3)T
Liquid (550 to 900°C)
Hp = Hycoe =16.8 +0.28357 ~ (0.84 x 107%)7?
¢, =0.2835 - (1.67 x 10-9)T

NaF-ZrF, (57-39 mole %)
Solid (120 to 495°C)
Hp = Hygoo ==6.6 +0.1914T +(2.97 x 10~3)12
c, =0.1914 + (5.93 x 10-3)T
Liquid (550 to 900°C)
Hp — Hygoc =60.2 +0.15307 + (8.48 x 10~5)T2
¢, =0.1530 + (16.95 x 10=3)T

NaF-ZrF , (53-47 mole %)
Solid {120 to 515°C)
Hop = Hygoo =-5.2 +0.1847T +(2.38 x 10=%)T2
c, =0.1847 + (4.76 x 10=5)T
Liquid (560 to 905°C)
Hp = Hygoe =555 +0.1884T + (4.65 x 10~3)72
c, =0.1884 + (9.30 x 10=3)T

In these expressions,

H = enthalpy in cal/g,

c, = heat capacity in cal/g . °C,

T = temperature in °C,

Table 7.2 gives a comparison of the heat ca-
pacities in the liquid state in the following units:
cal/g:°C, cal/g-atom-°C, and cal/cm3.°C. The
volumetric heat capacities and the heat capacities
on a per-gram basis have been found to be nearly
constant over a wide range of compositions,

In a report® published recently the heat capac-
ities of 17 fluoride mixtures are listed and com-
pared. Equations were formulated so that the
enthalpy and heat capacity of solid and liquid
fluoride mixtures could be predicted from the
composition,

YISCOSITY
S. |, Cohen

Reactor Experimental Engineering Division

A study was initiated to determine the effect of
fission products on the viscosities of fuel mix-
tures, The mixture used to simulate the ART fuel
after exposure to reactor conditions was NaF-
ZrF ,-UF , (50-46-4 mole %) with 11 g of RbF, 17 g
of BaF,, and 52 g of TaF; added per kilogram of
pure fuel, A sample of this fuel was prepared by
dry mixing of the constituents, Measurements were
made on the mixture and on a sample of the pure
fuel from the same batch to furnish a control.
However, the solubilities of the additives were
apparently poor, and it was not possible to obtain
a solution of the various components. Hence, the
data obtained are felt to be invalid. A second
sample of this proposed mixture is being prepared
that will be hydrofluorinated after the additives

 

5W. D. Powers and G. C. Blalock, Enthalpies and Heat
Capacities of Solid and Molten Fluoride Mixtures, ORNL-
1956 (Jan. 11, 1956).

TABLE 7.2. HEAT CAPACITIES IN THE Nch-ZrF4 SYSTEM

 

Composition (mole %)

 

Heat Capacity

 

 

NaF ZrF 4 cal/g «°C cal/g satem -°C cal/em®2C
65 35 0.278 7.82 0.812
61 39 0.272 7.79 0.809
57 43 0.272 7.9 0.824
53 47 0.253 7.50 0.776

 

178
have been introduced, The resulting mixture will
be analyzed, and a second attempt to obtain vis-
cosity data will be made.

An important observation was made during this
study: the viscosity measurements on the pure
fuel used as a control yielded values which were
about 15% lower than those obtained previously
for this mixture.® Consequently, measurements
were made on a second sample of fuel from the
batch used previously. The results obtained were
in satisfactory agreement with those previously
reported. Yisual comparison of the melts from the
old and new batches of fuel indicated that the new
material was of much higher purity. These ob-
servations clearly show the influence of fuel purity
on the viscosity,

Data on three other fluoride mixtures were also
obtained, The results are presented in Table 7.3.
The NaF-ZrF, data were obtained with the modi-
fied Brookfield and capillary viscometers, and the
results were in satisfactory agreement with capil-
lary viscometer results determined over a year
ago. The two beryllium-bearing salts were studied
in the beryllium dry box with the capillary vis-
cometers,

 

65, |. Cohen and T. N. Jones, Measurement of the
Viscosity of Composition 30, ORNL CF-55-3-62 (March
2, 1955).

PERIOD ENDING MARCH 10, 1956

PERFORMANCE COMPARISONS OF
SEVERAL FLUORIDE FUELS

H. F. Poppendiek
Reactor Experimental Engineering Division

There are many ways in which the performance
or effectiveness of fuels in a circulating-fuel re-
actor can be compared. In the comparisons pre-
sented here it is presumed that the reactor power,
fuel inlet and outlet temperatures, and reactor core
and heat exchanger geometries are given and fixed.
Also, it is considered that the core Reynoids num-
ber is about 100,000 and that the heat exchanger
Reynolds number is above the transition region,
Under these circumstances it is possible to ex-
press the heat and momentum transfer in the reactor
core and the heat exchanger in terms of simple
relations involving only the thermal properties,
namely, heat capacity, thermal conductivity, vis-
cosity, and density. These relations have been
evaluated for several of the more important ANP
fuels, and the data are presented in Table 7.4.
All quantities were normalized, that is, all numbers
in a column have been compared with the lowest,
which has been called unity., It is noted that the
major differences between the fuels occur in the
heat exchanger. The last two mixtures, which
may be called alkali-metal-base fuels, have radial
fuel temperature differences that are from 20 to
40% lower than those in the zirconium-base fuels,
Also, the heat exchanger pressure drops for the
alkali-metal-base fuels are only about’ half as
great as those for the zirconium-base fuels,

TABLE 7.3, FLUORIDE VISCOSITY DATA

 

 

Composition Temperature Absolute Viscosity Kinematic Viscosity B/T(°K)

(mole %) (°C) {centipoises) (centistokes) = Ae Reference
NcJF-ZrF4 550 11.3 3.45 p o= 0.07294]60/T -
(50-50) 850 2.9 0.96
Nc:F-LiF-BeF2 600 55.65 2.65 po= 0.195e2930/T o
(49-36-15) 800 2.95 1.44
NaF-LiF-Ber-UF4 575 8.4 3.45 p= 0.07364010/'1‘ s
(56-21-20-3) 850 2.7 1.18

 

*S. 1. Cohen and T. N. Jones, Measurement of the Viscosity of Composition 31, ORNL CF.55-12-128 (Dec. 23, 1955).
**S |, Cohen and T. N. Jones, Measurement of the Viscosity of Compositions 97 and 98, ORNL CF-55-12-127 (Dec.

23, 1955).

179
ANP PROJECT PROGRESS REPORT

TABLE 7.4, PERFORMANCE COMPARISONS OF SEVERAL FLUORIDES

 

Composition

Comparison Ratios™

 

 

 

 

 

Reactor Heat Exchanger
(mole %)
dq
At — Ay Ap
om dA T
c
NaF-ZrF ,-UF 1.16 1.15 1.25 2.04
{50-46-4)
LiF-NaF.ZrF ,-UF , 1.36 1.18 1.45 1.53
(55-20-21-4)
RbF-ZrFA-UF4 1.06 1.00 1.33 2.74
(48-48.4)
LiF-NaF-KF-UF 4 1.00 1.25 1.00 1.17
(45.3-11.2-41-2.5)
NaF-BeFZ-UFJ‘ 1.06 1.26 1.07 1.00
(70-28-2)
*Atom = radial fuel temperature difference in the core for an uncooled wall,
dq
— | = wallecooling flux required to lower the wall temperature to the fluid temperature,
dA c
Atr = radial fuel temperature difference in the heat exchanger,
Ap = the fuel pressure drop through the heat exchanger.

180
PERIOD ENDING MARCH 10, 1956

8. RADIATION DAMAGE

G. W, Keilholtz
Solid State Division

REVIEW OF TESTS OF CORROSION
OF INCONEL AND STABILITY OF FUEL
UNDER IRRADIATION

G. W. Keilholtz
Solid State Division

The effects of irradiation on the corrosion of
Inconel exposed to a fluoride fuel mixture and on
the physical and chemical stability of the fuel
mixture have been investigated by irradiating
Inconel capsules filled with static fuel in the
MTR and by operating in-pile forced-circulation
Inconel loops in the LITR and in the MTR. The
relatively simple capsule tests have been used
extensively for the evaluation of new materials.
The principal variables in these tests have been
flux, fission power, time, and temperature. In a
fixed neutron flux the fission power is varied by
adjusting the uranium content of the fuel mixture,
Thermal-neutron fluxes ranging from 101} to 104
neutrons/cm?.sec and fission-power levels of 80
to 8000 w/cm3 have been used in these tests,
Almost all the capsules have been irradiated for
300 hr, but in some of the recent tests the irradi-
ation period was 600 to 800 hr. After irradiation
the effects on the fuel mixture are studied by
measuring the pressure of the evolved gas, by
determining the melting point of the fuel mixture,
and by making petrographic and chemical analyses.
The Inconel capsule is also examined for corrosion
by standard metallographic techniques.

In the many capsule tests made to date, no major
changes that can be attributed to irradiation, other
than the normal bumup of the uranium, have
occurred in the fuel mixtures. However, the
analytical method for the determination of chromium
in the irradiated fuel mixture is being rechecked
for accuracy, The metaliographic examinations of
Inconel capsules tested at 1500°F for 300 hr have
shown the corrosion to be comparable to the
corrosion found under similar conditions in unir-
radiated capsules, that is, penetration to a depth
of less than 4 mils. In capsules that briefly
reached temperatures of 2000°F and above, there
was penetration to a depth of more than 12 mils
and grain growth,

Three types of forced-circulation in-pile loops

have been tested. A large loop was operated in
a horizontal beam hole of the LITR, The pump
for circulating the fuel in this loop was placed
outside the reactor shield. A smaller loop, in-
cluding the pump, was opetated in a vertical
position in the lattice of the LITR, A third loop
was operated completely within o beam hole of
the MTR. The operating conditions for these loops
are presented in Table 8.1, and results of chemical
analyses of the fuel mixtures circulated are given
in Table 8.2.

The LITR horizontal loop operated for 645 hr,
including 475 hr at full reactor power. The loop
generated 2.8 kw, with a maximum fission power
of 400 w/em3. The Reynolds number of the
circulated fuel was 5000, and there was a temper-
ature differential in the fuel system of 30°F. The
volume of the loop was large, and therefore there
was a large dilution factor.  Metallographic
analyses showed less than 1 mil of corrosion of
the Inconel walls of the loop., Chemical analyses
showed the irradiated fuel mixture to contain
180 ppm Fe, 150 ppm Cr, and 30 ppm Ni. Therefore
there was no evidence of accelerated corrosion in
this experiment,

The LITR vertical loop operated for 130 hr, with
only 30 hr of the total operating period being at
full reactor power. The loop generated 5 kw, with
a maximum fission power of 500 w/cm3, The
Reynolds number of the fuel was 3000, and the
temperature differential was 71°F. The surface-
to-volume ratio was 20, and the dilution factor
was 10. Chemical analyses of the irradiated fuel
showed 370 ppm Fe, 100 ppm Cr, and 50 ppm Ni.
Metallographic analysis of this loop showed less
than 1 mil corrosion at the curved tip.

The horizontal loop inserted in the MTR (MTR
in-pile loop No. 3), described previously,! operated
for 467 hr, including 271 hr at power. The loop
generated 20 kw, with a maximum power of 800
w/em3, The Reynolds number of the fuel was
5000, and the temperature differential was 155°F
for 103 hr and 100°F for 168 hr. The dilution

factor was about 5,

 

]D. B. Trauger et al., ANP Quar, Prog. Rep. Dec. 10,
1955, ORNL-2012, p 27.

181
ANP PROJECT PROGRESS

REPORT

 

 

 

 

 

 

TABLE 8.1. OPERATING CONDITIONS FOR INCONEL FORCED-CIRCULATION IN-PILE LOOPS
i . LITR Horizontal LITR Vertical MTR In-Pile
Operating Yariables
Loop Loop loop No. 3
Fuel composition (mole %) NaF-ZrFA-UF4; NqF-ZrF4-UF4; NqF-ZrF4-UF4;
62.5-12.5-25 63-25-12 53.5-40-6.5
Maximum fission power, w/c:m3 400 500 800
Total power, kw 2.8 5.0 20
Dilution factor 180 10 5
Maximum fuel temperature, °F 1500 1500 1500
Maximum temperature differential, °F 30 71 155
Reynolds number of fuel 5000 3000 5000
Operating time, hr 645 130 467
Time at full power, hr 475 30 271
Depth of corrosion ottack, mil <1 < <1
TABLE 8.2. CHEMICAL ANALYSES OF FUEL MIXTURES CIRCULATED
IN INCONEL FORCED-CIRCULATION IN-PILE LOOPS
Minor Constituents (ppm)
Loop Designation Sample Taken
lton Chromium Nickel
LITR horizontal lcop Before filling 80 ¥ 10 1015 200 * 100
After draining 180 T 40 150 £ 10 30 t5
LITR vertical loop Before filling 90 £ 10 80 £ 10 145 T 20
After draining 370 * 20 100 * 20 50 10
MTR in-pile loop No. 3 Before filling 40 £10 60 £ 10 40 T 10
After draining 240 1 20 50 = 10 100 £ 20

 

Chemical analyses of fuel from MTR in-pile loop
No. 3 showed 240 ppm Fe, 50 ppm Cr, and 100
ppm Ni. The high iron concentration probably
resulted from a sampling difficulty. Examinations
of unetched metallographic sections of the Inconel
tubing showed no corrosion penetration. Etched
sections showed no attack that was to a depth
of more than 1 mil. A slight amount of inter-
granular void formation was noted, but this was

neither dense nor deep. Measurements of wall

thickness showed no variations attributable to
cotrosion, The loop was examined carefully for
effects of temperature variations between the

182

inside and outside walls of tubing at bends, but
no effects of overheating were observed. Samples
from the inlet, the center, and the exit side of the
loop showed less than 1 mil of corrosion. The
low corrosion is credited to careful temperature
control of the salt-metal interface and to the
maximum wall temperature being below 1500°F at
all times.

Future loops will be operated at higher fission
powers and therefore greater temperature differ-
entials, The dilution factors will be kept low.
New fuels aond new alloys are being considered
for testing in future loops.
EXAMINATION OF THE DISASSEMBLED
MTR IN-PILE LOOP NO, 3

M. J. Feldman

C. Ellis R. N. Ramsey
E. J. Manthos A. E, Richt
W. B. Parsely E. D, Sims

Solid State Division

The sectioning of MTR in-pile loop No. 3 was
completed in the General Electric Company hot-
laboratory facilities at NRTS, and three carriers
containing the different sections of the loop were
received at ORNL. The first carrier received
contained the nose coil and two 14-in.-long pairs
of fuel tubes from behind the nose. The cobalt
foil, thermocouples, Calrods, and insulation were
removed from the nose coil in the ORNL hot-

. PERIOD ENDING MARCH 10, 1956

laboratory facilities. Five of the original six
cobaToils were found and will be analyzed. The
stripped nose coil and part of the heat exchanger
assembly are shown in Fig. 8.1. The nose coil
was sectioned with a hack saw, and the fuel was
melted out of the coil and the two pairs of 14-in.-
long fuel tubes in an argon atmosphere at a
temperature of approximately 650°C. Thirty-five
metallographic samples were obtained from the
nose coil, heat exchanger, bellows, and fuel tubes.
The locations of the specimens cut from the nose
coil are shown in Fig. 8.2, and the locations of
the other specimens are described in Table 8.3.
The eleven specimens cut from the nose coil
were examined metallographically in the as-
polished unetched condition, and no evidence of
attack was found. Wall-thickness measurements

TABLE 8,3, NUMBER DESIGNATIONS AND LOCATIONS OF METALLOGRAPHIC SPECIMENS
CUT FROM MTR IN-PILE LOOP NO. 3

 

Specimen No,

L ocation of Specimen

 

280

281

283

284

285

286

287

288

289

290

291 Radial section
295 Radial section on
296 Radial section on
297 Radial section at
303 Radial section on
304 Radial section on
305 Radial section on
306 Radial section on
307 Radial section on
308 Radial section on
309 Radial section on

Section of weld from heat exchanger U-joint

Transverse section of bellows on heat exchanger from outlet side
Transverse section of bellows on heat exchanger from inlet side
Radial section at location of thermocouples 5, 6, and 7 on outlet side
Transverse section of heat exchanger weld on outlet side

Transverse section of end of helium sniffer line on outlet side

Radial section at location of thermoccuple 21

Same as specimen 284 except on inlet side

Same as 285 except on inlet side

Radial section ot location of thermocouple 18

on cutlet side ]0% in, from end of helium sniffer

outlet side 151/2 in. from end of helium sniffer

outlet side 20 in. from end of helium sniffer

ocation of thermocouple 23 on outlet side

outlet side 30 in. from end of helium sniffer

inlet side 7 in. from end of helium sniffer

inlet side 12 in. from end of helium sniffer
inlet side 17 in. from end of helium sniffer
inlet side 271 in. from end of helium sniffer
inlet side 25 in. from end of helium sniffer

inlet side 30 in. from end of helium sniffer

 

183
ANP PROJECT PROGRESS REPORT

ey
PHOTO 16072

 

Fig, 8.1, Stripped Nose Coil and Part of Heat
Exchanger Assembly of MTR In-Pile Loop No, 3,

were taken on four of the samples from the coil
and on the original tubing. Statistical variations
in the wall thickness of the original tubing and
variations resulting from bending of the tubing
were established, Measurements taoken on the
four samples from the irradiated nose coil fell

184

UNCLASSIFIED
ORNL~LR-DWG 13033

 

    
 

264
1455 OF
263 261
262
269
1435 oF
270
1490 °F 271
273
1415 °F — 1390-1490 °F
274
¢ OUTLET
Fig, 8.2, Locations on Nose Coil from Which

Metallographic Specimens were Taken. The number
designations of the samples and the operating
temperatures at various points are shown,

within the established variations of the unir-
radiated coil. The results of attempts to etch
these specimens chemically and electrolytically
have not been satisfactory. The etched specimens
show what appears to be moderate intergranular
attack to a depth of | mil, but this may be an
etching effect rather than true corrosive attack.
Specimens 264 and 271 are shown in Figs. 8.3
through 8.6 in the polished and etched conditions.
Two etching techniques are being studied that
show promise of providing more definable micro-
structures,

Specimens from the pump and the two straight
sections of fuel tubes will be examined as soon
PERIOD ENDING MARCH 10, 1956

 

Fig. 8.3. As-Polished Specimen 264 from Nose Coil of MTR In-Pile Loop No. 3. 50X.

 

Fig. 8.4, Etched Specimen 264 from Nose Coil

UNCLASSIFIED
S RMG 311

of MTR In-Pile Loop No, 3. 250X.

185
ANP PROJECT PROGRESS REPORT

 

Fig, 8.5. As-Polished Specimen 271 from Nose of MTR In-Pile Loop No, 3, 50X.

 

Fig, 8,6, Etched Specimen 271 from Nose Coil of MTR In-Pile Loop No. 3, 250X.

186
as two cells can be decontaminated for cell-
equipment servicing. The fuel from the two
sections of fuel tubing will be analyzed chemically,
and specimens of the tubing will be examined
metallographically.

EFFECTS OF RADIATION ON THE MECHANICAL
PROPERTIES OF STRUCTURAL MATERIALS

J. C., Wilson
Solid State Division

Stress-Corrosion Tests

W. W, Davis J. C, Zukas
J. C. Wilson
Solid State Division

Eight Inconel, helium-pressurized, tube-burst,
stress-corrosion specimens were exposed to radi-
ation in hole HB-3 of the LITR in a helium atmos-
phere. The circumferential stress on each
specimen was 2000 psi, ond various test temper-
atures were used, Similar specimens are being
tested out-of-pile to obtain control data.

A plot of the log of the time-to-rupture vs the

operating temperature during irradiation, Fig. 8.7,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

—_——
ORNL-LR-DWG {1772
T T T
0.040-in. WALL, 0.191-in. ID 3
|
1600 _ INCONEL TUBING IN LITR ! . | | |
STRESSED BY INTERNAL GAS \ |
PRESSURE TO A CIRCUMFERENTIAL
L STRESS OF 2000 psi { }
~ o |
L : ] |
2. 1550 OO ; 1 l
uJ | ' :
o«
5 |
= i
g
o
L .
o
E I
© :
1500 : A A4 A —
N
w
o
1430 o ; L ‘ o
| L
10 20 50 100 200 500 1000 2000

TIME TO RUPTURE (hr)

Fig., 8,7, Results of Tube-Burst Stress-Corrosion
Tests of Internally Pressurized Inconel Tubing
Exposed to Radiation in the LITR at a Stress
Level of 2000 psi in a Helium Atmosphere,

PERIOD ENDING MARCH 10, 1956

revealed an unusual degree of scatter in the data
and led to an investigation of the condition of the
specimen tubing prior to testing. Extensive non-
destructive tests are being conducted on tubing
for tube-burst experiments. All but four of 28
specimen tubes examined by Zyglo inspection
were rejected because of pinholes and spongy
regions. Metallographic inspection of 25 sections
from the 24 rejected tubes resulted in the location
of only one appreciable defect, Fig. 8.8, on the
inside of the tubing. The types of defects that
show up in nondestructive testing may have been
partially responsible for the unusual scatter of
the data obtained in the experiments described
above. Several specimens that pass the non-
destructive tests will be irradiated, after the
out-of-pile stress-corrosion tests are completed,
to study the effect of irradiation on rupture time
in these inspected tubes. It is believed that the
data will aid in estimating the seriousness of
defects in the tubing in terms of creep life. The
better quality tubing stock obtained for radiator
fabrication will probably be used in subsequent
experiments if the radiation effects indicate the
need for more accurate determinations. The typical
structure near the fracture of one of the in-pile
tube-burst specimens is shown in Fig, 8.9. This
specimen ruptured after 84 hr at 1500°F. The
structure near the fracture of the only out-of-pile
specimen available is shown in Fig. 8.10. This
specimen fractured after 356 hr at 1500°F.

A cantilever stress-corrosion rig similor to that
used in the LITR? has been in operation out-of-pile
for over 250 hr. Two earlier out-of-pile tests were
not completed because of a weld failure in one
case and a thermocouple failure in the other.
Static sodium is used as a heat-transfer medium
on the outside of the Inconel tube being tested in
this experiment. The Inconel tube is filled with
the molten fuel mixture NOF-ZFF4-UF4 (50-46-4
mole %). Preparations have been made for opening
the similar apparatus irradiated previously in the
LITR and for examining the Inconel specimen
metallographically, Two more rigs of this design
have been filled with fuel and will be operated in
the future.

Metallographic examination of a similar canti-
lever apparatus that had been operated out-of-pile
with helium on both sides of the tubes (rather than

 

2W. W. Davis et al., ANP Quar. Prog. Rep. Sept. 10,
1954, ORNL-1771, p 142.

187
ANP PROJECT PROGRESS REPORT

 

UNCLASSIFIED
RMG 1314

Fig. 8.8. Defect Found by Metallographic Examination on inside of Inconel Tubing from Stock Used

in Tube-Burst Specimens,

sodium or fuel) showed that some surface oxidation
had taken place. There was some oxidation pene-
tration into the grain boundaries, but this was
only noticeable on the tension sides of the
specimens.

Two water-cooled, finned, tube-burst rigs for
stress-corrosion testing of Inconel tubing, de-
scribed previously,? have been filled with the fuel
mixture NqF-ZrF4-UF4 (63-25-12 mole %). One
of these will be operated in the LITF to check

heat-transfer design calculations.

Alternate Stress-Corrosion Apparatus

C. D. Baumann W. E. Brundage
Solid State Division

A design of an alternate apparatus for stress-
corrosion testing was described previously.4 With
this apparatus the maximum temperature of the fuel

 

3J. C. Wilson et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL.-1947, p 165.

4w, E. Brundage and C. D, Baumann, ANP Quar. Prog.
Rep. Dec. 10, 1955, ORNL.2012, p 184.

188

and of the specimen would be limited to 1500°F,

The preliminary design was based on calcu-
lations of the expected fission power, heat flow,
and the limitations of fabrication. Since the
thermal-neutron flux in hole HB-3 of the LITR was
not known accurately, a mockup of the fuel and
container was constructed, as shown in Fig. 8.11,
and was irradiated for a short time at a position
about 2 in. from the inner end of the hole. A
cadmium-magnesium alloy (9 wt % Cd, 91 wt % Mg)
was used to simulate the cross section of the fuel,
and cobalt foils were used as the monitors., After
irradiation the foils were removed, and the activity
was measured in a 100% geometry counter, The
measurements indicated a flux of 9 x 1012
neutrons/cm2.sec at the outer surface of the fuel
and 5 x 1012 neutrons/cm2.sec at the inner surface
of the annulus. Calculations based on these
measurements indicate an unperturbed thermal-
neutron flux at this position of 2 x 1013 npeu-
trons/cm?.sec,

The flux measurements permitted a refinement of
the design, and the equipment now being fabricated
PERIOD ENDING MARCH 10, 1956

 

 

 

 

Fig. 8.9. Structure Near Fracture in Inconel Tube-Burst Specimen That Was Internally Pressurized
with Helium to a Fiber Stress of 2000 psi and Ruptured After 84 hr at 1500°F in Hole HB-3 of the LITR.
300X.

 

Fig. 8.10, Structure Near Fracture in Inconel Qut-of-Pile Tube-Burst Specimen That Was Internally
Pressurized with Helium to a Fiber Stress of 2000 psi and Ruptured After 356 hr at 1500°F, 300X.

189
ANP PROJECT PROGRESS REPORT

ALUMINUM

 

(LTI T ’

UNCLASSIFIED
ORNL-LR-DWG 11773

POCKETS FOR
COBALT FOIL

 

 

 

T A T NTTT

 

i

 

- O.2821in. ,-{
—

 

~— 0.360in.

 

 

i \\\ \\ \\\\\\\\\\\\\\\\ \\\\\\ \\\\\\\ \\ \\\

‘-4—*4OOOIH —-———"1

 

 

 

Figo 8-]].

9% Cd-Mg ALLOY

Design of Capsule Used for Measuring Thermal-Neutron Flux in Hole HB-3 of LITR and

the Flux Attenuation To Be Expected in the Stress-Corrosion Apparatus,

is shown in Fig. 8.12. The platinum heater
surrounding the fuel container will be used to
supplement the fission heating to maintain the
operating temperature,
the equipment will act as a safety container in
case of a leak or rupture of the specimen. Power
generation in the fuel will be about 750 w/cm3.
Postirradiation measurements will be made on the
outside of the specimen tube, and the specimen
will be examined metallographically for corrosion.
It is expected that this apparatus can be used in
the MTR with little modification.

The outer container for

MTR Tensile Creep Tests

W. W. Davis N. E. Hinkle
J. C. Wilson
Solid State Division

The MTR tensile-creep-test apparatus in which
two lnconel tubes were stressed in tension to
1500 psi at 1500°F (in helium) was irradiated in
hole HB-3 of the MTR for 33 days at an average
power level of 30 Mw, Postirradiation sectioning
revealed that both specimens had fractured. Tem-
perature measurements taken during the test
showed an abnormal temperature distribution after
518 hr for specimen 1, which failed outside the
gage length, and after 188 hr for specimen 2,
which failed within the gage length. Throughout
the high-temperature region, both specimens had

dark surface films. The absence of metallic

190

reflection from the fracture surfaces indicated that
the specimens had not fractured during disassembly
of the apparatus. Metallographic examinations
indicated that attack by an unknown contaminant,
alone or in combination with temperature ex-
cursions, was responsible for the short rupture
life. The corresponding out-of-pile control test
has been delayed by leaks in the apparatus.

Metallographic examination of irradiated specimen
2 revealed an intergranular fracture and necking
at the point of failure. Photomicrographs of
longitudinal sections showed considerable internal
grain-boundary separation. The incidence of grain-
boundary separation appeared to be greatest ad-
jacent to the fracture and to decrease with
increasing distance from the fracture, Surface
grain-boundary separation was observed at the
inside surface to a maximum depth of 4 mils in o
few places.

Photomicrographs of sections adjacent to the
fracture show the presence of an unidentified film
on the specimen surfaces. The film on the inside
surface was thicker than that on the outside
surface of the specimen tube. Fracture surfaces
and surface grain-boundary separations exhibited
the film material, but internal grain-boundary sepa-
raticns did not,

Sulfur attack, because of its well-known effects
on nickel alloys, was thought to be a possible
cause of the surface film and the intergranular
UNCLASSIFIED
ORNL-LR-DWG 12000

 

    
  

  
   
    

FILLING TUBE

1
|
GAS TUBE ——-| m E

 

 

 

 

 

O {

 

INCHES

oUT 1 p N

R

Fig., 8,12, Alternate Stress-Corrosion Apparatus,

attack observed. It is important that the source
of the contaminant be found so that it can be
eliminated in future tests, No indications of sulfur
were found in the insulation for the wiring or the
furnace. An attempt to reproduce the surface film
on the irradiated specimens was made by heating
a similar Inconel tube in contact with sulfur for
a short time at a temperature near 1500°F. Obser-
vation of the region contacted by the sulfur
revealed a large amount of grain-boundary pene-
tration compared with the limited amount of surface
film and surface grain-boundary separation in the
irradiated specimen. Sulfur printing techniques
were applied to the control sample without
SUCCess.

The Inconel tubing used as the specimen stock
was subjected to ultrasonic and eddy-current tests,
and indications of defects were obtained. Samples

PERIOD ENDING MARCH 10, 1956

were cut from the regions classed as most de-
fective, but metallographic examination did not
confirm the defects.
mental results of the tests will be continued.

The analyses of the experi-

Pneumatic Stressing Device

A. S. Olson J. C. Wilson
Solid State Division

A pneumatic stressing device that will provide
a powerful, compact means of stressing creep
specimens with loads as great as 10,000 Ib has
been designed and is being constructed. The
device can be used for both laboratery and in-pile
tests of creep, stress corrosion, and stress relax-
ation. Work continues on the design of a suitable
recording creep extensometer.

Ductility of Nickel Alloys

J. C. Wilson
Solid State Division

T. C. Price
Pratt & Whitney Aircraft

The effect of temperature on the ductility of
nickel alloys in the temperature range 800 fto
1400°F is being studied, with the effect of strain
rate vs temperature being measured first, To date,
fairly complete data have been obtained for Monel,
and work on Nichrome and Inconel is under way.

The gage-length portions of the Monel tension
specimens were 0.1500 * 0.0005 in. in diameter
and 0.060 * 0.05 in. in length. Strain rates of
2 and 0.002 in./min were used, and the test
temperatures ranged from 800 to 1400°F. The
ultimate and yield strengths at both strain rates
appeared to vary linearly with temperature. There
was, however, a large nonlinear decregse in
elongation and percentage reduction of area in the
region 900 to 1000°F at the faster strain rate,
which was not observable at the slower strain
rate. These results are illustrated in Figs, 8.13

and 8.14.

At the slower strain rate the ductility values
were lower than those found at the faster strain
rates. At about 1000°F the ductility values nearly
coincide for both strain rates, and therefore the
processes responsible appeared to be independent
of strain rate at this temperature. At higher and
lower temperatures the curves tend to diverge,
with greater ductility being found at the faster
strain rate,

191
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR-DWG {14685

l \

STRAIN RATE 2.0 in./min
%, ELONGATION

% REDUCTION OF AREA |

AT FRACTURE
% REDUCTION OF AREA "
0.50 in. FROM FRACTURE

 

70

 

 

    

 

 

 

 

 

 

 

ELONGATION (%) OR REDUCTION OF AREA (%)

 

 

 

o ; 1 J

 

 

 

 

 

 

 

800 90¢ 1000 400 1200 41300 (400 4500
TEMPERATURE (°F)

Fig, 8,13, Effect of Temperature on Tensile
Ductility in Monel at a Strain Rate of 2 in,/min,

EXPERIMENTAL STUDIES OF REACTOR
MATERIALS AND COMPONENTS

W. E. Browning
Solid State Division

Effect of Radiation on Static Corrosion
of Structural Materials by Fused Salts

W. E. Browning H. L. Hemphiil
Solid State Division

Irradiations in the MTR of Inconel capsules
containing various fused-salt fuels have been con-
tinved in order to study the effect on the Inconel
of fissioning in the fuel. Since earlier tests® had
shown that irradiation for two weeks in the MTR
had no significant effect on corrosion, the irradi-
ation periods used for the capsules tested during
the past year have been of six or nine weeks
duration, Seven capsules filled with Nc:F-ZrF4
plus 2 mole % UF4, 2 mole % UF_, or 4 mole %
UF4 were irradiated at 1500°F :}or the longer
periods and are now being sampled for metallo-
graphic and chemical analyses, Additional Inconel
capsules filled with NaF-ZrF -UF, (50-46-4

mole %) are being prepared for irradiation in the

 

W. E. Browning, G. W, Keilholtz, and H, L. Hemphill,
ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 120.

192

UNCLASSIFIED
ORNL-LR-DWG {1686
70 T ‘ [
‘ w
I

 

i |
STRAIN RATE 0.002 in./min

60 O % ELONGATION
® % REDUCTION OF AREA AT
FRACTURE
S50 ———— 1 & % REDUCTION OF AREA

 

0.5Cin. FROM FRACTURE

 

 

 

40 T Ti |

30' 4 \ ‘

 

 

 

 

20

 

 

ELONGATION (%) OR REDUCTION OF AREA (%)

 

 

 

 

 

 

 

 

800 200 1000 HO0 1200 1300 400 1500
TEMPERATURE {°F}

Fig., 8.14. Effect of Temperature on Tensile
Ductility inMonel at a Strain Rate of 0,002 in,/min,

MTR at 1800°F. A duplicate capsule is being
tested out-of-pile for comparison, A multicapsule
facility has been prepared for exposures in a
high-flux position in the MTR.

The temperature-control system, de-
scribed previously,® has performed satisfactorily
for more than a year. Three identical units are
now in service, Various unforseen disturbances
have been handled, including one in which drops
of water were sprayed intermittently on a capsule
operating at 1500°F, without the occurrence of
excessive temperatures.

The improved thermocouples, also described pre-
viously,® have also served satisfactorily. Only
one of the 25 improved thermocouples that have
been used in the MTR has failed. The one that
failed was one of three that were sprayed with
water while hot after six weeks of irradiation; the
other two did not fail.

Mockup tests have been run on Hastelloy B
capsules under simulated MTR conditions to de-
termine whether oxide scaling on the outside can

revised

be prevented in order to control contamination of
the facility., Chrome-nickel plating of the outside

 

‘w. E. Browning, G. W. Keilholtz, and H. L. Hemphill,
ANP Quar. Prog. Rep. March 10, 1955, ORNL-1864,
p 146.
of the capsules gave an oxide coating in air at
high temperatures that was adherent during thermal
cycling. The coating was similar to that formed
on Inconel in that it did not scale, and it appears
to be promising for use with the present capsule
designs,

Holdup of Fission Gases by Charcoal Traps

W. E. Browning
Solid State Division

C. C. Bolta
Pratt & Whitney Aircraft

The apparatus described previously? is being
used for studying the various factors that affect
the performance of charcoal traps for the holdup
of fission gases. Both nitrogen and helium are
being tested as purge gases, and radiokrypton is
being used to simulate the fission gases. The
trap is a 13-in.-long, 2-in.-dia, sched-40 stainless
steel pipe filled with ¥ |b of 8- to 14-mesh

4
Columbia activated charcoal.

For each test the charcoal trap is filled with
purge gas, the manifold is evacuated, and radic-
krypton is allowed to flow into the krypton chamber,
The chamber is then sealed and the rest of the
manifold is evacuated. Purge gas is then allowed
to push the radiokrypton into the charcoal trap
at a constont flow rate of 5 ft3/hr. By measuring
the activity of the gas, the concentration of
radiokrypton in the effluent gas and therefore its
partial pressure can be determined. The relative
activity of the effluent gas is measured as it
passes through a small cell that has as one wall
the end window of a GM counter. The gas activity
is registered on a log counting-rate recorder.

A theoretical analysis of the adsorption process
in a number of theoretical charcoal-filled traps,
N, connected in series and occupying the total
volume of the trap has been made. The theoretical
problem is similar to the continuous-dilution-tank
problem in chemical engineering and applies only
to systems where dilution processes, and not
diffusion or adsorption, are rate limiting. The
rate of removal of radickrypton in each trap is first
order with respect to its partial pressure and
therefore its concentration in that chamber; that is,

dP F

dt  km
A series of N differential equations of this type
is solved simultaneously for the N chambers, and

PERIOD ENDING MARCH 10, 1956

the general equation for the Nth chamber is

N 4 p(N=1), (N 1)
N7 AF t
(]) Pn - e-—NFt/km ,
(N - D! (km)
where
P = partial pressure of radiokrypton, atm,

A = amount of radickrypton injected into first
theoretical chamber, em3.atm,

N = number of theoretical chambers,

F = flow rate of purge gas, cm3/min,

t = time after injection of pulse, min,

k = slope of the linear isotherm x/m = kp for
radiokrypton in the mixture of inert purge
gas and radiokrypton, em® (at STP) per
g-atm,

m = amount of adsorbent (charcoal) in trap, g,

x = amount of gas (radiokrypton) adsorbed iso-

thermally, cm® (at STP).

The term kim is a measure of the adsorptive
capacity of the charcoal trap for radickrypton in
the presence of purge gas at a given trap temper-
ature. Parameters km and N are chosen to fit
Eq. 1 to the experimental data.

The experimental results obtained with nitrogen
as the purge gas through traps at four different
temperatures are compared with analytical data in
Figs. 8.15 through 8.18. In these illustrations
the use of one trap is compared with the use of
two identical traps in series. The trap temper-
atures studied were 16, 5, =51, and ~110°C. The
deviations between corresponding curves are within
experimental error, In each case the activity
injected was determined by integration of the
activity-vs-time curves; all results were normalized
to the same area under the curve. |t may be seen
that, for a given trap geometry, the holdup times
(i.e., the time required for activity to break
through, to peak, and to disappear) increased with
decreased trap temperature. This result could be
predicted, because the energy of a gas molecule
is less at low temperature, and it therefore cannot
so readily desorb from the charcoal surface into
the moving gas stream. Also, it may be seen that
the maximum concentration of radiokrypton in the
effluent gas is lower at lower trap temperatures.

 

’G. W, Keilholtz, W. E. Browning, and C. C. Bolta,
ANP  Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012,
p 183 and Fig. 8.1.

193
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED

105 ‘ ‘ ORNL-LR-DWG 13046
& EXPERIMENTAL, ONE-TRAP SYSTEM AT16°C =
LA ANALYTICAL, ONE-TRAP SYSTEM
51— WN=7.83, 4m=1.487 x 10% cm? - -
@ EXPERIMENTAL; TWO-TRAP SYSTEM AT{5°C " —
0 ANALYTICAL; TWO-TRAP SYSTEM —
> | N=15,00, /m; 3.29x10%cm® |

 

 

 

 

 

 

 

 

 

 

 

n

o
™

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o,
Pt

 

 

 

 

 

 

 

 

 

RELATIVE EFFLUENT ACTWVITY (counts/min)

Q

M
h_tL+
T:‘_'_“:

>

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 10 20 30 40 50 60 70

TIME (min)
Fig. 8.15. Holdup of Radiokrypton in Nitrogen-

Purged Charcoal Traps as a Function of Temper-
ature, One trap held at 16°C; two traps in series

held at 15°C,

The use of two identical traps in series in-
creased the time required for the activity to peak
to almost twice that for one trap and increased
the break-through time to two to three times as
long as for one trap. The time required for the
activity to disappear from the effluent gas was
also increased, and the maximum radiokrypton
concentration was reduced.

Similar data were obtained with helium as the
purge gas. The shapes of the curves (Fig. 8.19)
are somewhat uncertain, however, because the
accuracy of the data was limited by the sensitivity
of the flowmeter in its lower range when used with
helium. Several runs were made at each temper-
ature, and the holdup times measured were sig-
nificant, even though the exact shape of the curve
may be in doubt. The curves show that radio-
krypton is held up for much longer periods of time

194

UNCLASSIFIED
5 ORNL-LR-0OWG 13047

s E—— F : J‘
T A EXPERIMENTAL; ONE -TRAP SYSTEM AT 5C |

& ANALYTICAL ; ONE-TRAP SYSTEM — -
— N=T.27, km =2.05 x 10% em® —— | ——

'@ EXPERIMENTAL; TWO-TRAP SYSTEM AT oc
2 -c ANALYTICAL ; TWO-TRAP SYSTEM ——t—— :
N=142, km = 4.82 x 10° cm® ‘

 

10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 — : \ ;X— - o -

i
K’FF’;‘AEP ,Two TRAPS
, §x 4%* —

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RELATIVE EFFLUENT ACTIVITY {counts/min)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 10 20 30 40 50 60 70
TIME (min)
Fig. 8,16, Holdup of Radiokrypton in Nitrogen-

Purged Charcoal Traps as a Function of Temper-
ature. One trap held at +5°C; two traps in series
held at 0°C.

in the presence of flowing helium than it is in the
presence of flowing nitrogen., Since the nitrogen
is adsorbed in large amounts on charcoal, the
charcoal surface is not so free to adsorb the
krypton. Helium will pass over charcoal with
very little adsorption, and hence krypton is
retained on the charcoal for longer periods of time
in the presence of flowing helium.

The experimental curves of Figs. 8.15, 8.17, and
8.18 are plotted together on Figs. 8.20 and 8.21]
for comparison. The time required for maximum
activity to be reached and the height and duration
of the peaks are clearly indicated as a function
of temperature. The analytical parameters N and
km for various temperatures, with nitrogen as the
purge gas, are compared in Fig. 8.22. |t appears
that the number of theoretical chambers N is a
characteristic of the geometry of the charcoal trap,
UNGLASSIFIED
5 ORNL-LR-DWG 13048

& EXPERIMENTAL, CNE-TRAP SYSTEM AT —51°C
4 ANALYTICAL, ONE-TRAP SYSTEM

N=7.43, km=1.09 x 10° cm®
® EXPERIMENTAL; TWO-TRAP SYSTEM AT ~50°C

o ANALYTICAL, TWO-TRAP SYSTEM
N=15.0, km =1.80 x10° ctm

ONE TRAP

—TWQO TRAPS

RELATIVE EFFLUENT ACTIVITY (counts /min)

 

0 40 80 120 160 200 240
TIME (min)
Fig. 8.17. Holdup of Radiokrypton in Nitrogen-

Purged Charcoal Traps as a Function of Temper-
ature, One trap held at =51°C; two traps in series

held at ~50°C.

since N does not vary for the different temperatures
studied. This should be true if the model used
to derive the expression for the curve has physical
significance. For two traps in series, N is twice
as large as for one trap, which indicates that N
may be a first-order function of the length of the
trap divided by the diameter,

It is indicated in Fig. 8.22 that &m increases
exponentially as 1/T increases (decreasing temper-
ature); actually, it is & that increases, since the
curves are for constant charcoal mass m. |f km,
the parameter obtained from holdup data by using
Eq. 1, is actually the slope of the linear isotherm
multiplied by the amount of charcoal, this straight-
line relationship should exist. |t therefore appears
that the parameter is proportional to the slope of
the isotherm, but whether the proportionality factor
is unity has not yet been established.

The time, t__ , required for the maximum partial

 

8. E. Guss, Solid State Div. Semiann. Prog. Rep.
Aug. 31, 1955, ORNL-1944, p 18,

PERIOD ENDING MARCH 10, 1956

UNGLASSIFIED
5 ORNL-LR-DWG 413049

0 ————
~— 4 EXPERIMENTAL; ONE-TRAP TRAP SYSTEM AT 10 c:
4 ANALYTICAL; ONE-TRAP SYSTEM T
N=8.03. 4m=6.74 x 105 cm ]
e EXPERIMENTAL, TWO-TRAP SYSTEM AT —110°C____
o ANALYTICAL, TWO-TRAP SYSTEM
al  wm=15.1, km=152x10%em®

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RELATIVE EFFLUENT ACTIVITY {counts/min)

 

 

 

 

  

 

1000 1500
TIME {min)

Fig. 8.18. Holdup of Radiokrypton in Nitrogen-
Purged Charcoal Traps as a Function of Temper-
ature, Traps held at ~110°C.

pressure  of radiokrypton to be reached in the
effluent purge gas can be derived from Eq. 1
(N - 1)km
(2) Lo = ————— .+
m NF
The maximum partial pressure, P _

oxr Can then
also be obtained from Eq. 1:

3) P _N(N — DY-14

-(N=1)
max (N = D! km

in Eq. 2 the time required for the activity to peak
depends on km and only very slightly on N, All
traps with the same amount of charcoal, m, at the
same temperature should have the same time-to-
peak, regardless of the shape of the trap. This
conelusion is consistent with the data of Guss.®
The height of the activity curve depends more on
N in Eq. 3 thon on km; therefore a long thin trap
would give a high narrow peak, and a short large-
diameter trap would give a broad peak with a low
maximum concentration.

Work is in progress to determine the relationship
between N and trap shape, and further studies of

195
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 43050

  
  
  
    
   
    
 
  

 

 

 

 

 

 

 

 

 

& EXPERIMENTAL; ONE - TRAP |
| SYSTEM AT —5%C

4 ANALYTICAL; ONE-TRAP —
SYSTEM

LA =121, |

/rm— 9.0 x 404 em? —

 

 

 

 

 

 

 

 

RELATIVE EFFLUENT ACTIVITY {counts/min)

 

 

 

A EXPERIMENTL\L L

- — Tt T
R STEM N T ‘A
. ONE ~ T AP SY TEM 3‘,"‘\ —14.85,
] AT . —/rm = 7.09 x 405 em?
A ANALYTICAL | - -

" ONE-TRAP SYSTEM :

 

  

 

 

 

 

RELATIVE EFFLUENT ACTIVITY (counts/min}

o 100 200 300
TIME (min]

500 800

Fig. 8.19. Holdup of Radiokrypton in Helium-
Purged Charcoal Traps at -5 and —~50°C,

other conditions are planned. The results should
provide engineering data for trap design and may
confirm further the theoretical relationship given
above.

LITR Vertical In-Pile Loop

W. E. Browning H. E. Robertsor
M. F. Osborne R. P. Shields
W. R. Willis
Solid State Division

D. E. Guss, U.S. Air Force

Fabrication, assembly, and preirradiation testing
of a new forced-circulation Inconel loop for oper-
ation in a vertical position in the LITR was
completed. This loop is essentially the same as
the one operated previously in the LITR,? with

196

UNCLASSIFIED
5 ORNL-L.R—DWG 13054

 

 

 

 

 

 

RELATIVE EFFLUENT ACTIVITY (counts/min)

 

0 100 200 300 400 500 600 700
TIME {min}
Fig. 8,20, Comparison of Experimental Data on

Holdup of Radickrypton in Nitrogen-Purged One-
Trap Systems at Various Temperatures,

minor modifications based on experience with the
first loop, The brush-type motor used previously
has been replaced with a three-phase induction
motor. The speed of the induction motor will be
regulated by the use of a variable-frequency power
supply. The tip of the loop will extend more than
4]/ in, below the position of maximum flux in the
reqctor to obtain a greater temperature differential
than that achieved previously. Operation of the
loop in the LITR is planned for April.

Instrumentation for ART Of-Gas Analysis

W. E. Browning
Solid State Division

Design conditions for the gamma-ray spectrometer
for the ART off-gas system have been established.
Four scintillation detectors will be used, one at

 

9G. W. Keitholtz et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 164,
PERIOD ENDING MARCH 10, 1956

UNCLASSIFIED
ORNL-LR—-DWG 13052

 

S —

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10° T
- [ ' . I — —
5 —— | ! e | } -
2 —
< L ; :
E — — — . I
5 T = -
£ - — | - '
3 .
3 o | ]
= ?\7 - .
E o
=
'_
Q
< _ _ ] _
= - ; _
= -
L |
0 . e I
L
re i - e o
Ll N e _
wl o
W I —10°C
z e ————
i : ~e__ ;
w : | X ‘
o e Do e~
e e | g — ]
. '\
\'\- =
. ] '\'_"'-
. e
S B S e
-
‘ 1 1
i 1 | | |
500 600 700 800 900 1000 GO 1200 1300 1400
TIME (min)

Fig. 8,21, Comparison of Experimental Data on Holdup of Radiokrypton in Nitrogen-Purged Two-Trap

Systems at Various Temperatures,

each of the inlets and outlets of the charcoal traps
in the off-gas lines from both the fuel system and
the reactor cell. Provision will be made for moving
the shielded detector up to 10 ft from the most
active off-gas sampling tube, under normal con-
ditions, and closer, if necessary. A fourechannel
scintillation spectrometer will be used to monitor
the detectors.

Inconel as a Thermal-Neutron-Flux Meonitor
D. E. Guss, U.S. Air Force

A study is being made of the feasibility of using
Inconel as a thermal-neutron-flux monitor in the
vertical in-pile loop and in static corrosion cap-
sules. If Inconel can be used as the monitor for
Inconel systems, it will not be necessary to add

monitors to the experimental system. Furthermore,

Inconel has the proper activation and physical
properties, whereas the various cobalt monitors
currently available either become too radioactive
during long-term irradiation or melt at the high
temperatures required by the experiments, The
cobalt monitors also require extensive activation
analysis before they can be used.

There are two constituents of Inconel which can
be used for flux monitoring: cobalt, about 0.14%
in lnconel, and chromium, about 15.5%. Cobalt
has the disadvantage that it must be chemically
separated from the Inconel after irradiation to
remove the Fe3? activity, which has gamma rays
of 1.1 and 1.3 Mev. Since cobalt is present in
such small quantities, there was, at first, some
doubt about its homogeneity, but activation
analyses of several pieces of Inconel stock show

197
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 13053

TEMPERATURE (°C}

 

 

 

 

° ~%0 -100
e 3%10
2 : 2
N FOR TWO TRAPS
o ; B )
- N FOR ONE TRAP —— -
5 5
W
"
5 2 2
&
~x
5
10 |
5 5
2 2
4
1o -
0.0030 0.0040 0.0050 0.0060
ek

Fig. 8,22, Comparison of Analytical Constants
N and km for One- and Two-Trap Systems as a
Function of Temperature,

that, within a given batch of Inconel, the cobalt
is homogeneous enough for monitoring purposes.

The Cr39 isotope, 4.4% naturally abundant,
captures neutrons and goes to Cr3', with a thermal
cross section of about 15 barns. The Cr3! isotope
has a half life of 26.5 days, and it decays by
K-capture, with an associated 0.32-Mev gamma ray
in about 9% of the disintegrations. Scanning of
irradiated Inconel with a gamma-ray spectrometer
has shown that the Cr3! peak is large enough for
a good value of the chromium agctivity to be
obtained by area analysis. The cadmium ratio of
chromium in the lattice of the LITR was measured
to be 29, which is consistent with the value to
be expected for a 1/v absorber. The cross section
is known to be about 15 barns, but a more accurate
measurement must be made.

The disadvantage in using chromium for long-term
irradiations is its relatively short half life. This
limits the time of irradiation for which chromium
could be effectively used as a flux monitor to one
or two months; after this length of time the activity
is too near the saturation point for chromium to be

198

effective as an integrated-flux monitor. For longer
irradiations, therefore, the cobalt must be used
to give a more accurate value for the thermal-
neutron flux.

Measurement of MTR Flux near Tip of In-Pile
Loop No. 3

D. E. Guss, U.S. Air Force

The thermal-neutron flux near the tip of MTR
in-pile loop No. 3 was measured by means of cobalt
foils attached to the nose coil at the positions
shown in Fig. 8.23. The values obtained at the
various positions are given in Table 8.4.

Fast-Flux Measurements in Hole 19 of ORNL
Grophite Reactor

J. F. Krause
Pratt & Whitney Aircraft

Two flux measurements were made in hole 19 of
the ORNL Graphite Reactor, which has been used
extensively for irradiations of organic materials.
The first traverse was made with the threshold
reaction S$32(n,p)P32 for the flux measurement
above 2.9 Mev, and the second was made with
Co°?(n,y)Co%? reaction for the thermal-neutron-flux
measurement. Each traverse was fairly constant
from the center of the reactor to a point 30 in,
from the center. Over this range the flux above

UNCLASSIFIED
ORNL-LR-CWG 13054

NO. 20 NO. 17

NO. 25

NO. 23

 

Fig., 8.23. Diagram of Nose Coil of MTR In-Pile
LLoop No., 3 Showing Positions of Cobalt Foils,
PERIOD ENDING MARCH 10, 1956

TABLE 8,4, INTEGRATED AND AVERAGE THERMAL-NEUTRON FLUX AT VARIOUS POSITIONS
ON NOSE COIL OF MTR IN-PILE LOOP NO. 3

 

Foil No.

Integrated Thermal-Neutron Flux

Average Thermal-Neutron Flux

 

 

(neutrons/cm2) (neufrons/cmz-sec)
15 3.85 x 1017 4.06 x 1013
17 7.08 x 10'7 7.46 x 102
20 8.20 x 107 8.64 x 1013
23 5.90 x 1017 6.22 x 1013
25 418 x 1019 4.41 x 1013
2.9 Mev was approximately 2.5 x 10° neu- and UFS"'. In all likelihood the bonding in the

trons/cmz-sec-w, and the thermal-neutron flux was
2.4 x 10° neutrons/cm?.sec-w. All measurements
were made in the ¥-in.-wall aluminum specimen
cans ordinarily used in hole 19. These cans were
surrounded by % , in. of flowing water, which acted
both as a hydraulic fluid for moving the cans and
Previous measurements in hole 19
were made during the irradiation of specific ma-
terials that may have altered the flux. Similar
measurements have been started in hole HB-3 of
the LITR, where the stress-corrosion tests are
being run on Inconel in contact with fluoride fuel
mixtures.

»

CHEMICAL EFFECTS OF NUCLEAR REACTIONS

M. T. Robinson
Solid State Division

as the coolant.

Effects of Fission Products on Properties

of Fluoride Fuels

M. T. Robinson
Solid State Division

J. F. Krause
Pratt & Whitney Aircraft

Three types of radiation effects can be dis-
tinguished in fused fluoride salt fuels such as
NoF-ZrF4-UF4. Radiation damage occurs as a
result of bombardment of the fuel with high-energy
particles and electromagnetic radiation; bulk
thermal changes are caused by the large amount
of energy dissipated in the fuel; and chemical
changes result from the replacement of uranium
by a mixture of more than 30 other elements. The
fuel consists primarily of ions, some simple, such
as Na* and F~, and some complex, such as ZrFs"

complex ions is largely ionic. For example, the
method of Pauling'® shows a Zr—F bond to be
about 75% ionic, In such a system, radiation-
damage effects should be very small, since there
are neither covalent bonds to sever (as in H,0
or organic compounds) nor a solid lattice to
disrupt. Bulk thermal effects, while they cause
a great deal of difficulty in the interpretation of
continyous in-pile measurements, are not par-
ticularly important in corrosion studies, where the
fuel-metal interface temperature is controlled, and
are even less significant in turbulently flowing
systems,

The yields of the various chemical elements in
thermal-neutron fissioning of U235 gs q function of
time have been calculated'? and used to calculate
the average valency of the fission-product mixture.
The chemical form each element is likely to adopt
in a high-temperature fused fluoride fuel is dis-
cussed below:

Elements in group 0, krypton and xenon, will
occur only in the elementary form,

Elements in groups |A, |IA, lIIB, and IVB will

adopt their usual group valencies:
+1: Rb, Cs
+2: Sr, Ba
+3: Y, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy
+4: Zr

 

]OL. C. Pauling, Nature of the Chemical Bond, p 70,
2d ed., Cornell Univ. Press, Ithaca, 1940.

”M. T. Robinson and J. F. Krause, Solid State
Semiann. Prog. Rep. Feb, 29, 1956, ORNL-2051 (to be
published).

199
ANP PROJECT PROGRESS REPORT

There is a possibility that, if the medium becomes
sufficiently oxidizing, some Ce(lV) will be present,
particularly since it may be stabilized by the
formation of some complex ion, like CeFS‘.
Similarly, in a strongly reducing medium, some
Eu(ll) is to be expected. However, the low yield
of europium makes this a relatively unimportant
point,

The transition metals beyond zirconium in the
periadic table are all relatively noble in com-
parison with chromium. Accordingly, it has been
observed that ruthenium and niobium deposit from
fluoride fuels onto the Inconel container.'? The
sparse thermodynamic data'? indicate that chromium
will reduce germanium, arsenic, and all the metals
from niobium through antimony to the elemental
state. Where valencies are required for these
metals, the following values have been chosen:

+1: Ag, In
+2: Ge, Te, Pd, Cd, Sn
+3: As, Nb, Mo, Ru, Rh, Sb

These are, in general, the lowest known valencies
which have been reported for
Fluorides that exhibit these valencies are not
known in every case.

The valencies to be expected for Se, Te, Br,
and | depend strongly on conditions in the fluoride

each element.

ORNL-LR-DWG {4720

IRRADIATION TIME (hr)
003 0102 05 41 2 5 10 20 50 100 500 2000

 

   

 

 

 

  

 

 

 

 
 
 
 

 

   
  

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T T
4.5 e I
4.0 —— ‘ ; : i :
O e T
uw —T i . : el
g 35 b, Ru, AND Mo LOST ] L
z N L ‘ ALC"NOBLE’iM TALS LOST]|!
: ) Hl :
> | sl
3.0 1 L _
2T LT | ‘ T
Ll & : ‘
2 T ] 1
g 25 = ‘ ] | 1B
20 1A/ —_ | T
1.5 L ‘ A | i
: T 1 T o :
L 11, L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

s 02 5 16°2 5 12 5 10

IRRADIATION TIME {sec)

02 5 10° 2

Fig. 8.24. Average Fission-Product Valencies
in Irrodiated Inconel Capsules Containing Fused
Fluoride Fuel Mixtures in the System NaF-ZF -
UfF

‘l

200

melt. Under moderate reducing conditions the
expected states are the normal anions:

-2: Se, Te
-1: Br, 1

If the conditions are mildly oxidizing, the ele-
mental forms are to be expected (valency 0).
Under stronger oxidizing conditions, positive va-
lencies may be anticipated.

The average valencies of the fission-product
mixture under various conditions are plotted in
Figs. 8.24 and 8.25 as functions of irradiation
time. The region above a valency of 4 presents
reducing conditions, whereas, in the region below
valency 4, oxidizing conditions exist, in com-
parison with the original uranium valency of +4,
The curves marked ‘‘inert container’’ are based
on the assumption that in capsules (Fig. 8,24)
all krypton and xenon descendants are retained
and that in reactors (Fig. 8.25) the descendants
of rare-gas isotopes of half life greater than 10 min

 

]2M. T.Robinson, T. H. Handley, and W. A, Brooksbank,
Solid State Semiann. Prog, Rep. Aug. 31, 1955, ORNL.-
1944, p 17.

131, Brewer et al., p 76-192, in The Chemistry and
Metallurgy of Miscellaneous Materials, Tbermo:ln_,?/namics,
NNES 1V-198, ed. by L. L. Quill, McGraw-Hill, New
York, 1950; W. M. Latimer, Oxidation Potentials,
Prentice Hall, New York, 1938.

"y

ORNL-LR-DWG 14712

IRRADIATION TIME (hr)
003 0402 05 1 2 5 40 20 50 1C0 500 2000

a5
|INeRT conTa

4.0 i
i No AP Ry 5T
5 | .
235 “|il* Nb, Mo, AND Ru LOST
> .
5 30 | — ALL"NOBLEMETALS LOS
= E ‘ !
u | : : I
_j i
425 . } } _

Kr AND Xe ISOTCPES LOST IF r‘/ =140 min
' | 2 i

 

5 02 5 1©0°2 5 1f2 5 10
IRRADIATION TIME {sec)

102 5 1002

Fig. 8.25.

Average Fission-Product Valencies

in Circulating-Fuel Reactors Operated with Fused

Fluoride Fuel Mixtures in the System NaF-ZrF,-

UF,.
are lost completely. The left-hand branch of each
curve, up to valency 4, is based on the elemental
state of the halogens and chalcogens. The right-
hand branch is based on the anionic forms of these
elements. In the intermediate branch, the halogens
and chalcogens are progressively reduced from the
elemental forms to the normal anions.

The curves marked ‘““Nb and Ru lost’’ are based
on the assumption that these two metals deposit
on the container walls as elements, Where the
valency is below 4, the halogens and chalcogens
are taken as the elements; where it is above 4,
these nonmetals are taken as the normal anions.
The great importance of molybdenum in determining
the average valency of the fission products may
be seen from the remaining curves, which show
the average fission-product valency for loss of
Nb, Mo, and Ru and for loss of all the ‘‘noble”

transition metals.

Curves showing the predicted excess ‘‘fluorine”
present in fluoride fuels as a result of the low
average valency of the fission products are pre-
sented in Figs, 8.26 and 8.27. The right-hand
axis is marked with the concentration of excess
Cr*? jon required to balance the excess oxidizing

-

ORNL-LR-DWG 11718

IRRADIATION TIME (hr)

03 4 35 10 20 50 100 500 2000

5000

2000
1000 ‘ 3 , 000
c ol D
§ "0 _TuERMALFLUX=2x10" 500
= o n/cm®-sec T —~
s O f
E 20 : 200 E_
o 100 L
S 100 1
T 50 5
: 50 o
o o

20 L
& 20 9
5 10 S
10

 

02 5 1%z 5 132 5 W2 5 0

IRRADIATION TIME {sec)

Fig. 8.26. Theoretical Prediction of Excess
“’Fluorine’" Present in Fused Fluoride Fuel
Mixtures in the System NaF.ZrF,-UF, os o
Function of f{rradiation Time in Inconel Capsules

in the MTR,

PERIOD ENDING MARCH 10, 1956

power of the fuel. It is very evident from these
curves that a substantial increase in the corrosive
attack of the fuel on the container may be expected
under irradiation, as compared with the ordinary
thermal corrosion, and that the attack will become
more serious as the fission rate (w/cm?) is
increased. This results from the instability, in
contact with Inconel, of ions of molybdenum and
ruthenium. Other things being equal, the effect
is expected to be more serious in systems in which
the fission gases are removed with high efficiency
than in systems where they are retained.

Contrary to these theoretical predictions, cor-
rosion in irradiated fluoride fuel-Inconel systems
has never been observed to be increased in com-
parison with corrosion in unirradiated controls. No
increased penetration of the irradiated metal had
been observed. The amounts of container metals
(Cr, Fe, Ni) found in irradiated fuels are, in
general, much the same as, or even less than, the
amounts found in unirradiated controls. It is
evident that the excess oxidizing power of the
irradiated fuel, which has been expressed here
in terms of ‘‘fluorine,’’ is accommodated without
increased attack on the container. Some of the

-

ORNL-LR-DWG 11717

IRRADIATION TIME (hr)

5 10 20 50 100 200 500 (000 2000
1000

1000
500 FUEL: 5.5mole % UF,

POWER: 60 Mw 500
200
200
100

50

20

EXCESS “FLUORINE” {ppm)
EXCESS Cr*t {ppm)

 

w* 2 5 0% 2 5 w0f 2 5 10
IRRADIATION TIME (sec)

Fig. 8,27, Theoretical Prediction of Excess
““Fluorine’”” Present in the System NaF-ZrF,-UF,
as a Function of {rradiation Time in a Circulating-
Fuel Reactor.

201
ANP PROJECT PROGRESS REPORT

ways in which this excess oxidizing power could
possibly be absorbed are discussed below.

MoF,. ~ If only a small part of the total mo-
lybdenum remained in solution as Mo(VI), it could
absorb the entire excess oxidizing power. The
required amount of molybdenum in solution would,
however, be several hundred parts per million and
would lead to far more dissolved MoF , than could
be expected to be stable in the presence of
Inconel.

Fluorides of Se, Te, Br, and I. — Since Se, Te,
Br, and | all form volatile fluorides, such as SeFé,
TeF,, BrF,, and ”:5’ a portion, perhaps two-
thirds, of the oxidizing power of the fuel could
be absorbed in such compounds. These com-
pounds, like MoF,, are strong oxidizing agents,
and their existence in the fuels can be only
transitory, unless they escape as gases. The
results of capsule irradiations argue against this
possibility.

CeF,. — As was suggested above, at least some
of the excess oxidizing power might well be
absorbed as Ce(lV), in the form of a complex ion,
such as CeF_~. Such an ion would be present
in comparatively small amounts and might well
serve as a ‘‘tracer’’ for the very similar ZtF_-—
ion. The yield of cerium is such that about
one-fourth of the oxidizing power could be thus
absorbed in systems from which the rare gases
were lost and almost one-half in systems in which
they were retained,

UF.. — Another way of absorbing the oxidizing
power of the fuel would be the oxidation of U(IV)
to U(V):

UF.~ + F = UF, + F-

In the fuel of the ART, this would require oxidarion
of up to 2% of the uranium after 500 hr of operation,

Solid Fluorides. — It is perhaps possible that
the ‘“‘noble’’ metal deposit is not metallic and that
it contains polymerized subfluorides,
(MoF)x. These materials would reduce the oxi-
dizing power of the fuel aobout one-half in the
reactor case.

such as

No one of the above suggestions can account for
all the excess ‘‘fluorine’” released by the
fissioning uranium. In every case, a substantial

202

oxidizing power would remain. That the excess
oxidizing power causes no increase in corrosion
argues convincingly that the deposited fission-
product layer effectively inhibits corrosion of
Inconel by the fuel.

The effect of the fission-product mixture on the
physical properties of the fuel has also been
studied. The well-known theory of Frenkell4
concerning the viscosity and electrical conductivity
of fused salts was used as a starting point, The
viscosities of the fused salts are determined by
the larger of the ions present, which in the case
of the fuels being considered are species such
as UFS' and ZrFs". The fission process de-
creases the number of these large ions by replacing
them with the small ions Rb*, Ba**, La**?, etc.,
none of which are likely to be complexed. There-
fore a small decrease is expected in the viscosity
of the fuel with increasing time of irradiation.
This decrease in viscosity will be exceedingly
small, however, because of the relative rarity of
UFS" ions (about one complex anion in four), the
low burnups envisioned in aircraft reactors, and
the high fission yield of zirconium, There will
be a similar small increase in the electrical con-
ductivity of the fuels. The effect of the fission
products on, thermal conductivity cannot be esti-
mated because of the lack of a suitable theory,
but any effect will certainly be very small, The
molar heat capacity of the fuel may be taken as
independent of composition, to a first approxi-
mation. The only effect of fission will be to
increase slightly the effective molecular weight
of the fuel through loss of the noble gases, mo-
lybdenum, ruthenium, etc. Thus the specific heat
of the fuel will decrease very slightly as irradi-
ation proceeds.

The eftect on corrosien is, therefore, considered
to be the only important effect of the fission
process on the properties of the fused fluoride
fuels, Theory predicts a substantial increase in
the corrosion rate as irradiation continues, but the
absence of such an effect in all experiments to
date leads to the suggestion that deposition of
niobium, ruthenium, and (presumably) molybdenum
on the container effectively inhibits the attack of
the fuel on Inconel.

 

]4.]. Frenkel, Kinetic Theory of Liquids,
Clarendon Press, Oxford, 1946.

p 439,
Use of Natural Lithium in Fluoride Fuels
Circulated in In-Pile Loops

M. T. Robinson
Solid State Division

J. F. Krause
Pratt & Whitney Aircraft

The use of lithium in the fluoride fuel system
NaF-KF-LiF-UF4 presents a problem created by
the nuclear reaction

6Li* + n—> 4He + 3Tt
The concentration of the tritium ion in the fuel
mixture NcF-KF-LiF-UF4 (11.2-41-45.3-2.5 mole %)
as a function of time and thermal-neutron flux is
shown in Fig. 8.28. The tritium-ion concentration

is expressed as the pT of the solution and is
defined by the following equation:

pT = ~log (gram-ions of T* per liter of solution) .

PERIOD ENDING MARCH 10, 1956

On the right-hand axis is shown the equivalent
concentration of chromium in the fuel, which would
be produced by the reaction

27 + Cr

(soln) {tnconel)

—> Crtt + T
(soln) 2(905)

[t is evident that such a fuel will be much more
corrosive under irradiation than out-of-pile, es-
peciaflly when irradiated in capsules in the MTR,
unless at least part of the Li% is removed from
the fuel.

Because of the large neutron absorption cross
sections of fuels containing natural lithium, there
would be substantial decreases in the effective
neutron flux in such materials in comparison with
similar fuels containing pure Li7. The decreased
effective neutron flux and the large dilution factors
normally used in in-pile loops will make the
chemical effects of tritium production less serious
than they would be in capsules. Nevertheless,

gy,
ORNL-LR-DWG 11824

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6 ] [ 1 T
~— <] 1T — T T TTTH:
I ] w7 = —log (g of T per liter of solution) i [
\\1\ | "-...__._.\ fx ot P 115
T Ney;, | f |
5 \\L\ i Zx 1ot ons/cme_ _ ; ; - 10
. \‘\\L:\\\\\ mi\&e\j\\ T 2
, T
g 4 ‘-\\\ ; h\:s% \\h“\ \‘\\\ ! g
5] \\\\ X f0f2 \\ ""'--.\ : \ . "‘"--..,_._._\ ] o
g T~ \\\\W\ \\\\\\\ \\\ 1 E
fi 3 \ M — Tt 2x 013 __\ -l | "'--\ -\\-.._ - "'-~..._._< 10 %
8 ) \"""'--. 5 \\\\""- | \ T~ T~ -
= \ \\ ~Sx 1513 — ~L ~ ‘\\ ~L z
< e — — N ™~ | | N~ 5 O
E \ \\ M~ X014 \ \\\ — \ \\ t
& 2 I~ ] 2x 1014 T~ "‘\._\ e~ T~ 107 7S
~
ng \\\ \\\ 5 \\ \h\\ \ \__\ \\\ I E
% \\ \\"‘-\ x4 015 \ \\\\.\\ \ N - 5 %[
; \l._ —— P _‘8 51073 \ \\\x \ "-~>...___= 103 B
> f T \ _ \-.\ | :\ : \* N
': \\.‘\:‘"\ 5 X fo 15 . . i 1 \ | h.__‘*
| | T [— "'--‘I_: 4
¢ ] | \"‘—x \ -\-_""-—«-_.___ —-fi‘_\_‘! : 10
0 — o
] 1T e
| o I | |
' 5
} L T | 1%
! 2 5 10 20 50 00 200 500 1000

IRRADIATION TIME {hr)

Fig, 8.28. Production of Tritium in the Fluoride Fuel Mixture NaF-KF-LiF-UF‘ (11.2-41-45,3-2.5 mole

%) Containing Natural Lithium as a Function of Time and Thermal-Neutron Flux,

203
ANP PROJECT PROGRESS REPORT

in order to obtain satisfactorily high fission rates
in the in-pile loops now being used, removal of
some of the Li¢ will be required.

In order to keep that portion of the corrosion
attributable to tritium to below 100 ppm chromium
in 300 hr, any lithium in fuel used in MTR capsule
irradiations must contain less than 0,2% Li6. In
in-pile loops, up to 1% Li% may be allowed without
excessive depression of the fission power at the
tip of the loop. This amount of Li% will also
restrict tritium corrosion to less than 100 ppm
chromium per 300 hr,

There seems to be no reason to expect the
apparent fission-product protection of Inconel found
in NaF-ZrF4-UF fuel systems to reduce the
attack of T* on :he metal surface. Furthermore,
it is not evident that this protective mechanism
will be operative in the lithium-containing systems.
The exact distribution of tritium among the various
species, T1, TF2', TF (soln), ond TF (gas), is
believed to be unimportant,

w.k.«"m
RADIATION DAMAGE TO BORON CARBIDE
0. Sisman
J. G. Morgan M. T. Morgan
R. M. Carroll
Solid State Division

A study of radiation effects on B,C and related
thermal-neutron shield materials is under way.
The first irradiations were made on uncoated
hot-pressed B4C (Norbide) and slip~cast B,C
bonded with SiC (The Carborundum Company) at
low temperatures {~200°F). The samples were
sealed under helium in quartz ampoules and were
irradiated in hole C-48 of the LITR for an exposure
of 6 x 10'? nut (as measured by cobalt-foil acti-
vation at the surface of the specimen). The
samples, the quartz ampoules, and the perforated
aluminum containers in which they were irradiated
are shown in Fig. 8,29, The samples and their
container were cooled by the reactor cooling water.
Final burnup of the samples will be obtained by
analysis of the B10.to-B11 ratio. After irradiation
the ampoules were broken, and the gas released
from the sample was measured, The gas-release
apparatus is shown in Fig. 8,30, The ram is
struck to breck the ampoule and release the gas
into a calibrated volume. A manometer with an
adjustable head is used to measure the volume
of the additional gas formed as the result of
irradiation,

204

HNCLASS|IFTED
PHOTO 15584

HOT-PRESSED

" COLD-PRESSED SINTERED

 

Fig, 8,29. Samples of B,C and the Containers
for lrradiation in the LITR.

UNCLASSIFIED
ORNL-LR-DWG 1306t

 

 

 

 

 

 

 

 

 

RAM

 

 

 

 

 

 

 

 

! ‘ AMPOULE -

 

| o /j
] | SPECIMEN — 7%
| “ "

TO MANCMETER

Fig. 8.30. Apparatus for Releasing Gas from
Quartz Ampoules in Which B4C Samples Were
Irradiated,
indicated that most of the
helium produced from the n,a reaction with B10
in hot-pressed B C would be released at temper-
atures above 1500°F, and therefore no helium
evolution was expected as a result of these tests
at about 200°F. The recent results confirmed the
prediction; none of the approximately 1 cm3 of
helium produced as the result of irradiation was
detected by the apparatus, which could detect
0.2 cm3, The hot-pressed samples also retained
their physical dimensions and bulk structure, as
shown in Fig. 8.31. At higher burnups, however,
the hot-pressed material may be expected to
crumble, With the cold-pressed and sintered B ,C

Previous work!3

 

Sw. . Yalovage, Effect of Irradiation on Hot-Pressed
Boron Carbide, KAPL-1403 (Nov. 15, 1955).

COLD-PRESSED SINTERED B,C

 

IRRADIATED

UNIRRADIATED

PERIOD ENDING MARCH 10, 1956

(bonded with SiC), a different effect was noted.
While the material appeared to retain its bulk
stability, there was gas release (under irradiation)
which far exceeded the total helium generation.
In addition, the walls of the quartz ampoule were
coated with a heavy black deposit, as shown in
Fig. 8.32, and the sample weight decreased 4%.
Unirradiated specimens heated for 72 hr ot 1800°F
did not show these effects, It is thought that some
foreign material (organic binder or pitch) introduced
during fabrication
sample,

may have remained in the

in order to determine whether the hot-pressed
samples will retain helium at elevated temper-
atures, gas release from samples irradiated at
temperatures up to 2150°F will be measured. In -

UNCL ASSIFIED
PHOTO 16273

HOT-PRESSED B4C

 

 

INCH

Fig. 8,31, Irradiated and Unirradiated B,C Specimens, 3X. Reduced 7%.

205
ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED
PHOTO 16275

  

Fig, 8.32. Coating Found Inside Quartz Ampoule
in Which Cold-Pressed Sintered B4C Was lrradi-
ated, 3X. Reduced 38%.

206

addition, other boron-containing materials will be
investigated, such as a 2% B-98% Fe alloy, o
6.6% B,C body bonded with copper and flame-
coated with copper and stainless steel, and an
89% B,C body bonded with SiC and flame-coated
with A|20 . Coating adherence, dimensional sta-
bility, cng helium evolution will be studied at
elevated temperatures under irradiation. Future
irradiations will probably be made at the MTR at
temperatures up to 2000°F with radiation exposures
of 1020 nut and greater.
PERIOD ENDING MARCH 10, 1956

9. ANALYTICAL CHEMISTRY OF REACTOR MATERIALS

J. C, White
Analytical Chemistry Division

DETECTION OF TRACES OF NoK IN AIR

A. S. Meyer, Jr. J. P, Young
Analytical Chemistry Division

R. G. Affel F. W. Manning

Instrumentation and Controls Division

An investigation was initiated in order to de-
velop an instrument for monitoring the concentra-
tion of alkali metais in the exhaust gases from
NaK-to-air radiators, This instrument is to be
used for detecting NaK leaks in the ART radiators,
as well as in the radiators now being used in test
installations, Tentatively, the sensitivity of the
detector to be used on the ART radiators has been
specified to correspond to an incremental increase
of 0.01 ppm (10 parts per billion) of alkali metals
in the air while it is crossing the bank of radiators.
An instrument of lower sensitivity, 10 ppm, will
be satisfactory for monitoring the exit gases from
the test stands, |t has also been specified that
the detector must have a rapid response time and
that the instrument for use on the ART must be
highly dependable, since no maintenance can be
performed during the period of the reactor experi-
ment. Effort is being directed toward the im-
mediate development of an instrument to menitor
the test-stand radiators as a step in the develop-
ment of the monitor for the ART,

Detection of Microgram Quantities

The proposed leak-detection method is based
on the measurement of the hydroxyl ion which is
formed when the alkali metal oxides and hydroxides
are absorbed from a sample of the contaminated
air in an aqueous solution. The pH of the aqueous
solution will be maintained at a constant value
by the coulometric generation of hydrogen ions
in o quantity equivalent to the quantity of hydroxyl
ion being absorbed from the sample of air, The
coulometric current required to maintain the pH
of the solution will be proportional to the rate of
addition of alkali metals, and, if a continuous air
sample is taken, it will be proportional to the
concentration of alkali metals in the sample.

Both photometric and electrometric methods for
the detection of changes of pH are being studied.,

To eliminate the buffering effects of carbon di-
oxide, the pH of the absorber solution will be
maintained between 4.5 and 5.0, In this region
the colored indicator bromcresol green (tetrabromo-
m-cresolsulfon phthalein) can be used as a sensi-
tive detector for increases in the alkalinity of the
solution, If 1 pg of NaK is added to 1 mi of the
solution, its transmittancy is decreased to 50%
of its original value.

A gas-scrubbing absorption cell of approximately
30-ml capacity has been fabricated (Fig. 9.1). The
gases enter the mixing chamber through a quariz
tube, which can be heated to temperatures as high

UNCLASSIFIED
ORNL-LR—-DWG 12608

   
 
  
 
   
   
 

LIQUID I
LEVEL

 

LIGHT *'-

BAFFLE \! i

CATHODE |
COMPARTMENTQ\%

 

 

 

PHOTOMETRIC
ABSORPTION

CELL \\

LIGHT BAFFLE

MIXING
CHAMBER

 

 

 

ANOCDE

 

1 O 1
INCHES
Fig. 9.1. Gas-Scrubbing Absorption Cell for

Detecting Atkali Metals in Air.

207
ANP PROJECT PROGRESS REPORT

as 1000°C, in order to prevent deposition of the
metal oxides on the walls of the vessel. When
bubbles of the gas pass through the helical ab-
sorption tube, this section functions as a gas lift
to produce circulation of the aqueous solution
through the photometric absorption cell. Air flow
rates of 600 cm®/min have been found to be prac-
tical. In 1 min, if @ concentration of 10 ppm of
NaK in the air is assumed, 6 pg of NaK will be
infroduced into the absorption cell. Since it has
been shown that 30 pg of NaK in the cell will
effect a decrease of greater than 50% in trans-
mittancy, 6 pug would cause a change of approxi-
mately 15%. Since the optical device planned for
use with this apparatus can easily detect changes
in transmittancy of the order of 5%, a predicted
response time of 1 min to a 10 ppm concentration
of NaK in the dir is conservative. The time re-
quired for the transfer of the solution from the
mixing chamber to the photometric absorption cell
is less than 3 sec, and it is thus well within the
above response period.

Photometric absorption cells for both inlet and
exhaust air will be required in the instrument for
the detection of NaK leaks from the test-stand
radiators, because other operations in the vicinity

of the test stands may occasionally contaminate
the inlet air with high and variable concentrations
of alkali metals. The instrument will be designed
to respond to the incremental increase of alkali
metal conceniration in the air while it is crossing
the radiator, For maximum stability, o double-
beam optical system will be used to compare the
absorbance of the solutions in the two cells, A
block diagram of the proposed instrument is shown
in Fig. 9.2,

Light of wavelength 630 my is separated into
two beams by a mechanical chopper. After passage
through the inlet and exhaust sample cells, re-
spectively, the two beams of light are recombined,
at phototube A. A portion of the light transmitted
through the inlet-air monitoring cell is separated
by a beam splitter and directed to phototube B.
When alkali metals are present in the sample of
inlet air, this light will be attenuated, and the
decrease in the response of phototube B will be
amplified and fed into a servomechanism which
will pass a coulometric current through the inlet
sample cell, This current will be regulated to
generate a sufficient quantity of hydrogen ions
fo maintain the bromcresol green indicator at the
original absorbance. Since the coulometric circuits

UNCLASSIFIED
ORNL-LR-DWG 12609

 

  

 

 

  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

I, —»
MECHANICAL CHOPPEF?/
i EXHAUSTAIR PHOTOMETRIC
7 ABSORPTION CELL
) e Ty
| -] | SERVO
2
INTERFERENCE | OPTICAL\ I, +1, (M )— | CURRENT
FILTER | WEDGE — | CONTROL
ll | 11
| PHOTOMETRIC |
| ABSORPTION CELL |
4
— T NN . TUNED
MIRRORA % AMPLIFIER
INLET AIR + 'h
|y~ PHOTOTUBE
I‘I
5
SERVO
CURRENT TUNED
CONTROL AMPLIFIER

 

 

 

 

 

 

Fig. 9.2, Diagram of Instrument for Detecting Microgram Quantities of Alkali Metals in Air.

208
for both the inlet-air and exhaust-air monitoring
cells are in series, the same current will flow
through both cells. A separate mechanical or
electronic device will be incorporated at the
exhaust-air cell to compensate for differences in
gas flow through the two cells. Any imbalance
of the light beams transmitted through the two
absorption cells will generate an a-c signal at
phototube A. The amplitude of this signal will
be a measure of the incremental increase of the
concentration of alkali metals in the air while
it is crossing the radiator, The output of photo-
tube A will be amplified and fed into a servo-
mechanism which will pass an additional coulo-
metric current into the exhaust-air monitoring cell,
The amount of additional current required to hold
constant the absorbance of the absorber solution
in the exhaust-air cell will be monitored, and if
this current exceeds a predetermined rate that is
indicative of a NaK leak, an alarm will be acti-
vated,

The preliminary investigations of electrometric
methods for determining the change of pH of the
absorber solution have indicated that glass elec-
trodes are somewhat more sensitive to changes
in pH of the absorber solution than are the photo-
metric detectors. The glass electrodes may be
less dependable, however, than the photometric
detectors because of variations in the asymmetry
of the potential of the electrodes. Tests are now
being made to determine whether the instability
of the potential of the glass electrodes is sufficient
to introduce false signals. Since highly stable
pH meters are available commercially, the instru-
mentation can be simplified if the electrometric
detector can be used. The over-all design of the
electrometric instrument would be functionally
similar to that of the photometric instrument,

Detection of Submicrogram Quantities

The measurement of the absorbance of light of
the frequency of a resonance line has been demon-
strated to be one of the most sensitive methods
for the detection of traces of vapors of certain
elements. A simple optical instrument has been
designed to detect concentrations of mercury vapor
of less than 1 ppm by the measurement of the
absorption of the mercury 2537-A resonance line.

 

). R. McNally, personal communication te A, S.
Meyer, Jr.

PERIOD ENDING MARCH 10, 1956

The absorption coefficient of the sodium doublet
at 5890 A is larger than that of the mercury 2537-R
resonance line, and, on the basis of the transition
probabilities and half lives of excited states,
cited by Mitchell and Zemansky,? the absorption
coefficient,? k, at the center of each of the lines
of the sodium doublet falls between that of the
mercury 2537-A line and that of the cadmium 2288-A
line and thus corresponds to about 10=12 N, It
can be calculated from the absorption coefficient
that a concentration of sodium vapor corresponding
to 0.01 ppm of sodium in air would produce a 50%
attenuation in a beam of sodium-doublet radiation
which traversed a 50-cm optical path,

Since sodium in compound form does not absorb
the resonance radiation, it is necessary to heat
the air containing the sodium to a temperature
sufficient to cause thermal dissociation of the
Na,O to sodium and oxygen ions. On the basis
of thermodynamic constants derived from measure-
ments of the volatility of Na,O at reduced pres-
sures, Brewer? has calculated that Na,O in air
would be appreciably dissociated at temperatures
near 1000°C,

An instrument based on these considerations
has been proposed for the detection of the sodium
which would be liberated by a NaK leak in the
radiators of the ART, A block diagram of the
instrument is presented in Fig. 9.3. The beam
of sodium-doublet light from a sodium-vapor lamp
or resonance bulb is divided inte two pulsed
beams of equal intensity by a mechanical chopper.
The beams are passed through heated absorption
cells which contain samples of the inlet and
exhaust air from the radiators and are recombined
at the phototube. When the intensities of the
transmitted beams are equal, a null a-c signal is
received at the phototube. If unequal concentra-
tions of sodium are present in the absorption
cells, an imbalance is produced in the transmitted

 

2A. C. G. Mitchell and M. W. Zemansky, Resonance
Radiation and Excited Atoms, Chap. I, p 92, The
University Press, Cambridge, 1934.

3The absorption coefficient, &k, is defined by the
relationship I = Ioe—kN, where [ and [ are the intensities
of the incident and transmitted light, respectively, and
N is the density of sodium in the light path in atoms

per square centimeter.

4L. Brewer and J. Margrave, . Phys. Chem. 59, 421
(1955).

209
ANP PROJECT PROGRESS REPORT

SODIUM VAPOR LAMP
INLET

AlR
SYNCHRONOUS ¢
CHOPPER

 
 

UNCLASSIFIED
ORNL—-LR-DWG 12610

ABSCRPTION CELLS

] —2m--

OPTICAL WEDGE

 

% MIRROR

 

 

 

 

 

 

 

 

 

___________ 'é—_*—-fl— PHOTOTUBE

 

 

 

 

 

 

 

/W FURNACE

;F  /
—_—— —_ - — _"l _____
INTERFERENCE
FILTER Y
MIRROR f

EXHAUST

AIR

 

 

 

 

 

 

METER AND
ALARM

SYNCHRONIZED
AMPLIFIER

 

 

 

 

 

 

 

Fig. 9.3. Diagram of Instrument for Detecting Submicrogram Quantities of Sodium in Air,

beams and a proportional a-c¢ signal is generated.
For low absorbancies, the output from the photo-
tube, at a given temperature of the samples in
the absorption cells, is proportional to the differ-
ence in sodium concentration in the two absorption
cells, An instrument of this type would thus give
a measure of the incremental increase of the
concentration of sodium in the air while it is
crossing the radiators,

The proposed double-beam instrument is readily
adaptable to stable instrumentation and does not
require careful regulation of sample flow rates.
If further investigation indicates that Na,O is
sufficiently dissociated at the temperature of the
exhaust gases, the absorption beam can be passed
directly through the stack gases to obtain a more
rapid response., It will be necessary to calibrate
the instrument empirically because its sensitivity
will vary with the temperature of the absorbing
gases and with the spectral distribution of the
incident radiation,

A test model of the instrument is being assembled.
This apparatus contains two quartz absorption
cells, 24 in. in length, which can be heated to a
temperature of 1050°C,
for the introduction of gaseous samples to each
tube or for the volatilization of sodium compounds
of known vapor pressure for calibration of the
system. Tests of the effect of temperature on the
sensitivity of the instrument, comparison of light

Provision is being made

210

sources, and tentative calibration should be possible
with this instrument.

SPECTROPHOTOMETRIC DETERMINATION OF
TITANIUM IN MIXTURES OF FLUORIDE
SALTS WITH TIRON

J. P. Young J. R. French
Analytical Chemistry Division

A method has been developed for the use of Tiron
(disodium-1,2-dihydroxybenzene-3,5-disulfonate) in
the determination of trace amounts of titanium in
NaF-ZrF ,-UF, and in other mixtures of fluoride
salts, The yellow color of the titanium-Tiron com-
plex is a very sensitive test for titanium.> The
molar absorbancy index is 13,900, The moximum
absorbance of this complex is at a wavelength
of 380 mu. The intensity of the color is essen-
tially independent of pH over the range 4.3 to
9.6. The uranyl, zirconium, and ferric ions in
the fluoride salts have been found to interfere
with the development of the titanium-Tiron color.
Zirconium ions interfere by forming a colorless
complex with Tiron and thereby consuming the
reagent. Ferric and uranyl ions form colored com-
plexes with Tiron., It has been reported that the
interference of the ferric ion may be eliminated
by reducing the ferric ion with sodium hydrosulfite

 

°). H. Yoe and A. R. Armstrong, Anal. Chem. 19,
100 (1947).
at a pH of 4.7;% the zirconium interference may be
eliminated by the addition of a large excess of
reagent;> and the interference of the uranyl ion
may be eliminated by the addition of iron as a
carrier and the precipitation of titanium with
sodium hydroxide in the presence of carbonate
ion® to complex uranium as the soluble uranyl
carbonate anion,

In the determination of titanium in NaF-ZrF ,-
UF,, quantitative separation of the titanium is
achieved by an ammonia precipitation in the pres-
ence of carbonate ion., Apparently, the precipita-
tion of zirconium, under these conditions, co-
precipitates the titanium completely, and it is
unnecessary to add a cartier, such as iron. The
precipitate is separated by centrifugation and then
dissolved in dilute hydrochloric acid. A large
excess of Tiron is added, and the solution is
buffered at a pH of 4.7, The solution is then
allowed to stand for 30 min, since the titanium-
Tiron complex develops slowly. Solid sodium
hydrosulfite is added to reduce the ferric ion, and
the absorbance of the complex is measured at a
wavelength of 140 mu. Since the absorbance of
a solution of sodium hydrosulfite is significant
below 400 myu, the absorbance of the titanium-
Tiron complex cannot be determined at its point
of maximum absorption (380 my).

Determinations of titanium in five samples of
fluoride fuels have been made by this method.
The coefficient of variation for these determina-
tions was 2%. Quantitative recovery of the standard
solution of titanium added to several of these
samples prior to analysis was obtained.

DETERMINATION OF TANTALUM IN FUSED
MIXTURES OF FLUORIDE SALTS

J. P. Young J. R. French
Analytical Chemistry Division

The determination of trace amounts of tantalum
in NaF-LiF-KF, as described previously,”® was

 

5p. A. Lee, B. S. Weaver, and J. W. Gates, Colori-
metric Determination of Titanium, Part 11 (.001% to 5%),
C-1.360.8 (Oct. 25, 1945).

7). c. White, Determination of Small Amounts of
Tantalum in NalF-L{F-KF and in NaF-Lz'F-KF-UF4,
ORNL CF-56-1-49 (Jan. 10, 1956).

8). P, Young and J. R. French, ANP Quar. Prog. Rep.
Dec. 10, 1955, ORNL-2012, p 192,

PERIOD ENDING MARCH 10, 1956

accomplished by modifying an existing method,
described by Dinnin,® for the colorimetric determi-
nation of tantalum with pyrogaliol. In the modified
procedure the fluoride sample is carefully digested
in dilute sulfuric acid in order to hydrolyze TaF,
to Ta,04 without loss of tantalum by volatilization
of the pentafluoride. The solution is then evapo-
rated to dryness, and the residue from the evapo-
ration is fused with potassium pyrosulfate. The
melt is then dissolved in ammonium oxalate. The
tantalum-pyrogallol complex is formed in a solution
that is 0.175 M ammonium oxalate and 4 M HCI,
after which its absorbance is measured at 330 mp,

A satisfactory method for the determination of
small amounts of tantalum in NaF-LiF-KF-UF
was also developed. Since uranium interferes with
the determination of tantalum by the pyrogailol
method, it was necessary to develop some means
interference,  According to
Hillebrand et «l, ' tantalum is quantitatively

of removing this

separated from vranyl ions by cupferron in a solu-
tion of sulfuric acid {(5%) which contains tartaric
acid. Since the samples were composed of alkali
fluorides and uranium tetrafluoride, it was neces-
sary to dissolve them carefully to prevent the loss
of TaF ¢ by volatilization and to oxidize the tetra-
valent uranium to uranyl ions. The samples were
dissolved in a mixture of sulfuric acid (5 vol %)
and aqua regia at 100°C. After the dissolution
of the samples, tartaric acid was added, and a
cupferron precipitation was performed at ice-bath
temperature. The precipitate was removed and
then ignited to obtain Ta,Og, which was fused
with potassiumpyrosulfate. The tantalum-pyrogallol
color was then developed, as described previously,

The coefficient of variation was 2% for the re-
sults of the determination of tantalum in both the
uranium-containing alkali fluoride mixtures and
the alkali fluoride mixtures without uranium. In
the separation of tantalum from uranium, a sample
which contained at least 1 mg of tantalum was
taken. It is expected, however, that the separation
is applicable for smaller amounts of tantalum,

 

9], 1. Dinnin, Anal. Chem. 25, 1803 (1953).

0. F. Hillebrand et al., Applied Inorganic Analysis,
p 120, 2d ed., Wiley, New York, 1953.

211
ANP PROJECT PROGRESS REPORT

DETERMINATION OF OXYGEN IN ZrF,
BY BROMINATION

J. P. Young M. A, Marler
Analytical Chemistry Division

A study of the possibility of applying the bromi-
nation method of Codell 't 4o the
determination of combined oxygen in metallic
oxides and in fluoride salts was continued, The

and Norwitz

bromination method consists in the high-temperature
reaction of bromine vapor, in a carrier gas of
helium, with an intimate mixture of the sample
and graphite. The products of this reaction are
CO and the metal bromide. The excess bromine
and the metal bromide are removed by a system
of cold traps, while the last traces of bromine
are removed by granular zinc maintained at 350°C,
The CO is oxidized by hot CuQ, and the resultant
CO, is utilized as a measure of the oxygen present
in the sample,

Attempts to determine oxygen in the presence
of fluoride salts have shown that consideration
must be given to the possible reaction of volatile
fluorides with the glass components of the appa-
ratus. In the study of the bromination of pure
Zr0,, it was found'? that it was necessary to
mix the oxide with a fluoride salt in order to re-
move ZrO, completely as a volatile reaction
product. It is believed that the fluoride salt
undergoes some type of exchange reaction with
ZrQ, that results in the formation of a volatile
zirconium fluoride compound and an oxide of the
metal in the original fluoride salt. The oxide
then reacts with bromine, In spite of the com-
plete removal of the ZrO,, the recovery of oxygen
as CO, was low. The apparatus and the procedure
used in the bromination of ZrQO, mixed with FeF,
and graphite were described previously,'?

In an effort to improve the recovery of oxygen
as CO,, mixtures of ZrO, and FeF; were placed
in small nickel bombs and heated to temperatures
of 550 to 700°C for periods of 1 hr. After the
bombs were cooled, the contents were mixed with
graphite and allowed to react with bromine, Again,
the recovery of the oxygen originally present in
the ZrQ, was incomplete, The effects of mixing
other fluoride salts with the ZrO, and graphite

 

"M, Codell and G. Norwitz, Anal. Chem. 27, 1083
(1955).

12y, p. Young and M. A. Marler, ANP Quar. Prog.
Rep. Dec. 10, 1955, ORNL-2012, p 191.

212

were also being studied. The addition of NiF,
did not improve the recovery of oxygen, but it
was found that the solid products resulting from
the bromination of the mixture of ZrO,, NiF,, and
graphite are large crystalline masses, whereas the
products obtained from the bromination of the mix-
ture of Zr0,, FeF,, and graphite are finely divided
solids, It is anticipated that analyses of the
crystalline products will provide a basis for a
better understanding of the reaction mechanism.
Spectrographic analysis has shown zirconium {and
nickel) in the crystalline reaction products, and
petrographic analysis has shown nickel bromide
and an unidentifiable phase. No zirconium has
been found in the graphite which remains in the
sample container after bromination,

DETERMINATION OF MICRO AMOUNTS
OF BORON IN FUSED FLUORIDE
SALT MIXTURES

A. S. Meyer, Jr, W. J. Ross
Analytical Chemistry Division

A method has been developed for the determi-
nation of small amounts of boron in fluoride salt
mixtures.
salts in an acidic solution of aluminum chloride
at room temperature, extracting the boron into
ether, and determining the boron by the carminic
acid method,'®  The determination of boron by
direct application of the usual procedures was not
feasible, because the high concentration of fluoride
prevented the formation of the chromogenic com-
plexes of boron. Therefore it was necessary to
develop a means of separating sufficient boron
from the fluoride for the carminic acid method to
be used satisfactorily.

is volatilized from hot acidic
therefore the dissolution of the
salts must be effected at a relatively low tem-

Boron trifluoride
solutions, and

perature, unless provision is made to trap the
volatile compounds of boron. It was found that
the dissolution of NaF-ZrF ,-UF, could be per-
formed efficiently by stirring the solid sample in
an acidic {(~2 M HCI) solution of 1 M AICI3~6H20
at room temperature., In this manner, boron was
retained in solution in the presence of fluoride

and metal ions.

 

133, T. Hatcher and L. V. Wilcox, Anal. Chem. 22,
567 (1950).

The method consists in dissolving the

'
| he method of Glaze and Finn'4 has been adapted
to the separation of boron from the other com-
ponents of solutions of these mixtures. This method
is based on the relatively high partition coefficient
of boron between ethyl ether and an acidic solution
of boron in ethanol-water (1:1). When the aqueous
solution is equilibrated with an equal volume of
ether at a controlled temperature, a reproducible
fraction (approximately 50%) of the boron is ex-
tracted into the ethereal layer. The presence of
fluoride is reported to decrease the extraction
coefficient. The original determination of boron
in the ether phase is performed titrimetrically on
a macro basis, The presence of certain cations,
AI3* Fed*, Zn?*, and Ba?”, reduces the accuracy
of the titration,

The extraction coefficient for micro amounts of
boron was established as 0.45, when the parti-
tioning was performed at 19°C by shaking the
mixture manually for 5 min. Practically all the
aluminum and the other metallic ions remained
in the aqueous phase, The reproducibility of the
extraction coefficient, +0.02, and the absence of
fluoride in the ether extract indicate that inter-
ference by fluoride is effectively eliminated through
the formation of stable complexes of aluminum and
Hluoride during the dissolution of the sample.

The concentrations of fluoride and metallic ions
which accompany the boron into the ether phase
are tolerable for the carminic acid method. This
method has been used with good precision for the
determination of boron in concentrations as low
as 10 ppm in NaF-ZrF ,-UF,, and it should also
be applicable to the determination of boron in

NaF-KF-LiF base fuels and in other fluoride salts.

SPECTROPHOTOMETRIC DETERMINATION
OF BISMUTH IN FUSED MIXTURES
OF FLUORIDE SALTS
A. S. Meyer, Jr, B. L. McDowell
Analytical Chemistry Division

A spectrophotometric method for the determina-
tion of bismuth as the tetraiodobismuthate(lll)
complex'® has been applied to the determination
of bismuth in NaF-ZrF ,-UF,. The absorbance of

the complex was measured at 337 mp in a volume

 

4 W, Glaze and A. N, Finn, J. Research Natl. Bur.
Standards 16, 421 (1936).

15N, M. Lisicki and D. F. Boltz, Anal. Chem. 27,
1722 (1955).

PERIOD ENDING MARCH 10, 1956

of 50 ml of 1 M H,S0,. Since the cited reference
includes a thorough report of variables such as
iodide concentration, acid concentration, and time,
the optimum conditions as set forth there were
used for the determination.

Test portions which contained 0.6 to é pg of
bismuth in @ solution of sulfuric acid were taken
for color development, The color was developed
by adding 10 ml of 5 M H, 50, and 20 ml of a
solution that was 14% Kl and 0.1% ascorbic acid
to the test portion and then diluting to a volume
of 50 ml with water, The absorbance of the solu-
tion was measured against a reagent blank,

The effects of several diverse ions on the ab-
sorbance of the tetraiodobismuthate(lll}) complex
were reported,'> but the effect of uranium was
not mentioned. The effect of uranium was there-
fore investigated for uranium-to-bismuth weight
ratios of up to 1000:1, and it was found that for
weight ratios of less than 30:1, the method was
applicable with an error of less than 3% in the
absorbance. For larger amounts of uranium a pre-
liminary separation is necessary. The coefficient
of variation was 2%, based on reproducibility of
results for standard samples and duplicate deter-
minations on samples which contained both uranium
and bismuth,

DETERMINATION OF DISSOLVED OXYGEN
IN LUBRICATING FLUIDS

A. S. Meyer, Jr, B. L. McDowell
Analytical Chemistry Division

The determination of dissolved oxygen in lubri-
cating fluids was carried out in a closed system
in which the dissolved or entrained gases in the
fluid were swept with CO, into a 25-ml azotometer
filled with a solution of KOH. The CO, was sup-
plied from a 2-1 Dewar flask filled with dry ice
and sealed with a Hershburg valve. The insoluble
gases collectedin the azotometer were equilibrated
with a solution of potassium pyrogallate, and the
loss in volume of gas was attributed to absorption
of oxygen by the potassium pyrogallate solution.
The loss of volume at standard temperature and
pressure was used to calculate the concentration
of oxygen in the lubricating fluid.

The oxygen content was essentially the same for
each of the lubricants studied; it ranged from a
maximum of 42 ug of oxygen per milliliter of fluid
to a minimum of 15 pg/ml. The coefficient of
variation of the results was 7%.

213
ANP PROJECT PROGRESS REPORT

DETERMINATION OF ZIRCONIUM BY THE
COMPLEXIMETRIC.-YERSENE METHOD

J. P. Young J. R. French
Analytical Chemistry Division

The compleximetric-Versene (disodium ethylene-
diaminetetraacetate) method for the determination
of zirconium was developed by Manning, Meyer,
and White.'® This method consists in adding an
excess of Versene to zirconium in a sulfuric acid
medium. The complex is formed within a period
of 10 min, at room temperature, after the pH of
the solution is raised to approximately 6, and the
excess Versene is then titrated with ferric ion to
a Tiron end point at a pH of 4,8, Very few cations
interfere with this determination, since zirconium
and iron both form very stable complexes with
Versene. This method was used successfully for
several months for the routine determination of
zirconium before difficulties were encountered —
the color change at the end point became sluggish
and indistinct, and the accuracy was found to
be unsatisfactory.

In a further study of the method, it was found
that the color change at the end point could be
greatly improved by digesting the test solution
on a steam bath, after the addition of the Versene,
and titrating the solutions while hot. The accuracy
of the procedure was improved by raising the pH
of the solution to approximately 6 before the ad-
dition of the Versene. A comparison of results
obtained by a gravimetric (mandelic acid) procedure
with the results of 21 determinations by the modi-
fied compleximetric-Versene procedure gave an
average difference of 0.6%.

DETERMINATION OF RARE-EARTH ELEMENTS
IN STAINLESS STEELS AND INCONEL

A. S. Mevyer, Jr. B. L. McDowell
Analytical Chemistry Division

J. A. Norris

Stable Isotopes Research and Production Division

The method previously described'’” for the de-
termination of traces of rare-earth elements in

 

Yop, ., Manning, A. S. Meyer, Jr., and J. C. White,
The Compleximetric Titration of Zirconium Based on
the Use of Ferric lron as the Titrant and Disodium-
1,2-Dibydroxybenzene-3,5-Disulfonate as the Indicator,
ORNL.1950 (1955).

7a. s. Meyer, Jr., and B. L, McDowell, ANP Quar.
Prog. Rep. Dec. 10, 1955, ORNL-2012, p 189.

214

stainless steel was modified to reduce the time
necessary for the deposition of reducible metals
at the mercury cathode from 4 to 1 hr by volatili-
zation of the chromium before electrolysis. The
chromium is removed as the volatile CrO,Cl, by
the addition of solid NaCl during the dissolution
of the sample and the precipitation of the oxides
of silicon, niobium, tantalum, etc. The results
obtained for samples of type 347 stainless steel
indicate that no loss of rare earths results from
this modification of the procedure, The meodified
method has been used for the determination of
lanthanum, cerium, samarium, europium, gado-
linium, and dysprosium in types 304, 316, 321,
and 347 stainless steel and in Inconel and inconel X,
Various electrolysis periods, from 1 to é hr, were
investigated, but no variations were noted in the
recovery of the rare earths,

The recovery of the rare earths for all types of
samples on which the method was used was quan-
titative, within the precision of the spectrochemical
procedure, There are some indications that the
recovery of lanthanum may vary, but the results
are inconclusive at this time. The lowest recovery
of lanthanum was approximately 85%.

DETERMINATION OF OXYGEN IN
METALLIC SODIUM

A. S. Meyer, Jr. G. Goldberg
Analytical Chemistry Division

Further study of the distillation method for the
determination of oxygen in metallic sodium'® has
revealed that excellent reproducibility is obtained
in the recovery of added Na,0 when distillations
are carried out for periods in excess of 1 hr at
temperatures between 800 and 850°F. No differ-
ences in oxygen concentration were found when
a series of replicate determinations of oxygen
were carried out over the extreme distillation con-
ditions of 1 hr at 800°F and 4 hr ot 850°F. Ad-
ditional distillations at higher temperatures have
confirmed the earlier observation that Na, O is
volatilized at temperatures above 900°F., The
exact mechanism by which Na,O is lost from the
system has not yet been determined. Flame-
photometric determinations of alkaline-earth metals
in the distillation residues have shown that the
concentrations of oxides of these elements are

 

187, s, Meyer, Jr.,, G. Goldberg, and W. J. Ross,
ANP Quar. Prog. Rep, Dec. 10, 1955, ORNL-2012, p 186.
not sufficiently large to contribute to the alkalinity
of the residues. '

Efforts to measure the recovery of a weighed
addition of oxide to a batch of sodium metal have
been unsuccessful, When a quantity of NaOH,
which was calculated to increase the concentration
of oxygen by 300 ppm, was added to a previously
analyzed batch of the metal, an increase of oxygen
concentration of only 70 ppm was measured in a
sample after the batch was agitated with helium
for 12 hr ot 1000°F. Since it appears that the
oxygen from solid compounds is difficult to dis-
perse in metallic sodium, a test is to be made
in which the contamination with oxygen will be
carried out by the injection of a measured quantity
of elemental oxygen into a stream of helium, which
will be bubbled through the molten metal.

ANP SERVICE LABORATORY

W. F. Yaughan
Analytical Chemistry Division

Analyses of 39 samples of mixtures of fused
fluoride salts were performed for the Wright Air

PERIOD ENDING MARCH 10, 1956

Development Center (WADC), The determinations
made on the WADC samples included total ura-
nium, ftrivalent uranium, iron, nickel, chromium,
and niobium,

The bulk of the work during this period continued
to be the analysis of fused fluoride salt mixtures
and alkali metals for the Reactor Chemistry and
Experimental Engineering Groups. A total of 1638
samples was analyzed for all ANP sources, and
an average of 3.4 results was reported per sample,
The backlog consists of 77 samples. A breakdown
of the work follows:

Number of Number of
Samples Reported Results
Reactor Chehistry 937 3139
Experimental 420 1904
Engineering
WADC 39 202
Miscellaneous 242 282
Total 1638 5527

215
ANP PROJECT PROGRESS REPORT

10. RECOVERY AND REPROCESSING OF REACTOR FUEL

H. K. Jackson

D. E. Ferguson H. E. Goeller
W, K. Eister

M. R. Bennett R. L. Jolley

F. N. Browder W. E. Lewis

G. |, Cathers J. T. Long

W. H. Carr R. P, Milford

S. H. Stainker

Chemical Technology Division

PILOT.PLANT DESIGN AND CONSTRUCTION

Completion of the engineering design of the pilot
plant for recovering fused-salt fuels has been de-
layed until March 1956, The remaining major items
of equipment, the electrical load center and the
instrument panel boards, are to be delivered about
mid-April, As of March 10, installation of equip-
ment is expected to be 60% complete, with 100%
completion scheduled for May 15, 1956.

ENGINEERING DEVELOPMENTS
Contactor

A percolator type of contactor was chosen for
the fluorinator to be used in the volatility pilot
plant, |In performance tests of the contactor, the
fused-salt fuel mixture was entrained in the product
gas line at a rate of 4.2 g/hr at 625°C and a
sparge gas flow rate of 3.3 scfm (Table 10.1),
which is the probable sparge-gas flow rate for

pilot-plant operation. Analysis of the entrained

TABLE 10.1. ENTRAINMENT OF FUSED SALT
MIXTURE IN PRODUCT GAS LINE FROM
FLUORINATOR VESSEL

Fused salt composition: NaF-Zer (5050 mole %)
Duration of run: 2 hr
Sparge gas: nitrogen

Vessel temperature: 625°C

 

Run Sparge Gas Flow Rate Entrainment Rate

 

No. (scfm) {g/hr)
17 0.0 0

18 1.1 0.75
19 2.2 , 3.13
20 3.3 4.20

 

216

material indicated that 70 wt % was particles of
the fused salt mixture and that 30 wt % was ZrF
sublimate.

4
A vapor trap has been designed for

eliminating entrainment in the product gas line,
but it has not yet been tested. It was also found
in these tests that the flange on the fluorinator,
which will be at 105°C when the fluorinator is
at 650°C, is leaktight against a gas pressure of
12 psig.

Freeze Valves

A freeze valve capable of sealing a ]/’Q-in.-dia
molten-salt transfer line against a gas pressure
of 20 psig was designed. A reservoir of molten
salt is used in one arm of a U tube to seal against
a frozen plug in the other arm. Both the inlet and
outlet arms are vented. In 50 melting and freezing
cycles no leakage was detected.

Several designs of internally cooled freeze valves
were tested, however, on the assumption that rapid
cooling was not effective, and it is now reasonably
certain that failure of earlier designs was the
result of minute fissures caused by shrinking of
the salt upon cooling from its freezing point to
the test temperature,

PROCESS DEVELOPMENT
Fluorination Studies

In further studies on the fluorination step, it was
found that the fluorine-to-uranium mole ratio re-
quired for volatilization of more than 99% of the
UF, could be decreased by the elimination of
impurities but that it was not significantly affected
by the concentration of the uranium in the fused-
salt fuel mixture (Fig. 10.1). The method of intro-
duction of fluorine into the melt and the rate of
flow of the sparge gas had some effect on the
fluorine-to-uranium mole ratio {(Fig. 10.2), but the
results could not be correlated with the known
PERIOD ENDING MARCH 10, 1956

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sl
108 OCRNL-LR-DWG 12641
) : /7 / /
/ 7/
/ /
/ /
80 : /
_ /
3
D 60
N
2
e
5 ,
2 40 s —
w
L
o
20
0
1 2 3 4
Fo /U MOLE RATIO
RUN U CONTENT
NO. MOLTEN SALT (mole %)
1 AS-RECEIVED NoF-ZrF,-UF, (50-46-4 mole %) 4
6 NoF - ZrF,-UF, (50-46-4 mole %), HYDROFLUORINATED
4nr AT 600°C 4
7 PREFLUORINATED NcF-ZrF, (50/46 mole RATIO) PLUS UF,
(76.2 %" URANIUM) a
8 PREFLUORINATED NaF-ZrF, (50/46 mole RATIO) PLUS UF,
(76.2 %% URANIUM), AIR-SPARGED FOR 2hr BEFORE
FLUORINATION q
2 PREFLUORINATED NaF-2ZrF, (50/46 mole RATIO) PLUS UF,
(72.9%" URANIUM) 4
3 PREFLUORINATED NaF - ZrF, (50/46 mole RATIO) PLUS UFy
(72.9 % URANIUM) a
4 PREFLUORINATED NaF-ZrF, {50/46 mole RATIO} PLUS UF,
(72.9%" URANIUM) 5
5 PREFLUORINATED NaF-ZrF, (50/46 mole RATIO) PLUS UF,
(72.9 %" URANIUM) 1

¥ THEORETICAL = 75.8%

Fig. 10,1, Effect of Uranium Concentration and Impurities of the Fused Salt Fuel Mixtures on the
Fluorine-to-Uranium Mole Ratio Required for UF Volatilization. Conditions: 100 ml of F, per minute;
lflé-in.-diu sieve plate on dip tube that introduced fluorine to melt,

217
ANP PROJECT PROGRESS REPORT

 

 

S
ORNL-LR-DWG 12612
100 7 77T 77T
;7] /
S0
80 ya A

 

 

UF6 VOLATILIZED (%)
Mo D M
O O Q
AN \
Z‘.& ~
O
AN
@ N\
N
|
|
|
|

 

 

AN

 

 

 

 

 

\

\
\

 

 

 

 

 

 

 

 

0 1 |
{ 2 3 4
F, /U MOLE RATIO

RUN FLOW RATE (ml/min) GAS DISPERSION DEVICE ON END OF
NO. Fo No 1/,-in.-DIA DIP TUBE

10 100 0 NONE
5 40 200 SIEVE PLATE, 3/, -in.- DIA HOLES
1 100 0 SIEVE PLATE, 34 -in.-DIA HOLES
2 100 200 SIEVE PLATE, %4 -in.-DIA HOLES
3 150 0 SIEVE PLATE, 3/4-in.-DIA HOLES
4 150 150 SIEVE PLATE, 344 -in.~-DIA HOLES
6 300 0 SIEVE PLATE, 3/g4-in.-DIA HOLES
7 100 0 SIEVE PLATE, '/yg ~in.~DIA HOLES
8 100 200 SIEVE PLATE, !/y¢ -in.-DIA HOLES
9 150 0 SIEVE PLATE, Y4 —in.-DIA HOLES
{1 100 0 THREE SIEVE PLATES, 3/g4-in.-DIA

HOLES, '/, in. APART
{2 100 0 PERCOLATOR DRAFT TUBE

Fig, 10.2. Effect of Sparge Gas Flow Rate and Method of Introduction of Flyorine into the Melt on
UF6 Volatilization.

218
variables.  The experiments were performed at
600°C in a 2-in.-dia nickel reactor with a 375-g
charge of NaF-ZrF ,-UF,. In some of the tests
the salt was made by the addition of UF, to NaF-
ZrF, (50/46 mole ratio) that was believed to be
relatively free of oxide impurities as a result of
previous use in a fluorination run, Uranium tetra-
fluoride concentrations of 1, 2, and 4 mole % were
used to study the effect of concentration. In other
tests NaF-ZrF ,-UF, (50-46-4 mole %) was used
as received. Data were obtained by direct sampling
of the salt at intervals during the fluorination.
The curves were extrapolated to the 100% volatili-
zation point for comparison, but usually a sharp
break was observed in the curve between 95 and
100%, which extended the curve to higher fluorine-
to-uranium mole ratios for volatilization of the last
traces of UF .

Volatilization of more than 99% of the UF , from
as-received NaF-ZrF ,-UF, required a fluorine-to-
uranium mole ratio of about 3.3:1, which was re-
duced to about 2.2:1 by sparging with HF for 4 hr
before fluorination. In two tests with the fuel mix-
ture synthesized by adding UF, with a uranium
content of only 72.9% (theoretical, 75.8%) to pre-
fluorinated NaF-Zer, the fluerine-to-uranium mole
ratio required for more than 99% UF , volatilization
was about 2.4:1. When very pure UF,, uranium
assay of 76.2%, was used, the fluorine-to-uranium
mole ratio was 1.4:1, which represents a fluorine
utilization efficiency of about 70%.

Vapor Pressure of the UF ;-NaF Complex

Vapor-pressure data for the UF -NaF complex
were obtained by the transpiration method over
the temperature range 80 to 320°C (see Fig. 10.3).
The reaction involved is described by the equation

(1) UF,-3NaF(s) —> UF(g) + 3NaF(s)

The data are fitted with the analytical expression

log Py, = 1088 — (5.09 x 10%/T) ,

where T is the absolute temperature. Use of the
Clausius-Clapeyron formula with this equation
resulted in a value of +23.2 keal per mole of UF,
for the enthalpy change of Eq. 1.

The data were obtained by passing nitrogen at
a flow rate of 100 ml/min, or less, through a pre-
pared bed of the UF ,-NaF complex at any desired
temperature, trapping out the UF, in the nitrogen

PERIOD ENDING MARCH 10, 1956

oy
3 QORNL-LR-DWG 12613

 

Np FLOW RATE:

© 100 ml/min
o ® 50 mi/min

4 20 ml/min

" log Fym =10.88 - 508 x 10°/7
T c

o

€

£

E i

w

@

)

W

Wy

ul

[ag

o,

e 10

Q

a

g
102
1o
1074

1.6 1.9 22 2.5 28
103 x 17 (°K)

Fig. 10.3. Dependence of UF .NaF Complex

Vapor Pressure on Temperature.

stream in dilute AI(N03)3 solution, and measuring
the total volume of nitrogen with a wet-test meter,
The UF, hydrolysis samples were analyzed by
colorimetric or fluorimetric methods to an accuracy
of better than 15%. Temperature control of the
bed was maintained always to within £0,2°C. The
UF .-NaF complex was prepared by saturating a
30-g bed of 12-20 mesh Harshaw NaF in a 1-in.-dia
vertical nickel reactor with UF , at 100°C.

Crude adiabatic experiments were made with
100-m} batches of NaF to show that the sorption
heat of 23.2 kcal per mole of UF, produces a
large temperature rise, approximately 130°C, if
total saturation with UF, (preheated to about
100°C) is carried out in a period of a few minutes,

Uranium Losses on Desorption of UF ; from NaF

Further evidence was obtained that proper control
of temperature and flow would prevent excessive
uranium losses in the NaF absorption-desorption

219
ANP PROJECT PROGRESS REPORT

pro®®@&te developed for decontaminating UF from
fission-product activity. Successive use was made
of a two-bed NaF system for three complete process
cycles with UF .. The over-all uranium loss was

0.04% in the fjrst bed and less than 0.01% in the

220

second bed. In a single cycle test in the same
equipment with fresh NaF, the uranium losses
were 0.05 and 0.01% in the respective beds, The
percentage loss therefore does not appear to de-
pend on how many process cycles are used.
Part li

SHIELDING RESEARCH
PERIOD ENDING MARCH 10, 1956

11. SHIELDING ANALYSIS

F. H. Murray

C. D. Zerby

Applied Nuclear Physics Division

S, Auslender

H. S. Moran

Pratt & Whitney Aircraft

ENERGY ABSORPTION RESULTING FROM
GAMMA RADIATION INCIDENT ON A
MULTIREGION SHIELD WITH SLAB
GEOMETRY

S. Auslender

The code of a Monte Carlo calculation of energy
penetration and deposition resulting from transport
gamma radiation in a shield of slab geometry! has
been used in a parametric study of a two-region
lead-water shield. The code utilizes straight-
forward sampling techniques, except for a doubling
technique operating on the unscattered flux.

For the parametric study the radiation was 1-Mev
gamma rays incident on the slab at 0 deg, 60 deg,
70 deg 32 min, and 75 deg 31 min. The first region
of the slab was composed of water 1,5 mean free
paths (mfp) thick at the initial gamma-ray energy,

 

Ic. b. Zerby and S, Auslender, ANP Quar. Prog. Rep.
Dec, 16, 1955, ORNL-2012, p 203,

UNCLASSIFIED
2-01-059-47

 

 

 

 

 

 

 

 

 

 

 

 

 

{28 I ] |
FRACTION OF THE TGTAL ENERGY =

>
© . REFLECTED —0.0345
o 112 ——— -———1— ABSORBED —0.7240
5 TRANSMITTED —C.2415
>
£ oeel — — | - _ -
T e L WATER — = LEAD:‘
< o
= a
=
3 i aso — —_—

x
L \

=
83 \
Syess | — —1- —
= o _l._i
z 4
EQo4s —0u — —t
= m o - — \

— e

& § -‘-_"'_‘_'T___ — ; \\
= .

o 0.32 . == o e
S D P =

|
o
LL_ Q46— “_ - -
>
|
O L |
0 0.5 10 1.5 20

Ho s NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE
PATHS AT INITIAL ENERGY

Fig. 11.1. Fraction of Energy from 0-deg Inci-
dent Gamma Rays Absorbed in a Water-Lead Slab.

and the second region was composed of lead 0.5
mfp thick. Preliminary resuvits of the calculation
are shown in Figs. 11.1 through 11.4. The dashed
lines are a fit by eye to the data. The large rate
of change of heat deposition near the front surface
of the lead can be explained by the rapid absorption
in the lead of the low-energy degraded gamma rays.

The coding has been extended to calculate energy
flux and tissue dose rate, as well as energy depo-
sition. Plots of the buildup factor for the dose
rate, from the gamma rays that penetrate the slab,
as a function of the composition of the slab are
presented in Figs. 11.5 through 11.9. The buildup

factor, B,, is defined as:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

D
B r = 5 '
= Hpr sec
D_e
¢
UNCLASSIFIED
2—0i—-059-48
T 144 ‘
& \
im b
‘ . i
wl i
E oo | _ FRACTION OF THE TOTAL ENERGY = |
; ' REFLECTED - Q.0910
g ABSORBED — 0.8578
24| | TRANSMITTED-0OSt2 |
Lel
< ot WATER - *#—"L'F LEAD -]
@ N
£ 0.96 _
S RN
a
SO I
L | i
o 0.80 N I ‘
L) .
g N\
5 0.64 N\
T '_‘—'i — 1
!
8 1
g 048 _ ~ L
- N i
3 N \
g o032 3
= -
5 . \
2 |~ N
5 S
2 _ N\ ——
016 = —T
o \~___‘
& i
5 o | ‘
x 0 05 1.0 1.5 2.0
~ po7» NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE

PATHS AT INITIAL ENERGY.

Fig. 11.2. Fraction of Energy from 60-deg Inci-
dent Gamma Rays Absorbed in a Water-Lead Slab.

223
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-059-49

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

=
g 1.60 |
L
W ;
w4449 WATER tea— | EAD ——mf
a \ !
Ll \ :
= \ | L
r 1.28 -
ul X -
W \ \
O
o A
2 o2 { | , ; -
S FRACTION OF THE TOTAL ENERGY =
@ ! REFLECTED-0.1338 "
> 098 \ ABSORBED ~0.84143
o \ TRANSMITTED —0.0249
¥
w —\
> 080 |- - \\
=
) =\
<I
g 0.64 A\ I
<1
(D —
- \ “
g oas = — R _
e — 1
2 N ‘
z _ e
-4 Q.32 =N | e A
= N \
E —\\.~ \
\
s L~ \ \
g 046 [ e T e -
5 i 1
0
& 0 0.5 10 1.5 20
~

po» NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE PATHS
AT INfTIAL ENERGY

Fig. 11.3. Fraction of Energy from 70-deg 32.
min Incident Gamma Rays Absorbed in a Water-
l.ead Slab.

where
D = calculated dose rate (mr/hr),

D, = dose rate with the source unchanged
but with the slab missing,

por sec @ = oblique thickness of the slab in
mean free paths at the source energy.

The composition of the slab is denoted by P and
defined, at the source energy, as

thickness of lead (mfp)

 

thickness of entire slab (mfp)
(ko) e

Ho
where

Lot = (gr)py + (Fo’)Hzo

224

UNCLASSIFIED
2-04-059-50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.92
| \
I
T L
T 176 — S
L
= \
= 1.60 Y ‘
<
< \ FRACTION OF THE TOTAL ENERGY=
o \ REFLECTED —0.1448
w .44 —\ ABSORBED —0.8375
a TRANSMITTED—Q.0477
3 \
1.28 |——t
3 -
< \
o A4z —— \ —
o
g \
> \
-
0.96 - ~»<l
3 \
' - —— WATER LEAD —
<[
= \ _
= 0.80 \
I \
o
- -\
=
W 064 \ -
2 A\ .
1
&’ c.48 \ { |
— _\ ‘
O
- \— z
b ._\_
© 032 =S iy
z S — \
F N \
I >~ \
xr 016 —_— e T Y ———— - ‘ N\ -
L. \ \
._; —.h--.___W ;\\
1 -
0
0 05 1.0 15 20

Hery NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE PATHS
AT INITIAL ENERGY

Fig. 11.4. Fraction of Energy from 75-deg 31-
min Incident Gamma Rays Absorbed in a Water-

Lead Slab.

The plots of Figs. 11.5, 117, and 11.9 are for
shields with the lead in front of the water, and the
plots of Figs. 11.6 and 11.8 are for water in front
of lead. The sources for Figs. 11.5 through 11.8
are plane monodirectional, The source for Fig,
11.9 is a surface source having a cos™ 6 angular
distribution, where 6 is measured from the normal
to the surface. This source is normalized to one
gamma ray per unit area of surface per 27 steradians
(in the forward direction):

1 = f S(Q) 4Q
2T

n + 1

27

cos” @

S(Q) =
When n = 0, the source is isotropic. For n = o,
the source is plane monodirectional and normal to
the surface.

It is obvious from the plots that lead is more
effective in stratified slab shields when it is
placed behind a good scatterer, such as water,
This is especially true for the lower source ener-
gies. The explanation for this is fundamentally

the same as for the large energy deposition rate

UNCLASSIFIED
2-01-059-%4

 

3.5 ‘
| | A

 

 

 

 

 

8,, DOSE BUILDUP FACTOR

 

 

 

 

 

 

 

 

1.00 0.75 0.50 0.25 0

Fig. 11.5. Buildup Factor for Dose Rate in a
Lead-Water Shield Resulting from 3-Mev Incident
Photons from a Plane, Monodirectional Source.

PERIOD ENDING MARCH 10, 1956

in the front edge of the lead when it is behind
water. The scatter in the answers is, of course,
due to the method of solution. Since the answers
are estimates of the correct answer, any one
problem may be calculated several times, each
If this is
done, the average of the several estimates can be
accepted as the answer, and a standard deviation
can be calculated for it.

time with a new set of random numbers.

UNCLASSIFIED
2-01-059-52

 

3.5

Bor =1mfp

 

 

 

 

2.5

 

 

 

2.0 o

 

8, , DOSE BUILDUP FACTOR

 

  

 

   

 

 

 

 

 

 

 

1.0 -9 2 /Pb
Py r
o0 7‘4
o
0.5
0 0.25 0.50 0.75 1,00

P = (pgrpy /iror AT INITIAL ENERGY

Fig. 11.6. Buildup Factor for Dose Rate in a
Watet-Lead Shield Resulting from 3-Mev Incident
Photons from a Plane, Monodirectional Source.

225
ANP PROJECT PROGRESS REFORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UNCLASSIFIED UNCLASSIFIED
2-01-059-53 2-04-059-54
4.0 T 4.0 f
b 3
| |
Kor =1imfp Hor =imfp |
3.5 35 . .
| |
| |
3.0 | - . e
» L
-
o a
= o
2 o
o < 2.5
a. o
po )
2
S b3
5 =
@ @
§ &
& 8
- -
o @
’
% HZO—F- - H Ob%/ljb//
1.0 i 10— 4 2 /‘
t Y
P#o’fi; Pral
/ s
I
//#io 'ljof |
0.5 0.5 | l
1.00 0.75 0.50 0.25 C 0 0.25 0.50 0.75 1.00
P=Apgrlp, /tigr AT INITIAL ENERGY P=(pgrlpy/Hor AT INITIAL ENERGY
Fig. 11.7. Buildup Factor for Dose Rate in o Fig. 11.8. Buildup Factor for Dose Rate in a

Lead-Water Shield Resulting from 1-Mev Incident Water-Lead Shield Resulting from 1-Mev Incident

Photons from a Plaone, Monodirectional Source. Photons from a Plane, Monodirectional Source.

UNCLASSIFIED
~ 2-01-059-55

  

 

  
   

 

 

 

 

 

e 7= 0 -
x
g A=t
5 n=2
E ‘fl =4 —
o n =8
8 '1 n==
0 !
>
m -
Ll
w
o 4
O
-~ n+1 a
T, 3 2, C088
ANGULAR SURFAGE |7
(2 _ . SOURCE DISTRIBUTION-" |77 , ,
1.1
1.00 075 0.50

P={por)py/ror AT INITIAL ENERGY

Fig. 11.9. Buildup Factor for Dose Rate in a Lead-Water Shield Resulting
from 3-Mev Incident Photons from a Plane cos” 9 Surface Source.

226
S. K. Penny

PERIOD ENDING MARCH 10, 1956

12. REACTOR SHIELD DESIGN

F. L. Keller

D. K. Trubey L. B. Holland

Applied Nuclear Physics Division

C. A. Goetz

H. C. Woodsum

Pratt & Whitney Aircraft
R. M. Davis, Glenn L. Martin Co.

GAMMA-RAY HEATING IN A 300-Mw
CIRCULATING-FUEL REACTOR

R. M. Davis C. A. Goetz

A study of the gamma-ray heating in the lead
and alkylbenzene shield of a 300-Mw circulating-

fuel reflector-moderated reactor (Table 12.1} was
completed. Throughout the calculation, the latest
experimental data from the LTSF mockup of the
circulating-fuel reflector-moderated reactor and
shield (RMR-shield) were incorporated insofar as
possible. The heating in the shield was resolved

TABLE 12,1. PARAMETERS OF A 300-Mw CIRCULATING-FUEL REACTOR USED IN
GAMMA-RAY HEATING CALCULATION

 

Reactor or Shield Region

 

Thickness {in.) Radius (in.)
ane Beryllivm island e 6.700
Sodium passage 0.187 6.887
Inconel-X cladding 0.125 7.012
Core fuel region 5.988 13.000
Inconel-X cladding 0.156 13.156
Sodium passage 0.187 13.343
Beryllium reflector 11.887 25,230
Inconel-X cladding 0.010 25.240
Sodium passage 0.066 25.306
Inconel-X cladding 0.010 25,316
Boron-10 0.200 25.516
Inconel-X cladding 0.010 25.526
Sodium passage 0.066 25,592
Inconel-X cladding 0.250 25.842
Heat exchanger 7.030 32.872
Inconel-X cladding 0.125 32,997
Thermal shield 1.035 34.032
Pressure shell 1.000 35.032
Insulation 1.000 36.032
Insulation Inconel-X shell 0.032 36.064 (91.60 cm)
Alkylbenzene passage 0.375 36.439 (92.56 em)
Lead 1.000 37.439 (95.10 cm)
Alkyibenzene passage 0.375 37.814 (96.05 cm)
Lead 1,76 (rear) 39.574 (100.52 cm)
4.56 (front) 42.374 (107.63 cm)
Alkylbenzene 7.25 {rear) 46.824 (118.93 cm)
14.10 (front) 56.474 (143.44 cm)

 

227
ANP PROJECT PROGRESS REPORT

(1) heating by
primary gamma rays otiginating in or near the
reactor core, (2) heating by fission-product-decay
gamma rays from the heat exchanger region, and
(3) heating from thermal-neutron captures in the
lead and borated (2% boron) alkylbenzene. The
third component was subdivided to take into ac-
count the secondary gamma rays from lead, hydro-
gen, and boron captures. Alpha particles from
thermal-neutron captures by boron were also con-
sidered and were the only source of heating other
than gamma rays included in the calculation. The
results of the calculation are plotted in Figs. 12.1
through 12.6. The methods of calculation are
described below.

into three principal components:

Primary Gamma-Ray Heating

The primary gamma-ray source includes both
prompt and fission-product gamma rays from the

2-01-059-556

 

 

 

 

 

 

 

 

 

 

TOTAL HEATING IN SHIELD (watts/g)

 

 

 

 

 

N 95 99 103 107 11 15 419
REACTOR RADIUS {cm)

Fig. 12.1. Total Heating in the Rear Portion of
the Lead and Alkylbenzene Shield of o 300-Mw

Circulating-Fuel Reactor,

228

core, capture gamma rays from the Inconel core
shell cladding, and capture gamma rays from the
beryllium. The heating in the alkylbenzene and
lead shield that results from these gamma rays
was determined by the procedure outlined below.

The heating in the alkylbenzene was obtained
by use of a recent analysis' of LTSF measure-
ments in which the important source contributions
to the gammo-ray dose rate in water beyond an
RMR-shield mockup were determined for various
lead and water thicknesses. The usual material
and geometry transformations were applied to the
resulting curves to obtain the dose rate in alkyl-
benzene behind various lead thicknesses. The
dose rate was then converted to heat, with the

 

TR. W, Peelle e al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 210.

2-04-059-57

 

i I —
— 1 PRIMARY GAMMA-RAY HEATING.

5| ~——=— 2 HEAT EXCHANGER FISSION-PRODUCT
U GAMMA -RAY HEATING. “‘“'fl’—_

3 HEATING BY THERMAL-NEUTRON |
CAPTURE IN SHIELD., ———A— | -

 

 

 

 

 

   
 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

HEATING IN SHIELD (watts /g)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I

 

9 95 99 10 107 444 s {
REACTOR RADIUS (cm)

-

9

Fig, 12,2, Components of Total Heating in the
Rear Portion of the Lead and Alkylbenzene Shield
of a 300-Mw Circulating-Fuel Reactor.
following expression:

where

Hokyt, Plaliyrrtpy) = Ky — Dp ; pla,zitp,) Hia,z) —

alkyl, P

tqlkyl

PERIOD ENDING MARCH 10, 1956

Ay
2-01-059-58

5 CAPTURES IN LEAD.”

HYDROGEN CAPTURE GAMMA-RAYS FROM
ALKYLBENZENE.

BORON CAPTURE GAMMAS FRCM ALKYLBENZENE.
107! ALPHA PARTICLES FROM BORON CAPTURE. —

3
n

)]

M

2
N

3a+3b + 3¢

HEATING IN SHIELD (watts /g)
n

3
o

[

40—5 3a

 

Toptd
9 95 99 103 107 M 15 149
REACTOR RACIUS (cm)

Fig. 12.3. Components of Heating by Thermal.
Neutron Captures in the Rear Portion of the Lead
and Alkylbenzene Shield of a 300-Mw Circulating-
Fuel Reactor,

r o
c

 

r . 19263
heating by primary gamma-ray sources {w/g),

thickness of alkylbenzene (cm),

thickness of lead (cm),

tactor to convert from dose rate to heat

2.58 x 10~7 (w/g)/(r/hy),

outside radius of reactor core (33.0 cm),

radius to a point in the reactor at which heating is computed (em),

229
ANP PROJECT PROGRESS REPORT

2-01-059-59

n

TOTAL HEATING IN SHIELD (watts/g}
w

~

1o~

 

87 gt 95 99 103 107 " "5 12 123 127 134 135 139 143
REACTOR RADIUS (cm)

Fig. 12.4. Total Heating inthe Forward Portion of the Lead and Alkylbenzene Shield of a 300-Mw Circu-
lating-Fuel Reactor,

D = primary dose rate (r/hr) in the LTSF at a distance z from the LTSF source plate

p,LT(@zilpp) =
with a radius @ (@ = 35.5 cm),

H(a,z) = transformation from the dose rate of a finite disk (radius a) source to that from an -
infinite plane source

E (uz) .

 

Eip2) - E (2)(1 + (a/2)7]'/2

 

 

O’R/O'LT = ratio of reactor to LTSF source plate equivalent surface source strength for neutrons,
o °r
R = ‘
4171'3_
P, = reactor power (3 x 108 w), ’
o PLT
T ma? ' -

230
PERIOD ENDING MARCH 10, 1956

aid
- 2-01-059-60C

PRIMARY GAMMA-RAY HEATING
HEAT EXCHANGER FISSION-PRODUCT
GAMMA-RAY HEATING

HEATING BY THERMAL - NEUTRON
CAPTURE IN SHIELD

HEATING IN SHIELD (watts/g)

 

87 ot 95 9% 103 107 1 115 "o 123 127 134 135 139 143
REACTOR RADIUS (cm)

Fig, 12,5. Components of Total Heating in the Forward Portion of the Lead and Alkylbenzene Shield of
a 300-Mw Circulating-Fuel Reactor,

P,r = LTSF source plate power (3.5 w), 2

c, = change in attenuation that results from replacement of the 4-in.-thick test heat
exchanger mockup with a particular reactor heat exchanger composition and thick-
ness (based on exponential attenuation at 6.8 Mev),

¢, = change in attenuation that results from the addition of small claddings, correct
pressure shell thickness and composition, etc.,
c, x c, = 0.460,

c, = change in attenuation that results from the substitution of alkylbenzene-350 at
330°F, assuming that p ), ). = 0.8 HH 0 based on electron density.

 

2The latest calibration of the new LTSF source plate has yielded a value of 3.5 w for the old source plate
(see Sec. 13).

231
ANP PROJECT PROGRESS REPORT

2-01-059-#61

30 CAPTURES IN LEAD.

3b HYDROGEN CAPTURE GAMMA-RAYS FROM ALKYLBENZENE
3c BORON CAPTURE GAMMAS FROM ALKYLBENZENE.

3d ALPHA PARTICLES FROM BORON CAPTURE.

30+3b+ 3¢

3c

HEATING IN SHIELD {(watts /g)

 

9% 105 115 125 135 185
REACTOR RADIUS {cm)

Fig. 12,6, Components of Heating by Thermal-Neutron Captures in the Forward Portion of the Lead and
Alkylbenzene Shield of a 300-Mw Circulating-Fuel Reactor,

The determination of the primary gamma-ray heating in the lead was made in a slightly different
manner. Since there were no LTSF curves of primary gamma-ray dose rate as a function of lead thick-
ness, it was necessary to use the transformed LTSF curves mentioned above, that is, plots of the
primary gamma-ray heating in the reactor as a function of alkylbenzene thickness for various lead thick-
nesses. 1he lead thicknesses were 0, 1.5, 3, 4.5, and 6 in. By integrating over r from the lead surface
to = infinity, the total heating in the alkylbenzene behind 1.5 in. of lead was evaluated and subtracted
from that behind no lead to obtain the total amount of energy deposited in the first 1.5-in.-thick lead
layer in the reactor, as follows:

 

Hpp pl0 < fpy < 1.5 in.)

 

 

> M L2 > _ : 2
470 4111 ";o Hateyt, platiytifpy = 0) r%dr — dap -£0+1.5 - Haleyt, PUalkyitpp = 1.5 in) r%dr
dmepy 3 _ 3
3 [(ro + 105 ino) - ro]

232
PERIOD ENDING MARCH 10, 1956

where

 

Hpp p(0 £ tp, £ 1.51in.) = average heating by primary gamma-ray sources in the first 1.5-in.-thick
lead layers (w/q),

ro = inside radius of lead (cm),
Pp, = lead density (9/cm®),
Palkyl = alkylbenzene density (g/cm?),

qukyl,P(tulkyl’th) = heating in the alkylbenzene at a distance tolkyl behind a thickness tpy
of lead (w/g).

Repetitions of this procedure for the other lead layers resulted in a histogram of primary gamma.ray
heating in the lead. A smooth exponential curve was then fitted to the histogram.

Fission-Product Gamma-Ray Heating

In the absence of experimental data for the contribution of fission-product gamma rays from the heat
exchanger region, it was necessary to calculate the heating from the fission-product gamma rays. The
results of recent spectrometer measurements of fission-product photons in the rotating-belt experiment at
the LTSF were used in the calculation.!

The general procedure foliowed was to evaluate the heating contribution from each of several ener-
gies and then to perform a numerical integration over the energy range. The energies chosen were 0,10,
0.25, 0.50, 1.0, 2.0, 3.0, 4.0, and 6.0 Mev. The surface source strength (at the outer surface of the heat

exchanger region) of photons of energy E was determined first. The energy deposited at o distance from

an infinite plane with this source strength was then computed, after which a geometrical transformation
was applied to convert to spherical geometry.

The choice of a buildup factor in determining the dose rate from an infinite plane was critical. The
preponderance of Inconel between the heat exchanger and the lead suggested the use of the energy
absorption buildup factor for nickel when the energy deposition at the inner surface of the lead was
computed. (In the actual calculation, the buildup factor for iron was employed, since that for nickel
was not available.}) When the heating at a point a few mean free paths inside the lead was computed,
however, the buildup factor for lead was used. The buildup factor for lead was also applied for the
heating in the alkylbenzene at the last lead-alkylbenzene interface. Water buildup factors were employed
for results at greater distances into the alkylbenzene.

The following expression was used in the calculation of the heating from the fission-product gamma
rays in the heat exchanger:

o) = fEHHX(r,E) dE

 

 

 

 

where
b (E) . SE)
O
HHX(T:E) = o r '—2_ Ba [iznui(E)ti} E] (?Fi(E)ti} '
., (E)
= mass energy absorption coefficient for photons of energy E,
p
Tux = outside radius of reactor heat exchanger (32,872 in.),
0
PO
S(E) = K(E) L(E) ,
w(E)
K(E) = EN,(E) 3.1 x 10'° (fissions/sec)/w,

233
ANP PROJECT PROGRESS REPORT

 

pr(E) = number of fission-product photons emitted per fission per unit energy interval centered
about the energy E,
p R
= v
o ;I
Vi

P, = reactor power (3 x 108 w),
V, = total fuel volume (20.11 %),

/
v, = volume fraction of fuel in heat exchanger (0.274),
)

 

t(E} = total gamma-ray absorption coefficient for photons of energy E,
-~ E)¢t

r HX

HX. ~ulE)t e -
LE) = 1 = %1 ;~HBhx ,

T x pS(EY s

o O

Tyx. = inside radius of reactor heat exchanger (25.842 in.),

I

HX heat exchanger thickness (7.03 in.),
enérgy absorption buildup factor for energy E and 2 (E)t . mean free paths,
7
f dx
x

2 E)t,
z

o
8
-
M
&=
™
k=
~,
I

 

t
—
M
&
™
=
—
Ik

Heating by Thermal-Neutron Captures in the Shield
Heating in Lead by Capture Gamma Rays Produced in Lead. — The heating in lead by secondary

gamma rays from thermal-neutron capture in the lead was calculated by using curves of the thermal-
neuvtron flux in lead which were derived from LTSF curves of thermal flux in water behind various thick-
nesses of lead. The general method of calculation was to construct a series of spherical shells that
were concentric about the point at which the heating was being determined and to sum up the contribution
from each, calculated by means of the following expression:

#a x2 62 , ea—,u,x )
Hpp,s () = KpF 2 |— f f B(r7) Blpx) ——— 2m® sin 0 d6 dx
\p Pb x.l 8] A

 

where
HPb.Sl(r) = heating in the lead by secondary gamma rays produced in the lead (w/g),
K, = 1.6 x 10- 13 w/Mev/ sec,
E = energy of lead capture gamma ray (7.4 Mev),
% = lead capture cross section,

u,/p = mass energy absorption coefficient,

¢(r’) = thermal-neutron flux at point r* where capture occurs,
B{ux) = point isotropic energy absorption buildup factor for lead,
x = distance from point 7 “ to point r where the capture gamma ray is absorbed.

Heating in Lead by Capture Gomma Rays Produced in Alkylbenzene. — A rough determination was
made for the heating in the lead from absorption of hydrogen and boron capture gamma rays produced in
the borated alkylbenzene. Heating contributions from several angles 6 about the point of absorption

234
PERIOD ENDING MARCH 10, 1956

were computed so that a curve of heating per unit angle vs angle could be drawn. An integration over 6
then yielded the desired result. The equation employed was

 

K, /2 X1 o= HX
HPb,Sz(r) = KyE Ec 0 f f H(r ") B(ux) ——2217962 sin 8 d0 dx ,
pp O X4 Arx
where
HPb,SZ(r) = heating in lead by secondary gamma rays produced in alkylbenzene (w/g),
E = energy of boron or hydrogen capture gamma ray (0.5 or 2.2 Mev, respectively),
%_ = capture cross section for hydrogen or boron in alkylbenzene.

Heating in Alkylbenzene by Capture Gamma Rays Produced in Lead and Alkylbenzene. — The analy-

' of the LTSF measurements which gave the secondary gamma-ray dose rate in water beyond an

sis
RMR-shield mockup was employed in calculating the secondary gamma-ray heating in the alkylbenzene.
These secondary gamma rays consisted of capture gamma rays from lead and from the boron and hydrogen

in the alkylbenzene. The heating was determined by means of the following expression:

r
c —-tl/)ll
Halkyl.53 = K1f5f6f7cm_r— ¢ Hla,zpp) ,
where
Hc!kyl ¢ = heating in alkylbenzene by secondary gamma rays produced in lead and alkylbenzene
- P W),
fs = dose rate (r/hr) in alkylbenzene from an infinite plane source of secondary gamma

radiation having the same surface source strength as the LTSF RMR-shield configuration
containing 12 in. of beryllium and 4.5 in. of lead

= DS,LTHTZ in. Be, 4.5 in. Pb, ta”(yl) H(a”tqikyl)'
Ds 1 7B epbrtalkyl)
fa - : : , as tabulated in Table 12,2, taken from work by
D¢ 7(12in. Be, 4.5 in. Pb, £, )

J. B. Dee and co-workers,

 

f; = correction for the differences in composition and thickness between reactor and LLTSF
heat exchangers

(2 -2

Hx,R=%Hx, LT 10:16 e-Z

. ux, (tHx=10-16)

C = correction factor to account for small material differences between the LTSF and reactor,

i

thickness of alkylbenzene cooling layers in the basic lead shield {(em),

TABLE 12,2, CORRECTION FACTOR /g APPLIED TO SECONDARY GAMMA-RAY DOSE RATE
FOR YARIATIONS IN LEAD AND BERYLLIUM THICKNESSES

 

 

 

ls{tportpy)
Lead Thickness (in.)
8 in. of Beryllium 12 in. of Beryllium 16 in. of Beryllium
1.5 4.7 1.6 0.32
3.0 4.3 1.3 0.26
4.5 3.9 1.0 0.21
6.0 3.5 0.76 0.17

 

235
ANP PROJECT PROGRESS REPORT

A, = neutron relaxation length in alkylbenzene cooling layers (determined to be approximately
6.0 cm),
H{a,zpy} = transformation from the neutron dose rate (at the lead surface) from a disk source (LTSF

source plate) to the neutron dose rate from an infinite plane source, taken to be

~84/
V(1 = e PR,

Heating in Alkylbenzene by Alpha Particles Produced in Alkylbenzene. — In the calculation of
the heating in the alkylbenzene from 2.3-Mev alpha-particle production by thermal-neutron capture by
the boron, it was assumed that the alpha particle gave up all its energy at the point at which it was
produced. The alpha heating in the alkylbenzene was determined by means of the following equation:

E'c:qSR(r'IF’b)

qukyha(rltpb) = (2.3 Mev) (1.60 x 1013 w/Mev/sec) ,

Palky!
where

H

i

alkyl, ot pp)

 

EC = thermal-neutron capture cross section for boron in alkylbenzene,
o, T
R c —t,/A
(;SR(r,th) = QSLT(z;th) H(a,z) > —e ! lCmf7 ’
Lt 7

brrlzilpy)

DOSE RATE OUTSIDE THE ART SHIELD
J. B. Dee C. A, Goetz D. K. Trubey

The gamma-ray and fast-neutron dose rates at a
distance of 50 ft from the ART have been calcu-
lated as a function of the thickness of the lead
shield.  The calculations were based on the
recently reported shield design procedure® and on
recent LTSF RMR-shield mockup tests and their
analysis (see Sec. 13, this report). {The analysis
of the L TSF data was based on an effective source
plate power value of 2.1 w.) The dimensions and
compositions used for the ART were taken from
previously reported descriptions.4+3

The gamma-ray dose rate was divided into three
parts: (1) primary gamma-ray dose rate originating
in or near the core, D,; (2) the dose rate from
secondary gamma rays originating in the shield,
D; and (3) the gamma-ray dose rate from the heat
exchanger, D, . The resulting dose rates are

plotted in Fig. 12.7. The details of the calcu-
lation have been published,®

 

3). B. Dee et al., ANP Quar. Prog. Rep. Sept. 10,

4A. P. Fraas, ORNL-1947 op. cit., p 15.
5W. L. Scott, Jr., Dimensional Data for the ART,
ORNL CF-55-11-148 (Nov. 21, 1955).

4. B. Dee, C. A, Goetz, and D, K. Trubey, Gamma-
Ray Dose Rate from the ART, ORNL CF-56-1-181 (Jan.
9, 1956).

236

GAMMA-RAY DOSE RATES AT 50 ft FROM ART (r/hr)

Figo 12.7.
ART,

thermal-neutron flux in LTSF behind a thickness tpy, of lead.

heating in alkylbenzene by alpha particles produced in the alkylbenzene (w/g),

2—01—055-39A

NEUTRON DOSE RATE (= 5.2 x 1072 rem/hr)
IS INSENSITIVE TO LEAD THICKNESS

4 5
LEAD THICKNESS (in.)

Op(50 ft)

 

Gamma-Ray Dose Rate 50 ft from
PERIOD ENDING MARCH 10, 1956

13. LID TANK SHIELDING FACILITY

R. W. Peelle
J. M. Miller
Applied Nuclear Physics Division

W. J. McCool

J. Smolen

Pratt & Whitney Aircraft
D. R. Otis, Consolidated Vultee Aircraft Corp., San Diego
W. R. Burrus, U.S. Air Force

Analysis of the dynamic 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. The analysis is
based on an effective neutron power of 2.1 w for
the old LTSF source plate {since removed). It had
previously been assumed that the power of the old
source plate was 3.6 w, but a tentative calibration
of the new source plate has indicated the 2.1-w
value.

ANALYSIS OF THE DYNAMIC SOURCE TESTS
ON MOCKUPS OF THE REFLECTOR-
MODERATED REACTOR AND SHIELD

H. C. Woodsum

One of the main purposes of the dynamic source
tests in the RMR-shield experiments was to meas-
ure the dose rate resulting from fission-product

 

]Shield Design Group, Pratt & Whitney ot ORNL,

 

 

 

 

 

 

 

gamma rays emitted in the mockup of the fuel-to-
NaK heat exchanger.
sources of radiation in the heat exchanger, a belt
of MTR-type fuel plates? was rotated from the
ORNL Graphite Reactor core hole (with the ILTSF

source plate removed), where the thermal neutrons

In order to mock up the

induced fissions, to a slot between two heat ex-
changer mockup tanks, where the fission-product
gamma rays comprised the source to be studied.
Neutron and gamma-ray dose rate measurements
were made in the water beyond the mockups.

The basic mockup (configuration 17) included a
3-in.-thick lead gamma-ray shield (Fig. 13.1). (The
dimensions of all the components in the mockup
are given in Table 13.1.) Only two changes in
this basic configuration were made throughout
these tests: ]]/2 in. of the lead was removed
(configuration 17A), and ]/2 in. of the boral next

 

 

2R. W. Peelle et al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL.2012, p 210.

2-01-057-66-258

 

 

Fig. 13.1. RMR Shield Mockup with Dynamic Source (Configuration 17).

237
ANF PROJECT PROGRESS REPORT

to the beryllium was replaced with 1 in. of poly-
ethylene {(configuration 17F).

TABLE 13.1. BASIC RMR-SHIELD MOCKUP
{CONFIGURATION 17)* USED IN
DYNAMIC SOURCE TESTS

 

 

 

Thickness

Component (cm)

Air 2.40
Aluminum window 0.95
Air 2.23
Fuel plates 0.20
Air 2,92
Inconel 0.32
Air 2.99
Beryllium 30.48
Boral 5.08
Ni-MaF tank (heat exchanger mockup) 5.08
Air 2.92
Fuel plates 0.20
Air 1.27
Ni-NaF tank (heat exchanger mockup) 5.08
Baral 5.08
Nickel 2.54
Air (distributed) 1.50
Lead 3.81
Water 3.10
Lead 3.81
Total 81.96

A,

 

*Configurations 17B . through 17E had infinite, 2.5-,
1.5-, and 1.25-sec transit times, respectively. {n con-
figuration 17A (infinite transit time) the last 3.81 cm of
lead was removed. In configuration 17F (infinite transit
time) the 5.08 c¢m of boral adjacent to the beryllium was
replaced with 3.81 e¢m of polyethylene.

 

g

|

J'“’ n N(E
D = —
S 2 C(E)

——

The cycle time, that is, time for the fuel belt to
make one complete revelution, was varied as
follows: infinite time (belt not rotating) and 2,
1.5, and 1.25 sec, Shorter cycle times were pre-
cluded by the danger of damage to the rig. The
rotation speed was set by means of o stroboscope,
and the constancy of the speed was checked with
a tachometer.

The results of the experiment were compared
with two calculations of the dose rate; for one of
the calculations it was assumed that all the
fission-product gamma rays were of a single energy,
and, for the other, the spectrum of gamma rays from
the fuel belt, as previously measured at the LTSF,?
was used. The two calculated values agreed, but
they differed by about 30% from the measured
value. It was found that this difference could be
attributed to the dose buildup factor for water
having been used in the calculation; that is, the
buildup factor was chosen as if the lead were an
equivalent thickness, in mean free paths, of water,
Substitution of a new buildup factor for the total
mean free paths (lead and water), based on Monte
Carlo studies of laminated shields, resulted in
agreement between the measured and calculated
dose rates,

Calculated Dose Rate from Fission-Product
Gamma Rays from the Heat Exchanger

Based on 2.7 Mev/Photon

The first calculation of the dose rate behind the
lead-water shield was based on a single gomma-
ray energy and an overage number of gamma rays
per fission. As a first step in arriving at the
proper values for the energy aond the average
number of gamma rays, the dose rate from an
infinite plane, isotropic, surface source with a
spectrum expressed as N(E) was calculoted from
the following relation:

](zfliti) Br(z‘#i‘ti) dE ’

E=0
where
n = number of fissions per second per watt = 3.1 x 1019,
N{E) = number of gamma rays of energy E emitted per fission per Mev,
C(E) = flux-to-dose conversion factor for gamma rays of energy E,

238
PERIOD ENDING MARCH 10, 1956

 

E(Zp;t;) = exponential integral for 2y t. mean free paths in the experiment,® tabulated previously?
as [ (e Vdy/y),
B (Zp;t;) = dose buildup factor for gamma rays of energy E in water through E“iti mean free paths.
The dose rate calculated on the basis of a single gamma-ray energy F “ would then be
n ['(E’)
m =9 CED B Gpg) B Zu,t,)

where ['(E ) is defined as the average number of ‘‘penetrating’’ gamma rays of energy F “per fission and
all other terms apply to the single energy. Then, from setting D equal to D

 

J‘wN—(EzE(ZzE)B(EtE)dE
B0 C(E) 1 y’i i r ‘ui i
INE") =

9 E (g E ) B (St )

 

C(E ")

The value of the numerator was obtained by numerical integration over E, using N(E) = 7.0e~ '*2F gamma
rays/Mev-fission.® For relatively thick gamma-ray shields the characteristic or average energy of the
gamma rays was found to be about 2.7 Mev; hence E” = 2.7 Mev. By substituting the proper values in
the equation, a value of I'(E’) = 0.69 was obtained for a 3-in.-thick lead gamma-ray shield.

If it is assumed that all the gamma rays emitted by a source of strength S, are of 2.7 Mev energy,
the calculated dose rate behind the lead-water shield can be obtained from the following relation:

e..zp.z.Rz.(x,y, £)
(n Do pelt = So cos O(y) dA(x,y) ————Br[z,uiRi(x,y,t)]
4nR2(x,y,t)
Area of belt

where
cos 0 = a term to take into account the longer period of exposure of the center portions of the
belt, than of the outer portions of the belt, to the activating neutron flux,
= 1 ~ (y2/4?) (see Fig. 13.2a),
dA = element of area on the belt {cm?),
.-.Z,u,.R , .
e ' ' = exponential attenuation of 2,7-Mev gamma rays from the belt in the heat exchanger
through the various materials (nickel, sodium fluoride, boral, lead, and water) of slant
thickness R with a gamma-ray absorption coefficient of p. at energy E,
47R? = inverse-square spreading for a point source from point of emission to point of detection,
(Zp,R,) = dose buildup factor for 2.7-Mev gamma rays from a point isotropic source through Zp R,

mean free paths of water,

x,y,t = coordinates,® as shown in Fig. 13.25,

 

3Acfuo|[y, the value of g used in this calculation was the same as that given inref. 4, and is slightly different

from the t for the present experiment,

Tables of Sine, Cosine and Exponential Inteials, vol. ll, National Bureau of Standards, prepared by Federal
Works Agency, Works Project Administration, New York ]940

5J. B. Dee et al., ANP Ouar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 185.

SFor these coordinates, ¢ rather than z is used to avoid confusion with the common usage of z ot the LTSF for
representing the distance from the source plate (that is, from the core hole); z is used as the distance from the core

hole in Figs. 13.1, 13.3, 13.4, and 13.5.

239
ANP PROJECT PROGRESS REPORT

SECRET
2-01-057-66-259

EXPOSED AREA OF BELT

  

:\1_(J,2/g2)

o. DERIVATION OF COS 6.

T~ FUEL-PLATE BELT

 

4. BELT GEOMETRY,

Fig. 13.2. Geometry Used in Calculations of the
Goamma-Ray Dose Rate Beyond the RMR-Shield
Mockups with the Dynamic Source.

If the length of the belt in the heat exchanger region is L and its height is 2a (Fig. 13.2¥), then
Eqg. 1 can be written in terms of x, y, and ¢:

(2) D

v,belt
=L/2 (Lt /t Wrxl+yZ+s2 '
f fx SV ~ (y/a)? e (& Y 1-31[(2;;,1.352./1!)\/x2 + y2 4 tzldx dy
dr(x? + y2 4+ £2)

where ¢ is the thickness of each slab of material. This integral can be separated as follows:

 

) Dy belr
L/2 =(>pt, W24y 442
A ’ Br{(zpzfi/t)\/xz +y2 4+ t2]
" =0 x2 + y2 + tz

 

Then, since

= cos ¢ (see Fig. 13.26) ,

240
PERIOD ENDING MARCH 10, 1956

x dx /.2 2
d(cos ¢) = Byt = sin ¢ do ,

(x2 + y2 + t2)3’/2

 

 

or
dx = sin ¢ m (x2 ¥ y2 +t2) de
and, since '
x
= sin¢ ,

\/x2 -+ y2 + £2
x? + y2 + 2 \/yz + 12

dx = ——————————— d¢p =———dod
cos ¢ cosqu

When terms in ¢ are substituted for x, the second integral of Eq. 3 becomes

: -.'I
[(L/2)/m] Z,ut /1) (\/y2+t2/cos ®) E‘u.t.
dqu< — V2 4y +12>

 

 

 

 

(4)
0 V2 + 12
Now, let
2, ,2 <EH#?
y< o+t = b ,
t
so that
l
Vy2 o+ 12 = b ,
sztz
and let

., L2 ]

(L/2?% + y2 + tzJ

Further, it is assumed that B_is a function of ¥ and ¢ only, which is a slight underestimate of the build-
up. Then the first portion of the integral of Eq. 4 becomes

qu e-b sec ¢ d
(5) —_—
. (bt/Zpt)
Since (bz/Z#iti) is a constant for constant y and ¢, the integral of Eq. 5 becomes

2\ 1\ %o
©) ( t >(—b—>f mbsecd 4y |

0

By = Sin

 

and, since f¢° e—bsec® 44 has been tabulated” as Fl,,b], the total integral, in terms of this
0

 

7). Moteff, Miscellaneous Data for Shielding Calculations, APEX-176 (Dec. 1, 1954).

241
ANP PROJECT PROGRESS REPORT

function, becomes

 

(7) D’y,belt =

5, f“ VU= 0@ B[S vy + 2| Flag, Gpz b + 2] ay
7 0 /y2 + 12

The following constants were used for a numerical integration of Eq. 7:
a = 14in. = 355cm,
t = distance between the belt in the heat exchanger and the detector
= 74.22 cm for this case (Table 13.2),
2ut. = 7.74, as shown in Table 13.2,

— L/2
by = sin” .
(L/D% + y2 + 2
L/2 = 29.5in. = 75 cm.

An upper limit of 45 deg for ¢ is obtained by letting ¥ = 0, and a lower limit of 41 deg is obtained by
lefting ¥y = a. For these two values of y,

 

 

E,u-t-
- iti /———-—yz L
= 7.74 , for y = 0 ,
= 8.58 , for y = a

For ¢o =41 to 45 deg and b = 7.74 to 8.58, Fl¢,,b] is essentially independent of ¢; hence, F[45 deg, &1
was used. (qSG was assumed to be independent of y.)

The values of the various terms used in the numerical integration of Eq. 7 are given in Table 13.3.
The integration gave a value of

D 1.16 x 1074 s, (me/hr) .

v, belt =

TABLE 13.2, NUMBER OF MEAN FREE PATHS FOR MATERIALS BETWEEN FUEL PLATES
IN HEAT EXCHANGER MOCKUP AND DETECTOR ATt = 74.22 em

 

 

 

 

1/p (2.7 Mev), pit (2.7 Mev), -
Component Thickness p, Den sity Mass Ab.sc.orption Number of
{cm) (g/cm>) Coefficient Mean Free
(cm?/g) Paths -
NaF 3.81 1.59 0.0366 0.22
Ni 3.81 8.90 0.0382 1.29
Pb 7.62 11.3 0.0418 3.60
Boral (Al 5.08 2.53 0.0372 0.48
H,0 51.1 1.0 0.0420 2.15 ’
Air 2.8 0.00293 0.04 0.000
2, = 74.22 2pt, = 7.74

 

242
PERIOD ENDING MARCH 10, 1956

TABLE 13.3, VALUES OF VARIOUS TERMS USED IN THE NUMERICAL INTEGRATION OF EQUATION 7

 

 

/ 2
1 = (y/a)* B (b) F[45 deg, b]
y V1 — (y/a)2 \/y2 + 2 b Fl45 deg, 5] B (b) for H,0

 

2, 2
0 1.000 74.20 7.740  1.68 x 10~* 7.38 1.671 x 10~3
5 0.9901 74.31 7.751 167 x 10~* 7.28 .64 x 1073
10 0.9595 74.87 7.809  1.55 x 10~* 7.40 1.470 x 103
15 0.9063 75.71 7.897  1.40 x 10~*4 7.50 1.251 x 10~3
20 0.8211 76.85 8.015 1.27 x 10~* 7.55 1.025 x 1073
25 0.7100 78.30 8.167 1.07 x 10™* 7.65 0.742 x 10~3
30 0.5345 80.04 8.348  0.92 x 10~* 7.82 0.481 x 107>
35.5 0 82.26 8580 0.63 x 10~% 8.40 0

 

Measured Gamma-Ray Dose Rates With and Without Rotation of the Belt

The measured gamma-ray dose rate without the belt rotating (which is analogous to the measured
dose rate in the static source tests) includes

(8) D‘)/,WOR = DP + DC + Dfp,core ’
where
D,y wor = total gamma-ray dose rate without rotation of the belt,

D, dose rate from prompt gamma rays,

i

D. dose rate from capture gamma rays produced in the mockup,

dose rate from fission-product gamma rays from the core.

It

/p,core
The measured gamma-ray dose rate with the belt rotating includes

(9 Dywe =Pp * Dt Pppug + Py core-he +
where
D,y wr = total gamma-ray dose rate with rotation of the belt,
/b HE = dose rate from fission-product gamma rays from the heat exchanger,
Dfp core.yg = dose rate from fission-product gamma rays in the core that were not removed by the
belt.
The term D, is of course the same in both Eqs. 8 and 9, and it is assumed that the term D - is the

same, since the small increase in the number of capture gamma rays from the rotating belt is ignored.
Therefore, the difference between the two measured dose rates is

D - D = (

¥, WR v, WOR Dfp,HE + Dfp,core-HE) - Dfp,core

Since the belt length is much greater than the core-hole diameter, Dfp,core-HE is negligible and can be
ignored; thus

(10) Dywr = Py, wor = Pppue = Pppcore

243
ANP PROJECT PROGRESS REPORT

Calculated Dose Rate from Fission-Product Gamma Rays from the Core Based on 2.7 Mev/Photon

Since Dy belt in Eg. 1 represents the calculated dose rate from the fission-product gamma rays from

the heat exchanger, it can be substituted for Dfp gg in Eq. 10. Then, if the value of D is calcu-

fp,core
lated, the difference between the two measured values (i.e., ny,WR - D’y,WOR) should be equivalent to

the ditference between the two calculated values (i.e., D%beh - Dfp,core)'

In the mockup the core source was a 28-in.-dia disk, as defined by the boral iris in the LTSF core

hole. In the calculation of Dfp core’ 1he source was considered to be isotropic. Hence,
S'I ar?
(1) Dfp,core = 7 El(zflzxi) - E"I 2'uit:!' T+ (_ri:) Br(z'uzii) '
where
§, = equivalent isotropic source strength from 2.7-Mev fission-product gamma rays emitted
from the activated portion of the belt at the core hole {mr/hr),
E,(Zp,t,) = exponential integral for 3yt mean free paths, as defined previously,
Zpt, = total number of mean free paths for materials between the fuel plates at the core hole
and the detector, 3.14 + 7.74 = 10.88 (see Tables 13.2 and 13.4),
a = radius of exposed area of the fuel plates = 35,56 cm,
T = total distance between the fuel plates at the core hole and the detector

= 49,88 cm + 74.22 ¢m = 124 cm (Tables 13.2 and 13.4),

dose buildup factor for 2.7-Mev gamma rays in water for E_pziz. mean free paths = 10.

i

Br(zyz't i)

Self-absorption was neglected, since the source was thin, When the values are substituted in Eq. 11,
it is found that

Dy core = 3:05 x 10765 (mr/hr)

TABLE 13.4. NUMBER OF MEAN FREE PATHS FOR MATERIALS BETWEEN FUEL PLATES
AT THE CORE HOLE AND FUEL PLATES IN THE HEAT EXCHANGER

 

 

 

 

1 p (2.7 Mev), frt (2.7 Mev),

Component Thickness P, Dens;fy Mass Ab'scfrption Number of

(cm) (g/cm®) Coefficient Mean Free

(cm?/g) Paths
U235 0.01 18.7 0.0442 0.008
Al 0.09 27 0.0373 0.009
Ni 1.27 8.90 0.0382 0.432
NaF 2.81 1.59 0.0366 0.222
Boral (Al) 5.08 2.53 0.0372 0.478
Be 30.48 1.85 0.0334 1.885
Inconel 0.32 8.5 0.0382 0.103
Air 8.83

2, = 49.89 2yt = 3.14

 

244
PERIOD ENDING MARCH 10, 1956

W Evaluation of Source Sfrengifl"§"“3”0 and S,

Since D can now be written as

v,belt — D[p,core
(1.16 x 107%5,) - (3.05 x 10~¢5s.) ,

the source strength S  and S, must be evaluated in terms of the gamma-ray dose rate. The source
strength S, of the belt of the core hole is

 

12 . PBnN
1Ak
where
P, = power of the belt (watts),
n = number of fissions per second per watt = 3.1 x 1019,
N = number of 2,7-Mev gamma rays emitted per fission = 0.469,
A = exposed area of the belt = 3970 e¢m?,
K = flux-to-dose conversion factor for gamma rays of 2.7-Mev

energy = 2.45 x 102 (photons/cm?2.sec per mr/hr).
Thus Eq. 12 becomes

§, = 2.2 x 104 Py (mr/hr) .

The source strength S, for the center of the belt in the heat exchanger region is

L'l
(13) S =— S, .
LZ
where
L, = diameter of exposed area of belt = 71 cm,
L, = total length of belt = 400 cm.

Thus Eq. 13 becomes
Se = 3.91 x 10° P (mr/hr)

Determinction of the Power of the Fuel Belt

The power of the fuel belt P, was determined by comparing the fast-neutron dose rate (Fig. 13.3),
the thermal-neutron flux (Fig. 13.4), and the sodium-activation curves for configuration 17 (no rotation)
with the corresponding curves for a similar configuration (3-C) used in the static source tests with the
old LTSF source plate.2 The ratio of the fast-neutron dose rates was found to be 1.65 and that for the
thermal-neutron flux was 1.37. These two values, when averaged with the ratio of the sodium acti-
vations2:8 (1.46), give a final average of 1.49. Thus, the power of the belt? is 1.49 times the previously
assumed effective LTSF source power of 2.1 w, or

Pg = 3.1w .

Comparison of Measured and Calculated Gamma-Ray Dose Rates

The data accumulated above give

D = (1.16 x 10=4(3.91 x 10%)(3.1) (mr/hr) ,

v,belt

 

8. 1. Chapman et al., ANP Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 197.

9This new power value was also determined by the LTSF staff by comparing neutron measurements taken in
water.

245
ANP PROJECT PROGRESS REPORT

2—01-0587-86-260

 

 

 

2—-01-057-66—261

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 T T T
1. { T T
; ‘ '[ CONFIGURATION 174 ]
® CONFIGURATION {7A 7=, 1.5 in OF Pb T=o,15in0F Pb 7
5 0 CONFIGURATION 17B t ==, 30 in. OF Pb _ (IZ_ONFIG%ROA;I%I} iF:be —
i = m _ LD -
HA CONFIGURATION 17C r = 25sec, 3.0in. OF Pb CONFrE;UHAT\ON e 1
! A CONFIGURATION 17D v =15 sec, 3.0in. OF Pb - T=25 sec, 30 in OF Pbi;,,,, B
a CONFIGURATION 17E 7 = 1.25 sec, 3.0in, OF Pb CONFIGURATION 17D b T
5> |0 CONFIGURATION {7F t =, 3.0 in. OF Ph, fin. OF _ T =15sec,30in OF Pb @ =
POLYETHYLENE CONFIGURATION 17E ]
T=125sec,30in.OF Pb .|
= TRANSIT TiME OF FUEL BELT CONFGURATION 17F
10— - = T=o,30in OF Pb,1in. ———Ho
‘j -] e oo AN OF POLYETHYLENE ‘
T . B N. 5 — T - \
\ - g W ‘T = TRANSIT TIME OF FUEL BELT |
3 - v \ A ‘ - |
- \ . _ S 2 g
£ A =
TN B |
e \ER‘—" £
= 2 AR s : %
E \ | &
b \ ‘
x ! (ZD
Lt
& t \ - =
T >
o N\ o
= N . =
o R A 4
E 5 4 s -
> - = o
2 \“{\ 5 ]
|
h : - _
& \\
2 .
A 1
i/ !
Ao
—f
10 SN
~ N 3
AN
5 ——— ' \\
R \ :|
5 l \ 70 20 110 130 150 170
| z, DISTANCE FROM LTSF CORE HOLE (cm)
152 |
80 30 100 110 120 130 . ..
7 DISTANCE FROM LTSF CORE HOLE (om) Fig. 13.4. Thermal-Neutron Flux Beyond RMR-
Shield Mockups with Dynamic Source.
Fig. 13.3. Fast-Meutron Dose Rates Beyond

RMR-Shield Mockups with Dynamic Source.

Dfp,care (3.05 x 10=%)(2.2 x 104)(3.1) (mr/hr) .

Thus the difference in the calculated values is

1.18 mr/hr

D’y,belf Dfp,core

The actual difference in the measured gamma-ray dose rates at z = 130 cm (z is the distance from the
core hole and z = 130 cm corresponds to T = 124 cm), as may be seen from Fig. 13.5, is

4,55 - 3.65 = 0.9 mr/hr ,

which is a factor of 1.3 less than the calculated value.

Calculated Dose Rate Based on Fission-Product Gamma-Ray Spectrum

In order to check the calculations, a separate calculation was made for which the fission-product
gamma-ray spectrum measured at the LTSF2 with the same rotating fuel belt system was used. A

246
2-0t-057-66-262

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

103 — . — — :
[ | 777 T ® CONFIGURATION {7A ==, 15in. OF Pb ——
4 o CONFIGURATION 178 7 ==, 3.0in. OF Pb -

Sr—— A GONFIGURATION 176 7 =2 sec, 3.0in.OF Pb_
. & CONFIGURATION 17D T =15sec, 3.0in. OF Pb __|
m CONFIGURATION 17E T =1.25sec, 3.0in. OF Pb__
2 "\ y T D CONFIGURATION 17F
o | T=wm, 30 in. OF Pb, 1in. OF POLYETHYLENE]
10° TN e e
E — - TIT - LTI oo
-
w
w
8 N
> ? -
<
& i
! 10 b——- [
2 "= —“BACK OF LEAD ==
= - I ]
= - ‘
Q 5
; T = TRANSIT TIME OF FUEL BELT ~ o
2 f— e ‘ T NG |
! : |
i | \
1
70 BO 90 100 "o 120 130 140 {50 160

2, DISTANCE FROM LTSF CORE HOLE fcm)

Fig. 13.5. Gamma-Ray Dose Rate Beyond RMR-
Shield Mockups with Dynamic Source.

integration was carried out over the
The contribution to
the dose rate from gamma rays with energies out-
side this range is negligible, and, by neglecting

numerical
spectrum from 1 to 5 Mev.

it, only a very small error should be introduced.
The method of calculating the gamma-ray ottenu-
ation in this case was the same as that used for
the 2.7-Mev gamma rays. Atfenuations for discrete
energies of 1, 2, 3, 4, and 5 Mev were calculated.
If the attenuation of the gamma-ray dose rate at
a distance z is represented as A(E,z), the total
gamma-ray dose rate at some particular distance
may be calculated as

5Mev N(E) A(E,z) dE

D(z) = C _ 7,
1 Mev K(E)
where
N(E) = number of gamma rays of energy E,
as given by the fission-product-decay
gamma-ray spectrum,
K(E} = flux-to-dose conversion factor for gamma-

ray energy E,
C = constant related to the power per unit
area of the belt,
The integrals for the dose rate both from fission-
product gamma rays emitted in the heat exchanger

PERIOD ENDING MARCH 10, 1956

region (rotating belt) and from those emitted in
the core region (no rotation) were evaluated by this
method, and the difference between the two should
be equivalent to the difference in the experimental
gamma-ray dose rate with and without the belt
rotating, The difference between these more exact
integrations was actually 1.17 mr/hr, which was
almost the same as the 1,18 mr/hr previously cal-
culated on the basis of a single (2.7-Mev) gamma-
ray energy and a slightly different spectrum,

Discussion of Results

in order to resolve the 30% discrepancy between
the calculated and measured dose rates, it was
necessary to investigate the assumptions involved
in the calculation. The most obvious uncertainty
was in the choice of a buildup factor through a
multiregion shield. Since the dose buildup factors
in lead and in water differ by a factor of 2, a
considerable error could be introduced by using
either one for the total number of mean free paths
in the shield. The buildup factor for water for the
specified number of mean free paths was used in
all the calculations. It was thought that the total
buildup would be more characteristic of the water,
since about 50 cm of water (2.15 mfp) followed the
3 in, of lead (3.6 mfp). However, some recent
Monte Carlo calculations by S. Auslender!® indi-
cate that for 3-Mev gamma rays through 8 mfp of
lead and water (4 mfp of lead followed by 4 mfp of
water) the buildup factor is about 30% lower than
the buildup factor through the same number of
mean free paths of water alone. Auslender’s calcu-
lations indicate that a correction should be applied
to the buildup factor that would reduce it to account
for the presence of the lead. Such a correction
would bring the calculations and the experimental
results into closer agreement. Furthermore, if it
were the choice of the water buildup factor alone
that gave a calculated dose rate that was higher
than the experimental dose rate, the discrepancy
should decrease if more water were placed be-
tween the lead and the detector (or the point of
calculation). This theory was tested with another
calculation for a z distance of 160 cm, which
would place about 3.4 mfp of water behind 3.6 mfp
of lead. In this case the calculated gamma-ray

 

‘OShielding Analysis Group, Pratt & Whitney ot ORNL,

247
ANP PROJECT PROGRESS REPORT

dose rate obtained by using a buildup factor charac-
teristic of water alone was 0.26 mr/hr, which is
only about 20% higher than the experimental dose
rate of 0.22 mr/hr.

Thus, since the gamma-ray dose from the fission
products in the heat exchanger of a circulating-
fuel reactor is only about 30 to 40% of the total
dose in the crew compartment, the error incurred
by using the water buildup factor for the total
number of mean free paths is not great for moderate
lead thicknesses. Of course, as shown by these
dynamic source experiments at the LTSF, it would
be better to use buildup factors more characteristic
of the lead and water combination in the shield.

tn the experiment, the gamma-ray dose rate from
fission-product gamma rays emitted in the heat
exchanger region did not change (within the limits
of experimental error) as the transit time was
decreased from 2 to 1.25 sec (Fig. 13.5). This
indicates that the fission-product gomma-ray
spectrum did not change significantly, which
substantiates the conclusions reported previously.?

The thermal-neutron flux increased as the transit
time of the belt decreased (Fig. 13.4). The in-
crease was, of course, due to more of the delayed
being emitted in the heat exchanger
rather than in the core region. Since the average
energy of the delayed neutrons is lower than the
average energy of the prompt neutrons, the delayed

nevtrons

248

neufrofii would, essentially, all have been lost in
the shield if they had been emitted only in the
core. However, when they were emitted in the
heat exchanger, they penetrated the shield and
added to the measured thermal-neutron flux. The
maximum increase was about 50% at the outer
surface of the lead gamma-ray shield. This per-
centage decreased to about 10% beyond about
20 cm of water and to a negligible fraction beyond
50 cm of water.

The discontinuity in the thermal-neutron flux
curves at z = 90 ¢m is due to a change in instru-
ments (from a 3-in. fission counter to a 121/2-in.
BF, counter) at this point. Although both instru-
ments were normalized to the same value in ploin
water, the center of detection of the ]2'/2~En. BF,
counter moved toward the front of the chamber (to
the left in z) as the slope of the resulting thermal-
neutron flux became steeper. It is believed that
this shift in the center of detection completely
accounts for the difference in measurement by the
two instruments.

The fast-neutron dose rate (Fig. 13.3) beyond
6 cm, or more, of water was independent of transit
time, and hence the delayed neutrons did not
contribute significantly to the fast-neutron dose
rate. The 1 in. of polyethylene plastic inserted in
the reflector region (configuration 17F) decreased
the fast-neutron dose rate by 20 or 30%.
PERIOD ENDING MARCH 10, 1956

14, BULK SHIELDING FACILITY

F. C. Maienschein

T. V. Blosser E. B. Johnson

G. deSaussure J. D. Kington

G. Estabrook T. A. Love

K. M. Henry E. G. Silver
W. Zobel

Applied Nuclear Physics Division
A. T. Futterer, Pratt & Whitney Aircraft

GAMMA-RAY STREAMING THROUGH THE NaK
PIPES THAT PENETRATE THE ART SHIELD

T. V. Blosser

Preliminary calculations by the ORNL Shield
Design Group' have indicated that gamma-ray
streaming through the NaK pipes that penetrate
the ART shield would increase the dose rate in
line with the pipes by a factor of roughly 2 x 103
for the original 7-in.-thick lead shield. For the
currently planned 4.3-in.-thick lead shield the dose
rate in line with the pipes would be only a factor
of 40 more than it would be without the pipe pene-
trations. An experimental program has now been
initiated to measure the dose rates in a mockup
of the ART shield and NaK pipes and to determine
the most effective pipe arrangement for reducing
this high leakage.

The experiment will be performed in the aluminum
thermal column tank atop the ORNL Graphite Re-
actor, The gamma-ray source, which will have a
large energy spread, will consist largely of 4.5-
Mev capture gamma rays from the thermal column
graphite plus a spectrum of capture gamma rays
(0.01 to 7 Mev) from a 0.040-in.-thick (12-in.-dia)
cadmium disk placed in the bottom of the tank
(Fig. 14.1). A 0.25-in.-thick boral plate, placed
over the cadmium, which will cover the bottom of
the tank, will prevent a background of capture
gamma rays from originating in the lead shield and
in the water; however, the boral will produce 0.5-
Mev capture gamma rays in the source area, The
boral capture gamma rays produced outside the
source area will be sufficiently shielded by the
4.3 in. of lead and will not cause a disturbing
background. There will be no appreciable attenu-
ation by the boral of cadmium capture gamma rays.

 

p. K. Trubey, private communication, to T, V,
Blosser,

The bottom of the aluminum tank will also produce
a spectrum of capture gamma rays in the source
area.

Various shield and pipe configurations will be
inserted in a two-step square opening in the lead
at the bottom of the tank. The steps of the opening
will be about 2 in. high, the lower section being
12 in. square and the higher section 16 in. square.
Aluminum powder (~0.8 g/Cm3) will be used to
simulate the NaK in the pipes. The following
configurations will be tested:

1. solid lead;

2. 3.8-in.-dia pipe perpendicular to bottom - air-
filled; aluminum-filled;

3. 3.8-in.-dia pipe at a 45-deg angle to bottom -
air-filled; aluminum-filled;

4. ART mockup (Fig. 14.2).

The design of the support for the shield mockup
has been completed, and the materials are on order.

2-01-058-0-4

  
 
 
 
 
 
   

 

12-in.- SQUARE OPENING
FOR DUCT MOCKUP

TGP OF
REACTOR SHIELD~
CONCRETE i S A e e R oy

-in-DIA CADMIUM DISK
BOTTOM OF TANK
Y4 in. OF BORAL

SEPARATIONS ARE
INCREASED TO SHOW PARTS \

   

Fig. 14.1. Thermal-Column Tank Arrangement for
the ART Shield Mockup Experiment.

249
ANP PROJECT PROGRESS REPORT

PATCH TO BE BUILT UP
AS REQUIRED FOR THIS
EXPERIMENT

    

 

  
 

3.625-in.-0D %
0.25-in.-WALL PIPE

ofb
FLREKS
SN
QR %fi,

o —in
-

  

 
    
 

’ 2-1-058-C-5
<7
SELKERN
OSSR
4%%5%
SO N
CINSULATION >

  
  
  
 
 
  
 

-

 

  
 

i ORI OO R
KIS

SRR
B BSOS SN

   

 

 

  

2Y>-in. SCH 40 PIPE

\ (2.875~in. OD x 2.469-in. D)

Fig. 14.2. ART Shield and NaK Pipe Mockup Arrangement.

DECAY OF FISSION-PRODUCT GAMMA
RADIATION

T. A. Love
R. W. Peelle

It was pointed out previously? that information

aMéut*the decay characteristics and the photon
U235

W. Zobel

energy spectrum of the fission products of
for short times after fission is essential in the
design of an optimum shield for a circulating-fuel
type of aircraft reactor, In such a reactor the fuel
is circulated through a heat exchanger, located
within the reactor shield, which thus constitutes
a secondary source of radiation. Preliminary
measurements of the time dependence® of the
fission-product gamma-ray mixture in the range from
5 to 150 sec, as well as some information on the

 

2R, W. Peelle et al.,, ANP Quar. Prog. Rep. June 10,
1955, ORNL-1896, p 203.

3R, W, Peelle et al.,, ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 201.

250

energy spectrum of the mixture,* have already been
reported., An extension of the first experiment to
other time periods is reported here.

The equipment used was described previously.®
Samples of enriched uranium, which weighed about
2, 7, and 16 mg, respectively, were irradiated in
the ORNL Graphite Reactor for periods of 0.8 or
32 sec. Energy calibration of the multiple-crystal
spectrometer was carried out in the usual fashion,
with Hg293, Na22, Zn85, Na24, F20, 7,90% ThC”
and N'6 sources. The time analyzer was calibrated
by using a 60-cyele pulser., To determine the
efficiency of the spectrometer as a function of
photon energy, the absolute strengths of Hg203,
Na?2, Cs'37, Zn%5, and Na24 calibration sources
were determined with the aid of the calibrated
high-pressure ion chamber of the ORNL Radio-
isotopes Control Laboratory. A correction was
applied to this efficiency to take into account the

 

4R. W. Peelle et al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 223.
peak-to-total ratio of the spectrometer. The cor-
rection factor is the inverse of the peak-to-total
ratio.

The data were corrected for counting losses in
the electronic equipment, a correction which in
some coses amounts to as much as 25%. Since,
as before, energy groups were used and since the
efficiency of the spectrometer varies considerably
over a single group, the average efficiency for an
energy group was determined by weighting with
the energy spectrum determined in the experiment
carried out at the LTSF.4 It is expected that this
average efficiency is more correct than that used
in the evaluation of the previously reported datgq,
but it must be realized that, if subsequent experi-
ments show a large variation of the spectrum with
time, a renormalization of the data may be neces-
sary,

The determination of the number of fissions
occurring in a sample depends on the length of
time the sample was in the reactor and on the
thermal-neutron flux available. The error in the
length of time amounts to at least 15%. The
effective thermal-neutron flux available, as well
as the macroscopic cross section of the sample,
was taken to be the same as in the previous
experiment; that is, the value obtained by Moteff®
by using the gold-foil and cadmium-difference
technique was used for the effective thermal-

neutron flux, and thermal cross sections were used
for gold and U233,

The decay rate is shown in Fig. 14,3 as a
function of time after fission for various energy
ranges. It includes a re-evaluation of the previ-
ously reported data. Again, the energy group be-
tween 1.62 and 2.3 Mev was investigated both with
the Compton (two-crystal) and the pair (three-
crystal) spectrometer. |t should be noted that the
agreement between these curves is reasonable if
the large statistical error associated with the pair-
spectrometer data is considered.

Figure 14,4 is a cross plot of the data presented
in Fig. 14.3, and shows the photon energy spectrum

 

3], Moteff, private communication, to R, W, Peelle.

PERIOD ENDING MARCH 10, 1956

as measured 1.5, 10, 100, and 1000 sec after
fission. The results obtained by integrating the
curves of Fig. 14.3 between 1.25 and 1600 sec are
given in Table 14,1. The uncertainty in the total
number of photons per fission and the energy per
fission is probably about £25%.

This work concludes the preliminary analysis
of the first part of the investigation of fission-
product gamma rays, that is, the detailed time
behavior of relatively large energy groups of fis-
sion-product gamma rays, The next phase of the
experiment will concern the detailed energy spectra
at certain time intervals after fission. As was
mentioned above, this phase may have an influence
on the phase just concluded through the weighting
factor used for the efficiency. It is expected that
after the conclusion of this second phase of the
experiment a more detailed report will be prepared
that will cover the entire investigation on fission-
product gamma rays. It is hoped that an expression
relating the gamma-ray flux to the energy of the
emitted gamma rays and the time after fission can
be included in the report that will make possible
calculations of the flux for any energy and time,

TABLE 14.1. MEASURED VALUES OF PHOTON

INTENSITY PER FISSION AND TOTAL ENERGY

RELEASE PER FISSION INTEGRATED BETWEEN
1.25 AND 1600 sec AFTER FIS5SION

 

Energy Photons per Energy per

Range Fission Fission (Mev)

 

Compton Spectrometer

0.28~0.51 0.696 0.275
0.51-1.12 1.103 0.894
1.12-1.62 0.428 0.586
1.62-2.31 0.210 0.412

Pair Spectrometer

 

 

(1.62-2.31) (0.197)* (0.385)*
2.:?—3.5 0.178 0.515
3.5=-5.0 0.038 0.057

Total 2.67 2.92

 

*Not included in total.

251
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
2-01-058-0-6

o —

0.28~0.51 Mev (COMPTON)

1

Pl I
{12 -1.62 Mev (COMPTON
- b

e, o.5h—‘1.g2 Mev (COMPTON)
'\( | | ] | .
X L

DECAY RATE (photons/ Mev - sec - fission )

1.62— 2.3 Mev (PAIR)

1.62-2.3 Mev (COMPTON

-2.3~3.5 Mev (PAIR)

——3.5—-5.0 Mev (PAIR)

 

{ 2 5 10 2 5 10 2 5 100 2 5 10 2 5 10

T, TIME AFTER FISSION (sec)

Fig. 14.3. Fission-Product Gamma-Ray Decay Rates as a Function of Time After Fission for Six Photon
Energy Groups.

252
PERIOD ENDING MARCH 10, 1956

UNGLASSIFIED
2-01-058-0-7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

t | ’
——COMPTON SPECTROMETER
F———PAIR SPECTROMETER
1o ! l——! [
L
_ e |
0 2 -1___._;..__! S
|
I |
c ' —Il | i
£ ; , mmm———= - 11.5 sec
i f i
Lot | S
a . L 10 sec
2 =
= — e ——
~ |
g |
5 10"4 ) Sy o
0.::“_’ | Lo — 10%sec
T
T
|
_5 1
1077 — Sommmme
I
|
I
L 3
T -1 10" sec
10 € C e
1077 y ;
0 1 2 3 4 5 &
ENERGY (Mev)

Fig. 14.4. Histogram of the Fission-Product
Photon Energy Spectrum for 1.5, 10, 100, and 1000

sec After Fission.

253
 

THE AIRCRAFT NUGLEAR PROPULSION PROJECT

AT

THE OAK RIDGE NATIONAL LABORATORY

MARCH 1, 1956

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANP PROJECT DIRECTOR W. H. JORDAN RD
CO-DIRECTOR S. J. CROMER RD
ASSISTANT DIRECTOR A. J, MILLER RD
D. HILYER, SEC. RD
REPORTS
PRATT & WHITNEY AIRCRAFT
A. W, SAVOLAINEN ARE
E.R.DYTKO, ASST. PROJ. ENG. PWA N. J. HARPER ARE
A. GIANGREGORID, ADM. ASST. PWA
H. C. GRAY, DESIGN ENGINEER PWA LITERATURE SEARCHES
A. L. DAVIS ARE
AIRCRAFT REACTOR ENGINEERING DIVISION SUPPORTING RESEARCH
S. J. CROMER, DIRECTOR RD ANP STEERING COMMITTEE W. H. JORDAN
A. B. LIVINGSTON, SEC. ARE W. H. JORDAN, CHAIRMAN A. J. MILLER
E. P. BLIZARD
W. F. BOUDREAU
G. E. BOYD
S. 1. CROMER
W. K, ERGEN
A. P, FRAAS
W, R, GRIMES
ASSISTANT TO DIRE
CTOR G. W. KEILHOLTZ
R. 5. CARLSMITH ARE W. D. MANLY
A. J. MILLER
A M. PERRY METALLURGY
H. F. FOPPENDIEK W. D. MANLY, STAFF ASSISTANT M
H. W. SAVAGE
E. D. SHIPLEY
J. A. SWARTOUT
A. M, WEINBERG
EXPERIMENTAL ENGINEERING CHEMISTRY
H. W. SAVAGE ARE W. R. GRIMES, STAFF ASSISTANT MC
NOTE: THIS CHART SHOWS ONLY THE LINES OF TECHNICAL COORDINATION OF THE ANP PROJECT. THE
VARIOUS INDIVIDUALS AND GROUPRS OF PEOPLE LISTED IN THIS AND THE FOLLOWING CHARTS ARE EN-
GAGED EITHER WHOLLY OR 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,
POWER PLANT ENGINEERING IS ALSO RESPONSIBLE TO ITS DIVISION DIRECTOR FOR THE DETAILED PROGRESS OF THE RESEARCH SHIELDING
A. P. FRAAS ARE AND FOR ADMINISTRATIVE MATTERS, E. P BLIZARD. STAFF ASSISTANT AP
THE KEY TO THE ABBREVIATIONS USED IS GIVEN BELOW.
AC  ANALYTICAL CHEMISTRY DIVISION ~ ORNL
INSTRUMENTATION AND CONTROLS AP APPLIED NUCLEAR PHYSICS DIVISION — ORNL RADIATION DAMAGE
. ic ARE  AIRCRAFT REACTOR ENGINEERING DIVISION — GRNL
BMI  BATTELLE MEMORIAL INSTITUTE G. W. KEILHOLTZ, STAFF ASSISTANT 55

 

 

 

 

 

ENGINEERING DESIGN
H. C. GRAY PWA

 

 

 

 

 

REACTOR PHYSICS
A. M. PERRY ENR

 

 

 

 

 

 

7503
W. F. BOUDREAU ARE

 

 

 

 

 

 

 

BP BENDIX PRODUCTS, DIVISION CF BENDIX AVIATION CORPORATION
C CHEMISTRY DIVISION — ORNL

CT CHEMICAL TECHNOLOCGY DIVISION - ORNL

cv CONSOLIDATED VULTEE AIRCRAFT CORPORATION

ENR  ELECTRONUCLEAR RESEARCH DIVISION ~ 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

MP MATHEMATICS PANEL - GRNL

ORINS 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

sl STABLE ISOTOPE RESEARCH AND PRODUCTION DIVISION — ORNL
$5 SOLID STATE DIVISION — ORNL

USAF UNITED STATES AiR FORCE

WADC WRIGHT AIR DEVELOPMENT CENTER

*PART TIME

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

HEAT TRANSFER AND PHYSICAL PROPERTIES
H. F. POPPENDIEK, STAFF ASSISTANT REE

 

 

 

 

 

REACTOR PHYSICS i
W. K. ERGEN, STAFF ASSISTANT ARE

 

 

 

 

 

FUEL REPROCESSING
H. K. JACKSON cT

 

 

 

 

 

 

 

 

255
 

 

 

QFFICIAL USE ONLY

THE AIRCRAFT NUGCLEAR PROPULSION PROJECT
THE OAK RIDGE NATIONAL LABORATORY

MARCH 1,1956

 

 

AIRCRAFT REACTOR ENGINEERING DIVISION
S,

J. CROMER, DIRECTCR
A, B, LIVINGSTON, SEC.

RD
ARE

 

 

 

 

ASSISTANT TO DIRECTOR

 

 

 

 

el “"’F'

.

BRI Sl T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R. $. CARLSMITH ARE
R. E. THOMPSON ARE
REACTOR PHYSICS POWER PLANT ENGINEERING 7503 EXPERIMENTAL ENGINEERING ENGINEERING DESIGN IMSTRUMENTATION AND CONTROLS
A. M. PERRY ENR A P. FRAAS ARE W, F. BOUDREAU ARE H. W. SAVAGE ARE H. C. GRAY PWA E. R. MANN i
M. WILSON, SEC, ARE P. HARMAN, SEC. ARE J. PARKER, SEC. ARE D, ALEXANDER, SEC. ARE 1. ZASLER ARE V. CUMMINGS, SEC., Ic
L, E. FERGUSON, SEC. ARE
H. W. BERTINI .:};E ADMINISTRATION CONSTRUCTION ENGINEERING
W, E, KINNEY REACTOR DESIGN
R. B. STEVENSON PWA W. L. SCOTT ARE M. BENDER ARE E. VINCENS Ic
M. TEAGARDS ARE W, R. LOCKHART, DRAFTSMAN  ARE M, OVERTON, SEC, ARE D. C, BORDEN PWA M. C. BECKER ic
5. V. PRICE, COMPUTER ARE . J. PLATZ, RECORDS ARE 1. M. CORNWELL PWA C. F. HOLLOWAY Ic
' HYDRODYNAMICS AND THERMODYNAMICS R. C. DANIELS ARE G. C. GUERRANT Ic
7503 AREA
W. T. FURGERSON ARE DEVELOPMENT TEST OPERATIONS :’ j g'si'f\'gsggl( ig’é T. A HERRELL Ic
CONSULTANT M, E, LACKEY ARE F. R. MCQUILKIN ARE - Y-
G. SAMUELS ARE S. M. JANSCH, SEC. ARE E.R.DYTKO PWA W, B, MCDONALD ARE N. J. GEBERT ARE
P, H, PITKANEN, UNIVERSITY OF SOUTH M. W. HINES, COMPUTER ARE R. CORDOVA ARE L. 5. RUSSELL, SEC. ARE D. HARRIS, SEC. ARE . E. LEUTLOFF A DEVELOPMENT
CAROLINA J. J. TUDOR, DRAFTSMAN ARE W. F. FERGUSON ARE K. LANE, SEC. ARE A. MONTGOMERY, SEC. ARE F. L. MAGLEY ARE
‘ V. J. KELLEGHAN ARE C. MANTELL PWA R. G. AFFEL Ic
+J W. 1. NELSON PWA G. H. BURGER I
APPLIED MECHANICS AND 1. M. MILLS, JR. ARE PUMPS TEST FACILITIES ASSEMBLY
STRESS ANALYSIS G. C. ROBINSON ARE COORDINATION W. A. SYLVESTER PWA F. C, HAZZARD GLM
R. D, STULTING ARE A. G. GRINDELL ARE J.W. TUMAVICUS PWA J. W. KREWSON I
R. ¥. MEGHREELIAN ARE C.F. WEST ARE S. M. DECAMP ARE E. STORTO ARE J. R. LARRABEE PA H.J. METZ Ic
C. A. PROAPS, SEC. ARE T.E.CRABTREE ARE W. L. SNAPP PWA J. L. CROWLEY ARE W. R. MILLER IC
L. E. ANDERSON PHA J. J. SIMON PHA x H‘SE@R i&g PUMP DESIGH A, L. SOUTHERN Ic
F. A, FIELD USAF 9201.3 AREA H. C. YOUNG PWA - J. 5. ADD{SON ARE
B. L. GREENSTREET ARE H. C. SANDERSON PWA W, ?.TCODBUBDLE ARE 1. R, CROLEY ARE
D. M. MILLER PWA G. . WHITMAR ARE HEAT EXCHANGERS CORROSION LOOPS -l Y ARE 1, R. JONES Ic
W. J. HARRIS PWA
5. E. MODRE PWA G. W. PEACH ARE 2. E. MACPHERSON ARE R e J. W. STARKEN Ic
D, H, PLATUS USAF W, C. TUNNELL ARE e s ARE C. P. COUGHLEN ARE -E. ARE C. W, WRIGHT Ic
. C. A, D. JARRATT PWA K. M. YOUNG I
D. L. PLATUS USAF I W COOKE owa A, G, SMITH PWA W 1. POND P . M. YOUN
C.H. WELLS PWA, PROCUREMENT - W L H. A
L. H. DEVLIN PWA L.V, WILSON ARE
W. R. DSBORN ARE L. R. ENSTICE PWA COMPONENT TESTS ¥. C. GEORGE ARE
HEAT TRANSFER - H . HOPKINS oA P, G SMITH ARE NUCLEAR AND SPECIAL DEYICES
R. D. SCHULTHEISS ARE INTERNAL J. G. TURNER PWA J. W, KINGSLEY ARE ELECTRICAL DESIGN C. 5. WALKER I
J. W, ADERHOLDT GLM 1. M. EASTMAN BP
D. L. GRAY ARE SPECIAL EQUIPMENT T. L. HUD3ON ARE to
. FOS ARE ELECTRICAL SERVICES
; TDG;FMR, PWA D. STOREY, SEC. ARE A. H. ANDERSON ARE 5. C. SHUFORD PWA
M. M. YARQSH ARE F. LEWIS, SEC. ARE R. CURRY PWA E. M. LEES ARE J. KERR ARE
- M. 1. R. EVANS ARE A, M. SMITH ARE D. L. CLARK ARE B. O, GARRETT PWA
E. MAEYENS ARE D. R. WARD ARE C. E. MURPHY ARE J. Q. NICHOLSON ARE SIMULATOR
SPECIAL PROBLEMS T, H. MAYES ARE §. F. HOWELL ARE P. GREEN Ic
W. B. COTTRELL ARE $. R. ASHTON ARE SPECIAL COMPONENTS TESTING TECHNICIAN GROUP F.P.G
h' QiKDNN "55 g: g: SEQE&EER A D. B. TRAUGER ARE R. HELTON ARE GENERAL DESIGN
. T. ARNWINE ARE A. A, ABBATIELLO ARE
EXTERNAL IN-PILE TESTS G, S, CHILTON ARE €. F. JURGIELEWICZ Pwa
J. M. COBURN ARE J. T, MEADOR PWA
CONSULTANTS #. D. GOOCH ARE A CoR - J. M. CUNNINGHAM ARE N. E. WHITNEY PWA
A, H. FOX, UNION COLLEGE R. 8. CLARKE, WELDED ARE C. W. CUNNINGHAM ARE R.E.DIAL ARE W. I GALYON ARE
W. LOWEM, UNION COLL EGE au T R. A. DREISBACH PWA - R. DUCKWORTH ARE G. R, HICKS ARE
M. AUBUCHON, SEC. ARE . H. DUCKWORTH ARE G. G, MICHELS0ON ARE
R. L. MAXWELL, UNIVERSITY OF P. A GHADT ARE
B.E. BLACK ARE H, FOUST ARE C. A. MILLS ARE
TENNESSEE D. M. HAINES PWA
A, 5, THOMPSON, STUDEBAKER-PACKARD W, H. HOOVER ARE R. H, FRANKLIN ARE C. F. SALES ARE
o . E, A, JAGGERS ARE W. D. GHORMLEY ARE
CORPORATION SPECIAL TESTS
G. F. WISLICENUS, PENNSYLYANIA STATE T. A. KING ARE C. 4. GREEN ARE RECORDS AND PRINTING
coLLEcE J. W, TEAGUE ARE L. P. CARPENTER ARE T. L. GREGORY ARE
T. K. WALTERS ARE M. H. COCPER PWA F. M. GRIZZELL ARE 5. J. FOSTER ARE
L. ¥. LOVE, SPECIAL PWA R. D. PEAK PWA R. A, HAMRICK ARE J, C. CABLE ARE
EQUIPMENT P. P. HAYDON ARE
E. B. EDWARDS ARE CORE HYDRODYNAMICS C. G. HENLY ARE
R. L. HEATHERLY ARE B. L, HOUSER ARE
D. R. WESTFALL ARE W. J. STELZMAN ARE B. L. JOHNSON ARE
T. ). BALLES PWA J.R. LOVE ARE
D. E. MCCARTY ARE
OFFICE SERYICES G. E. MILLS ARE
B. H. MONTGOMERY ARE
J. P. LANE ARE G C. NANCE ARE
D. G. PEACH ARE
H. E. PENLAND ARE
CONSULTANTS R. REID ARE
F. A. ANDERSON, UNIVERSITY OF F. J. SCHAFER ARE
MISSISSIPPI 1. R. SHUGART ARE
J. F. BAILEY, UNIVERSITY OF TENNESSEE C. E. STEVENSON ARE
1. FRENCH, UNIVERSITY OF TENNESSEE A G, TOWNS ARE
R. L. MAXWELL, UNIVERSITY OF C. A WALLACE ARE
TENNESSEE 8. C. WILLIAMS ARE
1. M. TRUMMEL, STATE UNIVERSITY OF G. W. WILSON ARE
I0WA
W. K. STAIR, UNJYERSITY OF TENNESSEE

 

 

 

 

 

256

 

 
 

THE AIRCRAFT NUCLEAR PROPULSION PROJECT
AT
THE OAK RIDGE NATIONAL LABORATORY

MARCH 1, 1956

 

 

SUPPORTING RESEARCH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

W. H. JORDAN
A. J. MILLER
STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT
W, K, ERGEM ARE E. P. BLIZARD AP . R. GRIMES MC ¥, D. MANLY M G. W, KEILHOLTZ 55 H. F. POPPENDIEK REE
CHE METALLURGY
REACTOR PHYSICS SHIELDING HEMISTRY AMALYTICAL CHEMISTRY € G METALLOGRAPHY RADIATION DAMAGE HEAT TRANSFER AND PHYSICAL
W, R, GRIM .D. OPERTIES R
A. SIMON* AP f D.RIE.ECSALDWELL, SEC. - b ARFECKLY SEC 5 agri USAF R. J. GRAY™ M G. W. KEILHOLT?Z 35 i S RESEARCH
SHIELDING RESEARCH 5. R, CUNED e . M. ZARZECKI,* SEC. W, H. BRIDGES M H. F. $0EPE:EI|_ES';HTH sec :EE
R R BATE USAF E.P. BLIZARD® AP A.L, ATCHLEY, SEC. MC 1. MORRISON, SEC. M 5. cro WTR Liatson - K. , SEC.
C. L. BRADSHAW MP R. RICKMAN, SEC. AP RESEARCH AMD DEVELOPMENT R.S. KEGLliSEEY :
R, &, COVEYOU* AP L.. ABBOTT AP SPECIAL PROBLEMS A. 5. MEYER, JR AC GENERAL CORRASION D. F. STONEBURNER M M.V KLAlS 5 HEAT TRANSFER
L. DRESNER" AP b AU ENDER P R, F. NEWTON RD J. P, YOUNG AC E. E. HOFEMAN M E.R. BOYD M HOT LAB FACILITY C. M. COPENHAVER REE
J. H. MARABLE" AP W LavERnD o C.C. BEUSMAN ORING G. GOLDBERG AC R. CARLAMDER PWA B, F. DAY M N. D. GREENE REE
3. W.NIESTLIE USAF P porERN e B E. MINTURN LUSAF B. L. MCDOWELL AC W, H, COOK M B. J. REECE M A, E. RICHT 55 D. P. GREGORY PWA
M. L. NELSON® AP it ar W. J. ROSS AC D. H. JANSEN M C. ELLIS ss H. W, HOFFMAN REE
C.D. ZERBY AP C.D. ZERBY* AP PHASE EQUIL!BRIUM $TUDIES J. R. FRENCH AC 4, W, HENDRICKS M E. J. MANTHOS 3 F. E. LYNCH REE
C. 1. BARTON e 1. E. POPE M T. 5. NOGGLE 53 G. L. MULLER PWA
.. SERVICE M . .D.
CONSUL TANT SHIELD DESIGN L ATCHER e L. R, TROTTER W. B. PARSLEY 55 L.D. PALMER REE
M . E. VALGHAN A R. N, RAMSEY 55 J. L. WANTLAND REE
J. A. NOHEL, GEORGIA TECH. C. E. CLIFFORD* AP R. 1 SHEIL MC RE APPLE he DYNAMIC CORROSION CERAMIC RESEARCH E.§.SCHWARTZ S5 R. M. BURNETT REE
F.L. KELLER AP R. E. THOMA MC D. E' CARPENTER aC ). H. DEVAN M E. D. SIMS §$ J. LONES REE
R. DAYIS GLM R. E. CLEARY PWA "R C.BRYANT ac E. A. KOVACEVICH M L. M. DONEY~ M R.L. MILLER REE
C. A GOETZ PYA A FRIEDMAN Me b c. srva AC o BRany " A. HOBBS,” SEC. M RADIATION METALLURGY G. M, WIKN REE
CRITICAL EXPERIMENTS ? Eéflfism F’:; B. A. SCDERBERG MC L. E. IDOM AC E. J. LAWRENCE M 1. C. WILSON® 5§
A. D. CALLIHAN AP D. K. TRUBEY AP R. E. MOORE MC A. H. MATTHEWS AC M. A. REDDON M . C. D. BAUMANN 5§ PHYSICAL PROPERTIES
M. L. RUEFF,* SEC AP H. C. WOODSUM PWA H. DAVIS WA A.D. WILSON AC C.E.CURTIS M W. E. BRUNDAGE s
e ‘ ' o R. E. MEADOWS MC C. M. WILSON AC FABRICATION A S TAYLOR" M W, W DAVIS 55 S. 1. COHEN REE
P M G, 0. WHITE" M
LID TANK SHIELDING FACILITY R.P.METCALF MC J.H, £CoBS M A " N NKLE 5 WD DONERS K REE
E. DEMSK| Pwa R.W. PEELLE AP FUEL PREPARATION RESEARCH M, R, D'AMORE PHA T.C. PRICE Pw 5. J. CLAIBORNE REE
J.F.ELLIS* AP v sEC ap H. INOUYE M R A T. N. JONES REE
W. J. FADER PWA - GLIDEWELL, SEC. G. M. WATSON MC V. KOLBA SLM CONSULTANTS - Ar WEEKS 55 N
V. G. HARNESS* AP ¥. R. BURRUS UsAF C. M. BLOOD e R. E. MCDONALD Pwa 1. C. ZUKAS 55
LoJ LYNN AP W. J. MccooL PHA D J. SASMOR e SPECTROGRAPHIC AMALYSIS I P Pace " CARBORUNDUM COMPANY
‘e J. M. MILLER AP -3 L T GRAPHITE SPECIALTIES CORPORATION REACTOR EXPERIMENTS
E. R. ROHRER* AP pE cv F. W, MILES MC 1. R MCNALLY* 51 T. K. ROCHE M L INSLEY
E. V. SANDIN ng J.SN.\OLEN PWA W. T.WARD MC 1. A. NORRIS* AND OTHERS sl R. W. JOHNSON o Y MCVAY w. E. BROWNING $5
D. SCOTT, JR. Al : T D. E. GUSS USAF
‘ TR
5. SNYDER PWA B M o CHEMICAL EQUILIBRIA NGNDESTRUCTIVE TESTING 31 S,?EB:,T?SszfggfifiNEi;gf,{nii.r H. L. HEMPHILL 55 FUEL PROCESSING
MK, GHT* AP - R . 5. ,
K ALBRIGHT H. JaRVIS ic L. G. OVERHOLSER MC R. B. OLIVER M STATION b R = H. K. JACKSON* cT
T. B. MILLICAN AP E.£.KETCHEN MC J. W ALLEN M R. P. SHIELDS 58
1. R. TAYLOR AP J. D. REDMAN MC ROMETRY R. W. MCCLUNG M o
B. J. STURM MC MASS SPECTROM 1 K. WHITE M CONTRACTOR SPECIAL PROJECTS CHEMICAL DEVELOPMENT
BULK SHIELDING FACILITY ———— . R BALDOCK* sl R, Q. CARDEN M NORTH CAROLINA STATE UMIVERSITY D. E. FERGUSON® cT
£. C. MAIENSCHERN AP CORROSION STUDIES 1. R. SITES sl Q. E. CONNER M M. T. ROBINSON 55 " G. 1. CATHERS e
.C. W. I MASON M 1. F. KRALSE PWA ‘R
C, BOUNDS, SEC, AP F.KERTESZ MC B. J. WARD M - M, R. BENMETT cT
2 ANND BMI H. J. BUTTRAM mMc FLUX MEASUREMENTS R. M. DUFF er
T. Y. BLOSSER* AP F. A, KNOX MC WELDING AND BRAZING
W, R, CHAMPION LAC 6. F. SCHENCK PWA 0. BINDER s UNIT CPERATIONS
G. DESASSURE AP D. ZUCKER MC P. PATRIARCA M METALLURGY CONSULTANTS WK, EISTER or
G. ESTABROOK AP J. M. DIDLAKE MC R. E. CLAUSING M N. CABRERA, UNIVERSITY OF VIRGINIA PHYSICAL FROPERTIES "1 T. LONG cr
A. T. FUTTERER PWA A, E. GOLDMAN M N, J. GRANT. MASSACHUSE TTS S STANKER o
K. M. HENRY AP PRODUCTION OF PURIFIED FUELS G. M. SLAUGHTER M INSTITUTE OF TECHNOLOGY C. C. MEBSTER 5 " G. JONES, JR cT
E. B, JOHNSON AP B. MCDOWELL H J. L. GREGG, CORNELL UNIVERSITY D.W.LEE cT
J. D. KINGTON AP G. J. NESSLE MC C. E. SHUBERT M ELECTRICAL COMPONENTS - W
5 5 MC W. D. JORDAN, UNIVERSITY OF ALABAMA 1. 5. WATSON cT
T. A LOVE AP M. R, SWAN, SEC. R.G. SHOOSTER M F—1 E.F.NIPPES, RENSSELAER JCF e
E. G. SILVER AP J. P, BLAKELY MC L. C. WILLIAMS M "o . - C. PIGG 55 DESIGN
W Z0BEL AP C.R. CROFT MC s POLYTECHNIC INSTITUTE C. C. ROBINSON WADC
. AR FI A' DOSS MC . W, F, SAVAGE, RENSSELAER H. E. GOELLER* cT
DL KREY e L EORGAN e SURFACE STUDIES POLYTECHNIC INSTITUTE ENGINEERING PROPERTIES F. N. BROWDER T
« E. Md NV SRITH e G. P. SMITH W G. SISTARE, HANDY & HARMON R. P. MILFORD ot
B oS v 3 TRUITT MC "C R BOSTON " F.G. TATNALL, BALDWIN-LIMA-HAMIL TON 0. SISMAN® 55
D. SMIDDIE Ic R. K, BAGWELL MC J. V. CATHCART M E.C. WRIGHT, UNIVERSITY OF ALABAMA ?.gi.Mcgggbe :g PILOT PLANT
G G.STOUT ic 1. P, EUBANKS MC H. W, LEAVENWORTH PWA M. T HORGAN b H. K. JACKSON® c1
H. WEAYER AP W, K. R, FINNELL MC J. J. MCBRIDE M : W, H. CARR cT
J. W. WAMPLER AP B. F. HITCH MC G. F. PETERSEN M W. H. LEWIS* cT
W, JENNINGS NC M. E.STEDLITZ M
TOWER SHIELDING FACILITY G. A. PALMER MC K L. GRIFFITH M
B THOMAS M L L HALL M METALLURGY CONTRACTORS
C. E. CLIFFORD AP R. G. WILEY MC BATTELLE MEMORIAL INSTITUTE
E. McBEE, SEC. AP MECHANICAL PROPERTIZS —1 BRUSH BERYLLIUM COMPANY
L UL ap MOLTEN SALT THERMODYNAMICS SLENN L MARTIN COMPANY
F. J. MUCKENTHALER AP F. F. BLANKENSHIP MC D. A. DOUGLAS M METAL HYDRIDES, INC.
G. RAUSA GLM 1. BLANDER MC C. W.DOLLINS M NEW ENGLAND MATERIALS TESTING
F. N, WATSON AP 5. CANTOR MC C. R. KENNEDY A LABORATORY
E.C. CARROLL < B. H. CLAMPITT MC J- R MEIR, IR. M RENSSELAER POLYTECHNIC INSTITUTE
I D. CONNER AP 5. LANGER MC J- ¥. MOODS M SUPERIOR TUBE COMPANY
J. N, MONEY AP M. B. PANISH MC K. #. BOLING a UNIVERSITY OF ALABAMA
G. G. UNDER¥OOD Ic L. E. TOPOL s 4. T. EAST M UNIVERSITY OF FLORIDA
R.E. WRIGHT AP il DG- Tfl%’” ’L: UNIVERSITY OF TENNESSEE
o . YIT RPORATION OF AMERICA
SUPPORTING STUDIES 3. MCNABE, R " (TRO CO *PART TIME
CONSULTANT P. A. AGRON c té. E s&m@s, JR. M
J. E. SUTHERLAND c K. M
H. A. BETHE, CORMELL UNIVERSITY o W, WALKER "
CONTRACTORS CONSULTANTS INSPECTION
METAL HYDRIDES, INC. A INSLEY A. TABCADA M
NUCLEAR DEVELOPMENT CORPORATION - N McYAY R. L. HEESTAND PWA
OF AMERICA R. M. EVANS M
E. B. PERRIN M
CONTRACTORS

 

AMES LABORATORY

BATTELLE MEMORIAL INSTITUTE

CARTER LABORATCRIES

A, R.NICHOLS, SAN DIEGO STATE

COLLEGE
UNIVERSITY OF ARKANSAS

 

 

 

 

 

 

 

 

257