ORNL-2274.txt

 

 

2274, Parts 1=5

 

1
OAK RIDGE NATIONAL LABORATORY LIBRARIES

aec Reseagc anp | WLANEARALL

3 445k 05kLL29E 7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I
Uit

  
     
 
  
  
  
   
  

L2730

UATH

AIRCRAFT NUCLEAR PROPULSION PROJECT

TAT.E
ILNITIALS

QUARTERLY PROGRESS REPORT

" .:U_f;.—%T_"f

FOR PERIOD ENDING MARCH 31, 1957

o] DO 1979 RITIEW OB
DECLASSIFIED BRPORTS

e This Dooument 1s Properly Declasaifiod.

3 Roviewed by P. g, Bak

ORNL Classification Qffyses ! }; 1979
=t

  

OAK RIDGE NATIONAL LABORATORY

- OPERATED BY
UNION CARBIDE NUCLEAR COMPANY

A Division of Union Carbide and Carbon Corporation

POST OFFICE BOX X * OAK RIDGE, TENNESSEE

 

 

| Features of Aircraft Reactors™
I_,!'

M

-
 

‘
' 3 | .;I

il

ORNL-2274, Parts 15
C-84 — Reoctors-Special Fectures of Alrcraft Reactors

This document consists of 338 poges.

Copy f/f of 254 copies. Serims A,

Contract No, W-7405 eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT

For Periad Ending March 31, 1957

W, H. Jordan, Director
3, J. Cromer, Co-Diractor
A. J. Miller, Assistont Directar

DATE ISSUED

JUL 51957

whief, Declassification Branch G!EE
ODAX RIDGE HATIONAL LABORATORY ” s T .
Oporoted by
UNION CARBIDE NUCLEAR COMPANY
A Divigion of Union Corblde and Corbon Corporatien
Post Qfflce Box X
Dak Rldge, Tannessee

 
ORNL-528
ORNL-629
ORNL-768
ORNL-858
ORNL-9219
ANP.60
ANP-65
ORNL-1154
ORNL-1170
ORNL-1227
ORNL-1294
ORNL-1375
ORNL-1439
ORNL-1515
ORNL-1556
ORNL-160%
ORNL-1649
ORNL-1692
ORNL-1729
ORNL-1771
ORNL-1816
ORNL-1864
ORNL- 1896
ORNL-1947
ORNL-2012
ORNL-2061
ORNL-2106
ORNL-2157
ORNL-2221]

 

Reports previously issued in this series are as follows:

Period Ending November 30, 1949
Period Ending February 28, 1950
Period Ending May 31, 1950
Period Ending August 31, 1950
Period Ending December 10, 1950
Period Ending March 10, 1951
Period Ending June 10, 1951
Period Ending September 10, 1951
Period Ending December 13, 1951
Period Ending March 10, 1952
Period Ending June 10, 1952
Period Ending September 10, 1952
Period Ending December 10, 1952
Peried Ending March 10, 1953
Period Ending June 10, 1953
Period Ending September 10, 1953
Peried Ending December 10, 1953
Period Ending Mosch 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
Period Ending March 10, 1956
Period Ending June 10, 1956
Period Ending September 10, 1956
Period Ending December 31, 1956

el
02 e Erd L0 B

PO EMUr -0 MOV ECOITANDPTNTEMTOEOMMNOTND

= -

CERPIMOM> X P D

 

INTERNAL DISTRIBUTION

. G, Affel

. J. Barton

., Bender

. 5, Billington

. F. Blankenship
. P. Blizard

. J. Borkowski

. F. Boudreau

. E. Boyd

. A. Bredig
. Ju Breeding

. E. Browning

. R. Bruce
. D. Callihan
. W, Cardwell

. E. Center (K-25)
. A, Charpie

. L. Clark

. E. Cliftord

H. Cocbs

. B. Cottrell

J. Cromer

. 5. Crouse

. L. Culler

. R. Cuneo

H. DeVYon
. M, Doney
. A. Dougles
. R. Dytke
. K. Eister
. B. Emlet {K-25)
. E. Ferguson
. P. Froas

H. Frye, Jr.
. T. Furgerson
. J. Gray
. T. Gresky.
. R. Grimes
. G. Grindel!
., Guth
. S. Harrill
. E. Hoffman
. W, Hotfman
. Hollgender
. 5. Householder
T. Howe

--'F g

ORNL.-2274, Ports 1=5
C-84 — Reactors-Special Features of Aircraft Reactars

47.
48.
49.

2.
52.
33+
54-55.
56.
57.
8.
59,

61.
62.
63.

63.

67.
68.
6%,
70.
71.
72.
73.
/4.
73,
76.
77.
78.
79.

8l.
82,

86.
87.
88.
89.
90.
21.
Q2.
93

DAPITIZPIO-0PN DN QT -MACDPDOT-EMrMENADIDDIT - IOOXE

. Jordan
. Keilholtz
. Keim

. Keller
. Kelley

—Sr U =TI

. Kertesz
. J. Keyes

A, Lane

. B. Lindauer

. 3. Livingston

. N. Lyon

. G. MocPherson
. E. MacPhersen
. C. Maienschein
. D. Manly

. R. Mann

A. Mann

. B. McDonald

R. McNally

. R. McQuilkin

. V., Meghreblion
. P. Milford

. J. Miller

. E. Moore

G. Morgan

. Z. Morgan
. J. Murphy
. P. Murray (Y-12}
. L. Nelson
. J. Nessle

B. Oliver

. G. Overholser
. Patriorca

K. Penny

. M. Perry
. Phillips

C. Pigg

. M. Reyling
. E. Richt

T. Robinson

. W. Savage

. W, Savelainen
. D. Schultheiss
. Scott

o
£

L. Scott
D. Shipley
 

-
w

4. A. Simon 107. J, C. White
@5. O. Sisman 108. G. D. Whitman
Q6. J. Sites 109. E, P. Wigner (consultant)
97. M. J. Skinner 110, G, C, Williams
98. A, H. Snell 111, J. C. Wilson
9. C. D. Susano 112, C. E, Winters
100. J. A, Swartout 113. W. Zobel
101. E. H. Taylor 114-116. ORNL - Y-12 Technical Librory,
102. R. €, Thoma Document Reference Section
103. D. B. Trouger 117-124. Laboratory Records Department
104. D. K. Trubey 125. Lobeoratory Records, ORNL R.C.
105. G, M, Watsen 126—-128. Central Research Library
106. A. M. Weinberg
EXTERNAL DISTRIBUTION
129, Aerojet-General Corporation
130. AF Plont Representative, Baltimore
131. AF Plant Represenfative, Burbank
132. AF Plant Representative, Clifton
133, AF Plont Representotive, Long Beoch
134-135. AF Plont Representative, Marietta
136-138. AF Plont Representative, Sonta Monico
132--140. AF Plant Representative, Seattle
141. AF Plant Representative, Wichito
142. Air Material Command
143. Air Reseorch and Development Command (RDGN)
144. Air Technical Intelligence Center
145-147. ANP Project Office, Fort Worth
148. Albuguerque Operations Office
149. Argonne Notioncl Laboratory
150. Armed Forces Special Weopons Project, Sandia
151. Armed Forces Special Weapons Project, Washington
152. Assistant Secretary of the Air Force, R&D
153—158. Atomic Energy Commission, Woshington
159, Atomics International
160. Botrelle Memorial Institute
1561-162. Bettis Plant (WAPD)
163. Bureau of Aeronautics
164. Bureau of Aeronautics (Code 24)
165. Bureau of Aeronavtics General Representative
166. Chicago Operations Office
167. Chicago Patent Group
168. Chief of Naval Research
169. Convair-General Dyromics Corporation
170. Curtiss-Wright Corporation
171. Engineer Research and Development Laboratories
172--175. General Electric Company (ANPD)
176. General Nuclear Engineering Corporation
177.
178.
179-180,
181.
182.
183.
184.
185.
186.
187.
188-190.
191.
192,
193.
194,
195.
196.
197.
198-201.
202.
203.
204,
205.
206.
207,
208.
209.
210,
211~228.
229-253.
254,

i
-—c. |

-

-

Glenn L. Martin Company

Hartford Area Dffice

Headquarters, Air Force Special Weapons Center

ldaho Qperaticns Office

Knells Atomic Power LLaboratory

Lockland Area Office

Los Alamos Scientific Laboratory

Marquardt Aircraft Company

National Advisory Committee for Aeronautics, Cleveland
National Advisory Committee for Aeronautics, Washington
Naval Air Development and Material Center

Noval Research LLoboratory

New York Operations Office

Narth American Aviation, lnc. (Missile Development Division)
Nuclear Development Corporation of America

Office of the Chief of Naval Operation (OP-351}

Potent Branch, Washington

Patterson-Moos

Pratt & Whitney Aircraft Division (Fox Project)

San Francisce Operations Office

Sandia Corporation

School of Aviation Medicine

Sylvania Electric Product, Inc,

Technical Research Group, New York

LISAF Heodquarters

USAF Project RAND

U.S. Novol Radiological Defense Laboratory

University of Californic Radiation Laboratory, Livermore
Wright Air Development Center (WCOS|-3)

Technical Information Service Extension, Oak Ridge
Division of Research and Development, AEC, ORO
 
 

FOREWORD

This quarterly progress report of the Aircraft Nuclear Propulsion Project at QRNL
records the technical progress of the research on circuloting-fuel reactors and other ANP
research ot the Loboratory under its Contract W-7405-eng-26. The report is divided into
five mojor parts: ). Aircraft Reactor Engineering, 2. Chemistry, 3. Metallurgy, 4. Redia-
tion Domage and Fuel Recovery and Reprocessing, ond 5. Critical Experiments and
Reactor Shielding.

The ANP Project personnel are engaged in many phases of reseorch directed toward
the achievement of nuclecr propulsion of circraft. A considercble 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 o circuleating-fuel type of reactoer.

The design, construction, and eperation of the Aircraft Reactor Test {ART), are the
current objectives of the project. The ART is to be a power plant system that wiil in-
clude o 60-Mw circuloting-fuel reflector-moderator reoctor ond adequate means for heat
disposal. Operation of the system will be for the purpose of determining feasibility and
for studying the problems associated with the design, construction, and operation of a

high-power circulating-fuel reflector-moderated aircroft reactor system.

k3

i
i

ey

vii
 

 
FOREWORD
SUMMARY

L1

1.2,

1.3.

1.4,

 

CONTENTS

PART 1, AIRCRAFT REACTOR ENGINEERING

AIRCRAFT REACTOR TEST DESIGN. ... e e

Applied Mechanics and Stress Analysis. ..o
Tests of ART Siructural Design Criteria. oo e
Thermal Stress Analyses of Shells. . e e e s
Design Analysis for Creep Bending ... e
Radiator Stress Analysis Problems ...,
Core-Shell Thermal Shield .o e e e e et e

Core HydrodyMOmigs .o i e o et e e et e et e e e s £t one e eh et e et et e e £t e
ART Shi@lding oot et et et ettt et ettt ettt et et st e eb £ ekt ee et et e et e e ere e

ART INSTRUMENTS AND CONTROLS ..
Flow Decay in ART Primary NaK Syslem Followmg Power Loss ..

Liquid-Metal-Level Transducers ..ot et e en et aene e
-0 Level Probes oo e e e e e e e e e e et

Fuel-Expansion-Tank Level Indicotor ... e vt cree e e e e e ean
DM Cr e @ ve] T 818 (e e e s e et
SO C-L Vel TBEES it et s et et eh e b e e e e e en e s e en e e en e eee e 2 e

Pressure Transmitters for Use at High Temperatures oo
MOgretic Flowmeters ..ot e et e e e s e e bt i s s e
Remotely Adjustable Pressure Regulators. .o e e e e e e e
Turbine Flowmeters for Use at High Temperatures ... e

AR T Thermocouples e ettt e e
COMPONENT DEVELOPMENT AND TESTING L. e e et e e e e e

Pump Developmeant TeSts ..o et e e et s ee e e e e e annians

Bearing and Seal Tests ...
Fuel Pump Development Water Tesrs

Fuel Pump High-Temperature Performance T SIS o e

Fuel Pump Endurance Tests ..
Sodium Pump Development Wmer Tesls :
Sodium Pump Barrel Seal Test ..

S0dium Pump Endurance Tests oot mnie e e et vt eevn e eas b st epe e e
Primary NoK Pump Development Tests et e
Auxitiary NoK Pump Development Water TeSIS .o e e e e e

Reactor Component Development Tests ..
Heat Exchanger and Radiator Development Tests
Woter Flow Tests on Aluminum North-Head Mockup

Dump Volve Development Tests i e i ers ot s et s b e ernabe s e

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

---------------------------------------------------------------------------------------------------------------------------------------------
1.5.

1.6,

 

Ersteritn Shall THermal SRTBIITI TBEE s s o il e o
Sadiym Tircuid Water Flow TeBlE o v oo s s sass s s i i s a

Auxiliory Component Development TeSts .o i i i i et s e b 2o e s o
High-Frequency Thermal-Cycling Water Tests o s s,
HighFigieney LhatlaCueling TESELAED wommmmnmmmemmms e s s s
Cold Ticos and PLuGhing It iBBEEE ez s xooms sy om0 o sy s o i s s
L gii - Matal Vopor T on B SBIE: s s s s o e oA 0 1 e S ST EAN 54 A SRS A
Tivconitim - Blugride Vopol T ROPE s oo i v s s s o e s s s R S s e

Tnsulotion Toms b v s S S i A S B S S B T S L S B 0

REACTOR AND FACILITY CONSTRUCTION ..o iiiiitie it e cee e e et ene ettt be ey snss e b oo
ART.ETL Reaétor Canstrottion and Asaembly . . cnmmmmmimmmmmi i ki s as s sssassims s

North-Head Assembly ...

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

Reflactor-Modarator Aasembly:cs e o s s i s s i s S i g b s
lstond and Pressore: Shell Liner.... s sm s A
Fuel-to-Nal Heat Exchangers ... e e vt st s e s e e s
Assembly Methods Devel OPment ..o i e e e st e e e s

ETU Facility «oooovrreerre.

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

Toast Toll Compoments s s s s o i s S S s e S M B B et

AET Cacilityonusssmnaam
Test Facility oo
Disassembly Facility ...

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

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

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

AR, B LU AN IRGELE R AT RS o cmmmummomnosnsmms o e s R S S
AR T I DBt TN P A TR cosnccmns s o s i 8 S A S e RS O R85

ETU Operation Planning......

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

ART Disassembly ProC@dures . eieeiis et e eriesaears e stes sba s s1s st beassesren sees enssemsesneaeres sras eeee e

In-Pile Experimentation .....
In-Pile Loops....cceen...

Frototype In-Pile Pump

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

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

iR el LaBn e ERCHOGET TaB By oo e sty sos 60 samissi e s s O B St
liak thal T R PG HERIOE. DR TTEN S s issssmisints comass o o a5 S RS o R B e S B s

¢ HEAT TRANSFER STUBNES o s o o mms satos oo nssa iess Si t es e hS r emsvi i

ART Fuel-to-NakK Heot Exchanger woiiiimuiii viniiitinimi i s i s inssasio sssiveivaissivsissoss

ART Hydrodynamics ..........
Core Hydrodynamics ...

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

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

Instantaneous Velocity Profile Meusurements ................................................................................

Heat Tronsfer in Reflector-Moderated Recctor o ores . oo e et eee e et e e e ere s e e e e e e e

ART Core Heat Transfer

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

Vortex Tube Heal TronSler uems o ms i s B s fa fa e s e i S T i ST S

Thermol Cyeling Experiments csin s et oo e ien B ras sanmss some s mpmsas s sssas nsompns

Physical Properties............
Enthalpy and Heat Cnpnciiy

Yiscosity ..

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

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

Thermal Canduchwty

Electrical Conductivity

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

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

50
51

53
53
55
57
&0
61
62

65

635
65
65
b6
56
&7

67
69
70

/0
70

73
73
74
75

75
75
75
76
76

7

79
79
80

81
81
82

82

87
87
87
88
88
 

PART 2. CHEMISTRY

. PHASE EQUILIBRIUM STUDIES. .. i i et e cienees s e O
Modification of Apparatus for Thermal Anolysis. ..o i . 9

Visual Observation and Differentiol Thermal Anclysis of Small Quentities of
Fused Solts oo e e e e e e e et e e e e e e e 91

The System LiF-UF ..o e s s G2
The System NaF-HIF o e 93
The System LiF-Cef i et s e 93
The System KE-Bef, .o s 93
The System NaF-RbF-Bef, .ot e, 96
The System NaF-LiF-UF, e 96
NaF-ZrF ,-Steuctural Metal Fluoride Systems oo 96
Opticol Properties and X-Roy Diffroction Patterns for Compounds in Fluoride Systems .............. 99

Subcontract Work of Mound Laborotory and ot Bottelle Memorial lastitute oo, 101

. CHEMICAL REACTIONS IN MOLTEN SALTS oottt et ettt e eanr e 102
Equilibrium Reduction of Niff, by Hy in NaF-ZeF e 102
Solubility of NiF, in NaF-ZrF o e 103
Stability of Chremium Flyorides in LiF-ThF, ond LiF-BeF,-ThF, Mixtures ..o 103

Reduction of UF‘1 by Structural Metals....... o et e 104
Reduction in Molten RbF-ZrF, Mixtures ... s s s 104
Reduction in ThF -Bearing Melts. ... e 103

Measurements of Solubility of Rore-Earth Fluorides in Molten NoF-ZrF -UF, Mixtures ............ 105
Solubility of CeF g i 104
Solubility of Li::F3 . VDD ORGP P UT U COR A PROPUPPOTPRURPN |
Solubilities of CeF, c:-nd LaF in Combmu?lon e ettt e e e e e e senene e 100
Solubility of SmE,.... . et ettt et et e s e snennnina e 1O
Etfect of BaF, or SrF on SO'UbIIITY of CeF et te ettt et 108

Solubility Relotionships Arnong Rare-Earth Fluorides in Molten NoF-ZrF -UF, .0 109
Solubility of Individual Rare Eorths ... e s et 109
Solubility of Binary Rare-Earth Mixtures..........cccciiiiiiiin s e, 109

Solukility of Xenon Gos in NuF-Zer-UF_{ .................................. e e e e e et e ke SRR
Comparison of the Solubilities of Helium and Xenon Gases in NaF-ZrF -UF . 112
Selutility of HF in NaF-ZrF, e tt et £ r e r e ek he sere b e ebeb e b ene e enen sren e ene e e s e rnenne e 1 1O

EMF Measurements in Molten Fluorides. .. . i i e 118
Concentration Cells with ZZeF, Eiectrodes ............................................................................... . 18
Concentration Celis with N|°|N|F Electrodes. ..ot e 118
Daniell Cells . e it s et ettt e e s e e s tees e e rieae essnen e st nes s snraesnseneecrsannices | 1P

Activities from EMF Measurements of Molten and Solid Solutions of Salts oo 120

Electromigrotion in Molten Salts as a Possible Separotion Technigue for
Processing Fuel . e e ettt e e e ennne e 2]

. THEORY CF MOLTEN FLUQRIDE BEHAVIOR.. ... i s 123

Correlations of Molten Fluoride Behavior Based on Charge-te-lon Radius Reties ... 123

 
2.4,

2.8,

2.6.

2.7,

xii

 

Some Aspects of Solution Behavior of Fluoride Melts ..o,
A Structure Model and an Acid-Base System for Fuel Mixtures ...,

PHYSICAL PROPERTIES OF MOLTEN MATERIALS. ..o

Vapor Pressures of Molten Fluorides s isi s st e it i s ssbass s

Surface Tension and Density of Molten UF i

PRODUCTION OF PURIFIED FLUQRIDE MIXTURES ... e

Praodiatich Bl VoriBoE P ube TPl aBritles e ommrs somssensssssiio s ssosia g 9o boisss (s s essos
Preporation of ¥ Fy o s s i s s s s s s e
Prepcration of Fluorides of Molybdenum ... oo
Erepargtion-ot Uther Molerial susmsmmmuseninsiaunssma o o s ss e s s s s ne s e lisiiss

Pilot-Scale Purification Dperations . ... cieesieieiries s eersasessreorr e e snsesesns rresnenss e sreast e s e s nmnenvrme s
Production=5cale Oparations ..o .o it s s e seee e et et sereeabe e seeaes seb e e bene £ ebes e hene £ ea e rmnte e
Bitrhins it DitpErrsing DEePiitiie o cmommmssse i i i s s S s e 846

OB B B TG <. voviss ot s S R 3 A A S R A AN i S S i
Fillihg, Droining; ondSampling Operttions cossuwiaim s s s e
Filling of Pratt & Whitney High-Temperature Critical Assembly ...

COMPATIBILITY OF MATERIALS AT HIGH TEMPERATURE .. ...

Penetration of Graphite by Molten Fluorides.. ..o e e e
Elesteglyiais Bahia NaOH I SR S o e 5 s s s s s s e s S A
Physical Properties of Elastomers Exposed to Attack by Liquid Metals ...co.ovoiviiiicccinccnn e,

AN A LY TH AL CHEMIST RY Lo e ettt e et e s ees Sensase s eees esb et s sn e ees e ebean smsans
Crssiiatntl LEitEs ar PSS T B cnmmmsmoonieosm e i s e s S S e D e G O SR O
Betérmination of Oxygen i Mebtallit L iBRTEI oo mssivemmsmiis s i mss s s i v s iiiiss soioass sivdass
Apparatus for Sampling and Analyzing Alkali Metals.......ocii i e

Anolyses of Lithium-Base Fluoride Salts for Metallic Impurities e
AU DI o ossinse e eesuscn st ae s ae ae e ea e ases s rvmsass o e saasaes res e eae s eae s sten sam semsnt e eesee e nat s e e e aaes saes erts eemmenn e sen sres
W OO UM o ittt er e e ettt es te s e et ey e e bt et e s e et ae £t s £ e et e et ran et e e e e en ettt e eeeeeeneee e ean en eaen e en eereeeeemnanes
Titanium

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

Spectrophotometric Determination of Nickel in Alkali Metals.. .o

Determination of Trace Amounts of Zirconium in Sulfate Solutions with
Pyrocatechel Violet

H R R R R R R AR R L R R L AR AR S LA by r b4 E S Lt T4 aaad eyt E e m TR PR F N R TR R A N AR R AR EE R4 AN AL EE REAMyE REALEEEE & Aad R EEa

Determination of Sulfur in Fluoride Fuels oo,

Oxidation of Chromium with Argentic Oxide

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

Solvent Extraction of Molybdenum from Acidic Solutions with
Tri-n-octylphosphine Oxide

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

Preparation of Rare-Earth Fluoride Tracers

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

Service Laboratory

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

Vapor-Phase Polymers in Alkali-Halide Salts

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

 
3.1

B

3.3.

 

PART 3. METALLURGY
DYNAMIC CORROSION STUDIES ..

Forced-Circulation Loop Tests ..

Fuel Mixtures in Inconel und in Husielloy B ........................................................................

Nak in Inconel ..

Sodium in I'nconen' and in Stomless Sreel ...................................................... e B, A
Long-Duretion Tests of NaK and Sedium in Inconel oo e

Thermal-Convection Loop Tests ..

NaF-KF-LiF-UF, ond Sc:-dtum in Nlckel Mo!ybdenum AHoys ..............................................

Void Migration in Inconel Exposed to NaF-ZrF -UF,

NaF-KE-LiF-UF, end Sedium in Nishium . vwnnimnnms s amnins
GENERAL CORROSION STUBIES ... o bk 5550 i85 ne ik mmncingms sons ammers e dhams silizsn grnn oo
Tests of Inconel Tube-to-Header Joints with Recrystallized Welds .. et e
Tasts of Hoviries Mo, S84 Braming BUIE8 s s s s ot s o5 i s s s S0 :
Compatibility of Molybdenum and Nickel-Molybdenum Alleys in NaF-KF-LiF-UF, ...
Thermel-Convection Loap Test-of Sodivm: inHedtelloy Buvsninmnnesasnnuyasea.
Sodium-Berylfium-lnconel Compatibility in o Stotic System {Test No, 3} oo,
Effect of Zirconium on Corrosion in on lnconel-Sodium System .............. e e e
Ml ydenimes Bl gt G ammaiti b T i SOREEIN s s o cmm e s S R SRS e
Titanium Carbide~Nickel Cermets in Molten Sodium. .. oo i
Boron Carbide and Boron Nitride in Melten Sodium ... v,
SHC S0 11 MOlt@Nn SOGTUM oot ettt e eet et e e ee et e e et s e e s rsm s saen s eteeees e
BT L T A L T BT cmssssomn somnsonic s s o A R B TN N
Valoti ity PrHot Plant T oRrE ST STUGTBE umeri s svms e ovieisss 10 647880880 F0E 58 S50 LR MR VR

Exgerimental Studies with'Molten Lithium.cnaaianmusisasaiminsonmaiminsma g

Nickel-Molybdenum Alloy Development ... i et e v e e e s
General Status of Development Program ... s e
Propariies OF AUEE HOREE oo mmmommonms smseses i mieinsss s o0 s s 5600 i kit S S b
Proparbias Bl B ew IO umrmmummonssenmssn im0 s s b s e s

Fabrication of ORNL Experimental Alloys ...

Production of Commerciat-Size Heaots of Nlckel Molybdenum Alloys ................................

Phase Diagrom Studies at the University of Tennessee ..

Alloy Development at Battelle Memarial Tnstitute ..o
Fabrication of Composite Tubing at Superior Tube Co. oo
Shield Plugs for ART Pumps ..o i it e et sty et e et e e s e et et st it e
WEUTEEEE STl WG EERRE rurmommansmsinens s onsoess rosmn e i 5oy s v S PO 5 SOV Y SR A
Tiubulor Control Rage e im0 s i i S S R s S A A S RS
Evaluation of Arc-Cast ond Powder-Metallurgy Niobium ....ocoooiiiiiii

Metal Hydrides for Use as Moderators in High-Tempercature Reactors ... vovicnnn e
Capsules for InePile Tests o i e an et e saes et e sae e sees e s srere srerne

HYdriding: B GUIBMBIT e msiom s svemsms s e s s s s isasis

 SECRET]

147

147
147
150
154
155

157
157

162

164
164
164
165
166
166
169
169
171
173
174
174
175
176

177

177
177
182
183
183
183
186
187

188
188
188
189
120

192
192
193

Xiil
3.4.

3.3,

3.6.

4.1.

XY

UM IV ETO ] P ro@Ue b iMoo oo oo oo e et e s ete e eeate s e oee er eere b e et e e e e e e e e aeae e et ra e mreaan
G TR oo e N S S S

WELDING AND BRAZING INVESTIGATIONS v s v s ivss s siais s siisass soiicassianss

Heat Exchonger Fabricatioa...

ART Sodium-to-NoK Heat Exchqngers

lnvestigation of Rapid Heoting Cycles
OB ERIIaH BF Vol e GO T SIS oo oms R s 550 55000 99 SX E0 S D030 e SRS SR e
ART Fill-and-Drain Tank Fabricotion ... R S S O S S S e
Welding ond Brazing of Nickel-Molybdenum Alloys ... .. o B T R

Studies of Hastelloy B ond W Welds. . :
Welding and Brozing Studies of Experlmenfcl Nlckel Molybdenum A]toys

Hot-Ductility Tests of Experimental Nickel-Molybdenum Alloys ..o,

NaK-to-Air Rodiator Fobrication .,
Effects of Brazing Yariables on Broze Adherence
Assistonce to Yendors ..

Studies of Grain Growth in Incone]l T ubBe s ot e et et ee e reeasts eerarar tene sraeas sree smeearmranreas

Influence of Brazing Temperatures...

lnfluence of Time at Temperature... .
Influence of Tube- Strmgh'lenmg uncl Tube F’ollshlng Opemhons
Influence of Cold Work .. :

Subcontract Work «ommnmims oo oo o o s S T T S S S R TR S B N S SRR

MECHANICAL PROPERTIES STUDIES...o i s et e b e e

Dinamie Sirain Propeities of TREBRB e rmimnissammimsanmsis o v v s w5 s
B T T
Festingof Beryl i v s s s s T s s s R e

NONDESTRUCTIVE TESTING AND INSPECTION OF MATERIALS AND COMPONENTS ... .

Eddy-Current Inspection T e hmiGuUes oo oo iee o e oeeeeer s ee e et ot e e et e e et e en e e e et s et oo oo oo
T UBING I SPECIION oottt e et e e e e e et et et tees s nes e a et bt o e

Clatding T hicRAsEs MeuiremBiiiel o e s b o s

Radiographic Inspection Techniques .o o e et et et et e e e e e e

Examination of Leok in Sump Tank of Pratt & Whitney High-Temperature

Critieal Assembly ... it s e e e e e et bttt skt et

Moterial In5pecfion e b
Tubing...
Plate ..

Berylllum Hc:-f Pressmgs

PART 4. RADIATION DAMAGE AND FUEL RECOVERY AND REPROCESSING
RARIATION DEAMAGE oarmmrmsmnsmssosiag ssws s =

Examinations of Compenents ond Materials Tested in the MTR ... iviies o1
In-Pile Loops... e
Irradicted Moderutor Muiermls

'h-'

193
194

195

195
195
195

198
200

201
201
201
205

205
205
208

208
209
209
209
211

215

216
216
220
224

225

225
225
227

230

231

232
232
236
236
239

243

243
243
243
4.2.

5.1,

3.2,

3.3,

Creep ond Stress-Rupture Tests of [nconel oo

Apparatus for Stress-Rupture Tests of Tubing in the MTR ...
Tube-Burst Tests in the LITR s e e et e s s

Effect of Radiation on Corrosion of Structurel Motericls by Static Fused Salts ... ..

Inconel Capsules.....ooconein

Hastelloy B Capsules.. . e e e e e e

In-Pile Loop Experiments....

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

Qut-of-Pile Test of F'ump fc:-r LITR Verhcaf Loop ...........................................................

Assembly of LITR Vertical Loop ...

Irradiation Facilities in Cok Ridge Reseurch Rec:-ctor

Investigation of Sulfur Contamination of a Dry Box by Neoprene Rubber ...

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

Effect of Radiation on Thermal-Neutron Shield Materials oo e e e

Fused Solt Polaragraphy . ... i e e e e e er e et ssaeae s e eenas et et sete e stbe et et sraaaessae s ara .

Neutron Flux Studies...

Effects of Radiation on Electronic Components ...

Mincrity Carrier Lifetime Measurements in n—Type Germunlum

o Irr O OO S oo oot e oo e e e tes e e er e vensoes rre e mte e e aeeeeeaeee s eeaaneeene aeeanaeeeeaaeens

Transistors..

A Safe for Removul of Samples from 1he Reucmr .............................................................................

Modifications to Hole 30 of the ORNL Graphite Reactor ..o e

Dielectric Constant and Loss Tangent of lrrediated Plastics ot Microwave Frequencies

FUEL RECOVERY AND REPROCESHNG ..o, e

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

P i ot P It e oo e e e e e e et e e e et et e ae e e ae ettt eea e e e,

Sulfur Embrittl ement of Nickel . o e e e e e e e e

Oxidotion Resistance of Alloys in Flucride Systems. ...

Porosity and Density of NaF .. e e e e

R U e UM o Om i ma i 0M oo ciioee et e cee ettt e eet e e e ee e e s eeee eeee sen s ser s eaes eesen te s aes eare e se s veensenssen s tnsntee vete ernennense,

PART 5. CRITICAL EXPERIMENTS AND REACTOR SHIELDING
CRITICAL EXPERIMENTS e s

Reflector-Moderated Reactor Experiments at High Temperatures. ..o e

S DING THE O RY et e et et a b e a e b eees

Summary of Monte Carlo Calculations of Gamma-Ray Penetration in Multiregion Shields
with Slab Geometry o oo e e e et e

LID TANK SHIELDING FACILITY et e et e e e

______________

New Instrument Carriage at the LTOF L e e e e e

Investigation of the Physical Properties of Lithium Borohydride—Ammonia Under

Subcontract No. 300 e ST e e e by e e eyt e

Pressure-Temperature- Composifion Re|ut¢onsh|ps

Density Relationships ...............
Survey of Shielding Materiols Under Subcontroct No. 931 ..

248
249

250
250
250

254
254
256
2537

257
257
260
260

265
265
266
268
269
269

269

274
274
274
278
279
281

285
285

286

286

290
290

290
292
292

293

M
5.4, TOWER SHIELDING REACTOR T oot e s evesasns s svsvssnn e ee 29
NUCTEor Parameter Sty ..o i et s st e e e e e e e £ sbss e ki ek e b e 294

 

ST R amEa ] SUUHRL s sasmmvns s s e 5 s A A B S S 299
e e o T G S G 300
 

ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART 1. AIRCRAFT REACTOR ENGINEERING

1.1, Aircraft Reactor Test Design

Tests ore under way or have been proposed
with which to evaluate the ART structural design
criteria, since experience in the design of
structures for operation af temperatures in excess
of 1000°F is limited and in general not directly
applicable. The high operating temperature will
greatly affect the creep ond relaxation of the
structure, and temperature cycling will cause strain
cycling.  The only test completed in this study
thus far was the initial core-shell low-frequency
thermol-cycling experiment. The results of this
one test indicate that the strain-cycling criterion
used tor the core shells is conservative. In-
formatien on strain-cycling and creep character-
istics of lnconel, as well as on the effects of
“‘cumulative damage,'’ is to be obtained in heat
exchanger tests that ore now wunder woy., An
attempt to predict the life of the heat exchanger
revealed the need for creep measurements of
Inconel in fused sclts at the corrosion rates
expected in the reactor,

The distribution of stresses in the various shells
that will resuvlt from design-point temperature and
pressure profiles is being analyzed. Shells V
and V| (shell ¥V separates the boron-tile curtain
around the heot exchanger from the fuel, and shell
V| is the pressure shell liner) have been studied,
and final designs for the junctions of these shells
have becna prepared on the basis of the onolyses.

Efforts are being made to predict the stress-
rupture life of tubes subjected to a bending stress
on o short section., The results of this stdy
indicate that the calculoted maximum stress is
quickly reduced to about B0% of the predicted
elostic value and remains essentially constont for
the life of the member. The study alse indicoted
that constant-sttess data are required for calcu-
lations of this type, rather than constent-load data.
An experimental attempt to verify the “reduced-
stress effect’’ is being made.

Detailed stress analyses of the ART fin-type
NaX-to-air rodiators are under way. Several critical
areas in the design hove been identified, and

alterotions required by stress considerations have
Studies were made of weight ond air
drag loads, relative thermal expansion between
fins ond header, and relotive thermal expansion
between adjacent tubes in the event of flow
stoppage as a result of oxide plugging.

Series of tests ond experiments are under waoy
in order to determine the fatigue life of Inconel
under simulated ART conditions, since the core
shells might be subjected to surface temperature
oscillations os large as BPF ot frequencies of
from l,;; to 4 cps. It it is found that lncenel does
not possess adequate fatigue strength, design
modifications moy be required to alleviate the
temperature oscillations.  One proposal being
studied provides for mounting o pair of thin Inconel
shells in the core to serve as '‘thermal shields™
for the structural core shells,

been made.

Flow studies of the full-scale 21-in. plastic
model of the ART core were continved in order
to investigate woll temperoture fluctuotions, which
may be analogous to the fluid temperature fluctu-
ctions observed in the one-half-scale volume-heot-
Varicus types of probes and
thermocouple arrangements are being invesfigufed
for use in these studies. The fiow-guidance ex-
periments continued with tests of the core entrance
region alone, the core entrance and equatorial

source experiment,

regions simultaneously, and the core island bellows
region within the center volute of the cxial-flow
type of header.
bution, both radiclly and circumferentially, were
obtained with an inlet guide vane, a cenical baffle
plate, and o slotted disk, which wos loceted & in.
above the equator.

Some improvements in flow distri-

1.2, Reactor Physics

The gamma-ray dose rate at the outer surface of
the ART neutron shield was computed as o first
step in a re-evoluation of the shielding of the ART
ond auxiliary equipment, Gamma-roy  source
strengths were obtained through the use of two-
dimensional neutron-diffusion calculations. Sec-
ondary gamma rays from the lead shield were

neglected, since Lid Tank Shielding Fecility

‘- 3
XV
 

(LTSF) experimenis showed that the dose varied
exponentially with lead fthickness up to cbout
4.5 in.; the ART lecd is 4.3 in. thick. The LTSF
data ore interpreted ta meen that the intensity of
gamma rays penetrating the 4.3-in.-thick leod
shield of the ART will be much greater than the
intensity of gomme rays preduced in the lead,
The colculated dose rate at the surface of the
shield was found to be 1840 r/hr. The relotive
importance of each source is indicoted by the
percentfage contribution teo the total dose rate:
core, 22%; outer core shell, 9%; reflector, 12%;
reflector shell, 35%; heat exchanger, 22%.

The fast-neutron dose rote at the shield surface
was computed from LTSF data. It was found thar
ot a distence of 50 ft from the center of the ART
operating at 60 Mw with o 4.3-in.-lead and 31.5-in.-
water shield, the calculated gamma-ray dose is
22.8 rem/hv and the fast-neutron dose rate is
0.05 rem/hr.

1.3. ART Instrumenis and Controls

The ART NaK circuit flow-decay characteristics
were investigated as an initial step in evaluating
the significance of a pump power failure in terms
of temperoture and pressure transients in the ART
heat-dump system. Experimente] data were ob-
tained for o loop in which an ARTY prototype NoK
pump was being tested ot high temperatures.
Curing shutdown tests, complete tabulations of
pump shaft revolutions as a function of time were
obtained from which speed vs time dota could ke
obtained through numerical differentiation. The
data thus obtained were correlated with computed
velumes of segments of the ART primary NoK
heat-dump circuit.

Tests of commercially availoble resistance-type
liquid-metal-level tronsducers were continued, and
when the tests were terminated at 3732 hr, during
which time the NgK level of the system was cycled
65,000 times, oaly one of the four instruments was
in operating condition. The other three units failed
because NeK seeped into the insulation threugh
cracks at the U-bends. The reaction of the Nal
with the insulotion caused visible deformation.
ORML-designed units, which will be clemped into
place rather than welded, ore being prepared for
further tests. The ORNL units will hove improved
temperature compensation, and the weld design has
been modified.

xviii

All-welded resistance-type on-off level probes
are being developed as replecements for con-
ventional spark-plug probes, which are not suitable
for pressurized systems and temperatures above
450°F. An especially fabricated spark plug with
a beryllivm oxide insulator was tested in NaK as
an on-off level probe, but it failed in 20 hr of
operation because of short circyiting of the in-
sulator. Several special plugs with insulators high
in olumino ond mognesium conient were received
for testing.

Tests of the helium-bubbler-type tuel-level in-
dicetors were continued, and tests to determine
the cause of bubbler plugging were initiated. It
is thought that the wmateriol which plugs the
bubbler tubes is Zr0,. Special static-level tests
are to be run therefore in which the oxygen content
will be carefully contrelled. It is hoped that the
present average life of 1100 hr in a static-level
system can be increased to the specified 3000 hr.

Evaluation tests of commercial pressure trans-
mitters for use at high temperatures were continued.
Units that employ differential transformers were
found to be difficult to use becavse the field of
the Calrod heaters used to maintain the operating
temperature has on effect on the output readings.
A unit that employs @ varioble-permeance type of
pressure transducer was unsolisfactory because
of lack of svitable aligning guides. Additional
tests of the promising four-legged strain-goge
transmitters are being maode. Also o second set
of pneumatic-output fransmitters completed a suc-
cessful 3000-hr life test in o NaK-filled test loop.

Six of the SIé-in. magnetic flowmeters purchased
for the ETU and the ART were installed in the
flowmeter test loop for calibration. Preliminary
indications are that ot least four of these flow-
meters will be ccceptoble for ETU installation,
Six 2-in. magnetic flowmeters are aveilable for
calibration, and the 3%-in. models designed for
plug indicotors and cold trops are being fobricated.

Two prototype models of remotely adjustable
pressure regulators were tested. These units were
found to be stable under line-voltage, supply-

pressure, and normal ambient-temperature vari-
ations. Life tests are to be run on these
instruments,

A 1-in. turbine flowmeter ceased to operate after
340 hr of an endurance test scheduled for 3000 hr

in a circulating fused salt. X-roy exomination
showed that the unit was not broken, and it is
thought that oxidation resulting from previous
operaticn in NaK systems may have caused the
unit to bind. The unit will be removed for ex-
aminaticn when the loop is drained. A 31,5-En.
turbine flowmeter operating in a NaK system failed
some time during the first 40 hr of operation.
Disassembly of the loop showed eight of the ten
blades 10 be missing; the remaining two blades
were bent; and both shafts were broken. The
cause of this failure is being investigated.

Tests of closure welds of Inconel-sheathed
thermocouples in sodium and in fused salts are
under way, The Heliorc welding process has been
found to yield 30 to 50% acceptable welds. Bend
tests were made, ond single U-bends on a 3,:’1—ir:.
radius were found to have little effect on accuracy,
Repeated bending caused a 2 to 3°F shift. The
drift of data obteined with sheathed and with
beaded Chromel-alumel thermocouples as a function
of time and temperature is being studied.

1.4, Component Development and Testing

Yrradiation tests of a full-scale ART fuel pump
rotary assembly are to be run in a gamma-irradiation
facility in the MTR cenal. The final selection of
a lubricant-coolant for the reactor pumps will be
based on the results of these tests.

On the basis of further satisfactory test per-
formance, modified Durametollic seals have been
specified for use in the ART NaK pumps. Hf the
irradiation tests show either UCON LB-140-X or
Gultcrest 34 lubricants fo be satisfactory, buna-N-
base O-rings will be specified for the reactor pump
¥ the MTR tests prove these lubricants
to be uvasatisfactory, the seal material mey ogain
become o problem.

The development 1tests with water of an ac-
ceptable impeller configuration (designated model
32) for the ART fuel pump were completed., The
tests showed that flow rate, head, and pumping
power remained fairly constant when the liquid
level in the expansion tank was reduced from 3 to
g.g in. Further reductions in the level resuited in
definite decreases in flow rote, head, ond pumping
power, Tests for determining degassing times
were alse conducted with the liquid level in the
expansion tank ot 3 in. and the pump eperating of
the design speed of 2700 rpm. Degassing of an
injected air sample was complete in less thon

seals.

1 min. Tests of the pump at high temperatures

with NaF-ZrF -UF, (50-45-4 mole %, fuel 30) as
the circuloted fluid were started, The pump head
is 1 to 2 ft lower than that cbtained with water,
ond this discrepancy is beiny investigated. Pers-
formance and cavitation dota are being obtained.

A similar pump that operated with fuel 30 tn an
endurance test was shut down after a total of
2587 hr of operating time and 120 thermal cycles
in the temperotuee range of 1400 to 1100°F, Ex-
ominaticn showed the pump to be in good condition,
ond it has been clecned, reassembled with new
seals, ond put back into operation., There was
no trouble with the hydroulic drive during the lost
1240 hr of operation because a back pressure was
mointained on the discharge side of the motor to
ensure that the piston shoes were seated on the
wobble plate af oll times.

The ART sodium pump is now being subjected
to development testing with water on the same
basis as that used for the fue! pump. Tests to
date hove been hampered by ingassing of the loop
in which the pump is instalied. Since free move-
ment between the sodium pump barrel and volute
will be required in the ART to ollow for radial
movement, a seal is being developed for this
region to minimize leakoge of sodium between
the heat exchanger inlet and the pump centrifuge
outlet. Two types of multiple-ring seals are being
considered, as well 0s a simpler seal with o single
Fing.

An ART sodium pump is also being fested in an
endurance test with sodium ond has accumulated
1584 nr of operating time with 300 thermal cycles
in the temperature range of 1250 te 1050°F. The
oump priming difficulties encountered previously
were solved by stopping the pump when transferring
sodium to the pot, The off-ges line plugging
problem was resclved by lowering the liguid level
in the pump poft.

High-temperature testing of an ART primary
NoK pump was continued, and performance and
efficiency curves were obtoined ot o NaK temper-
ature of 1200°F. The data are in good agreement
with the daofa obtained with water, An ART
auxiliary NaK pump is being tested with water
and an Inconel impeller in ploce of the brass
impeller used previously. Cavitation performance
of the Inconel impeller has been found to be
entirely satisfoctory. After a vent was provided
for gos trapped during filling, the pump primed
satisfactorily, A specific technique for starting
and siopping the pump was developed.
Fuel-toNaK  heot and NaK-to-air
radiator test operations were continued. One test
was terminated by a MNoK leck in the circulating
cold-trap economizer after ¢ total of 1845 hr of
operation. The heat exchanger and the radiotor

exchonger

in this system hod been operated noniscthermally
for 1456 hr, including 552 hr ot o fuel inlet temper-
afure of 1600°F and 29 thermal cycles. Dato
obtained in this test are being analyzed as part
of a study of mass transfer in a NaK circuit,

A 20-tube semicircular heat exchaonger was de-
signed ond is being fabricated thot simulates the
tube geometries of the ART main heat exchangers.
This heat exchonger will be subjected to o thermal
cycling progrom in order te wverify the stress
analysis of the ART heot exchongers,

The ART twin fuel pump system was tested
further as part of the water flow tests of the
aluminum north-head mockup. With the improved
model 32 impellers in the pumps, the liquid
spraying and sloshing noted previously in the
expansion tank and the resultant wetting of the
tank ceiling were greatly reduced. The madel 32
impeller olso resolved the problem of liquid rising
in the shoft annulus of o stopped pump if the other
pump was stilt running.

Tests of ART prototype dump valves were con-
tinued. VYalves which have rigidly mounted plugs
are now considered to be essenticl for satisfactory
aoperation. The seat and plug materials which have
been found to be the most saotisfactory are the
Kentanium cermets K 151A vs K-152B and KM vs
K-162B8. All tests to date have been made with
fuel 30, but a test stand is now being modified
for tests with NaK,

The second of the two proposed thermal-stability
tests of the outer core shell was started under
conditions similar to those used previously. Di-
mensions were carefully taken for later comparison
with dimensions to be taken after the test,

Orifice plugs were tested for controlling sodium
flow through the cooling holes in the ART be-
reflector in order to obtain o uniform
temperature throughout the beryllium. A full-scale
wooden mockup of one-half of the symmetrical
sodium circuit in the reflector entrance region was
The head and flow distribution
measurements obtained in these tests are being
used in design anclyses, A water test was also
conducted on a proposed byposs slot to be cut
between the sodium-to-NaK heat exchanger shell

ryllium

olsc fested.

and the external pressure shell of the reactor.
This slot will allow some sodium to byposs the
present s5lots between the pressure shell and the
upper-deck outer shell and will provide a more
direct entrance for flow into the heat exchanger
inlet.

The high-frequency thermal-cycling test loop
designed for studying the effect of thermal cycling
The loop has
operated satisfoctorily ot design specitications in
water tests, and data are being taken on temper-
ature amplitudes at the test section as a function

on Inconel tubing was completed.

of freqguency ond flow rate. The water tests will
be continued wuntil thoraugh fomiliarization with
the operotion is achieved.

The plugging indicaters that are to be provided

in sach of the eight NeX cold-trap circvits of the

ART to monitor the sodium oxide level were
calibrated against chemical onalyses. The coli-
bration curves obtoined include most of the

praciical range of plugging-indicater application,

Tests were made on modified liquid-metal-vopor
condensers. A satisfactory NaK cendenser is now
ovailable, but further developmental effort is
required to obtain o satisfactory sodium condenser,

The design of an experimental prototype of the
ART zirconium flucride thermal-condensation vopor
trap was completed, and a unit is being fobricated.
Traps poucked with A|203 have operated satis-
factorily in several tests.

Tests were run on three bronds of insulation
being considered for use on piping corrying high-
tempergture NaK, The thermal conductivifies and
heat losses were obtained.

1.5. Reactor and Facility Construction

The detailed design of the reactor is essentially
complete, with some work remaining to be done
on the design of sodium annuli surrounding the
fuel overflow, off-gas, and fill-and-drain lines,
MNumerous design changes haove been made based
on further design studies, development
component  fabricatien difficulties, ond
shrinkage tests.

The schedule for ossembly of the ETU end ART
reactors was revised on the basis of delayed
delivery of compenents being produced by outside
vendors. Work on the north-head assembly is
awaiting delivery of the sodium-to-NaK heat ex-
changers.  The components for two reflector-

tests,
weld

moderator assemblies are now available, ond the

xx
u
 

 

assembly for the ETU is progressing satisfactorily.
Work is clso progressing on the strut-load ring
assembly. Most of the beron carbide tiles required
for two reactors are ovailable, aad no difficulty
is anticipated in obtairing the remainder. Some
difficulties are being experienced by the fabricator
of the stoinless-steel-clod boron-copper shielding
material thot may be resoived by a relaxetion in
tolerances.

The fobricotion of the Inconel shells by the
spinning process remains o major problem, but
progress is being made. An alternate effort is
being made tc fabricate the thin core shells by
hot forming oversized ports and then finish
machining them to final thickness. It is doubtful,
however, that shells produced in this manner will
have satisfactory metallurgical structure,

One set of beryllivm island parts is complete,
and work is under way on the pressure shell liner.
The dual effort on fuel-to-NaK heat exchanger
fabrication was evaluated, and Black, Sivalls &
Bryson was selected as the fabricator.

Extensive tests of reactor assembly methods
have been performed, and the assembly procedures
have been developed. The assembly area aond
fixtures are olmost completed.

Much installetion work on the ETU assembly has
been done, and prelimincry estimotes show an
expected ETU focility completion date, exclusive
of the reactor and its associated equipment, of
September 1, 1958. A further study of the items
fo be ploced within the ETU cell that simuleotes
the ART cell brought about several changes,
including the elimination of the equatorial and
south-head lead shielding, Design of the cell
compenents {other thon the reactor) is well under
way, and fabrication of many items has been
initiated.

Much of the outside contract work on the ART
focitity has been completed. The work to be done
by ORNL forces on the installation of process
piping ond equipment remains, and design work
on this phase of the project is well under way.
Procurement was initiated on o few items,

Plans for removing ond sectioning the ART after
operation were studied further. A Ix‘iz-scale modef
of a proposed hot cell is now avsoilable as an aid
in visualizing the complex techniques that will
be involved and in designing the tools and tixtures
that will be required.

1.6. ART, ETU, and In-Pile Operations

A critical review of the reactor system operability
is being made, ond conflicts among the design,
construction, and instrumentation phases of the
project are being resolved. A study of manpower
required for various phases of the program in-
dicated thot the peok work foad during operation
of the ART will require 40 engineers ond 18
technicians.

As the result of o design review, additional
provisions have been made for the ceontainment
of octivity at the ART focility in the event of
concurrent heat exchanger and radiater lecks; @
number of design and instrument changes have
been incorporated into the system; ond addisional
component development tests have been schedufed
to demonstrate the odequacy of the system for the
prescribed test program. Procedures for filling
the various systems were prepared.

The planning for ETU operation is paralleling
that for ART operation, with the maintenance of
ETU flow sheets, instrument lists, and valve
tabulations being emphasized, A preliminary study
of ART emergency conditions to be simulated with
the ETU is under way.

Procedures for disassembling the ART were
studied further, Possible methods for performing
the remote cutting operations that will be required
within the cell were studied, and several dis-
odvantoges of a proposed underwater operation
have developed.

Developmental work is under woy on the higher
performance pump and heat exchanger that will
be required for in-pile loop experiments with the
new nickel-molybdenum alloys and lithium-base
fuels, A pump is now availoble that will be
suitable for a loop with maximum temperatures in
the range of 1600 to 1650°F, which correspond to
o pump temperature of 1500°F. |In on effort to
increase the copacity of the heat exchonger, a
vnit with machined nickel fins has been designed
and is being fabricoted for testing.

In-pile tests of moderctor materials continued
with the preporation for insertion in the MTR of
copsules conteining beryllium oxide, graphite, and
zirconium hydride specimens.

XX
1.7. Heat Tronster and Physical Properties

Heat transfer studies with the delta-array heot
exchanger were continved. Flow in the laminar
region wos investigated to establish the effect
of natural convection on the total heat tronsfer,
No effect could be observed over the Reynolds
modulus range exomined. The fluid friction choroc-
teristics on the fuel side of a straighi-tube model
of the ART fuel-10-NaK heat exchanger were tested
further, with measurements being token down to
a Reynolds modulus of 1530. The early transirion
in the data indicates the effect of the turbulent
spacer wokes on the system pressure drop. FPre-
liminary designs have been made for an apparatus
to study the heat transfer and velocity struciures
in a single fuel-side channel of an ART type of
heat exchanger.

Additional data on the
screens in reactor cores ore to be obtoined with
the use of the swirl entronce system. A mathe-
matical analysis indicated that even under fuel
conditions reactor core screens will not
reach dangerously high temperatures. FPhotogrophic
records have been mode of lominar and surbulent
velocity profiles i o

stabilizing effect of

dump

number of geometries.
Quantitative data on the turbulent velocity strycture
of woter flowing through a circular tube were
obtained by superimposing the velocity profile
photograph on a grid photograph.

A volume-heot-source experiment is plonned with
which to establish the effect of screens on the
thermal structure in the ART core. A series of
vortex tube heat transfer and hydrodynomic ex-
been outlined for determining
quantitatively the characteristics of this system.
Therma)-cycling studies of the strength and cor-
resion characteristics of Inconel in a fluoride salt
mixture envircnment were continued. The sig-
nificonce and limits of some of the experimentol
variables have been determined, There is some
indication that high-frequency cycling
causes greater corrosive attack thon does low-

periments have

thermal

frequency cycling.

The enthalpy and heat capacity for the mixture
NaF.ZrF,-UF, (50-25-25 mole %) were determined
in both the solid ond liquid states. Data on the
pure compound U‘F‘t have been obtained for the
solid stete only. Viscosity measurements hove
been made with three mixtures in the NaF-RbF-
ZrF,-UF, system. The viscosities ronged from
7.5 to 9.5 centipoises ot O00°C down to 3.3 to

3.5 centipoises at 800°C,
studies have concentrated on the problems of
eliminoting from the experimentol data the effects
of variable heat flows. An experimental study
has been initiated to determine the electrical con-
ductivity of {iquid soedium and of sodium-potassium
alloys.

Thermal conductivity

PART 2, CHEMISTRY
2.1. Phose Equilibrium Studies

A simple modificatian wos mode ta the thermal
analysis apporatus in order to minimize losses
by wvaporization.  With the modified opparatus
it hos been possible to analyze samples containing
up to 80% ZrF , with negligible loss in weight of
the sample.

A new vocuum furnace that is odaptable to
microscopic observations was developed ond
fabricated for use in visual observation of liquidus
and solidus points and for differential thermol
analyses of quantities of fused salts.
indicote that

accurate melt temperctures

small

Preliminary visual observations

reasonably caon be
measured with only 150 mg of material.

Quenching studies of LiF-UF, systems thot
contain 20 to 32 mole % UFA confirmed the
liquidus ond solidus relationships previously
reported. |t was found that the compound -d'L.iI:-Ui:‘1
undergoes very slow subsolidus decomposition to
LiF and 7LiF-6UF4 at about 470°C. Thormal
anolysis  studies showed thot these mixtures
undercooled readily if the melt wos first heoted to
aver 800°C.
results characterized the equilibrium conditicns,
a seeding experiment wos conducted on g mixture
contoining 24 mole % UF .
accordance with

in order to be sure that the quenching

The results were in
the results of the quenching
studies, except thot the temperatures of the thermal
ettects were about 10°C lower,

A study of the similarities of the systems

NuF-Hde and NaF-ZrF, has revealed three
NaF-Hde compounds that ore onologous to
NaF-ZrFi compounds. The high-temperature

vsed for visuol observations of
liquidys temperatures of mixfures in the LiF-CeF,
system. A diogrom of the system was prepared.
Examination of slowly coeled melits showed only

furnace was

L.iF and CeF, to be present and contirmed thermal
indications of the cbhsence of complex compounds.

Thermal analysis of the system KF-BeF, indi-
cated on incongruent melting point aof 390°C and o
peritectic composition with about 52,5 mole %
Ber. The phase equilibrivm diagram for the
system was prepared.

Quenching studies of the system NoF-RbF-Ber
were initiated. For mixtures containing 33.3 and
40 mole % NoF, the ternary compound is the
primary phose, the solidus temperature is 452°C,
ond RbF.BeF, is the other phase below the
solidus. The phose relationships between the
liquidus and solidus have not yet been established.

A large number of quenching experiments were
completed thot showed the tentative phose equi-
librivm diagrom presented previously for the
system NoF-LiF-UF  to be in eror in several

respects. A revised diogram was therefore
prepared.
Phase equilibrium studies of the systems

NaF-CrF ,-ZrF ,, NaF-FeF -ZrF,, ond NaF-NiF -
ZrF, were undertaken in order to determine the
formyla and phase relationships of the ternary
compound in eacch of these systems. Only one
ternary compound wos observed in each system.

Optical properties ond x-ray diffraction data
ore presenfed for the following cempounds: S-
KF-?Ber, QKCinFeCIZ, KCl-FeClz, INaF-HIF
2MNaF-BfF ,, NaF.CrF,.2ZrF,, NoF-.FeF,.2ZrF ,
and NaF.NiF _.27rF ,.

2.2, Chemical Reactions in Molten Salts

The investigation of the equilibrium reduction
of NiF2 by H2 in Naf-Zrf ot 550°C was con-
cluded, ond the dato were combined with these
previously obtained. The doto indicate that the
equilibrium constant, K, is independent of the
NiF, concentration over the concentration range
studied, which included values not for from the
concentration of NiF2 in the soturated solution at
this temparoture.

In o carefu! determinction of the solubility of
nickel as NiF, in NaF-ZrF, (53-47 mele %) of
55(G°C, a volye of 352 + 11 ppm was obtoined.
The dota indicoted that the solubility remcined
constant when the omount of saturating phase
changed from a trace to a total of 1310 ppm Ni T
Further, the value of 352 is in substantial agree-
ment with o value of 400 ppm previously reported
for the solubility obtained with an oddition of
1760 ppm Ni**,

Additionul experiments on the reduction of UF
by Fe® in the reaction medivm RbF-ZrF  (60-40

mole %) showed that equilibrium wos established

in 5 hr and that the equilibrium concentration of
Fe** is significantly higher at 400 than of 800°C.
The concentration values obtained fell in the ronge
of those obtained in other alkali Hluoride-binary
mixtures, satisfoctory explanation is
ovailable for the lack of elfect of the recction
medium. In controst, the studies of the Cro-UF4
reaction showed that the chromium concentration
increases in the order of use of RbF, KF, NaF,
LiF, which is the order of decreasing ion size.
Additionel data on the reaction of UFA with Cr®
in the RbF-Zrf, reaction medium ot 600 ond at
800°C indicate that the fluoride

increases only slightly when the RbF concentration

ond no

ion activity

is increased from 50 to 58 mole %.

Rare-earth fluorides labeled with Ce 141, [La 140,
ond Sm133 were used to study the solubilities of
CeF3 angd Lu:II:3 ¢lone and in combination and
SmfF, clone and in combinaticn with CeF3 in
NcF-Zer-UFd {50-46-4 mele %) at varicus tem-
peratures.  The odditien of both LaF; and SmF,
decreased the solubility of CeF,. Similorly addi-
tions of BoF, and 5rF, reduced the solubility of
CeF,. The data indicate that the solubilities of
the rare earth flyorides are sufficiently high to
permit quite long-time operotion of o reactor before
precipitation of these materials becomes a problem.

The determination of the solubility of xenon gas
in molten NoF-Zer-UFd {50-46-4 mole %) os a
function of pressure and temperoture was suc-
cessfully completed. As in the cose of helium,
the solubility wos found to increase with in-
creosing temperature.

A direct equilibrotion method was used for
determining the solubility of HF in NoF-ZrF
(53-47 mole %) at 600°C. The dato obtained,
which will be extended to higher temperatures
ond pressures, indicate that the solubility is
1.25 x 10~ moles of HF per cubic centimeter of
melt af 1 atm. The data of the solubility of HF
will be valuable in studies of methods of salt
purification, some aspects of corrosion problems,
and processes for dissclution of solid fuel
elements for uranivm recovery.

Preliminory measurements of concentrotion celfs
with l[‘»li':’/'l*lilz2 electrodes and the solvent KF-LiF
(50-50 mole %) indicote the sclubility of Nin
in this sclvent to be obout 10 times that in the
solvent MaF-ZrF  (53-47 mole %), This difference

4
can probably be oftributed to the basic noture of
the KF-LiF solvent in which the weokly acid NiF,
would be more soluble thon in the more acid
NaF-Zer mixture. In order to obtain more infor-
mation regording the chonge of solute activity
coefficients with solvent composition, efforts are
under way to develop more reliable types of Danieil
cells and to measure a series of these cells in
different media with vorying but dilute concen-
trations of the metal fluorides, NEFz, FeFE,
Cer, and possibly ather solutes.

The possibility ef obtaining activities from emf
measurements of melten and solid sclutions of
salts was studied further with meosurements of the
LiCl-AgCl system. The main deterrent to the
application of similor techniques to fluoride fuels
is the lack of o suitable substitute for a fluorine
electrode.

Electromigration in molten salts has been shown
to be a successful means for separoting isotopes,
ond therefore it is of interest to determine the
usefulness of this technique for the presyumably
easier problem of seporating i1ons of different
elements in fuel Some preliminary
experiments with LiT and Ni ** ions in a chloride
melt showed that if the chloride lattice is con-
sidered to be stationary, the Li ¥ ions moved about
20 times faster than the Ni** jsas.
electromigration should be a very efficient means
of recovering Li' from o spent fuel, but the
engineering arrangements for working in fluorides
rather  than
formidabie probiems. Even the chloride systems
seem |ikely to be very awkward from an engineering
standpoint.

mixtures,

In principle,

chlerides preseat some extremetly

2.3. Theory of Molten Fluaride Behaviar

An attempt hos been made to correlate the
behavior of molten Hluorides on the basis of chorge-
to-ion radius (Z/ R) ratios. If the Z/ R ratios of
the salts in o mixture differ considerobly ond
especially if the salts are of widely different
valence states, the extra stobility of the mixture
will be large, and the structure will be considerably
distorted in comparison with the structures of the
component salts, that is, the larger the difference
in Z/R rotios, the more stable the mixture relative
to its compenents. For components which mix
without o change in coordination number and with
little distortion, the negative deviation from ideal
solution behavior is accounted for, in general, by
o dacrease in cation repulsive energy upon mixing.

It is alse convenient to correlate the behavior of
the complex melten fluoride mixtures in terms of
the attraction between the fluoride
nelghboring cations. The ions in a selution tend
I in positions of lowsest

ions ond

to arrange themselves
energy. Accordingly, the mutucl energy value
determines the positien of the fluoride ions with
respect to oltarnative cations, and fluoride ions
will prefer as necrest neighbors those cations
which yield the largest Z/R ratios.
in the difference in the Z/R raties of the cations

An increase

of a fluoride solution will resvit in an increase
in the negative deviations from ideal solutien
behavior shown by the mixture.

A useful qualitative model for describing the
liguid structure of fuel mixtures based on the
NaF.ZrF, system was devised in which liquid
ZeF , is considered to be a somewhat disorganized
lottice of flucride bridges cornecting zirconium
ions. The addition of NaF to ZrF4 results in the
breaking of the bridges to give zirconium ions
which are surrounded partially by bridging fluorides
and partially by nonbridging fluorides. Since the
complexing potentialities of o fluoride melt depend
on the fluoride ion activity, it is very important to
know how this activity, which cannot be measured,
varies with composition.

The fluoride ion activity con be correlated in «
qualitative manner by an acid-base scheme in
which catiens are rated according to their tendency
to behave os fluoride donors or acceptors. If
fluoride denors ore defined as bases and fluoride
acceptors as acids, the cations fall into the same
general pottern as thot used for the classification
of oxides as acidic or basic. The NaF compound
is an exomple of o sirong base and ZrF4 of a
strong acid,

In order to map the relative acidity of o fuel
solvent, Z/R ratios are plotted vs solvent compo-
sition. Such a plot permits estimates of the
reletive negative deviations of solutes, and, hence,
also relative solubilities in case the pure solute
is the saturoting phase. An important use is the
selection of which will
reduce corrosion by complexing the corrosive
agents rather thon the corrosion products.

solvent compositions

2.4, Physical Properties of Molten Materials

The wvapor pressures of the less volatile

compenents in fuel mixtures are being studied.

XXlv
 

In order to determine the vopor species, the
absolute vapor pressures of molten alkali fluorides
were measured by a quasi-static method. A suf-
ficiently wide temperature range was covered to
give curvature on a log p vs 1/T plot. Equations
were estoblished ond constants evaluated for the
compounds LiF, NaF, KF, RbF, ond CsF. Tento-
tive estimates were olso moade of the percentage
of monomeric molecules in the saturated vapor of
the various molten olkoli fluorides.

A preliminary value for the density of mohen
UF . was obtained for use in o determination of the
sur‘f‘oce tension, At 1075°C, a density of
6.80 t 0.09 g/cm® was found by using graphite
pycnometers. The density of the molten material
being greater than that of solid UF, (6.63 g/cm?®
based on x-ray data) and the behavier of the salt
surface level upon freezing in the pycnometer
indicate that UI:‘1 either contracts or expands
only very slightly upon melting.

2,5. Production of Puritied Fluoride Mixtures

The hydride of yttrium is being studied for use
as a moderator in high-temperature reactors.
Considerable quantities of YF, (™~1400 g} were
therefore prepared for use in manufacturing the
metallic yttrium from which the hydride is prepared.

The product of the reduction of MoF with FeF
was collected in o condenser and redlshiled uné
the yellow product was found to melt at 78°C
rather than the 60°C previously reported for the
moterial before distilletion. |t has not yet been
determined conclusively that the material is MoF ..

In the pilot-scale equipment, 42 hatches were
processed that totaled 1575 lb of various fluoride
compositions, with o mojor portion of the material
being NuF-KF-LiF-UF4 (11.2-41-45.3-2.5 mole %).
Nickel storage receivers are unsatisfactory for
the lithium-bose mixture becavuse of the large
exponsion of the material upon melting. Incenel
cans are now being used for 50-lb batches, but
when inadvertent exposure of the mixture to air or
water occurs in Inconel cans, extensive contami-
nation of the melt with CrF, and FeF, results.
Since these contominants cannot be removed
economically by retreatment of the batch, the
material must be discarded.,

In the production-scale facility 446 batches
totaling approximately 11,500 Ib were processed.
Some 18,000 of the 30,000 Ib of NaZrF_ ordered

from a vendor was received and found to ie satis-

factory. A total of 15,700 Ib of processed fluorides
was dispensed in botch sizes ronging from 1 to
25¢ lb.  Approximotely 4000 |b of processed
fluorides and 2500 Ib of ligquid metals were charged
into test units.

2.6. Compatibility of Materials ot High
Temperature

The study of the penetration of grophite by
molten tluerides was continved. An experiment
was performed in which commercial APC graphite
was impregnated with NaF-ZrF, (53-47 mole %)
and then subjected fo a sublimotion step in order
to raise the melting temperature of the residual
mixtyre to above that of the uronium-bearing
mixture. The sublimation procedure increased
the NnF-to-ZrF4 ratio in the graphite, with the
final rotic being obout 16:1. This material will
be tested in o molten uranium-bearing mixture to
determine whether a sclid impregnating material
will resist penetrotion.

Two types of graphite (code Nos, R-28 and M-80)
obtained from the National Carbon Compony were
immersed in NaF- ZrF -UF | (53.5-40-6.5 mole %)
ofter impregnation with Nulé ZrF (53-47 mole %).
Examinations revealed that these specimens were
not penetrated to ony consideroble degree by either
of the salt mixtures.

it hos been established through electrolytic
studies of the NoOH-Ni system that in o non-
isothermal system the hot nickel electrode is
usvally enodic and the cold electrode is cathedic.
This situation can, however, be reversed by the
use of a hydrogen atmosphere. Since NeOH is o
potential moderator-coolant material, ottempts are
being made to balance the current in an effort to
find @ way to eliminate the moss-transfer effects
observed in NaOH circuloted under a temperoture
grodient.

In service tests of valve seat materials exposed
to helium scturated with NaK vopor, G-E SE-550
and SE-450, Dow-Corning Silastic-80, and Dy Pont
5570 withstoed NaK fairly well, with SE-550
possibly being superior to the other materiols.

2.7. Analytical Chemistry

The apparctys for the detection of submicrogram
omounts of sodium in oir wos assembled and tested.
In preliminary tests 100 ppb of sodium could be
easily detected. By means of appropriate re-
finements in the Iinstrument it appears that the
 

E

amounts of zirconium in

desired limit of detection of 10 ppb is feasible
with the present apparatus. Zirconium was found
to be a suitable moterial from which to fobricote
orifices for the injection of NaK inte air. These
orifices will be used in the testing of the NoK
leck detectors.

The methancl-dissolution method for the deter-
mination of oxygen in lithium was investigated.
In this method the lithium oxide torms an equivalent
amount of water upon dissolution in methanol.
The then titrated with Korl Fischer
reagent. The reliahbility of the method is dependent
on the degree to which the methanol can be dried.
Methano! was produced with ¢ concentration of
10 ppm of water.

An improved device for sampling alkali metals
wos developed which includes a Jomesbury valve
modified so that it also serves as a sample holder,

water is

The valve is connected to the operating apparatus
by means of a metal-to-glass stondard taper jeint.

Methods for the determination
amounts of agluminum, venadium, and titanium in
lithium-base fluoride salts were modified in order
to apply these methods in the presence of in-

of microgram

terfering metals. In the method for aluminum
analysis, zirconium, iron, uranium, titanium, and
vanadium are removed by extracting them with
This particular ex-

traction essentially isolotes the aluminum in the

tri-n-octylphosphine oxide,
oquecus phase. For the vonadium determination,
the interfering ions, including nickel, iron, chro-
mium, eond molybdenum, among other ions, were
removed by deposition at the mercury cathode
prior to determination of the vanadium os the
benzohydroxamote complex. Titanium was separated
from molybdenum by a basic carbonate precipitation
and from niebivm by extraction of the thiocyonate
salts with diethyl ether,

A sensitive, spectrophotometric method for the
determination of nickel with 4-isopropyl-1, 2-
cyclohexaredionedioxime was developed. When the
nickel-dioxime complex is extracted with xylene,
the ultimate sensitivity is less than 0.5 ppm of
nickel in a 5-g somple of alkali metal. The inter-
ference of iron is masked by use of the fluoride
ion, This method is also applicobie to the deter-
mination of nickel in solutions of zirconium- or
lithium-bose fused salt fuels.

A study of interferences of vorious cations in
the spectrophotometric determination of microgram

xxvi

sulfate solutions with

pyrocateche! violet was completed,  Vanadium,
titanivm, and aluminum, os well as the anions
that form relatively strong complexes with zir-
conium, were found to interfere seriously.

The methylene-blue method for the determination
of sulfur in fluoride salts was improved by the
substitution of a reducing mixture of stannous
chloride and dehydrated phosphoric acid for the
standard mixture of red phosphorys, hydriedic
acid, and formic acid. As much as 1 g of the
fluoride salt con be dissolved in 20 ml of the
new mixture, whereas the fluoride salts are only
slightly soluble in the red phosphorus mixture,
The improved method is thus much more relioble,
sul fur,
Argentic oxide was studied as on oxidant for
This oxidant is efficient in solutions

particularly for low concentrations of
chromiym,
of all acids except hydrochloric, Excess oxidant
is destroyed by heating the solution. The use
of this reagent materially shortens the procedure
for the spectrophotometric determination of chroemium
with diphenylcorbazide with apparently no sacrifice
in precision,

The extraction of melybdenum from acidic solu-
tions with tri-n-octylphosphine oxide was investi-
goted.  Under fovorable conditiens as much as
20 mg of molybdenum(Yl) can be extrocted quan-
titatively in a single extraction by 0.5 mmoles of
reagent.

A satisfactory effusion cell was developed for
use in mass-spectrometer investigations of the
compositions of the wvapor phases of fused-solt
fuel mixtures. Preliminary measurements were
made on alkali-halide salts.

PART 3. METALLURGY

3.1. Dynamic Corrosion Studies

Three forced-circulation leops fabricated of
Hastelloy B were examined that had circulated
NuF-KF-LiF-UFd {11.2-41-45.3-2.5 mole %, {fuel
107} at a meximum fluid temperature of 1850°F.
Two of the loops operated with maximum wall
temperotures of obout 1750°F oand showed no
svidence of mass transfer after 407 and 1000 hr,
respectively. A third loep, which operated with o
maximum  wail temperature of 1710°F and at @
relatively low flow rate, showed slight evidence of
deposits near the pump. The hot legs of all loops
were found to be badly pitted after the tests, but
 

much of the surfoce roughness could be attributed
to the condition of the inner surfaces of the tubing
before the tests,

An Inconel forced-circulation loop was operated
with NaF-ZrF -UF (50-46-4 mole %, fuel 30) at
@ maximum f?md temperature of abowt 1700°F
and a maximum wall tempercture of 1840°F. After
1000 hr of operation the loop wos found to have
been attacked to a depth of 9 mils, and there were
some deposits in cold-leg sections, The fuel
mixture No. 70, NaF-ZrF -UF (56-39-5 mole %),
wos tested in another |ncone1 forcad-circulation
foop operated at o fluid temperature of about
1500°F Loop operotion wos terminated after
716 hr, and ettack to a depth of S mils was found
in the hot leg. A thin metallic layer wos aiso
noted in the cold leg. An examination of another
Inconel forced-circulation loop that operated
500 hr with fuel mixture No. 30 also revecled hot-
leg attack to a depth of 5 mils, but no cold-leg
deposits were found.

An evalustion was made of the effectiveness
of cold traps in the removal of oxide impurities
from circutating NaK and the resultant effect on
mass transfer in Inconel-NeK forced-circulation
loops. Cold traps maintoined ot temperatures of
100, 300, 600, and 800°F, respectively, were
utilized in cenjunction with loops operoting at o
maximum NoK temperature of T1500°F and a fluid
temperoture drop through the loop of 300°F. A
definite increase in the amount of depasit ma-
terial was noted in the loep which contained the
cold trap operated ot 800°F os compared with
the lcop having the cold trap operated at 600°F
The change in the amount of deposited material
accompanying a decrease in the cold-trap temper-
ature below 600°F was less than the change from
800 to 600°C. The average sizes of the porticles
comprising the deposits clso varied with the
cold-trop temperature, becoming coarser as the
cold-trap temperature increased. The deposits
found in these NaK systems were [ess thon one-
half the weight of deposit found in sodivm systems
operated under similar temperature and time con-
ditions.

Two lnconel-sodium forced-circulation loops
which operated at flow rates of 1.5 and 3.0 gpm
were also examined. The maximum sodium temper-
oture in these tests was 1500°F, ond the temper-
ature gradient was 300°F. Operation of the loops
for 1000 hr ot the higher flow rate produced

1'-
"

deposited material that weighed 14 g compared
with 10.4 g for the loop operated with the lower
flow rate. The average thickness of the deposit
was alse higher in the loop operated with the
higher flow rate.

A type 304 stainless steel foreed-circulation
loop after operation with sodium at 1500°F showed
only slight traces of mass-transfer deposits. This
loop was operated for 1000 hr with a thermal
gradient of 300°F.

Long duration tests of Inconel forced-circufation
loops were made with both NaK and sedium circu-
lated at @ maximum Ffluid temperature of 1500°F.
The test with NoK, for which a temperature
grodient of §00°F was used, waos terminated ofter
2760 hr.  The coocled section of this loop was
found to contoin extensive crystclline deposits
that ronged te 14 mils in thickress. The loop
operated with sodium for 4000 hr with o temper-
oture gradient of BQ0°F showed considerably larger
deposits that reached a thickness of 50 mils in
some areos.

Thermol-convection loops fobricoted from various
experimental nickel-melybdenum alloys were tested
with sedium ond with NeF-KF-LiF-UF , (11.2-41-
45.3-2.5 mole %, fuel 107), for 1000 and 5CG0 hr,
respectively, at maximum fluid temperatures of
1500°F. There was attack by the fuel mixture to
on average depth of 1.5 mils in all loops except
those construcied of alloys containing aluminum
additions. For the alloys containing oluminum a
maximum depth of attack of 3 mils was found.
Correspondingly the pickup of aluminum by the
fuel was high in the loops fabricated from alloys
Similarly
picked up by fuel circuloted in alloys centaining
tungsten; no cppreciable increase in
ottack accompanied the buildup of tungsten in the
fuel, Vaonadism, titanium, and niobium additions
to the alloys had no apgarent effect on corrosion;
but the fuel picked up slight amounts of titanium
and niobium.

An average hot-leg attack of 1 mil was found in
several nickel-molybdenum alloy leops operated
with sodium, ond deposits were visible in the cold
legs. These deposits did not adhere tightly to the
loop woll, and consequently difficulty hos been
encountered in examining them metallogrophically.

A series of standard Inconel thermal-convection
loops was operated with l‘~~!t:1|‘:-Zt'F4-UF_4 (50-46-4
mole %, fuel 30) at 1500°F to determine whether

containing aluminum. tungsten was

however,

xxvii
 

I l""J

the migration into lnconel of voids instituted by
the teaching of chromium during the corrosion
process would continve if o loop were left at test
temperature ofter the fuel was dreined. The
results showed that the migration of voids into
Inconel did not proceed when the leoching of the
chromium was stopped.

Three stoinless-steel-clad niebium loops showed
onfy slight hotleg cottack in tests with NoF-KF-
LiF-UF‘$ (191.2-43-45.3-2.5 mole %, fuel 107) at
1500°F; however, some cold-leg deposits were
observed in niobium loops which operoted 1000 hr
or longer. A nicbivm loop wos also operoted
1000 hr with sodium ot 1500°F. The attock wos
negligible, but cold-leg deposits were observed.
Chemical analysis of the deposited material
indicated that the sodium probably teoked into the
cladding Joop during operation.

3.2. General Corrosion Studies

Additional tube-to-header (oints with
receystallized welds were corrosion tested in NaK

(56-44 wt %) and in Na F-ZrF ,-UF, {50.45.4 mole %)

for 100 kr in seesaw-furnace apparctus with a hot-

Inconel

zone temperature of 1500°F, These samples were
fabricated by using a prechamfered header plate
which results in decreased material deformation
The tests showed
no greater attack along the tube-header interface
than on the tube and on the header,

Haynes No. 8244 brazing clicy (9.4% Cr—4.4%
Fe-3.7% S5i~2.2% B8--0.3% Mn-0.06% C-bo! Ni) on
Incone! tube-to-header joints showed good cor-
rosion resistance o NaK (56-44 wt %) in a Y00-hr
seesaw-furnace test at o hot-zone temperciure of
1500°F. Subsurfoce voids fo a depth of 3 mils
were observed on the alloy when tested in NaF-
ZrF UF, (50-46-4 mole %) under similor con-
ditions, '

A seesaw-fumace test of 500 ht at a hot-zone
temperature of 16050°F wos run to defermine the
compatibility of a 15% Mo—6% Cr—bat Ni alloy and
molybdenum when exposed to NoF.KF-LiF-UF,
{11.2.41-45.3-2.5 mole %, fuel 107). The presence
of nickel on the surface of the molybdeaum
specimen ofter the test indicated dissimilar metol

and greater weld surface orea,

mass tronsfer.  Similar results were found with
o melybdenum specimen exposed in o 17% Mo-2%
V_bal Ni alloy capsule to fuel No. 107 under
similar conditions,

A Hastelloy B

tested with sodium at a hot-leg tempercture of

thermal-convection loop was

xXxvirl

1600°F ond o cold-leg temperature of 990°F for
1000 hr. The ottock ond mass transfer in this
loop were quite similar to the resuits obtoined
with sodium in Inconel thermal-convection Joops.

A third sodium-beryllivm-Inconel competibility
test has shown that chromium plating of lnconei
reduces the amount of intermetallic compound
formation to approximafely one-third thot found
when Inconel and beryllium are placed in direct
contact. Thermal cyeling did not result ia spalling
of the Be Cr phase.

A zirconium specimen exposed to sodium for
400 br of 1600°F in an Inconel standpipe capsule
was found to have absorbed oxygen from the
sodivm. This removal of oxygen from the sedium
appecred, however, to have no effect on the eamount
of mass transfer os compared with that found in a
stondard Inconel stondpipe copsule containing neo
zirconism and tested under similar conditions.

Molybdenum and beryilium were tested for com-
patibility when exposed to sodium. Exposures of
100 and 500 he at 1500°F resulted in the formation
of two extremely hard intermetallic compounds,
MoBe., and MoBe, ,, ot the interfoce,

Four titanium carbide—nickel cermets, three with
10, 20, and 30% nicke! and one with 25% Ni plus
5% Mo, were attacked to a depth of approximately
0.5 mil during a 100-hr expesure to sodium at
1500°F in o seesaw-furnace. In previous similar
tests these cermets were not ottocked by Naf-
ZrF4-UFd.

Boron carbide specimens with theoreticol densi-
ties of 85 ond 90% were found to have fair
resistance 1o attack by static sodium in lnconel
containers at 1500°F. A dense (2.15 g/em3)
boron nitride specimen was severely ottacked in a
100-hr exposure to sodium at 1500°F in o seesaw
furnoce. Exposyre of an SiC-St specimen for 100 hr
fo sodium ot 1500°F in o seesow furnoce resuited
in complete removal of the silicon.

Lead-fithium alloys contoining 0.65 wt % Li
showed weight and dimensional increoses when
tested in water af vorioys temperotures. The
weight gains increased with the temperature and
time of the test.

In the lithium corcosion reseorch being cearried
out atNuclear Development Corporation of America,
base-line deta for plugging of type 316 stoinless
steel-lithium thermal-convection loops hos been
obtained. The addition ¢f titanivm and zirconiuvm
as getters has led to increosed loop life but has
 

not prevented mass transfer. The use of cold traps
for impurities hos also led to improved loop
performance.

3,3. Fabrication Research

The nickel-molybdenum alloys fobricated ond
tested thus far have, in general, significantly

better corrosion resistance and strength than
Inconel without sacrifice of fabricability and
weldability,  These alloys now appeor to offer

definite possibilifies of use as structural materials
of reactors that utilize more efficient fuels at
higher temperatures. The corrosion resistance of
most of the alloys is equivalent to that of Hastel-
loy B, and the stress-rupture data for oll olloys
tested show on average life of 40G hr at 1500°F at
a stress of BOOO psi. The fabricability depends
largely uvpen the kinds ond quantities of elements
added to the basic nickel-molybdenum alloy, and
with proper control reascnable yields of moterial
are obtained from taboratory heats,

At present an alloy designated INDR-8, which is
baosed on the Ni-Mo-Cr-Fe system, seems to best
fulfill all the requirements, In arder to investigate
the ronge of mechanicel properties developed in
alloys of this system, the tensile and stress-rupture
properties are being determined for a series of
compositions, with the maximum content of the
major components, other than nickel, being 20% Mo,
10% Cr, and 10% Fe. The alloys tested to date
with {ow and intermediate contents of Mo, Cr, and
Fe are oll stronger than lnconel and approach the
strength of type 316 stainless steel. There is
evidence that olloys with higher Me, Cr, ond Fe
content will respond to aging and show even
higher strengths.

A series of alloys designated INOR-9 are being
investigated, in which the elements Cr, Al, Ti,
ond W have been eliminated becouse of their poor
corrosion resistance in fuel 107. Alloys with o
maximum of 15% Mo and with significant additions
of niohium and iron were prepared, and the limits
of solubility and oxidation behavior were de-
termined.

Six tube blanks of ecch of three nickel-molybdenum
base olloys prepared at ORNL have been extruded
and submitted to Superior Tube Co. for redrawing
to tubing, These alloys were designed for studies
of the effect of 10% Cr and 7% Fe, singly and to-
gether, on the corrosion resistance of a 17% Mo—bal
Ni base alloy.

Sheet, plote, wire, and bar products fabricated
from the six 4800-1b heats designated INOR alloys
1 through & have been received. Alloys INOR-4
and INOR-& became virtually 100% scrap because
of fabrication ditficulties, while alloys INOR-2
and INOR-5 were processed with considerable
success, The tube shells of the INOR alloys, two
each of INOR alloys 1, 2, and 5 and one each of
INOR alloys 3 and 6, are presently being tube
reduced for conversion to small-diometer seamless
tubing at Superior Tube Co.,

The contract with the Westinghouse Electric
Corporation covering the fabrication of large pilot
heats of nickel-molybdenum base afloys has been
signed, and work will begin immediately. Initially,
one 3600-ib heat of alloy INOR-8 will be cast and
fabricated inte sheet, plate, wire, bor, and tube
products for eveluation,

Yarious techniques are being utilized in studying
the phose relationships of the Ni-Mo ond Ni-Mo-Cr
alloy systems under @ subcontroct at the University
of Tennessee, Resulis show that the beto and
gammo phoses fransform ot somewhot higher tem-
peratures than those previously reported.

A totat of 16 tube blanks of four Ni-Mo base
alloys which were extruded for Battelle Memorial
Institute last quarter failed during tube reduction
at Superior Tube Co. These processing failures
substantiate earlier evidence that these alloys
with 0.12% C are not amenable to tube-drawing
operations. Batielle has prepdred new olloys with
o lower corbon content and has submitted twelve
billets of each of two of the alloys to ORNL for
tube-blank fabrication, The tubing processed from
these blanks is scheduled for forced-circulation
loop tests at Bettelle,

Laorge amounts of composite tubing may be re-
quired to circumvent moss tronsfer in nickel-base
alloy systems., To demonstrate the feasibility of
producing the material, several composite tube
shells of Inconel over type 316 stainless steel
were processed on a production schedule. Ex-
cellent yields of souad tubing were obtained from
the coextruded material,

Two full-sized gomma-ray shielding pfugs com-
posed of tungsten carbide and Hastelloy C were
prepared for the ART fuel pumps. However, a die
failure produced loteral expansion of the plugs,
with the result that the parts are considerably over-
size ond will be used only for brazing tests, A
redesigned die has been obtained,

Xxi X
 

Continved difficulty in producing the stainless-
steel-clad Cu-B,C shield material at Allegheny-
Ludlum has required considerable coordination and
evoluotion of moterials, Recent work based on
recommendotions made by ORNL has resulted in
good yields of sound material,

In attempts to eliminate lack of ductility resulting
from stringering of fine dispersions of Lindsay
Oxide, a technique was developed for producing
dense, hard porticles of the desired size, Three
cores for extrusion billets were prepared with this
material dispersed in electrolytic nickel powder.

The recovery and recrystollization of arc-cast
and powder-metallurgy niobium are being studied
prior to comparing their mechanical properties,
Nearly ideaticel microstructures were finally pre-
pared from the two materials by an initial anneal
at 1450°C followed by severe cold reduction and a
leng annecl ot 1250°C, This material will be used
in determining the recrystallization kinetics of the
two types of niobium,

An in-pile test capsule of hydrided zirconium
clad with molybdenum was fabricated, Equipment
for the production of yttrium metal wos designed
ond built., Studies on the cladding of molybdenum
with nickel-molybdenum alloys were started,

3.4. Welding and Brazing Investigations

An invesfigation wos conducted to determine
optimum brazing procedures for fabrication of the
ART sodium-to-NaK heat exchangers. The mos?
influential variable was found t¢ be the rate-of-rise
to brazing temperature, A similar investigation is
being conducted to determine the minimum rate-of-
rise to lxazing temperature required to effect a
satisfactory back braze of the tube-tfo-tube sheet
joints of heat exchangers.

The techniques developed previousty for the
fabrication of valve components were used in the
preparation of various valve materials for testing
under simulated service conditions,

A study of the problems ossociated with the
fabrication of the ART fill-and-drain tank is under
way. lests are being run to establish welding and
brezing procedures and joint designs thot will re-
sult in crock-free fusion welds. A promising joint
design now being investigated is a trepanned type,

Additional results were obtoined in the investi-
gation of the mechanicol properties of Hastelioy B
and W welds on ]/’z-in. plate, and the results of aoll
the tensile test data have been anclyzed. The

‘fz-in.plqte required for welding and brazing studies
of some of the nickel-molybdenum experimental
alloys wos received ond tests were initiated, No
cracking tendencies were found in tests of INOR-7
and -8,

Results of Polytechnic
Institute on experimental ond commercial nickel-
base alloys indicate thet it is usualiy sufficient
ta evaluate the hot ductility s measured ot 2300°F
during cooling from some higher temperature and to
compere these data with the hot ductility measurad
at 2300°F during heating. The test results show
that none of the INOR alloys behave in as ductile
o manner os Inconel and thay they behave more
like Inconel X than like Hastelloy B,

tests at Rensselaer

The effects of brazing variables on braze ad-
herence in NaK-to-air radiators were studied further
in order to assist the vendor in fabrication prob-
Specific recommendations regarding the
cleoning methods, and brazing tech-

lems.
equipment,
niques were made,

The initial result of a study of the effect of
brazing ond hect-treating operations on the grain
growth of Inconel tubing indicates that stress
relief treatments at 1500°F do not result in grain
growth. Significant grain growth resulted, however,
from }é-hr brazing cycles at approximately 1922°F.
It was also noted that groin growth proceeded more
rapidly in the grains locoted along the ocuter wall
of the tube than in those located along the inner
wall, Straightening and polishing operations had
only minor effects on the grain growth, The in-
fluence of prior cold work is being studied.

In the molybdenum welding studies under sub-
contract at Battefle Memorial Institute the effect of
atmosphere purity, vorious types of shields, and
several cleoning technigues were investigated,
Acid etched specimens were somewhat better than
samples prepared by other methods, Also, the use
of a trailing shield in open air welding was found
to  yield weldments with significantly greater
ductility,

1.5. Mechanical Properties Studies

The behavior of Inconel when subjected to re-
peated thermal cycles is being investigated by
means of relaxotion tests and mechanically induced
cyclic strain tests at ORNL; the University of
Alabama is contributing to this study by conducting
thermally and the

induced strain cycle tests;
 

Battelle Memorial Institute is evaluating the effect
of dynamic loads superimposed on static loads,

Reloxation data were obtained for coarse-grained
Inconet at 1300, 1500, and 1650°F, Prier tensile
strains at temperature were shown o have little
effect on the relaxation behavior. Evidence that
coarse-grained Inconegl has poorer resistance than
fine-grained material to repeated strain cycles has
been obtained.

The evoluation studies of creep properties of the
varicus new olloys heing developed for reactor
use were continued. The results of the alloy de-
velopment program thus far indicate that consider-
able progress hos been mode toward achieving the
objective of developing a material that is as fab-
ricable as required for reactor
possesses adequate strength, and is corrosion
resistant in high-temperatyre fuel mixtures, liquid
metals, ond air. [t has been possible to improve
the creep characteristics by odditions of carbide-
forming elements, such as niobium and tungsten,

construction,

without seriously cffecting the other good charac-
teristics,  Several alloys of this type hove ex-
hibited a stress-rupture life of 1000 hr when
tested in fuel 107 at 8000 psi and 1500°F,

Studies were made of the elevated-temperature
properties of the beryllium blocks being fabricated
for the ART by the Brush Beryllium Co. The tests
indicate that the strength properties do not vary
appreciably from those predicted on the basis of
tests of smaller pressings.

3,6. Mondestructive Testing and Inspection of
Materials and Components

The recently developed Impedograph, which
measures both the a-c resistonce and the reactance
of its testing coil over a wide range of freguencies,
was applied successfully in eddy-current inspec-
tions.  This instrument simultaneously presents
inspection information from both the inner ond
outer surface of tubing, regordless of the materiol
or size of the tube,

Equipment and techniques were developed for
using an eddy-current proche coil to measure the
thicknesses of coatings or cladding., Measurements
made of the cladding thickness on Mark X MTR fue)
plates demonsirated the feasibility of the inspec-
tion techniques. With further experience it should
be possible to measure cladding thicknesses of
most olloys on o cermet core, of any of the oxida-
tion-resistant alloys over molybdenum or niobium

cores, of on austenitic stainless steel over a
nickel-molybdenum core, of o nickel-molybdenum
alloy over an austenitic stainless steel core, or of
the outer layer of duplex tubing. The ability to
measvre Inconel on austenitic stainless steel or
austenitic stainless steel on Inconel is currently
considered to be marginal.

Exposure dota are being compiled for the radi-
ography of various thicknesses of Inconel, stainless
steel, nickel-melybdenum, and similar alloys, Also,
test weldments in L{a*in. Inconel sheet are being
radiographed through 8 in. of beryllium to explore
techniques for the inspeciion of the closure welds
of the ART fuel-onnulus core shells,

A b-in.-dia cap, for use as a cotch basin, which
had been welded 1o the bottom of the sump tank of
the Pratt & Whitmey high-temperature eritical
assembly, was removed when leckage was detected
and was found to be severely cracked, It wos
learned that the cap had been machined from bar
stack rather than being a forged fitting, as speci-
fied.

[nspections construction of
reactor and test components resulted in rejection
rates for Inconel that voried from 0 to 63%, de-

of materials for

pending on type and lot of material, All the Hastel-
loy B inspected was found to be completely intferior
to requirements,

One of the six, lorge, beryilium hot pressings
fabricated for the ETU and ART reflector-moderators
was inspected under laboratory conditions to evalu-
afe the fechniques used for the field inspection.
The *‘squirter’’ technique being used in the field
was demonstrated to be satisfactory, if used with
care,

Welding inspections indicated a performance
record of B3% acceptance of critical welds and
over 95% acceptance of other standard welds.

PART 4. RADIATION DAMAGE AND FUEL
RECOVERY AND REPROCESSING

4,7. Radiation Damcge

Metallographic specimens were cut from in-piie
loop No. 4 which circulated o fluoride fuel mixture
in the MTR., MNone of the specimens examined thus
far have shown attack greater than L5 mils in
depth, and no signs of mass transfer hove been
found, Colored omorphous deposits were found on
the pump slinger and on the forward bellows that

were similar to deposits found in in-pile loop No. 3.

- XXX
 

The MTR-irradiated Inconel capsvle containing
beryllium oxide slugs was disassembled and ex-
omined, There was no evidence of BeO powder in
the caopsule, and ne significant changes wete ob-
served in either the lengths or diameters of the
slugs.

An opparatus for stress-rupture tests of Inconel
tubing under irradiation in the MTR was essentially
completed, Eight separote specimens will be
tested simultaneously. Duplicate specimens will
be tested out-of-pile at stresses of 2000, 3000, and
4000 psi ot 1500°F,

Rupture times of 1050 and 1650 hr were obtained
for two specimens of Inconel tubing tested at 2000
psi ond 1300°F while being exposed te a thermal-
and fast-neutron flux of 6 x 10'? neutrons/cm?: sec
in hole HB-3 of the LITR. These rupture times for
this 0.010-in.-wall tubing were in good ogreement
with out-of-pile data, A third in-pile specimen,
stressed to 1000 psi at 1500°F, did not rupture ia
1670 hr, ond the total creep deformation will be
measured uvpon disassembly of the apparatus.

Examination was completed on o tubular Inconel
specimen that was stressed in bending at o moxi-
mum stress of 1000 psi ond irradiated for 1120 hr
ot 1500°F while exposed to NmF-ZrF_‘-UF4 (63-25-
12 mole %, tuel 41) on the inside and sodium on
the outside. Corrosion on the fuel side was negli-
gible and did not oppear to depend on the sign or
magnitude of the stress, On the sodivm side there
was attack to a depth of cbout 3 mils,

Two MTR-irradiated and twe control copsules
fabricated of Inconel ond filled with fuel mixtures
of the NaF-ZrF -UF, system conteining 2 mole %
UF, were opened ond examined. The capsule
exposed to a flux of 1.5 kw/ecm?® for 676 hr ot
1500°F and its control capsule showed corrosion
penetration so negligible as to be confused with
surface faults, The copsule containing 4 mole %
UF ,, which was exposed to a flux of 3.7 kw/em®,
wos attacked at the mid-point to a depth of 5 mils
in 235 hr at 1500°F, The distribution of the attack
was that anticipated on the basis of the tempera-
ture profile. The contribution of irradiation to the
attack connot be judged until control date ond
chemical analyses of the fuel are available.
Hastelloy B capsules were tested out-of-pile, and
stmilor copsules have been inserted in the MTR,
Means for applying oxidatien resistent cootings
to the Hastelloy B capsules cre still being investi-
gated.

Xxxii

Qut-of-pile tests of the pump designed for circu-
lating fused-salt fuel in the LITR vertical in-pile
loop have demonstrated that the newly designed
ZrF ~vopor control features are satisfactory. An
in-pile loop incorporating a similar pump is being
prepored for insertion in the LITR. Facilities are
being plonned for the operation of similar foops in
the Oak Ridge Research Reactor.

During preparotions for filling the LITR wertical
in-pile loop, it was discovered that the interior of
the dry bux wus contominuted by on unknown
gaseous sulfur compound, Sulfur in the Necprene
rubber gloves was found 1o be the source of the
contamination,

Irrodiotion  exposures of nine Cu-B,C samples
were completed in the MTR. Exominations of six
of the sampies ore under way,

The applicability of palarographic techniques ta
the study of the fission-product chemistry of fused
salts is being investigated. Means for accommo-
dating the high electrical conductivity of the fused
salts, as compared with aoqueous solutions, are
being developed.

A survey was made of the relative variation of

nevtron flux with distence in hole 51N of the ORNL
graphile reactor by using sulfur ond cadmivm-
covered gold detectors. A plot is presented of the
relative activities of the detectors as a fuaction of
distance from the caoncrete shield.

The minority carrier lifetime of n-type germanium
has been found to change under irradiation according
tc a power law for both fost-neutron and gomma
irradiation. lrradiation with Co®? gamma radiation
and subsequent annealing show o complicated

the barrier change of diodes and
The results of tests indicate that the
properties of such devices would be changed by o
pulse of gomma irradiation ond that there would
not be immediate recovery upon removal of the

structure In
traonsistors,

pulse,

A safe was designed for the removal of samples
from the ORNL graphife reactor without subjecting
personnel to radiation, The device can be used
at the reactor loading face, or, by transferring it to
another dolly, at holes 50, 51, and 52,

An exposure device has been designed for use
in hole 30N of the Graphite Recctor with which o
sample can be moved from the external shield to
the irradiation position in about 10 sec. This
device will facilitote measuremenis on semicen-
ductor compoenents under irradiation.
 

The loss tangents of polyethylene, polystyrene,
Teflon, Nylon, and phenol formaldehyde were
measured before and after a series of Co®? irradic-
tions,
ofter exposures of 10% r; however, the dielectric

All samples showed measurable changes

constent changes of these materials were less than
the accuracy of measurement, that is, £5%,

4.2. Fuel Recovery and Reprocessing

Construction of the fused salt—fluoride volatility
pilot plant is essenticlly complete, except for the
faciliies for handling the ARE fuel, which will
not be installed until runs with nonirradiated
material are completed, Shokedown tests have
been storted,

The embrittiement of nickel and nickel alloys
with sulfur was studied to determine the allowable
sulfur contamination in the volatility process. A
nicke! foil fest wos devised which will detect
3 ppm of sulfur in o fused salt. Underordinary con-
ditions, however, in equipment with heavy nickel
walls, o sulfur contamination of 200 to 500 ppm is
necessary 1o produce any effect in ane exposure,
but the sulfur effect is cumulative, It appeors that
the scavenging of sulfur by nickel means mat if
the salt is once exposed in the fused state to
nickel equipment it becomes relatively free of
sulfur, The sulfur embrittlement of Inconel was
found fo be much iess than that of nickel.

The gas-phase oxidation resistance of types 316
and 347 stainless steel was found to be quite un-
satisfactory in tests of their use
as containers for fused fluoride salts., This sub-
stanticted pilot plant experience with the charge
melt vessel,

An over-all porosity of 65% was found for the
type of NoF to be used in the absorption-desorption
step of the decontamination process. The porticle
and bed porosities were 48 and 33%, respectively,
There was little difference between different lots
of Harshow Chemical Co. material,

Experiments were conducted to
plating out of ruthenium on metal surfaces could
be used to achieve additional decontamination, In
the initial tests, low decontaminction was achieved
in ¢ nickal trap, but there wos only o small amouat
of activity involved and the surface area of the
nickel was small, |t is anticipated that the re-
moval of ruthenium with a nickel column moy be

laboratory

tast whether

made more effective by using a greater surface area
or by operating at o higher temperature,

PART 3. CRITICAL EXFERIMENTS AND
REACTOR SHIELDING

5.1, Critical Experiments

A critical assembly mocking up the circulating
fuel reflector-moderoted reactor under study by
Pratt & Whitney Aircraft was operated at tempera-
tures near 1250°F, The fuel consisted of ¢ molten
mixture of sodium, zirconium, and uranium fluorides,
and the control rod contained a mixture of 30% rare
earth oxides and 70% nickel. With the control rod
out, the critical concentration was 10,97 wt % U
(5.1 mole % UF,) at 1258°F; with the rod inserted
to the midplane, the critical concentration was
12,20 wt % U, The temperature coefficient of
reactivity over the range of 1200 to 1350°F was
=3,6 x 10~ {Ak/k)/°F,

5.2. Shielding Theory

Several Monte Carlo calculations have been per-
formed on the Oracle to determine the dose rats,
energy flux, energy deposition, and in some cases
the energy spectra in laminated shields as a func-
tion cf the energy of the gamma-ray source and the
angle of incidence of the gamma rays on the shield.
Since it is impracticol to publish ali the results
from these calculations in o progress report, a
summary of the problems is presented along with
references to the published data,

5.3. Lid Tank Shielding Facility

A new instrument positioner has been installed
of the Lid Tonk Shielding Facility (LTS5F) which
will indicate the absolute position of any radiatien
detector in the tank water to within £0.5 mm, This
is o considerable improvement over the old posi-
tioner with which absolute pesitions could not be
determined to an accuracy greater than 15 mm.
The new system also provides for remote control of

the positioner,

Under Subcontract Neo, 390, Metal Hydrides, Inc.,
haes investigated the physical properties of lithium
borohydride-ammonia, which has been suggested as
o neulron shield. Pressure-temperature-composition
relationships hove been determined for materiol
temperatures ranging from 0 to 80°C, The gmmo-
nolysis reaction, which was found to occur to a
smoll extent, wos studied briefly hetween 50 and

XXX
 

100°C, The densities of liquid lithium borchydride
ammonicates were also determined,

A survey of shielding materials, which was per-
formed by the Technical Research Group under
Subcontract No. 931, confirmed that lithium hy-
dride and polyethylene are the most effective neu-
tron shields known to dote. Lithium hydride is
the only lightweight neutron shield material that
can be used in the temperoture ronge of 700 to
1200°F, which is the range thot may be expected
in some parts of the direct-cycle reactor shield.
The high-density materials required for gamma-ray
shields still center around lead, bismuth, tungsten,
uranivm, etc,

XHXIV

5.4, Tower Shielding Reactor |l

From @ nuclear parameter study and a cursory
examiration of the heat removal problem, the
dimensions and composition of the Tower Shielding
Reactor Il (TSR-Il) core have been established.
The diameter of the internal water reflector will be
17.5 in,, and the radial thickness of the core will
be 5.5 in,, with an aluminum-to-water volume rotio
of (.7, This will permit the use of the standard
Bulk Shielding Facility reactor fuel plate ond
water moderator arrangement, that is, 60-mil-thick
fuel plates separated by 120 mils of water, The
cooling water flow rate will be 1000 gpm,
Part

AIRCRAFT REACTOR ENGINEERING

S. J, Cromer
1.1. AIRCRAFT REACTOR TEST DESIGN
A. P, Fraas

APPLIED MECHANICS AND STRESS ANALY SIS
K. V. Meghreblian

Tests of ART Structural Design Criteria

Experience in the design of structures for opera-
tion ot high temperatures, that is, temperotures
in excess of 1000°F, is limited. Some experience
has been gained throughout the technology from
the design of jet and rocket engine components
and wvery high-performance steam power plants,
but little of this experience is directly opplicable
to the design of a system such as the Aircraft
Reactor Test {ART). Therefore considerable
effort has been devoted to establishing suitable
design criteria for this reactor.

The high-temperature phenomeno that will be
encountered in the ART are creep, reloxation,
and strain-cycling. The design problems associoted
with creep were discussed previcusly in connection
with the stress analysis of the north-head structure,
and the design criterio used for this onalysis
were given in a previous report.! The design
criteria for strain-cycling and relaxotion onalyses
were discussed in connection with the core shell
tests.’ The only test to date which could yield
any information on the volidity of the design
criteria selected for the ART was the core-shell
low-frequency thermal-cycling experiment.?  Only
one such test has, as yet, been completed, but
the resulss indicote thet the strain-cycling criterion
used for the core shells is conservative. It was
not possible from this single test to deduce the
degree of conservotism of the criterion.

Further tests ore under way or have been pro-
posed with which to evaluate the odequacy of
the selected design criteria. In the heat exchaonger
test progrom presently under way it should be
possible by suitable control of test procedures
ond  careful
obtain

selection of operating programs to
information on strain-cycling and
creep choracteristics of Inconel, as well as ins

useful

formation on the effects of ‘‘cumulative damage.’
Such ¢ program hos been initigted, ond the

 

'R, V. Meghreblian, ANP Quar, Prog. Rep. Jure 10,
1956, ORNL-2108, p 19.

2R, V. Meghreblian, ANP Quar. Prog. Rep. Sepi. 10,
1956, ORNL-2157, p 23-26.

first of these tests wos completed on Black,
Sivalls & Bryson heat exchanger type {HE-8 in
IHE-B (see Chap.
Development and Testing’’).
under

1.4, '"Component
This test was made
steady-state conditions with @ maximum

test stand

metal the heat exchonger tubes
of the order of 1650°F. Since only a few relatively

minor thermatl cycles were imposed on the system,

temperature in

the principal loads on the structure were those
dve to the fluid drag forces on the tubes, The
life of the heat exchanger wns estimated on this
basis, and the cafculotions indicated that the
system should have operared for 600 hr before
failure occurred. The test was shut down after
500 hr without o failure. This ottempt to predict
the life of a complex structure under creep and
corrosion conditions demonstroted the importance
of obteining creep measurements of Inconel in
fused salts at the corrosion rates expected in
the reactor.

Additional heat exchanger which will
include power-cycling, ore presently being pre-
pared, These tests will be conducted on both
the intermediate {IHE) and the small {SHE) straight-
tube types of test heat exchangers,?

tests,

In an attempt
to simulate the helicol shape of the octual fuel-
ta-NaK heat exchanger in the ART, o special
small heat exchanger (SHE-9) has been designed
and is now being fabricated (see Fig. 1.4,16 in
Chap. 1.4, *"Component Development and Testing”').
The tubes in this unit ore in the form of 180-deqg
arcs, and the composite design will incorporate
many structural features found in the ART heat
exchangers., When tested under conditions which
the ART design conditions, this heat
should demonstrate the odequocy of
the reactor compoenent.

simulate
exchanger

Thermal Stress Analyses of Shells

B. L. Greenstreet 5. E. Moore
D. M. Miller J. R. Tellackson

A detailed progrom of analyses is presently under
way in order to determine the distribution of stresses
in the various shells that result from design-point

 

*Ses, for example, L. H, Devlin and J. G. Turner,

ANP Quar. Prag. Rep. Sept. 10, 1956, ORNL-2157,
Fig. 1.4.5, p 50.3 £ ?
ANP PROJECT PROGRESS REPORT

temperature and pressure profiles. Particular em-
phosis is being placed on the determination of the
thermal stresses produced in the shells. The
temperatyre distributions summarized in the previous
progress seport! are being used as o basis for
these calculotions.

thermal

In the computations of the
assumed
This

is not a valid assumption in certain local regions

corresponding stresses, it is

that the entire structure remains elostic,

of the shells wherein the temperature profiles
are extremely steep.
deformation will occur,

in these regions, plastic

Nevertheless, it is ex
pected that except for a few regions of extreme
distortion the elastic onalyses will give adequate
estimates of the strains induced in the structure.
The strains thus obtoined are being used to de-
termine the strain-cycling life of the structure.
This procedure was applied in a study of shelis
V and V1.3 The analysis of shell V was completed®
ond is summorized here,

The function of sheli ¥ is to separate the boror-
tile curtain around the outer surfoce of the heat
exchanger from the fuel, and thus it serves pri-
marily @s @ container, lts importance as a
structural member stems from the foct that a rupture
in this shell would expose the tiles to the fused
salt, and it is possible that such exposure might
cause @ significant change in reactivity because
of the high boron content of the tiles.

Shell ¥V consists of three ports, as shown pre-
viously,® and each part is ottached separately to
shell W1 {the pressure shell liner), The boron
tiles are contained between these conceniric
shells, The inner surface of shell V is bathed
with fuel from the core, ond the outer surface of
shell V| is bathed with sodium from the island
circuit, At full power’ these fluids operate ot
markedly different temperatures, and this situation
gives rise to differential expansion be-
tween the two shells, Because of the apprecioble
difference in thickness of these shells (f in.

for shell V ond ?{B in. for shell VI), the growrh of

thermal

 

4R, v, Meghreblian, ANP Quar. Prog. Rep. Dec. 3!,
1956, ORNL-2221, p 3-12.

SFor shell structure of ART, see R. ¥. Meghreblian,
éNP1 uar. g'rog Rep. Dec. 31, 1956, ORNL-2221,
g . o] P

8B, L. Greanstreet of al., Thermal Stress Analysis of
Shell V intermediate, ORNL CF-57-3-6 {to be published).

7R, V. Meghreblian, ANP

uar. Prog. Rep. Dec. 31,
1956, ORNL-2221, Table 1.1.3, p 10.

shell ¥ is limited by shell V!, ond thus the stress
analysis s
shell V,
The most critical arec in this ossembly is the
intermediate section which extends from jus?t below
the equator to about 30 deg south latitude (Fig.
1.1.1).  This section is exposed to large tem-
perature variations relative to shell VI, which
produce free radiol expansions at the equator of
0.02) in. and at the 30-deg south latitude junction
of 0,033 in. (note that these growths represent in-
terferences between the shells V and VI}. In od-
dition, there is a relotive axial growth of 0,028 in.
Much of the distortion produced by these relative
expansions appears as locol effects in shell V ot

reduced principally to @ study of

the attachment points te shell V|. Because of
the smell thickness-to-radius ratic of shell V,
however, the stresses orising from the edge dis-
plocements and rotations diminish ropidly with
Jatitude ongle.

UNCLASSIFIED
DREHL-LR-DWG 20652

 

 

 

SHELL M
EQUATOR .,

= M —_—y

Y
SHELL ¥
TRANSITION PIECE
|
e %

AREA UNDER CONSICERATION
FQR ol TERNATE CONFIGUR&TIONS
SEE F1G 11.2 —= ===

Fig. 1.1.1. Intermediate Section of Shell V.,

The localized noture of the siresses allows
for the reduction of the stress level by decreasing
the difference between the mean shell temperafures
ond the temperature grodients in shell V neor the
Changes in the temperature dif-
ferences and gradients from those present in the

junction points,

configuration of Fig. 1,1.1 can be accomplished
by wvarying the shell thickness neor the junction
and/or by crecting small regions of stagnant fuel
around the junction., The lotter effect can be
ochieved by placing o baffle around the edge of
the shell.
Since the lorgest temperoture effects occur of
the lower edge of shell V {Fig. 1.1.2), this will
also be the locetion of the largest stresses. De-
tailed fwo-dimensional temperature calculations,
for which reloxation methods were used, were

carried owt for wvarious configurations of this

QRML -LR-DwG 208535

BOROM TILES
SHELL VI

  
    
   
   
 
 

SHELL V

THIS AREASHOWS ATTACHMENT
COMFIGURATIONS CONSIDERED

SHELL VI

SCDIUNM PASSAGE

A= FLAT PLATE JUNCTION
8 - CONMNE JUNMCTION
€~ CYLINDER JUNCTION

Fig. 1.1.2, Lower Junction of Shell ¥I and Intec-

mediote Section of Shell V.

region. [he mean temperatures in shell V for two
representative configurations considered in these
studies are shown in Fig. 1.1,3. The calculations
revealed that the largest reduction in relative
expansion between the two shells could be
ochieved by the use of the fuel baffie mentioned
chove. It was also found that both the gradient
ond the near the
iunction were affected by a chonge in the thickness
of shell ¥V, while a change in the thickness of

mean temperatyre di tference

the stagnant fuel [ayer affected only the ftem-
perature difference.

The temperature distributions obtained from these
studies were opplied to three geometric shapes
for the tronsition section between shells ¥ and
Vi, a cone, a cylinder, and a flat plate, as shown

in Fig. 1.1.2.

Stress analyses were performed

PERIOD ENDING MARCH 31, 1957

URCLASSIFIED
QRNL-LR-OWG 20854

 

 

 

 

 

      
  
  

 

 

 

 

 

400 ——

E %

< |_CONFIGURATION | e

: ; |

O i

2 300 = | ! | _! L

L) f CONFIGURATION [0 , _

é T — T I 1 [T 1§71 3

= 200 141 C1-SHELL V WITHOUT BAFFLE | | | .‘

E O-SHELL V WITH SBAFFLE SHOWRN

& N FIG L2 —F—+—F— —/

& | | i |

2

=100 f | Pl | N | | . 1 | . {

= Mol

&

= = 7 ' [

Ll

ol b i l - | | | . |
0 2 4 € a 10 i2

DISTANCGE ALONG SHELL W FROM LOWER JUMCTION WITH SHELL
VI INUNITS OF Yg INGH

Fig. .I-] .3|
Temperature of Sedium at Junction of Shells ¥ and VI,

Meon Temperature of Shell ¥V Above

for each of these geometries in which the tem-
perature gradient and the temperature difference
across the tronsition section were used aos para-
mefers. Secondery parameters in this study were
the section thickness ”Ilé and }'4 in.) and the
thickness of the stagnont fuel region in the vicinity
of the junction {0 to ]f'ré in,). The results of these
calculotions were compared on the bosis of the
maximum combined stress occwiring in each con-
figurotion (usually at the junction between the
transition piece and shell ¥), The combined
stresses were computed according to the octohedral
shear stress theory. In two dimensions the octahe-
dral shear stress is given by

_ 2 7
T = oy + 95 g, 0, ,

0l )

where o, and o, denote the principal stresses in
a bioxial system, The results for the three ge-
ometries with o }’lfi-in. layer of stagnant fuel are
summarized below:

T{psi}
Flat plate with baffle 20,500
Cylinder with baffle 40,400
Cone with baffle 100,700

Without the botfle, the maximum stresses were
significantly larger.
section, the octahedral shear sftress was 25,000

psi.

In the case of the flat-plate
ANP PROJECT PROGRESS REPORT

The final design for the junction, selected on
the basis of this study, is o lf’lé-in. flot plate with
a 74-in. region of stagnant fuel. This general
configuration is vsed at other juanction of shells
V and VI, but the balfle is omitted because the
differences are oppreciably lower.

A 20,500-psi value for the octohedral shear stress

temperatyre

corresponds to a fatigue life of about 300 cycles.
The ART will be power cycled, at mest, 25 times.

Design Analysis far Creep Bending

0. L. Platus

It was mentioned above that the Black, Sivalls &
Bryson heat exchanger type IHE-8 used in the
design criterion test was operated at steady state
and thot under the tubes are
subjected to a steady fluid drag force. These
forces ploce o bending sfress on o short section
of each tube (Fig. 1.1.4), and in a first approxi-

these conditions

mation the tubes behave os cantilever beams in
the wvicinity of the headers. In order to predict
the stress.rupture life of these tubes it is neces-
sary to toke into account the effect of creep on
the stress distribution over the cross section of
the tube. Calculations indicote that if this effect
is incfuded o reduction of about 20% con be
realized in the moximum fiber stress in comparison
with that predicted by the elastic theory. This
could result in as much as a fivefold increase in
stress-rupture life, as indicoted by the tensile
stress-rupture data. The calculations® were based
on a method® in which tensile creep data are

 

8p. L. Platus, Siress Analysis of a Heat Exchanper
Tgbe in' Creep Bending, ORNL CF-57-2.101 (Fab: 52

9G- H. MacCullough, ""An Experimental and Anclytical
Investigation of Creep in Bending,'' J. Appl. Mechanics
1, 35-40 (1923).

LWNCLASSIFIED
ORNL-_R-DwG 20855

—= FLID FLOW

— DRAG FORCE

SECTION I BENOING
STRESS
" hicapen

Fig. 1.1.4. Drag Load on Heat Exchanger Tube Stem,

utilized to predict the creep rate and stress distri-
bution in a beom under ¢ bending stress. In this
method it is assumed that o quasi-viscous state of
creep exists, by considering the siresses fo be «
function of creep rote ond time ond 1o be inde-
pendent of strain, and that the cross sections
remain in a plane during bending, The lotter
assfififlfifi leads to the relation

(= e,
where é(y) is the strain rate in the beam at distance
y from the neutral axis, €_ is the sirain rate in
the extreme fiber at y = b/f.r;, ond 5 is the outside
|f the tensile creep data

!

diameter of the tube.
con be ecsily expressed analytically, the equation
may be used in on integral statement of the moment
at any section to obtain an analytical solfution
of the stress distribution and the maximum strain
rate as o function of time and the applied moment,
If the creep deta connot be represented by a con-
venient mathematical expression, the equation
must be solved with the moment relation by a
For this purpose a set of
stress vs creep-rate curves with time as a pora-
meter is required, such as those in Fig. 1.1.5.

This general oppreach was used to onalyze an

aumerical technique.

Inconel heat exchanger tube in pure bending stress
at 1500°F in an NafF-ZrF ,-UF, (50-46-4 mole %,
fuel 30) environment, and @ numerical calculation
was mode for which the creep-rate data presented
in Fig. 1.1.5 were used, A determination of the
stress distributions and the moximum creep rate
for a constant moment gave a calculated maximum
elastic stress of 6500 psi {these conditions were
selected in order to shorten the time required for
o creep-rupture bend test).,  The results of the
calcolations are shown in Figs, 1.1.6 end 1,1,7,
Figure 1,1.6 gives the maximum stress and creep
rate in the tube, and Fig. 1.1.7 gives the stress
distribution at various time intervals.

Two important observations may be bosed on
this study. First, the calculated maximum stress
is quickly reduced to about 80% of the predicted
elastic value and remains essentially constant
for the life of the member. Second, constant-stress
creep data are required for calculations of this
type, rather than constant-load data.

An experimental attempt to verify the ''reduced-
stress effect’” is being made with lead specimens

subjected 1o pure bending loads. The immediate
PERIOD ENDING MARCH 31,1957

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

1072 ORNL-LA-0WG 20B56
4
L
5 |
5
—
E
s 2
c
e
=
= 4
) to
=
x
2
o
o
¥
o 5
o
5
2
ldfi
1000 2000 000 4000 5000 6000 o000 8000

STRESS (psi)

Flg, 1.1.5. Tensile Creep Dato for Annealed Inconel Tested at 1500°F in an NoF-ZrF‘-UFJ‘ {50-46 4 Mole %,

Fuel 30) Environment. (Secret with caption)
ANP PROJECY PROGRESS REPORT

C LASSIFEED
CRNL =L R-DwG 20857

 

{51075
$000 [ — - ——| — 500
\
8000 |+ — T % 800
|
. 7000 i - e %m/ - 700 T
g | \ l | /bflq [ ;§_
~ SO0 ".'\ oo — s _451.;:'4_. &0 E
@ | || MAXIMUM STRESS ) :
w5000 T T T e S 500 =
£ “redl | MAXIMUM CREEP RATE /1530 i
< 4000 T g e O] a00 A
5 i “-'-r--.i-_s.--;qaj &
£ Sigel _ . Cwl || 300 W
x Wi/ «
; :;Uf | [

£ ¥
S 3
|

!
“'-F.&l&.
fi

| |
g B

 

ol— — 0
o 1C0 200 320 350
TiAE Lir)
Fig. l.l.6. Maximum Stress ond Creep Rote in on
Annealed !nconel Tube Under Bending 5tress in an

NaF-ZrF‘-UF‘i (50-46-4 Mole %, Fuel 3)) Environment,

(Secret with captian)

UNCLASSIFED
CRHL-LR -DWG 20858

 

o
kel

o
oo

o
~

L
@

a5

o8

FRACTION OF MAXIMUM FIBER DISTANCE
ABOVE MEUTRAL AXIS

  

 

 

 

3000 4000 5000 6000 7000
STRESS {psi)

o
o oo 2000

Fige 1.1.7, Stress Distribution in an Annealed Inconel
Tube Under Bending Stress inan NnF-Zer -UF_‘1 (50-46-4
Male %, Fuel 30) Environment. (Secret with caption}

to predict the
lite of these specimens con the

objective of the experiment is
bending-rupture
basis of the tensile creep data.

Radiator Stress Analysis Problems

D. H. Platus R. V. Meghreblian

Detailed stress onalyses of the ART fin-type
NaK-to-gir rodictors are presently under
Although these anoiyses are not yet complete,

way.,

severa| critical areas in the design have been

identified, ond alterations required by stress con-
Three
problems have been examined, in particular, which
oppear to occount for some of the foilures ex-
perienced with these units. In these three analyses,

siderations have been made in the design,

studies were mode of weight and air drag loads,
relative thermal expansion between fins ond header,
and relative thermal expansion between adjacent
tubes as a resvit of flow stoppaoge from oxide
plugging,

The ART radiator consists of a butterfly arrange-
10 of fins ond headers with the fin banks
resting on support plates rigidly attached to the

ment

headers. Expansion of the hecders relative to
the fins will lift the tube-fin banks off the support
plotes during power operation and thereby cause
the tubes to behove as cantilever beams subjected
fo a uniform distribution of drag and weight loads,
Since the tubes are conmnected by the fins, the
tube bonk will The effect of

the fins is to produce a restoring moment at each

behove as a unit,

tube-fin connection that is proportional 1o the
slope of the tube at that point (Fig. 1.1.8). This
complex array of connected tubes may be treated
as @ homogeneous matrix if the system of discrete
restoring moments at each fin-tybe joint is replaced
by a continuous distribution. The problem is then
analogeus to the classical “beam on elastic
foundation.”” A colculation based on this model
was performed for the ART radiator, ond the air
and weight loads were found to give tube stresses
at the heoders of the order of 1500 to 2000 psi.

In the second problem, an analysis wos mede
of a thermal strain-cycling effect induced by ¢

 

”:ll... P, Carpenter, |, T. Dudley, and M. H. Cooper,
ANP Quar. Prog. Rep. Dec. 31, 1956, ORMNL-2221,
Fig. 1.4.18, p 48,

UNCLASSIFIED
ORNL-LR-0WG 20859

FINg _._H_,,“)

M FIM
\ DEFORMATION

TUBES}:‘

Fige 1.1.8, Tube-Fin Deformations in Radiotor.
temperoture difference between the hecoder drum
ond the nearby fins. The unfinned portions of
shown in Fig. 1.1.9, will be subjected
to bending as a result of cooling of the fins to
below the header drum temperoture. Based on
strain-cycling ond stress-strain chaoracteristics
1500°F, a plastic onalysis'' in-
dicated that temperature differentiols large enocugh
to be likely to cause foilure were produced in the
0.5-Mw test units, The radiator units presently
being fabricated hove been modified to occcom-
modate the temperature differenticls. The length
I indicated in Fig. 1,1.9 has been set at 1}4 in.,
and the entire fin bonk haos been sliced ot
3.]/3-in. intervals (Fig. 1.1,10). (See Fig. 1.4,19 in
Chap. 1.4, ““Component Development and Testing'’;
this radiator, now being prepared for testing,
will be provided with o thermocouple grid to meao-
sure the temperature distribution of the air lgaving

the radiator.)
In the third problem, an onalysis was made of

the effect of oxide plugaing of one or more tubes
in the matrix. The plugged tube(s) will be colder
than adjacent members and will thereby be sub-
jected fo thermal strain as a result of relative
expansion, Under these circumstances the fins
behave as diaphragms and each exerts an axial

tubes,

of Inconel at

force on the tube which fends to pull the tube
awoy from the header (Fig. 1,1.11). The mognitude

 

e & Field, Rectangular Cross Section Cantilever
Bgam; f;: the Plastic Range, ORNL CF-56:6.85 (June
20, 1956).

UNCLASSIFIED
ORNL-LAR-0WG 20863

= FIN AND TUBE
e = / MATRIX

LOWER FIN

 

     

HE ADER
INDIVIDUAL TUBE

 

 

 

 

 

i

i . +
HEADER DRUM

 

 

Fige 1.1,9. Differential Expansion Between Header

and Fins,

PERIOD ENDING MARCH 131, 1957

URCLASSIFIED
ORNL-LR-DWGE 20984

    

 

 

 

 

 

 

    

FINS
Fig. 1.1.10. Radiater Modifications.
UNCLASSIFIED
ORNL-LR-DWG 20HE0
HEADER

F

 

%

FIMNS

(TS5
Fra

3
-

e
PLUGGED TUBE [(COLD)

ADJACEMT TUBE [HOTH

Flgs 1.1.11. Fin Loods on Plugged Tube.

of the force is determined by the flexural stiffness
of the fin, by the stiffness of the tube in tension,
and by the distance of the fin from the header.
By trecting this discrete set of forces os a con-
tinuous force distributed along the tube, it is
possible to write a differential equation for the
axiol stress in the tube ond o corresponding
equation for the bending stress in the fins as a
function of the position along the tube. This
model indicated that the fins were relatively stiff
and produced very nearly the maximum strain
{0 AT), that is, the omount that would be induced
with a system of completely rigid fins. Never-
theless, it wos not possible, on this basis, to
ANP PROJECT PROCGRESS REPORT

demonstrate that the maximym temperature dif-
ferential conceivoble in the system could couse
failure of the tubes under the cyclic conditions
of the test or of plugging. Metallurgical experience
indicates that the resistonce of Inconel to plastic
strain cycling is seriously impaired by embritile-
ment in oreas wetted by the brozing olley. Since
two radiotors failed under conditions which in-
dicated tube plugging, the embristling effects of

the brazing alloy on Inconel are being studied,

Core-Shell Thermal Shield
D. H. Platus

Measurements obtained from the one-halfscale
ART-core-model volume-heat-source experiment!?
indicated that the core shells might be subjected
to surfoce temperature oscillations os large as
80°F at frequencies of from ]{" to 4 cps. Series
of tests ond experiments are therefore under way
in order to determine the fotigue life of laconel
tubing under simuloted ART conditions. If these
tests demonstrate that Inconel does not possess
fatigue strengths adequate to meet the opersting
conditions onticipated for the ART, it moy be
necessary to introduce design modifications to
olleviate the temperature oscillations in the core.

Several possible modifications have been sug-

gested. One proposal is to mount a pair of thin
Inconel shells in the core to serve as '‘thermal
shields'’ for the structurol core shells. '3 These
sheils would be similor in appearance to the

actual core shells and would define two reiotively
thin flow channels, as shown in Fig. 1.1,12, The
resulting hydrodynamic structure should be con-
siderably more stable than thot in the present
core configuration. The instability in the present
design results from several effects. One effect
is associated with lorge redial temperature dif-
ferences in a fluid flowing in o thick passage
when a lorge volume heat source is present in

the fluid.'* Another effect is the characteristic

 

'2N. D. Greene et al., ANP Quar. Prog. Rep. June 10,
1956, ORNL-2106, p 222.

13D, H. Platus, A Bajjle Design to Shield the ART
Core Shells From Temperalure Oscillations in the Fuel,
ORNL CF-57-1-134 {Jan. %, 1957).

44, F. Poppendiek and L. D. Palmer, Forced Con-
vection fleat Transfer Between FParalle! Plates and in
Annuli with Volume Heat Sources Within the Fluids,
QRMNL-1701 (May 11, 1954).

10

UNCLASSIFIED
GRNL-LR-0wS 2086)

N
BAFFLES = i\.‘:nl :
R
N 1
Fig., 1,1,12, Proposed ART Core Boffles. (Secret

with caption)

flow separation that occurs in o lorge-angle di-
verging channel.'® Ancther is the radiol pressure
variation associated with swirl obout the island.
These adverse effects would be alleviated within
the thin channels, ond the haffles would shield
the core shells proper from the temperature oscil-
letions in the thick central onnulus.

Calevlations indicate that the divergence angle
of the varicus channels can be maintained below
that likely to induce flow separation by careful
selection of the thermal shield configuration.
A further odvantoge to be gained from the use of
baffles is thet the moximum uncooled core-shell
temperatures predicted with the baffles are sub
stantially * lower than those in the present core,

Comparative sketches of the temperature profiles

 

ISJ, Nikyradse, Forsch. Gebicte lagenieurw., B 289,
1-49 {1929},
  

UNCOQLED wall
TEMPERATURE

POTENTIAL WALL,

 

  
   

PERIQD ENDING MARCH 31,1957

UNCL ASSIFIED
QORMNL=-LR-DWG 20862

UNCOOLED Wal L
TEMPERATURE

  
      

HAFFLE TEMPERATURE

 

B S A AT AR A

TEMPERATURE
VARIATION
MIXED - ME AN MIXED = ME AN
TEMPERATURE TEMPERATURE
\ MAXIMUM CORE-SHELL
TEMPERATURE
SODIUM S0DIUM

TEMPERATURE \y

 

Lel

TEMPER.&TURE\B’

 

 

YL 7S

LA}

Fig. 1.1.13. Radial Temperoture Profiles in ART Core {2) Without and (&) With Core Baffles. (Secret with

caplion)

for the two cases are shown in Fig, 1.1.13. More-
over, because of the smaller radial temperature
difference within the thin channels, the total power
extracted by the sodium circuit would be reduced
by opproximately 50%. Even if the sodium system
were to fail and no heat was extracted through ‘the
core shells, the steody maximum tempercture of
the shells would not be more thon 50 to 100°F
obove the mean fuel temperoture, depending on
the channel thickness. A comparison of axidl
temperature profiles with and without the boffles
for channels of Y-in. constant thickness is pre-
sented in Fig. 1.1.14,

Several disadvantoges and problems associated
with the use of boffles have been recognized.
The neutron poisoning effect of the Inconel baffles
would require o critical mass increase of 10 to
15%; the thin baffle structures must be designed
to withstand the static and dynamic fluid forces;
and design of the entrance region to distribyte

UNCLASSIFIED
ORNL-LR -0WS 20864

 

VTO0

1600

§ 00

1400

1500

TEMPERATURE ["F)

1200 =

1100

 

 

 

 

CISTANCE FROM CORE INLET {a}

Fig. 1.1.14, Axlal Temperature Profiles in the ART
With and Without Baffles In the Core. (Secret with

coption)

11
ANP PROJECT FROGRESS REPORT

the flow properly between the channels and the
central cnnulus may be o difficult problem. These
various factoers ore presently being considered,
and experiments are being prepared for deter-
mining the opfin.-num configuration for the boffle
design,

CORE HYDRODYMNAMICS

W. T. Furgerson J. J. Keyes
W. J. Stelzman

Flow studies of the full-scole 21-in. plostic
model of the ART core were continued in order to
investigote wall temperature fluctuations, which
may be onalogous to the fluid tempercture fluctu-
ations observed in the one-half-scale volume-heat-
source experiment, 216 aqad 10 test flow guidonce
devices in either the core entrance region or in
the fuel annulus that would reduce flow reversals
ond other instobilities. An ofttempt was mode to
measure a transient surfoce-heat-transfer coef-
ficient by utilizing o hot-wire probe located on the
ovter wall. The probe waos made by winding o
tight flat spiral of enameled Nichrome wire (0.006
in. in diometer) in the end of o 3,{‘ ¢ in.-0D metal
tube; epoxy resin woas used to hold the wire in
place. The face of the tube was then machined

 

G- W. Greene ¢! af.

ANP Quar. Prog. Rep. Sept. 10,
1956 ORNL-2157, p 227, g

LOW - TEMPERATURE
HELEUM IMLET

    

HELHOM OUTLET
GLASS-TG-METAL

 

 

Eriis e

 

 

—l—

Py,

 

{ NICKEL LEAD WIRE e

 

e |
AT T T,

 

 

A ]

|

e e —r— =

4.5 in

   

N

It shoyid
be possible from the variation in resistance of the
heoted wire with time to calculate the corre-
sponding fluctuations in the heat transfer coef-

to expese the surface of the spirai wire.

ticient arising from boundary layer fluctuations,
if the temperature coelficient of resistivity of the
wire were known. Attempts were made to measure
the temperoture coefficient of resistivity of the
wire, but unfortunately the resistance of the wire
was not stable at the temperature required to moke
a satisfoctory determination.

Improved probes are being developed in which
the sensitive elements will be thin strips of nickel
{vacuum evaporated) or silver {cpplied by painting
from sclution) on the ends of a plastic rod. It is
believed that these elements will hove good
response to high-frequency fluctuations and suf-
ficiently high temperature coefficients of resis-
tivity to permit operation ot reletively low probe
temperatures.
clternative
boundary-loyer fluctuations in the core, meosure-
ore made with o helium-cocled surface
thermocouple. The design of the thermocouple is
illustrated in Fig. 1.1.15.  The thermocouple
““‘gun-barrei’” type,!” and it is

In  an technique for measuring

ments

element is of the

 

7). E. Mott and A. G. Smith, Jr.

ANFP Quar. Prog.
Rep, Dec. 31, 1956, ORNL- 222],p54

UNCLASSIFIED
QANL-LR-DWG-20863

36 =n.- 0D, 0.00-in, - WaALL STEEL TURE

"Ya-1n.- 0D INCONEL TUBE

-

THERMOCOUPLE
JUNCTION

S5TEEL PLATE

. i
STEEL "GUN BARREL - TYPE THEARMOCQUPLE

Fia. 1.1.15. Schematlc Diagram of Helium-Coofed Surface Thermocouple,
mounted in a thin steel plate across which a
thermal gradient is maintained by means of a
steady blost of helium cooled by liguid nitrogen,
The resulting heat flux is sufficient to maintain
the probe surface at o temperature obout 60°F
below that of the ambient core fluid, whose temper-
ature is constant at 95°F, Data ore being obtoined
for several core geometries by the use of this
technique.

The flow-guidance investigations were continued
with studies of the core entrance region alone, the
core entrance ond equatorial regions simuite-
neous ly, and the core island bellows region within
the center volute of the axial-flow type of header.
The configuration used ot the core entrance con-
sisted of a combination of guide vanes, designoted
CA-1, and a series of annular tiners, designated
shrouds 1 and 2, which split the flow entering the
central annulus of the header into separate con-
centric ¢ylinders of fluid, The outer cylinder of
fluid was allowed to proceed downward at its
natural flow angle, whereas the component of
downward flow of the inner cylinder of fluid was
increased by the guide vanes. The purpose of
this configuration was to reduce the flow reversol
Complete probe traverses
of this system, os well as observations of dye
injections, indicated that no significant improve-
in flow characteristics were obtainad in

ot the core entrance.

ments
comparison with the charocteristics of the swirl-
type core header No. 2 without inlet vanes and
with a simuslated conica! island expansion joint.!8

The second configuraticn tested consisfed of o
combination of a core inlet arrangement, designated
GS5-2 inlet guide vane and GS-2-P10Q conical baffle
Flure,w in combinatien with a slotted disk, desig-

 

]‘BW. T. Furgerson ond W. J. Stelzmon, ANP Quar.
Prog. Rep. Sept, 10, 1956, ORNL-2157, p 26.

'9%. T. Furgerson, W. J, Stelzman, and D. B, Trauger,
ANFP Quar. Prog. Rep. June 10, 1956, ORNL-2106, p 25.

PERIOD ENDING MARCH 31, 1957

nated ‘‘flame ring’’ 1, which was located 6 in.
above the equator. The purpose of this combination
was to stabilize and distribute the flow evenly
across the annulus above the equator while pre-
serving as much of the rotational component as
possible.

Results of the test showed an increase in fuel
circuit resistonce equivatent to 1.3 ft of working
fluid. Claw-probe traverse data showed that the
flow distribution was improved both rodially and
circumferentially in comparison with that of series
17 tests of a contiguration which had GS-2 inlet
guide wvanes and a G$-2-P10 boffle plate but neo
“flame ring’" and in comparison with that of series
20 tests of a configuration which had o bare header
and core configuration with no inlet guide vanes
or botfle plotes. The degree of variation of oxial
velocity with time was decreased, aolso, os ob-
served by the decreased angle through which
injected dye traces fluctuated, These effects on
the flow distribution were more morked below the
“flame ring'' than above it.

The configuration of test series 24 consisted of
the bare hecder and the core used in series 20
tests, but with the center body in the north-head
region modified by an inverted conical section.
The purpose of this modification wos to decrease
the adverse axial pressure gradient along the upper
islond surfoce and consequenfly to increase the
axial flow component in this region.

Claw-probe traverses showed that more flow hod
been directed downward along the island surface,
In the region above the equator this was apparent
as a reduction in the amount of reversed flow. An
undesirable side effect was shown by dye traces;
nomely, a region of large low-frequency fluctuations
ol flow with time was apparent which could lead
to temperature fluctuations in a volume heat scurce.

13
ANP PROJECT PROGRESS REFORT

1.2. REACTOR PHYSICS

A. M. F'erry

ART SHIELDING
C. M. Copenhaver A. M. Perry

A program of calculations hos been undertaken
to re-evaluate the shielding of the ART ond such
auxiliary equipment os the fuel fill-and-drain
tank, fuel-recovery tank, etc. The first step was
to compute the gamma-ray dose rate at the outer
surfoace of the ART neutron shield, without taking
inte account the effects of the several penetrations
for MoK lines, fuel lines, etc. The effects of
such penetrations must be considered separately.

Gamma-ray source strengths in the ART *Z were
obtained through the use of two-dimensional
neutron-diffusion calculations. [n computing the
dose at the surface of the neutron shield (the outer
loyer of the water region), on infinite-slab model
was used thet waes based on the geometry ond
scurce strengths in the equatorial plane of the
ART. The plane-to-sphere transformation® was
then oapplied to each portien of the source.
Secondory gomma rays from the lead shield were
neglected. The justification for this is that in the
Lid Tank Shielding Facility (LTSF) tests of the
gamma-ray dose rates in reflector-moderated
reactor mockups with various lead thicknesses,?
the dose varied exponentiolly with lead thickness
up to about 4.5 in.; the ART lead is 4.3 in. thick.
These dota ore interpreted to mean that the
intensity of gaomme rays penetrating the 4.3-in.-
thick lead shield of the ART wiil be much greater
than the intensity of gamma roys produced in the
lead.

The heat exchanger wos regorded as a distributed
source of gamma rays, the reflector as a fomily of
five plane surfoce sources, and the core as a set
ot six plane surface sources. The shells were
treated as plone surface sources. The energy

 

IR. B, Stevensen, ANF Quar. Prog. Rep. Sept. 10,
1956, ORNL-2157, p 32.

24, W. Bertini et al., Basic Gamma-Ray Data for ART
Hedt Deposition Calcuiations, ORNL-2113 (Sept i7,
19 56),

*E. P. Blizard, Introduction 1o Shield Design, ORNL
CF-51-10-70, Part |, revs, p 52 (Jon. 30, 1952),

4G. T. Chopman, J, B. Dee, and H. C. Woodsum, ANF
Quar. Prog, Rep. June 10, 1955, ORNL-18%6, p 190,

14

spectro? were divided into energy groups with
upper limits at 0.75, 1.5, 3, 5, 7, and 9 Mev.

The buildug factors used were those for water,
since the outer layer of the shield is 31.5 in. of
water. The buildup factor was uvsed in the form

Blur) = a + bur ,

with the constants & aond & evolvated fo fit the
buildup-factor curves® for ur between 5 and 15
mean free paths.

The calculoted dose rote ot the surface of the
water bag {r = 170 cm) is 1840 r/hr. The relative
importance of each source is indicated by the
percentage coniributions to the total dose rate:
core, 22%; outer core shell, 9%; reflector, 12%;
reflector shell, 35%; heat exchanger, 22%.

In such a calculation the gquestion arises of
whether the buildup-factor method can yield volid
results for the reflector-moderatedreactor geometry.
The gomma rays penetrate @ substantial thickness
of lead, and, since the lead absorption cross
section has a minimum for gammea-roy energies
from 3 to 5 Mev, it might be supposed thot most
gomma rays entering the water would have energies
in this range, For the higher energy sources, use
of water buildup foctors might then lead to appreci-
oble errors.  In order to check this point, on
attempt was mode to determine the energy spectrum
of the gamma rays entering ond leaving each
component of the shield. Use was made of the
energy spectra from monoenergetic point sources in
infinite homogeneous media, as computed by
Goldstein and Wilkins.?> The sources were divided
into a set of point sources, ond the dose rate at
the inner surface of the lead shield was computed
for each source energy group. Attenvotion and
degradation of the rodiation in the lead were then
computed® as if the flux incident on the inner
surface of the lead were duve to a set of point
monoenergetic sources at the center of ¢ hollow
covity inside the lead, The same procedure wos
applied to the water with a new set of effective

 

SH. Goldstein end J. F. Wilkins, Jr., Calcuwlation of
the FPenetrations of Gamma Rays, Final Report, NYO-
3075 (June 30, 1954).
peoint monoenergefic sources determined by the
energy spectrum emerging from the |ead.

This method of calculatien yields a gamma-ray
dose rate of 1610 r/hr ot the shield surface. The
energy fluxes at the shield surface in each gamma-
ray energy group are listed below:

Gamma-Rey Eaergy Interval Flux
(Mav) {Mavfcrnz-sec)
x 10%
0.25-0.75 108
0.75-1.25 86
1.25-3 271
3--5 303
5-7 147
7-8 185

PERIOD ENDING MARCH 31, 1957

The fast-neutron dose rote ot the shield surfoce
was computed from LTSF data,? with the use of
the transformation from finite-disk surface source
to spherical surfoce source and the factors thot
accounted for the additional boral aftenuation and
source-to-shield surface ratic in the
LTSF configuration, as formulated by Dee and
Otis.® For an LTSF equivalent surface source
strength of 0.428 x 10~? w/em?, the calculated
fast-neutron dose rate ot the ART shield surface
was 3.86 rem/hr, Thus ot o distance of 50 ft from
the center of the ART operating at 60 Mw with
a 4.3in.-lead and 31.5-in.-water shield, the
calculated gamma-ray dose rate is 22.8 rem/hr and
the fast-neutron dose rate is 0.05 rem/ hr.

increased

 

6). B. Dee ond D. R. Otis, ANP Quar. Prog. Rep.
Dec, 10, 1955, ORNL-2012, p 204.

15
ANP PROJELUT PROGRESY REPORT

1.3. ART INSTRUMENTS AND CONTROLS

E. R. Mann

C. 5 Wolker

R. G, Atfel

FLOW DECAY IN ART PRIMARY MoK SYSTEM
FOLLOWING POWER L05%

W. F. Werts'

A NcK pump power foilure will produce temper-
ature and prassure tronsients in the ART heat-dump
S5y stem,
istics were therefore investigated for use as o

The Noll circuit flow-decay choracters

guide in evaluating the significance of such o
power loss, This investigation was based on the
premise that flow is proportionel to pump speed
when other fuid cirewit conditions are constant.
Experimentsl dota oo the transient chorocter-
istics of a NaK circuit were ohigined for o loop
in which an ART prototype NaK pump was being
tested at high temperotures. The loop contained
about & B of circulating Mak, During shutdown
tests, high-speed motion pictures were taken o¢f
the pump shoft and of a pointer atfached to «
synchronous This photographic
record provided o refiable and accurote method

fiming motor,
of determining the number of pump-shaft revelutions
gs a function of time, With o comgplete tabulation
of pumpeshelt revelutions vs time esfoblished,
speed vs time data could be obtained through
numericol differentiation,

Severol shutdown tests were mode from each of
two initial conditions, The NaK flow wos es-
sentiolly the some for both conditions, but the
‘nitial pump-shaft speed in one cose was about
three-fourths the initial speed in the other cose.
& hand valve was adjusted, as required, to provide
the desired flow for the pump-shaft speed vsed,

The date were compared ond evalucted aceording
to the foilowing system-operation-decay snolyticol
relotionship:

[ power 5upp|ied] {

power to
e svstem

cirentate fluid
T pawer to change
l kinetic energy

When the system driving power was terminated, the
relotionship became

( powveer o
civeulate fluid

() 0 =

E power fo chcnge]
P kinetic energy |

If N is rotor speed and k, ond k, cre squotion
constants, the

[ power fo

b 3
cirgulote fluid | = HE = & N

ond the

[kinetic energy] = KE = kzN?' :

axpressed in power-time units, Then the
. | power to change Z{KE) . an
(2] L kinetic energy | T T T Zky N A

¥ Eg. 2 is substituted into the original ex-
pression, £q. 1, the following expression for

speed vs time is obtoined:

an

¥

at

 

(3 0 = &, N% & 2k,

The selution 1o £q. 3 can be shown to be in any
of the following lorms, with the subscript O indi-
cating the initicl values at steady-stote conditions
bofore the start of operation decay:

 

 

 

» | kot }
{4} — = —,
N2k, W,
| N 1
(5} - ¥ e
My (¢ o/ KE /2y + 1
‘ IINQ ‘.‘ Irf Z'KEG\‘
(%) . k, -] - )
Y J \ HP,

ln Egs. 4, 5, and 6, N and ¢ are the only variables.

Plots of the speed decay vs time were made for
various initial kinetic energy-to-power rotics, as
1.3.1.  The curves shown are
thearefically volid for ony iype of pump, fluid
system, or davice, so long as the system power

shown on Fig,

proporfionel to the cube of the
mass-—velocity involved., The initial kinetic energy-
to-power rotio must be determined with the kinetic

consumption is

 

1 . .. Do
Or ossignment fraom Beodix Products Division,
 

o0

PERIOD ENDING MARCH 31!, 18257

UHCLASSFIED
CRNL~-LR—0OwWG 20867

 

o
——
=

    
 
 

b

SPEED DECAY (%% OF INITIAL SFEED)

 

 

|/ ANITIAL KINETIC ENERGY-TO-POWER RATIO (AT 100 % SPEED)
' |

=

* THE TIME UNIT MUST BE THE SAME AS THE
TIME UNIT OF THE KINETIC EMERGY EXPRESSION

 
 

 

| ' ' g

 

 

 

 

10 45

20 25 30 33 40

TIME AFTER LOSS OF POWER * {sec]

Fig. 1.3.1. Speed Decay Choracteristics of a Centritugel Pump upon Loss of Power,

energy expressed in power-time wunits {such as
horsepower-seconds, watt-hours, etc.} to give a
term having the same units of time as those
applied to the abscissa of the curve (seconds,
hours, etc.).

The data were analyzed in detoil for two shut-

downs from the following initial steady-state
operating conditions:
Condition Rua 1 Run 2
Shatt speed, rpm 3506 24678
Fluid flow, gpm 1220 1216
Dischorge head, ft 394 212
Fower supplied, ha 119.3 68.8

Partial checks of data from several other tests
indicated almost identical results for the con-
ditions invelved, All the analyzed test dota
closely motched the computed decay characteristic
shown on Fig. 1.3.1 for on initial kinetic energy-
to-power ratio of 2, that is, KE /HP; = 2, and
speed decay down to about 30%. Below the 30%
speed-decay region, the system mechanicol friction
becomes increasingly more significant, since the

system kinetic-energy driving force has dropped
to less than 10% of the original value.

The operating decay characteristics were sub-
stantially the same for both sets of conditions,
and therefore only the test data for the shutdown
frem a shaft speed of 3506 rpm were fully reduced
to completed plots, as shown in Figs. 1.3.2 ond
1.3.3. The pressure characteristic curve shown
on Fig. 1.3.2 was obtained by using the premise
that pressure varies as the square of the speed
and the flow rate, With other conditions constant,
the flow rote is considered te be proportional to
the pump-shaft speed, as previously indicated.

The analytical expressions for the two test
conditions would be identical if the initiol kinetic
energy-to-power ratios were the same. The power
ratio for the two conditions is 119.3/68.8 = 1,735,
and the experimental data indicated that the decay
characteristics for the two sefs of conditions were
identical; therefore, the kinetic energy ratic must
he neorly the same as the power ratic, An investi-
gation of the rotating kinetic energy ratio for the
two conditions reveoled that

[N

\ %,

%

_ ff?fi) . 1715 .
7678

7
ANP PRUOJECT PROGRESS REPORT

00

UMCLASSIFIED
ORNL=-LR— WG 20868

 

20 : —= . L

 

50 [——— wde
o |—
30 | e e — —

| --PRESSURE
20 | — .

CECAY CHARACTERISTIC [ % OF INITIAL WALUE )

19 |— — =

 

a0 | —_— i e -— 4 PPR

 

 

20 25 30 35 a0

TIME {sec]

Fig, 1.3.2. Flow Rate ond Pressure Decoy in MaK Circuit After Loss of Power to the Centrifugal Pump.

Thus it can be seen that the rotating kinetic
energy-to-power ratios were about the same for
the two test conditions.

The fluid kinetic energy cannot be included in
the analytical expressions, since the flow was
substantiolly constont, In view of this, it would
seem that the flow inertia is negligible, as com-
pared with the rotating inertia, in its contribution
to the system kinetic energy. If the foregoing
deductions are considered to be the case, the
major contributors to continuing system operation
after o power loss are the rotating elements of
the pump ond the electric drive motor. This also
means that the only appreciable contribution to
any extended continving fluid circulotion during
the decay period is received from the stored
kinetic energy in the mechanically rotating parts,
end that the pump is stili the driving force and
is inducing ftuid displacement in a manner sub-
stantially the same as during the previous steady-
state operation,

Since the flyid flow is dependent upon the pump
driving force, fluid travel around the circulating
loop should be substantially proportional to the
pump ongular displacement. The information pre-
sented on Fig. 1.3.3 was determined by opplying

18

this principle to the test data described above in
combination with the computed volumes of seg-
ments of the ART primory MaK heat-dump circuit,
The volumes used in the preparation of Fig. 1.3.3
were those designated circuit 1 in Table 1.3.1.

LIQGUID-METAL-LEYEL TRANSDUCERS

G. H. Burger A, M. Leppert
R. E. Pidgeon, Jr.2

Tests of the commercial resistance-type liguid-
metol-level transducers described previously® were
concluded,  Operation of the NoK level fest
focility was continued until February 21, ot which
time only one of the four transducers originally
instailed in the test focility remained in operating
condition. The performence of the transducer that
remained in operation wos excellent for a period
of 3732 hr, during which time the NoK leve! of
the system was cycled 65,000 times. The four
transducers were removed from the test facility
for examination, and the unit that remained in

 

204 loan from Radlo Corp. of Americo.

R, E Pidgecn, Jr., ond G. H. Burger, ANP Quar
Prog. Rep. Dec. 31, 1956, DRNL-2921, p 20.
PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED

ORML=LR-DWG 20869

 

|
[-‘ii__'_

 

 

T
h‘““-—-
T

1777

/ RADIATORS

o

 

 

 

/.

 

.

NN
\ \

 

Lol

 

L HEAT EXCHANGER

|

 

F O{W ER LO-’SS

/L

OLO PIFE

g )

 

1
/

sz-ml // / /
DESIGN POINT

_[ OPERATION

]

/1

~

o]

|
—
W

 

RADIATORS

 

—c ]

4
/

 

AN\

|~

 

HEAT EXCHANGER

 

 

 

FLUID MOVEMENT ARQUND CIRCUGT
RELATIVE LOOQP SEGMENT POSITIONS

  

POINT OF POWER Loss‘l""—-—-_______

/)Y [/
////7/

—~
\\
NN

A
Y
&

 

 

 

 

 

7/ RADIATORS

 

 

g
/

//

 

!
HOT PIPE

 

 

[/ /

HEAT EXCHANGER]

 

 

DR

\\\

N7

/
A)LD PIPE

\

 

T
[
[
b,

{ i

RADIATORS.

 

_3‘_

 

 

 

 

 

L

»-,____-'::
]
._\-

\

 

/]

 

"

 

 

—{Q -3 0 5

[//

|
15 Z
TIME [sec)

0 25 30 35 40

Fig. 1.3.3. NoK Flow Rote Decay Charocteristic After Power Loss Based on Loop Test Dota and ART System

Daesign Yolumes. {(Contidential with coption)
ANP PROJECT PROGRESS REPORT

Toble 1.3.1. Volumes of NaK in Vorious Segments of the Four NoK Clrcults of the Art Primary Heat Dump System

 

 

NoK Volume {f1)

 

Segmant

Total for ths

 

 

 

 

 

 

 

Circuir 1 Circuir 2 Circuit 3 Circuit 4 Fasi Eoreuiian
Radiator 3o pump 1411 1.411 1,411 1,411
Pump (circulating) 0.26 0.25 0.26 0.26
External cold pipe t.787 1.5446 1.305 1,044
Intermal cold pipe 0.762 D.646 0.4656 0.735
Branches (to header pipe) 0.123 0.123 0.123 0123
MNorth hecodsrs and connecting pipe 0.242 0.242 0.242 0.242
Total 4.585 4.228 3.997 3.835 16.645
Haot exchanger 0.722 0.722 0.722 0.722 2,888
South header and cennecting pipa 0.171 g.171 0.1 0.171
Branches (header pipe 1o commen line) 0.123 0.123 0.123 0.123
Interrial hot pipe 0.646 0.838 0.694 0.4687
External hat pipe 1.463 1.188 1.478 1.134
Pips {2 in.}) to elevation 834 1, § in. 0.454 0.561 D.748 0.631
Het pipe to rodictor tubes 1.292 1,392 1.292 1,392
Total 4.449 4.273 4.606 4,138 17.466
Rodiotor 0.476 0.476 0.474 0.476 1.904
Total Circulating Yolumse 10.232 2.699 2.801 2.171 38.903

 

*Circwits 1, 2, 3, and 4 are referred te on 7503 Facility

respectively.

operation for 3732 hr s shown in Fig. 1.3.4a.
Cne of the three units which failed scon after the
start of the test is shown in Fig. 1.3.46. The
failure of the unit was caused by MNaK seepage
inte the insulotion througn the crack visible neor
the bottom of the U-bend, and the other two
unsatistactory units failed in a similar manner.
The reaction of the NaK with the insulation coused
the visible deformation,

The NaK-level test facility is being medified
to provide facilities for testing CRNL-designed
and -fabricated resistonce-type level transducers.
These level tronsducers ore to be clamped into

20

heat-dump piping drawings as loops B1, A2, Al, and B2,

place in the surge tonks (instead of welded) to
facilitote installation and removal.

Fabrication of six ORNL-designed resistance-
type level transducers is under These
instruments will be similar to the commerciol
transducers, but they will have better temperature
compensation, and

way.

improved weld design. A
further design modification was made to increase
the range of level determination; that is, with the
ORNL-designed instruments, it will be possible
to measure the level closer to the bottom of the
tank. Design of a similar resistance-type level
transducer for use in the ART sodium expansion
tank was completed.
PHOTO I5 53

 

-— Tl T——

UNCL ALEIFIED

PERIOD ENDING MARCH 37,1957

UM ASSIFIED
PHQTO 28788

 

 

(e

{4)]

Fig. 1.3.4. Resistance-Type Liquid-Metal-Level Transducers After Operation for 3732 hr In o NaK-Level Test
Facillty. (a) Uait that remainad in operating condition throughout the test. (b) Unit that operated only 24 hr, but

remained in the test facility exposed to NaK for 3732 hr,

ON-OFF LEYEL PROBES

G. H. Burger A. M. Leppert
R. E. Pidgeosn, Jr.

Eight of the resistance-type on-off level probes
described previously! were constructed and in-
stalled in the surge tonks used in the fluid-flow-
facility for testing

measuring system in  the

 

44, M. Leppert, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-222%, p 29.

magnetic flowmeters in NaK. These all-welded
probes are being ftested os a replacement tor
conventional spark-plug probes, which are not
suitable for pressurized systems and temperatures
above 450°F. The surgetank level control system
that incorporates these on-off probes was in
operation on the test loop for about 2 doys, but
an apporent failure of the gas-pressure control
system ollowed helivm to be forced into the NaK

system. Some changes are being made in the

21
ANP PROJECT PROGRESS REFORT

on-off probe and its control circuit to give better
confrol and a better safety margin in operation.

The on-off level probe being tested consists of
a /fi in. -dlq copper wire contained in ond insulated
from a f -in.-0D, 0.025-in.-wall lnconel tube at
ali points, except the bottom, where if is welded.
A current of B to 10 amp, 60 cps, is supplied to
the probe. A relay across the probe is energized
when the NaK level is below the probe ond de-
energized when NoK touches the probe, A similar
on-off probe with an Inconel center wire and a
bridge amplifier for the control circuit is being
studied. The probe would con5|51 of a .-'"]6-”1 -dia
Inconel wire insulated in a ffi-m -00, 0.025-in.-
wall Inconel tube. The bridge circuit would be
balonced when the NaK touched the probe and
unbalanced when the NaK level was below the
probe. Since the tempercture coefficient of resis-
tivity of Inconel is low, the zero shift resulting
from a temperature change would not be so lorge as
in the present probe. A specification has been
prepared and submitted to manufacturers for o
packaged amplifier to be used in this bridge
circuit.

An especially fabricated spark plug with a
bery Hium oxide insuvlator wos tested in NaK as
on on-off level probe, The plug failed after 20 he
of operation becouse of short circuiting of the
insulator. The plug was removed ond reworked,
and it is to be retested when the NaK-level test
facility (referred to in previous section) is back
in ¢peration.

Several special plugs with insulators high in
alumina and magnesium content were received from
the Champion Spark Plug Co. for testing ond
evceluating. These plugs will also he tested in
the NaK-level test focility or other test rigs as
soon as possible.

The usefulness of the conventienal type of spark
plug as o reliable on-off probe is very doubtful
ot this time, The plug life is very short when the
probe is operated at temperctures above 450 1o
500°F, and there is always the danger of gas
leakoge through the insulator, This leokoge
problem could be very sericus if the plug were
used with radicoctive fluids, and it would be
necessory to completely con or seal the plug.
Therefore the use of the spark plug as a reactor
type of level element is not recommended, ond
the development and testing of the resistance type

of on-off probe is continuing. It is expected that

22

this unit will replace the spark plug except in
systems which can be operated safely ond reliably
within the spark-plug limitations, or in systems
in which the process fluid electrical conductivity
is too low for use with the resistive type of on-off
level probe.

FUEL-EXPANSION-TANK LEYEL INDICATOR
R. F. Ayland
Dynamic-Level Tests

Four helium-bubbler-type fuel-level indicators
completed 1000.hr tests at 1500°F in the dynamic-
level test rig, described previously,” without
plugging of the bubbler tubes, Two of the bubblers
were fabricated of 31’ -in.-0D, 0, 065-m.-wu|1 Inconel
tubing, ond the olher two were / -in,-QOD, 0.035-
in.-wall Inconel tubing. The hellum used for these
tests was passed through a Drierite filter and
through copper turnings maintained ot 1000°F in
order to getter the oxygen and remove the moisture.

After completion of these tests, the focility was
modified in an attempt to cbtain a better series
of level calibrations. The medifications included
a means for accurately vorying the fuel level
continugusly in the two vessels by using gos
pressure; a new bubbler design with a shallow V
in the bottom of the tube woas used; and the
bubblers were positicned within 0.010 in, of the
correct level while the vessel was at the operating
temperature.  An improved type of voriable coli-
brating probe was designed, installed, and cali-
brated ot the operating temperature. Also, the fuel
was thoroughly mixed prier to calibration to
preciude density errors due fo thermal gradients.
As a result of these modificotions, the firss
accurate, repreducible series of static level cali-
brations was obtained. The maximum level error
was 0.08 in, in 5 in., or 1.6%.
reproducible within 1.5%.

A dynamic-level test scheduled to operate for
3000 hr at 1500°F has completed 750 hr without
bubbler plugging. Prior to startup, however, with
the bubblers operating in static fuel ot 1350PF,
otl four bubblers plugged in a period of 24 hr, for
reasons unknown, Two bubblers were removed and
the contents sent for onalysis. The other two
were unplugged by roising the fluid temperature

The results were

 

SR F. Hyland, ANP Quar. Prog. Rep. Dec. 31, 1936,
ORNL-2221, p 29.
to 15300°F, The bubbler plugging problem hos
become so serious that a separate series of tests,
described below, is to be made to determine the
cause,

Static-Level Tests

The average life of a helium-bubbler type of
level indicater hos been found to be about 1100 hr
in g static-level system in which a NaK scrubber
is used to reduce the moisture content of the
helium ard in which copper turnings mcintcined
at 1000°F are vsed to reduce the oxygen content,
Since the average life falls far short of the 3000-hr
life specified, ottempts are being made to de-
termine the cause of plugging of the bubbler tubes.
Although the technigques for onalyzing fuel mixtures
for oxygen compounds are not precise, it is thought
that the moterial which plugs the bubbler tubes
is ZrO,. Therefore o special batch of oxyfluoride-
free fuel will be used in the next series of tests.,

It has been suggested that @ cold trap be used
in place of the NaK scrubbers for removing
moisture from the helium, since NaK sntrginment
might contribute to the plugging. Methods for
checking the oxygen content of the helium and
for reducing the oxygen content ore being studied,
Further, the vessels used for future tests are to
be hydrogen fired.

These recommendations are to be incorporated
into a simple series of bubbler tests in static
fuel, without level instrumentation, and the ad-
ditional precoutions described below will be taken,

1. An all-silver-soldered spectrometer-leak-tight
system, with bellows valves, will be employed.

2. The system will be flushed thoroughly by
alternately pressurizing with helium ond evacu-
ating.

3. The dewpoint and oxygen content of the
supply helivm will be continuously monitored.

4. The purge helium will be preheated to the
temperature of the fuel,

As a further check on the role of oxygen impurity
in the helium as a cause of piugging, one test
series will be run with heflium of known oxygen
content, as furnished in commercial gas, and
onother test series will be run with helium passed
over uraniym furmnings held at 600°C to reduce the
oxygen confent to as low a value os possible.

PERIOD ENDING MARCH 31, 1957

FRESSURE TRANSMITTERS FOR USE AT HIGH
TEMPERATURES

W. R. Miller

Evaluation tests of commercial transmitters for
use at high temperatures were continued. Two of
the six units mentioned previously® that employ
differential tronsformers were tested, The firs?
unit tested was not satisfactory in that the over-all
occuracy was only 3,3% full scale at 1200°F. The
other unit was somewhat better, with an over-all
accuracy of 2.1% full scale ot 1200°F. The
differential-transformer method of detection has
proved to be vausually difficult to use in that the
tield of the Calrod heaters used to maintain the
aperating temperature has on effect on the output
readings. shielding is
helpful but does not relieve the situation com-
pletely. At present two more units of the type
shown in Fig. 1.3.5 are undergoing tests.

The unit mentioned previously® that employs a
variable-permeance type of pressure transducer
was tested, but no datc were obtained. Lack of
suitable aligning guides allowed the core of the
transducer to rub the coil housing. This wunit
(Fig. 1.3.6), which was on consignment, has been

The use of magnetic

returned to the manufacturer,

Redesigned madels of the four-legged strain-
gage type of transmitters (Fig. 1.3.7) were also
tested. Armor was added to protect the capillary,
ond o shock-mounted sub-base was added within
the _ase for vibration isolation. The redesigned
caose eliminates zero shift when the cover is
tightened. Because of a poorly designed shipping
clamp the strain goges were domaged during
shipment, and therefore the test results for these
models were not up to the standord set by the
results obtained with the original six prototype
medels,  The shipping clamp hos now been
modified by the manufacturer.

The next lot of four-legged strain-gage trans-
mitters will feature a sensor housing with improved
stress choracteristics for use with more corrosive
process fluids. The details of this design change
have been satisfactorily resolved with the manu-
facturer.

 

Sw. R. Miller, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 30

23
ANP PROJECT PROGRESS REPORT

\ URCLASSIFIED
PHOYO 27350

 

 

Figl Ilals‘l

ducer.

Caollery Chemical Co, Pressure Trons-

24

A second set of pneumatic-cutput transmitters
completed a suvccessful 2000hr {ife test in a
NaK-filled test loop.
being examined metallurgically for weld defects
and corrosion,

These units are presently

MAGMNETIC FLOWMETERS
G. H. Burger C. L. Pearce’

Six of the 3)-in. magnetic flowmeters purchased
for the ETU and the ART were installed in the
tlowmeter test loop filled with NaK for calibration
agoinst a water-calibrated {(0,25% accurate) venturi.
As stated previously® the units will be calibrated
over the widest temperature and flow-rate ranges
possible with the test loop. These ranges will
cover, in general, the design ranges of the ETU
and the ART,

Operation of the loop began on February 12,
1957, and it had operated a totel of 40 hr when it
became necessory to drain the NaK for repairs
ond medifications of the loop. During the 40-hr
operating period, fluid flow rates of up to 1600 gpm
and fluid temperatures of up to 1350°F were
achieved.

The dota obtoined with the mognetic flowmsters,
while not ¢t all conclusive becouse of the short
operating time, did give general indications of
the operating choracteristics to be expected. [t
appears that ot least four of the six flowmeters
now installed in the loop will be acceptable for
the ETU if the accuracy specifications ore met
in subsequent tests.

Six of the ART-ETU 2-in. magnetic flowmeters
are being installed in another test loop for cali-
bratien against o water-calibrated {0.25% occurate)
venturi. Operotion of this loop will probably start
in May 1957,

The :’}é-in. magnetic flowmeters designed for plug
indicater and cold-trap flow rate meosurements
are being fobricated. Several of the units will
be tested for temperature effects on the magnets.
These units will not be calibrated because the
accuracy requirement is +15%.

 

?On toen from Radic Corp. of America.

BG. H. Burger, and C. L. Peorce, ANP Quar. Prog.
Rep. Dec. 31 1930, ORNL-2227, p 23.
PERIOD ENDING MARCH 31, 1957

LUNCL ASSIFIED
PHOTOQ 17347

 

 

 

 

 

MR gl;‘n o
ARNE I AN
——
AR AR AR KRR R NN R R R
Lt Jl'tfefrla||lr|1||1llt1|\|‘li“lt'la1‘1‘-
Fig, 1.3.6. Crescent Engineering & Rasearch Co. Pressuyre Transducer,
UNCL A5FIED
PHQTO 28160

 

Fig. 1.3.7. Taylor Instrument Companles Pressure Tronsducer.

25
ANP PROJECT PROGRESS REFPORT

REMOTELY APJUSTABLE PRESSURE
REGULATORS

¥, R, Miller

Two prototype models of remotely adjustable
pressure regulators were tested. These units were
from different manufocturers but were quite similar
in operation.
is possed through ¢ voicescoil type of trensducer
to position the pilot operotor of a pneumatic relay
system, The resulting output pressure is directly
proportional to the magnitude of the current in the
transducer coil. Tests performed thus far showed
these units to be stable under line-voltage, supply-
pressure, and normal ombient-temperature varis
ations. Tests will soon be initiated in which
these units will be subjected to extreme ambient
conditions in a 3000-hr life test,

A remotely adjusted direct current

TURBINE FLOWMETERS FOR USE AT HIGH
TEMPERATURES

G. H. Burger

The 1-in. turbine flowmeter, designoted unit No.
2-Rev. 3, described previously,? was installed in
a forced-circulation gos-fired Inconel leop con-
taining NuF-ZrF4—UF4 (50-46-4 mole %, fuel 30)
for calibration egainst an orifice by using Taylor
transducers as pressure sensors and for an en-
durance test of 3000 hr. Operation of the loop
was storted on January 8, 1957, and the turbine
operated satisfactorily for approximately 340 hr,
|+ was then noted that the turbine output frequency
was Considertu different from what it should
have been bosed upon the orifice-measured fiow
rate, and an investigotion showed that the unit
was not operating., An x-rgy examination of the
unit showed, as for as could be determined, that
the unit was not breken, |t was therefore believed
thot, since the unit hod been operoted in NakK
systems three times previously, the bearing had
been oxidized to some extent and that operation
in the salt mixture caused it to bind after a short
operating time,
the loop is drained ond exomined in order to
determine the cause of failure.

The unit will be removed when

A 3 .ia, turbine flowmeter was installed in a
forced-circulation loop filled with NaK in order
to test the unit under actual operating conditions.

 

t:"IG. H. Burger, ANP Duar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 25.

26

Prior to installation in the lcop the unit hod been
tested ond calibrated at a high flow rate in a loep
containing water, The unit performed satisfoctorily
during the calibration tests. The calibration data
are shown in Fig. 1.3.8, The only unusual
operating characteristic wos noted at a flow rote
of 600 gpm, at which point the tyrbine ‘'chattered’’;
at flow rates other than this no vunusual chatter
of vibrotion was noted. The total test time with
water was approximately 6 to 8 hr, and flow rates
of up to 1100 gpm were recorded. |t wos felt on
the bosis of the results of the water tests that
the unit wos satisfactory for calibration under
aperating conditions in a NaK system and that
any possible failure of the uait could not damage
the test loop.

UMCLASSIFIED
ORNL-LR—DWG 20870

T T T ]

 

n
o

 

B
o
|
|
&
o

|
|
™
&
PRESSURE DROP ACROSS TURBINE [psiol

 

LA
Lo
T
|
|
|
L
<

™
L)

i

|

|

   

 

>
|

1

| LT
PRESSURE DRGOP
o bz L 11| 5

o 200 400 ao0 BOOQ 1000 4200 1908
FLOW RATE f{gpm!

OUTPUT FREQUEMNCY
OF FLOWMETER |counts/fsec)
o

 

Fig, 1.3.8,
Turbine Flowmeter.

Water-Collbration Dota for o 3‘%-in.

Operafion of the loop wos started on February 12,
1957, ond the turbine flowmeter operated os
expected for approximately 12 min. The loop flow
was started at 420 gpm and was rapidly increased
te 1220 gpm, as indicated by the turbine flowmeter,
At some point between the flow rates of 970 and
1220 gpm, however, the unit failed in some manner.
At the approximate time of failure and possibly
during all or part of the operating time, the loop
opetators could hear chaitering of the unit., The
tempergture of the circulating NaK qt the time of
tailure was 155°F, The loop continued to aperate
ofter the turbine fallure and eventuclly reached
o flow rate of 1680 gpm and o maximum fluid
temperature of 1350°F, which wos held for only
a short time and then lowered. The loop operated
a total time of approximately 40 hr before pump
surges necessitated shutdown. The surges may
have resulted from ingassing brought about by

improper operation of the surge tank level system

while measuring venturi pressures.

On  Februery 14, 1957, before disassembly of
the loop, @ radicgraph wos made of the turbine
flowmeter which showed that the bearings were
broken and thas o number of blades were missing.
Cn February 18 the flowmeter was removed from
the loop, and the indications of the rodiograph
were confirmed, as shown in Fig. 1.3.9. Eight
of the ten blades were missing; the two remaining
blodes were bent opposite to the direction of
rotation; and both bearing shafts were broken.
The main throttling valve waos removed from the
system, and six of the blades were found therein,
A seventh blade was found under the volve housing

in the drain line. The eighth blade has not yet

PERIOD ENDING MARCH 31,1957

been found, but the seorch is continuing; the
piping is to be x-rayed, flushed, and swabbed.
The turbine hos been disessembled and will be
carefully examined in an effort to determine the
cause of failure.

ART THERMOCOUPLES
J. T. Delorenzo

incongl-sheathed closure
welds made by the Heliarc welding process are
being tested. Nineteen closure welds on sheathed
Chromel-alumel thermocouples that had previously
passed x-ray and dye-penetrant inspections were
exposed for 2100 hr ot 1500°F in sodium, and there
were no evidences of leaks. Thirteen similar
specimens are now being exposed to NaF.ZrF,-
UF, (50-46-4 mole %, fuel 30) at 1500°F. The

welds on three platinum~platicum-10%

thermocouples with

closure
rhodiym thermocouples are also being tested in

UHCLALLIFIED
PHOTO 2294

 

 

Fig. 1.3.9. Damaged 3},;3-in. Turbine Flowmeter as Compared with a Similar Undamaged Unit,

27
Tabkle 1.3.2. Results of Drift Tests of Sheothed ond Beaded Chromel-Alumel Thermocouples in Alr

Sheathed Assemblies®

 

at 1300, 1600, and 1800°F for Yarious Times

Special Beoded Assemblies? Mermol Seaded Assemblies®

 

 

Sheathed Assembl iesd

 

Test Peried
(hr} Deviction® Spreud‘{ Devigtion® Spreod-f Deviotion® Spreudf Deviation® 51:-r:a-ic;u:l-"r
(®F) (°F) (°F} (°F) (°F) (°F) (°F} (°F)
Test Temperatuce: Y300°F
a 5.0 3.3 -7.0 2.1 -5.3 1.1
1000 5.5 3.7 —~ 4.0 3.1 —4.4 2.2
2000 5.3 2.4 -3.8 5 —4,1 6.5
2500 9.3 2.2 ~0.3 3.0 -0.5 4.0
Test Temperature: 1800°F
0 1.8 2.8 6.0 2.0 -7.1 3.5 4.7 4.3
1000 6.4 2.6 3.9 2.5 -2.2 9.1 3.6 3.1
2000 7.0 2.7 6.2 5.5 —0.7 10.7
2500 9.1 3.3 9.4 5.3 -0.8 12.0
Test Temperature: 1800°F
0 1.0 2.6 -7.5 4.3 1.3 2.1 5.5 4.2
1000 7.7 15.1 15.2 19.2 8.7 14.3 10.2 10.5
2600 8.0 16.2 26.2 2.5 11.7 19.2
2500 11.5 20.3 31.4 20.3 19.8 31.0

 

9%ix ossemblies were tested and each assembly contained two thermocouples aged 24 hr ot 1350°F in heljum prior to testing.

bSix assemblies were tested and each gssembly contained two thermocouples fobricated with especially cleaned wire {aceuracy, iafs of 1% ever range

530 1o 2300°F) in carelully cleanad wells.

k
€Same as special assemblies except thot they received no speciol ¢cleaning.

ASume as assemblies described in footnote a sxcept that they ware aged for 200 hr ot 1350°F in helivm.

€ Deviotion of the averoge of the readings of the twelve thermacouples tested from the stondard test temperature of 1308, 1600, or 1800°F.

"’qu'rmum spreod of the twelve readings obtained.

IHOd 3y S5IHI0HS L3370 dNY
fuel 30. To date, the Heliarc welding process has
vielded 30 to 50% acceptable welds.

Bend tests have been mode in order to determine
the effect of bending the sheath on the output of
the thermocouple, Single U-bends on a ¥.in.
radius had very little effect, even when the Lent
area was exposed to o temperature grodient of
100 to 150°F per inch. The inaccuracies induced
were less than 0.5°F. The repeated bending of
¢ single specimen appeared to cause o 2 to 3°F
shift, but this change will be rechecked with
several more test specimens.,

The fabrication of the sheath weld similarly was
found to have little effect on the occuracy of the

thermocouple wire. Aijl deviations measured were

less than 0,5°F,

The drift of the data obtained with sheathed and
beaded Chromel-alumel thermocouples operating in
air as o function of time and temperature is being
studied. The results of tests at 1300, 1600, and
1800°F in Table 1.3.2, Data

are presented

FERIOD ENDING MARCH 37,1957

obtoined in the closure tests in static sodium,
described above, showed less than 3 or 4°F drift
of sheathed Chromel-alumel thermocouples at
1500°F in 2100 hr.

A program is being prepared for the Oracle tc
focilitate processing drift data. The emf vs
temperature curves for Chromel-alumel and plat-
inum—platinum-rhadium thermocouples have been
tabulated on the Oracle magnetic tape memory, and
o workoble program has been outlined in prepo-
ration for actual computations with the raw data.
The system will be checked with previocusly
processed dota before new data are used.

Fumaces in which the heating coils can be
moved at a uniform rate were designed and fabri-
catec for testing the uniformity of thermocouple
wire, Two fumaces are now ovailable which
give either symmetricol or asymmetrical gradient
patterns.  The maximuym gradients that can be
obtained with the two furnaces are 300 and 700°F
per inch, respectively.
ANP PROJECT PROGRESS REFPQORT

1.4, COMPONENT DEVELOPMENT AND TESTING

H. W. Savage

PUMP DEVELOPMEMNT TESTS

E. R. Dytko' A. G. Grindell
Bearing and Seal Tests

W. L. Snopp! W. K. Stair?

The dynomic tests of lubriconts under irradiation
thot were to be run in the LITR3 are, rather, to be
run in o gommo-irradiation facility in the MTR
canal. With the MTR facility it will be possible to
use o slightly modified fuli-scale pump rotary
assembly ond to examine the unit for damage
immediately upon removal from the gamma-irradi-
otion field. The present schedule colls for fest
operations to begin next gquarter,

The final selection of a lubricant-cootont for
the reactor pumps will be based on the results of
the gamma-irradiation tests in the MTR, Since both
orgonic liquids ond highly refined paraffin-base
mineral oils ore being considered at this time, all
parts of the pumps and the external lubricant-
pumping system are designed to be compatible
with either liquid., For exomple, copper is incom-
patible with organic liquids, and therefore copper
parts in the seals and journcal beorings ere being
reploced.  All-stainless-steel bellows-seal test
assemblies were ordered ond will be incorporoted
into the MTR test units. Journal bearings fabri-
coted from Alcoa alloy 750-T5 were selected to
replace the original bronze bearings. The small
percentage of copper (0.7 to 1.3%) in this alloy is
probebly insignificont with respect to the organic
liquids of interest, but a somple of the alloy will
be tested in UCON LB-140-X, an orgonic liquid,
before the MTR test is made. There are no copper
parts in the external lubricant-pumping system.

Two 1000-hr tests were conducted to determine
the suitobility of the aluminum olloy, Alcoo 750-T5,
for this particulor bearing application in which
the effective journal beoring arec is 2.4 in.%
Gulfcrest 34 was used os the lubricant in both
tests. Thetest conditions are given in Table 1.4.1,

 

1on assignment from Prott & Whitney Aircraft.
QCDHSU‘TDHT from the University of Tennessee,

3":':'- L. Snapp ond W, ¥. Stair, ANP Quar. Prog. Rep.
Dec. 31, 1956, ORNL-222), p 32

30

Neither test gave any evidence of bearing mal-
function, and both the bearings were equally as
bronze bearings tested under similer
conditions, Aluminum alloy becrings hove been
placed in rotary elemenis for use in various other
tests, and if satisfactory service is obtained the

aluminum alley will be used far all recctor pump

good os

journal bearings.

Table 1.4.1. Conditions of Test of Suitabillty
of Alcoe Alloy 750-T5 for Pump Bearlng Application

 

 

 

Test 7501  Teast 750-2
Shaft speed, rpm 3400 4400
Test temparoture, °F 180200 170-%20)
Operating period, hr, ot
specified journol bearing
loads:
300 1b 621 527
450 |b 287 214
00 b 166 407
Continyous operating 1074 1148
pariod, hr
Continued tests of the modified Durametallic

seals? being developed for use as seals in NaK
pumps have produced further encouraging results,
More than 5000 hr of operating time was logged for
such seols, including 2500 hr in a rotary element
pumping NoK ot 1400°F and 2500 hr in a me-
chanical shakedown stand with the seals subjected
to a wide range of diHferential pressures from
20 psig, which simulates a vacuum filling operation,

to 275 psig, which is more than a maximum
catastrophe pressure, and to rubbing speeds from
0 to 60 fps.

It wos found thot for the very high differential
pressures, obove 200 psig, a slight design modifi-
cation will be necessary to prevent relative motion
berween the positioning collor for the lower seal
aond the shaft, On the bosis of these gocd test
results, the Durametallic seals hove been specified
for use in the ART MNaK pumps.
Tests have shown the buno-N-base O-rings

recommended by outside venders for use with
UCON LB-140-X lubricant tc be quite satisfactory,
Similar tests showed the bunag-N-bose O-rings to
be satisfactory for use with Gulfcrest 34 lubricant,
and therefore this elastomeric secl can be used
for the reactor pumps if either of these lubriconts
is acceptable. [f, however, the MTR tests prove
these lubricants to be wnsatisfactory, the seal
material may ogain become a problem. For the
NaK  pumps, commercially aovailable
elastomer seals will be satistactory, since Gulf-
crest 34 will be the lubricant used.

similar

Fuel Pump Development Woter Tests

J. J. W, Simon*?

The development tests with water of an accept-
able impeller configuration® {designated model 32)
for the ART fuel pump were completed with the

 

4Dn_ assignment from Pratt & Whitney Aircraft,

SM. E. Lockey, G. Samuels, ond J. J. W. Simor, ANP
Quar. Prog, Rep. Dec. 31, 1956, ORNL-2221, p 33.

PERIOD ENDING MARCH 31,1957

tests described below. Performance tests were
run for liguid levels of ]/2 to 3 in. in the expansion
tank to encompass the level chonges expected
during the change from zero power to full power
operation of the ART. In general, operction is not
affected by a liquid-ievel change when the liquid
level is 2 in. or more above the floor of the
expansion tank, Below this level the fluid head
showad a slight decrease, the expansion-tank
fHlow rates decreased, the power required to drive
the pump decreased slightly, ond the fHlow rate in
the main loop decreased slightly as the liquid level
neared ]5 in.

The effects of decreasing the liquid level in the
system on the flow rate, head, and pumping power
ace shown in Figs. 1.4.1, 1.4.2, and 1.4.3, re-
spectively, It may be noted thaot the flow rote,
heod, and pumping power remained fairly constant
down to the %-in. level in the expansion tank.
Below that level, definite decreases in flow rate,
head, and pumping power occur, ond, also, although
not shown on the figures, fluctuations of all three

UNCLASSIFIED
ORBL~Li=DWs 20871

FLUMD REMOVED FROM LOCGP {lilers)

g 1 2 3 4 5 &

A A

 

    

 

 

3
o 18— - -
,:_: INITIAL COMDITIONS
; | 645 gpm [DESIGNI 535 g 735gpm
2 § Z100rpm & 2Z700eom x 2700pm K1
AR === o 2400¢pm =i
= =
Wil | o 2700 rpm .- =
b s 30G0rm S
& o
5 25 - a X300 rpm =%
. z
2 a
| ; ;
LOOP WOLUME 487 (~145 iters) (-,_5
30 |- f — — @
<f
&
i
| w
15 | — e L 1 | o ; e ] i ——] _ _ |
‘ | |
dsob—— 44—+ | 4 . | Lo | e 4 .
3 2 the ' 34 Q —a—— BELOW EXPANSION TANK FLOOR ———
STATIC LIGLAD DEPYH AROVE EXPANSION TeMK FLOOA LEVEL {in)
Fig. 1.4.1, Effact of Liquid Level Changes on Flow Rate in a Loop Test of an ART Fual Pump {Model 32

Impeller) with Water as the Pumped Fluid. (Secret with caption)

31
ANP PROJECT PROGRESS REPORT

URCLASSIFIED
ORHL—LR-DWGE 20872

FLUID REMOVED FROM LOOP {Nitars)

 

 

 

 

 

 

 

o ; 2 3 4 5 & ? 8 g 10 [} 2 13 ]
D e - n v - - — —
r— ! T | L | B =S i P ¥ I
1 ? I‘-“hm.
| ]
Bl - - - i b i 5 -
10 i"- = & i By
i ' THNITIAL CONDITIONS |
- G645 gorm 535 gpm 735apm 2|
5 | B 2100 rpm & 2700 rpm % ZFO0 epm
g 20 o 2400 rpm _ | - = bj
¥ o 2700 rpm L
= ® 30001pm =
9 & 3300 rpm =
& a
1&:_1 25 2= 5.
o LOOP VOLUME : 4 1134~ 113 liters } &
& - 5
x
30| N
35 [ —
40 | — — -
1
45 L = ok SRR = A . -
3 2k 2 tha 1 3 Ya D= —  BELOW EXPANSION TANK FLOOR —

STATIC LIOUID DEPTH ABOWVE EXPANSION TAnK FLOOR LEVEL (in |}

Fig. 1.4.2.

Effect of Liguid Level Chonges on Fluid Heod in o Loop Test of an ART Fuel Fump (Model 32

lmpeller) with Water as the Pumped Fluid. (Secret with coption)

varichles occurred when the level was changed
from ¥ to 0 in. The fluctuations remained, but at
o negligible level, in the next part of the continued
loop droining peried, during which the guantity of
tluid removed wos increosed from 8 to 11.5 liters.
With more than 8 liters of fluid removed, the liquid
level in the system was below the expansion-tank
The loop ingassed continuously, and the
ingassing was occompanied by large and repid
declines in flow rate, heod, and pumping power
after more fluid wos removed. Further draining
produced power fluctuotions which indicafed that

floor.

slugs of liquid and gas were surging through the
system.

32

Tests for determining degassing times were also
conducted with the liguid level in the expansion
tank at 3 in. and the pump operating at ¢ speed of
2700 rpm. It was found that degussing of an
injected air somple was complete in less thon
} min. The degassing time increased to about
10 min as the operoting speed was reduced to
about one-half of design speed (2700 rpm), as
shown in Fig. 1.4.4, line A. The degassing time
decreased with increased main loop flow rates,
as shown in Fig. 1.4.5, and the increcsed main
loop flow rates also increased the expansion tank
bypass flow rates tor the same operoting speed.
At the present design point of about 2700 rpm and
PERIOD ENDING MARCH 31, 1957

UNESL ASSIFIED
GREHL-LA-DWGE 20873

Pl REMOVL O FROM LOOE (Lilars)

   
  

RAaNGE \WuICATES SPREAD
OF DATA FROM ALL RUNS

e

DECREASE N PUMPING POWER [ %t

   

 

 

LOGP YOLUME. 412 (~ 113 1iters) i
=4
20 o
' 31
| 3
e
25 | i | "
. |
ol A1 L 0Ly ] . | |
3 2 1 1 %, g 0 —a——— BELOW EXPAMSION TANK FLODR ——=

STATIC LIQWD DERTH AGOVE EXPANSION TAMK FLOOR LEVEL lin)

Fig. 1.4.]. Effect of Liquid Level Changes on Pumping Power in o Loop Test of an ART Fuel Pump {(Model 32
Impealter) with Water as the Pumped Fluid, {Secret with caption)

UNCL ASSIFIED
DANL-LR-0WG 20874

 

 

 

 

3000 | - | — — - - .
|
| |
. |
2700| ! ! = - 1 P LI h—_ — 3
| CURVE A: SYSTEM RESISTANCE SET FOR
| | | 645 gpm AND 2700 rpm
2400 |— | L : e
| | ' CURVE B: SYSTEM RESISTANCE SET FOR
E | [ 845 gpm AND 2400 rpm
(=%
oy
Ll
uwl
[
N
=
2
5
i
Lt
o
O

 

|
Qa0 1 } ] | —: ¢ L B } 1 | — b i i =— ‘

|
600 L.___._ 1 b i b | - ' |
Q 60 120 180 240 300 360 420 480 540 600 580 T20 780 540
DEGASSING TIME [sec)

Fig. V.4.4. Effect of Operating Speed on Degassing Time for on ART Fuel Pump (Model 32 Impeller) Set for
Constant Resistance with the Liguid Level! 3 in. Above the Floor of the Expoension Tonk. (Secret with caption)

33
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL -LR-DWG 20875

 

 

0 ——————— ‘
70 - — i —
f'
80 | - .
o 580 gpm
¢ 50 | - i
—
0 \\I | |
=
& |
| _
2 40 ! A |
Ll
[ v .
' |
30 = 4*""—#
20 I : P
|
|
0 — ! = e
2400 2700 Elaee
OFERATIMG SPEED (rpml
Fig. 1.4.5. Eftect of Locp Flow Rore on Degassing

Time for an ART Fuel Pump {Model 32 Impellar) with

the Ligquid Level 2 in. Above the Floor of the Expansion
Tonk. (Secret with caption)

645 gpm, little, if any, increase in degassing time
cccurs with an increase in the exponsion tank
flow rate.

Fuel Pump High-Temperature Performance Tests

H. C. Young® P. G. Smith

Test operatien of an ART fuel pump with a
meodel 32 impeller and NflF-Zer-UF (50-46-4
mole %, fuel 30) as the circulated tlyid was started
in December 1956. At the cperating conditions of
645 gpm, 2700 rpm, and a fuel temperature of
1200°F, the head is 1 to 2 ft lower thon that
obtained for water with the some impeller, vofute,
and leop. Investigotions ore under way in order
to determine the reason for the discreponcy between
the woter and fuel performance dota, Considerable

difficulty has been experienced with the Moore

 

0 essignment from Frort & Whitney Alrcralt.

pressure-measuringdevices inthat the zero position
shifts from doy to day, and in order to obtain
relioble pressure data it is necessary to calibrote
these instruments before each test run by pressur-
izing the loop with helium to as high as 100 psig.

The off-gos riser connecting the pump pot to the
ZrF ,-vopor trap has been replaced twice during
the course of operation. The first time the riser
had to be replaced because a ZrfF, plug formed
when the fuel temperatures were
reach 1500°F. The

that was being maintained on the

and metal
inadvertently allowed to
temperature
riser line ond the vapor trop was low encugh in
comparison with the fuel temperature to couse
Zer to deposit and plug the riser. The line
was replaced without terminoting the test. The
off-gas riser hod to be reploced ¢ second time
when a leak developed. This time the test oper-
ation had to be shut down. Partial plugs of ZrF
have also occurred in the gas passage through the
pump barrel, possibly as the result of reverse flow
or back diffusion of gas through the pump borrel,
Performance and cavitation dota are being ob-
tained, and endurance tests are planned.

Fuel Pump Endurance Tests
P. G. Smith

The ART fuel pump being tested for endurance
wos shut down on November 29, 1956, with a total
of 2587 tir of Uperufing time and 120 thermal cycles
over the temperature ronge of 1400 to 1100°F.
The fluid circulated during these tests was
NoF-ZiF -UF | (50-46-4 mole %, fuel 30).

Examination of the pump impeller ofter this test
revealed that it wos in good cendition, and it
showed only very slight domage, if ony, as a result
of high-temperature operation.  Examination of
other parts showed that the upper metal O-ring seal
allowed fuel to leak into the annular space between
the pump barrel and the thermc! barrier can. This
cavsed the formaticn of o ZrF, plug in the riser
connecting the pump pot to the vapor trap, which,
in turn, caused the sluggish and errotic off-gos
flow noted during operation.

There was no evidence of lube oil legkage from
the pump seals during the tests, but upon disas-
sembly the lower seo! was found to be worn to
such an extent that it should have permitted o
small amount of leakage. There was evidence of
this leakage inside the thermal barrier con, which
was not filled with barrier material for this test
and which contained o small amount of carben
deposit.

The pump wos cleaned and reassembled with
new seals ond placed bock in operation for further
testing. It hos accumulated o total of 550 hr more
of continuous operation and 35 thermal cycles
over the tempercture range of 1400 to 1100°F.
The lube oil leakage rates are 5 cm3/day from
the lower seal and zero from the upper seal.

There had been no trouble with the hydraulic
drive for the past 1240 hr of pump operation. This
may be attributed to a back pressure of 100 to
120 psig being mointained on the discharge side
of the motor, os recommended by the manufacturer,
to ensure that the piston shoes are seated on the
wobble plate at all times.

Sodium Pump Development Water Tests
J. J. W. Simen

The hydroulic problems to be solved in the ART
sodium pump are not yet fully delineated. An
initial oppraisal ond the results of development
work tc dote indicote similarities between this
pump and the fuel pump, for which five major points
were investigated,’ nomely, system pressure
tluctuations, gas pressure grodient down the sheoft
annulus in the direction of purge gas flow, flow
through the centrifuge before ingassing occurs,
pressure levels based on ART stress limitations,
and cavitation. In considering eoch of these
oroblems in relation to the sodium pump, it is
apporent that, based on tests mode to dote, the
pressure fluctuations are small, the flow required
into the expansion tank is smell enough {5 gpm
maximum, per pump) not to cause ingassing through
the centrifuge, and the ART stress limitations do
not impose stringent requirements on the pressure
levels. The sodium pump will require a decreosing
gas-pressure gradient in the direction of gas How
{rom the oil catch basin down the shaft annulus
inte the exponsion tank (Fig. 1.4.6). Cavitation
date hove not yet been obtained, ond on investi-
gation must be made of the loop ingassing which
has cceurred at shrottled flow rotes.

A new pump barrel was designed and is being
fabricated that more closely mocks up the ART
sodium pump barrel ond the exponsion tank configu-
ation. The differences between the present con-
figuration and the new pump barrel, which mocks

 

71, 30 W Simon, M. E. Lackey, oand G. Somuels, ANP
Quar. Prog. REP- SE‘PL IO. 1956; ORNL‘E]&?; p dD'

PERICD ENDING MARCH 31, 1957

up the reactor system, are shown in Fig. 1.4.7.
The volute in the new configuration is now fully
instrumenied for measuring the radicl unbolance
force ot the centrifuge discharge area. Also, the
effect on pump performance ol fluid leakage be-
tween the heat exchanger inlet and the volute con
be determined.

The tests hampered by
ingossing of the system, which becomes more
pronounced whea the flow rate in the main loop is
reduced. Several runs were made fo obtain
performance data, however, and data obtained in a
low-speed run correfated reasoncbly well with
dato obtained in the eariier development tests.®
The data obtained in runs at speeds sufficiently
high to bracket the design head are shown in
Fig. 1.4.8.

No work has been done, as yet, toward the
solution of the gas-pressure grodient problem.
At present the expansion tank operaies at a pres-
sure level greater thon that in the rotary element
cil catch basin, A modified degussing disk will be
tested in the new pump barrel configquration.

to daote have been

Sodium Pump Barrel Seal Test

J. W. Cooke?

Free movement between the sodium pump barrel
and volute will be required in the ART to allow
for radial movement to accommodate thermol ex-
pansion and axiol misalignment resulting from
accumulative assembly tolerances. A seal must he
provided in this region to minimize leckage of
sodium between the heat exchanger inlet and the
pump centrifuge outiet regions. Two designs of
multiple-ring seols, Fig. 1.4.9, have been tested
with water.

The maximum leckage rote for seal design No. |
was 2.1 gpm at a pressure differential of 50 psi
and on axiol spacing of 0.015 in, The moximum
leckage rate for seal design No. 7 wos | gpm at a
pressure differential of 80 psi ond an axial spacing
of 0.0135 in. Both the pressure differential and
spacing at which these maximum leckage rates
occurred are more severe thon the estimoted ART
values. It is onticipated that the moximum leakoge
rotes given above will not seriously interfere

 

84, ¢, Young and M, E. Lockey, ANP Quar. Prog.
Rep, March 10, 1956, ORNL-2061, p 45.

70n assignment from Pratt & Whitney Aircroft.

35
ANP PROIJECT PROGRESS REPORT

UNCLASSIFIED
CRNL-LR-DWG 20876

fi——— LEAKAGE OIL PURGE DIP TUBE
' |

|| e~ — — GAS SUPPLY TO PUMP

PRESSURE GAGE

 

 

 

 

 

 

 

 

 

 

 

 

i . GAS FLOW — — — =
il . LIQUID FLOW ————e=

 

 

 

 

 

LUBE OIL—__ [ || | |

 

 

 
 
  
        
   

 

 

 

SEAL—— I

RADIATION BARRIER
——=— QFf ~ GAS LINE

 

 

 

 

 

 

 

 

 

 

LEAKAGE OIL CATCH BASIN

L
/

 

 

 

OPERATING LIQUID LEVEL

 

 

    
  
      

 

-

 

 

 

NSNS AR

 

    
  

_.‘_',,._-/ _ —
T TR

 

 

 

\

T PUMP DISCHARGE

 

 

 

PUMP SLCTION

Fig, 1.4.6. Schematic Diagram of Sodium Pumg,

36
PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED
ORNL-LR=DWG 20877

 
  
    
      
  

 

 

SURGE TaNK
Llguip LEVEL

 

 

 

E4

 

% N
R

  
     

O—

DETAILS WOT SPECIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[a} h\\.

@}.. s+~ HELIUM PURGE GAS
(B}~ o= GAS SEPARATED FRCM LIQUID

(C)—=e  LICUID FLOW THROUGH CENTRIFUGE

{(O)—=> man FLOW

Fig. 1.4.7. Comparison of (g) Sodium Pump Cenfiguration Presently Beling Tested and (b) Modificotions Being
Made 1o Mock Up the Recctor System, {Confidential with ception)

37
ANP PROJECT PROGRESS REPORT

UNHCLASSIFIED
ORNL-LAR=DWG 20878

 

   

 

210 : ‘
45300 rpm
.-."'-'-'_h
200 | e \L —
| aeatlls {4400 cpm
.- |
‘.
1o | s 5\‘ !
| |
180 — |
W
] ‘ gt 4200 rom ‘
; 1RO — l
a
w
g |
150 - -
| r\-
~
150‘ J-:fl_‘.'l_““ =4 e |
' \Boomm
| |
140 - |
|
120 L L 1 — e —
300 400 300 0o TO0 aoo
LOW RT lapmt
Fig- 1.4.8. Performance Dota for Sodium Pump Ob-

tained in Development Tests with Water.

sodium pump centrifuge operation;
will be stested in future

with normal
however, this point
eoperation,

Another seal design, which incorporotes larger
radial ¢learances between the inner surface of the
small ring ond the outer surface of the pump borrel,
will be tested with woter. The purpose of this
design modification is to reduce the assembly

problems associated with the multipie-ring seal.

Sodium Pump Endurance Tests
P. G. Smith

A sodium pump is being operated in an enduronce
test with sodium ond has occumviated o total of
1584 hr of operation, during which the pump pot
has been thermally cycled 300 times over the

38

UNCLASSIFIED
ORNL-LR- DWG 20879

      
    
  

F W o
AR SN
PR

s

—

L,

 

 

 

 

 

DESIG\I NG 2

 

 

 

 

Fig. 1.4.9.
Assembly Showing Proposed Multiple-Ring Seal.

Partial Cross Section of Sedium Pump

temperature ronge of 1250 to 1050°F. The oper-

ating conditions for this test ore given below:

1050 to 1250°F

Pump pot temperoture

Pump pot pressure 8 psi

Fump speed
Catch basin helium off-gas purge

Helium purge down pump shaft

3500 rpm
50 Vitars/ day
500 liters/day

It was found that the pump priming dificulties,

were

could be

which reported previously, !
eliminated by stopping the pump when transferring
the sodium into the pump pot. Previously, the
transfers of sodium to the pump pot were made
with the pump rotating at 500 rpm, and anywhere
from 2 to 4 hr were reguired to prime the pump.
The off-gos line plugging, also reported previ-
ously, ’? was found to be a result of too high o
liquid level in the pump pot. The pump probes
were found to be incorrectly sized.

The lube oil leakage rate varies from 5 to 20
cm3/day from the lower seal, and there is no
There is,

o slight amount of leakage of lube oil

meosurable leckage from the upper seal.
however,

 

10p, G. Smith, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 39.
somewhere in the upper seal region. This leakage
is not sufficient to be detrimental, and its source
cannot be located until the operation is stoppad.

Primary NoK Pump Development Tests

H. C. Young!! J. N. Simpson !

testing of an ART primary
NaK pump,'? designated PK-P-1, wos continued.
Performance and efficiency curves obtoined at
a NaK temperature of 1200°F are presented in
Fig. 1.4.10. As stated previously the data being
obtained are in good ogreement with dote obtained

High-temperature

 

 

    
   
 
   
  
 

 

 

 

with woter.
UMCLASSIFED
ORML—LR—DOWE 20880
s
r
|
075
4 ;
30 G.74-. i ~0 TS5
0.73- N |
0.?2 i 7 I-O.?45
g 3850 rpm 7‘77/ o4 —
= 340Q rpm
‘C'; A50 —— 300 rpm s——ce—— -
% 3200 rom I—_.,c
100 rpm —pe———pt_
a . o B
2 300 —gobt rom =
& 2900 rpm
250 2800 rpm _
(2700 rom
2600 - . |
rp I A
i m v SIGN RANGE
152

 

 

 

400 600 BOG 1000 4200 1400 1500 1800
FLOW RATE {gpm)

Fig. 1.4.10.
Frimary NoK Pump Operating ot 1200°F,

Performance and Efficiency Curves for

The results of cavitation runs at several speeds
and flow rotes at a NaK temperature of 1200°F
are presented in Fig. 1.4.11. The cavitation
data hod to be taken ot different speeds for each
of several censtont flow rates, 1000, 1220, and
1400 gpm, since it was not possible to maintain a
constant speed with the wound-rotor motor which
has o step-type resistance speed control. The
cavitation porameter, o, was plofted against speed

 

”Dn assignment from Pratt & Whitney Aircroft,

124, c. Young aond J. N. Simpson, ANP (uar. Prog.
Rep. Dec." 31, 1956, ORNL-2221, p 47.

FERIOD ENDING MARCH 31,1957

along lines of constant flow ond then ogainst flow
alorg lines of constent speed. From the latter
plot, the suction pressure
suppress cavitation over the normal pump cperating
range was plotted, as shown in Fig. 1.4.12, for
two NgK temperatures, 1200 ond 1400°F. The
vapor pressure of NoK increases by 5.33 psi from
1200 to 1400°F, ond the constant spread between o
given constant speed line for eoch of the two
temperatures is very necrly 5.33 psi. The gos
pressure on the pump surge tank is 0.5 psi less
than the static suction pressure and therefore con

be determined from Fig. 1.4.12,

In order to further show the effect of vapor
pressure on cavitation, dota were token at o given
head and flow rate for NaK temperotures from
720°F, where the vapor pressure is essentially
zere, to 1400°F, where the vepor pressure is
approximately 7.5 psia.  The results of these
tests in terms of the cavitation parameter, o, are
indicoted by Fig. 1.4.13. If covitation voried only
with wvapor pressure at a given pump operatiag
point, it would be expected that ¢ would remain
constant over the temperature range. As may be
seen in Fig. 1.4.13, there wos essentially o
constant ¢ at temperatures from 1180 to 1400°F,
and o slightly higher but constont ¢ ot 720 and
914°F. It is believed thot the results would
be more significant if speed and flow rote were
held constant instead of head ond flow rate in
order to hold constant impeller inlet conditions,
ond therefore additional cavitation data are to be
obtained.

static required to

This pump and its associated loop have now
been operated approximately 2100 hr ot temper-
atures from 1100 to 1400°F. The rotory element
of the pump was removed and inspected affer
1700 hr operction, ond it was confirmed that, as
suspected, some NaK had entered the oil caotch
basin when the NaK level rose in the pump tank
cs the result of ingassing after approximately
800 hr of operation. The pump hod, however,
continued to operate satisfactorily for the next
900 hr. A consideroble amount of carben from the
oil was deposited on the pump shoft and on the
radictor baffles in the pump tank, and thus it was
apparent that the oil seol leokage had run down
the pump shaft, even though there was a 50-liter/day
helium purge from the oil catch basin and «
500-liter/day helium purge down the pump shaft
annulus and out the pump tank reflux condenser.

39
CHANGE IM MAXIMUM HEAD {%)

 

 

UBCL ASSHIED
ORML-LAR-DWGE 20884

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

02 Q44 D46 0.48 Q.20 Q.22 024 0.26 o8 Q.20 o 32 0.34 0,36 Q.38
VALUE OF CAVITATION PARAMETER
r
RUM FLOW COMNDITION
| —| | 1220 gpm, 3280 rpm, 355-f1 HEAD
g Il ! 2 1220 gpm, 3447 rpm, 38{-fl HEAD
3 4220 gpm, 2690 rpm, 225-ft HEAD
‘ | 4 000 gpm, 28¢5 rpm, 281-ft HEAD, NG CAVITATION
5 4000 gpm, 2230 rpm, 338-f HEAD
Z & 40C0 gam, 3510 rpm, 405-0 HEAD
74400 gpm, 3150 rpm, 295« HEAD
- | . & 1400 gpm, 2580 rpm, 168-1t HEAD
I 1 | | L
044 0 46 .48 0.50 Q52 0.54 056

VALUE OF CAVITATION PARAMETER

Fig. 1.4.11. Cavltation Data for Primary NaK Pump at Yarious Flow Conditions with NaK ot 1200°F.

LHO4IY SSFHO0YL LDOIT0dd ANV
UNCL ASSIFIED
P f. Lh- NG 2088

1400°F Nak

 

 

e
§

- 1200 Nak

AL

 

 

RAINIRALIME STATIC SUCTION PRESSURE TO SURPRESS CAVITATION {psig)

300rpm pas f(

£ 7
oo I,_,,C—T ! ! i 1
29000pmy” 4 .
7 |
00 000 100 1200 1300 FMO00 1500
FLOW RATE (gpm)
Fig. 1.4.12. Minimum Static Suction Pressure Re.

quired to Suppress Cavitation of Primary MoK Pump.

I+ is believed that a cylindrical riser around the
shaft to obove the seal interface will direct the
leakage oil inte the catch basin ond prevent
leakage down the shoft into the NaK. On the
pump being tested the riser was below the seal
interface. Further examination showed that there
were no signs of rubbing of the impeller against
labyrinth seals, even though the pump wos oper-
ated cccosionally at conditions well off the volute-
balonce line.'3 The impeller ond volute appeared
to be in excellent condition.

The rotary assembly was reinstalled in the loop,
and a test run of 408 hr wos made without the
helivm bleed down the pump-shaft annulus. The
50-liter/doy oil catch basin bleed was maintained,
The rotary assembly was again inspected, and a
very smoll amount of NaK and oil mixture was

 

BH, ¢, Young and M. E. Lockey, ANP Quer. Prog.
Rep, March 10, 1956, ORNL-2061, p 48.

PERIQD ENDING MARCH 31,1957

UNCLASSIFIED
ORNL=~LR-DWEG 20883

 

 

 

 

 

O
|
|
'2 &
- I
2 4
2 |
L | |
= |
=
E & | =
3
-
=
= |
wog | | Ha¥X TEMPERATURE o L
4 RN AND FLOW COMDITIONS
7 {3 4400°F, 1200 gpm, 3417 rpm, 379-ft HEAD
k2 12 914°F, 1200 gem, 3405 rpm, 364-ft HEAD
b L 1 T20°F, 1200 gom, 3380 rpm, 347-ft HEAD .
10 1285%F, 1225 gpm, 3370 rom, 368-~1t HEAD
| 2 M8O°F, 1205 gpm, 3365 rpm, 37011 HEAD
|
ian S e _
o4 G1s 08 020 0.22 0.24 0.25
YALUE OF CAVITATION PARAMETER
Fig. 1.4.13. Cavitation Data for Primary HaK Pump

ot Yarious NaK Temparotures.

found in the oil catch bosin, but its presence had
not affected the oil seal leakage purge system.
The impeller surfaces had an etched appearonce,
but approval for further operation was given by the
metallurgists.  The critical dimensions of the
rotary assembly were measured before and after
this test run, ond it was found that all dimensions
had remained constant, except the volute mouth,
which had opened up from 0.003 to 0.010 in.,
probably because of stress relief and pressure
stresses, An additional test is plonned to provide
further data on operation without a heliuvm bleed
down the pump-shaft annulus.

Heat removal from the pump bearrel by the lubri-
cation and cocling oil was measured at several
NaK temperatures ond liquid levels in the pump
tank. At lubricating and coeling oil flow rates of
approximately 3.5 gpm, the heat removcl ranged
from 1.8 kw at 1200°F, the normal operating level,
to 2.3 kw at 1400°F, the maximum operating level.

A resistance type of level indicotor was vsed in
the pump tank, and the level recdings were com-
pared with the indicoted by spark-plug
probes. The resistance tfype of level indicator

gave very safisfeciory performance ot temper-
otures ahove §50°F.

level

41
ANP PROJECT PROGRESS REPORT

Auxiliary NoK Pump Development Water Tests
H. C. Young

Initial water tests were conducted on an auxiliary

J. N, Simpson

NoK pump with an Inconel impeller in o test loop
designated PK-A-1. A single-poss shell-and-tybe
heat exchonger was installed in the 4-in. Inconel
A 60-hp

d-c drive motor was used fo permit constant-speed

pipe loop to remove pump power heot.

measurements,

The performence dota obtained in these initicl
tests of the auxiliary NaK pump with an Inconel
impeller (No. 1) are compared in Fig. 1.4.14 with
dato obtained previously with o brass impeller
ond volute.'* At moderote flow rates, the dota
correlate very weli, but at high flow raotes the head
drops more ropidly for the lnconel impeller thon
for the brass impeller. It was suspected that
cavitation cauvsed the drop, and the tests were
repeated with higher pump tank gas pressures than
those which appeared to be normally odequate,

 

My, ¢, Young, J. G. Teague, and M. E. Lackey,
ANP Quarp Pfflg- Repo J'Mfl‘fi' IO; Igjfip ORNL'Z‘IOé. P 54'

UNCLASSIFIED
ORML—LR—~DWG 20884

 
    
    
   
  

480 — = |
|
400 |—— 3550 pm =a |
i - |
930 ! 3450 rpm
3250 rpm '
| DESIGN RANGE
300 — -- S — . ~——

305;0 rpm=1""
2950 rpm __.--"i' LA

 

 

 

E 260 - 2850 l’?l"fi__
a SPECIAL DESIGN POINT-.
T | ‘ 2650 rpm——:.‘
$ 200 | 2550 rpmeg = .
5
T
| WATER TEST WITH INCONEL
150 IMPELLER NO.1 |
——— WATER TEST WiTH BRASS |
IMPELLER
0o
180G rpm
50 |- |
0" Y N i 2
G 1Qa 200 300 400 LS00 600 [isle
FLOW RATE [gpm)
Fig. 1.4,14., Parformance Data for Auxiliary NakK

Pump Operating with Warer.

42

that is, 25 rather than 10 psig.
remained the
pressures.

The pump heac

same, however, at the highes

The discharge opening of Inconel impeller No, 1
was approximately 10% smaller than that of the
bross impeller, ond Inconel
impeller (No. 3), with dimensions very close fo
those of the brass impeller, was tested. The
results were the some as those obtained with
Inconel impeller No. 1. All performence dota were
token numerous times, and it is possible that the
lack of correlation was caused by a difference in
dimensions of the Inconel and the brass volutes or
by a slight change in impeller inlet geometry.
The wvariance is considered to be insignificant
over the normal opercting range of this pump.

More cavitation data were obieined for this
pump operating with water thon for any pump
tested thus far. The Inconel impeller cavitetion
performance,

therefore another

Fig. 1.4.15, compared fairly well
At 300 gpm the
cavitation parameters were practically identical
for the brass and Inconel impellers. At higher flow

with that of the brass impeller.

rates ond speeds, the parameter was somewhat
greater for the Inconel impeller and probably
reflected fobrication differences between the brass

UMCLASSIFIED
ORML-LR-DWE 20885

    

 

 

0 .
1 ferre—
2 — —
3 e - —
- {
F o4 L ]t
o I
w5 F
L I
=
5 6 }
z |
i
27T :
=
o B : e || ]
tr,J F INGONEL IMFELLER |
5 | 8 ——— BRASS IMPELLER
r 9 T | '
c|
10 - — - —
CURVE FLOW CONOITIONS AND WATER TEMPERATURE
" A 2960 rpm, 300 gpm, 280.0-1 HEAD, 190°F
12 8 3353 ipm, 430 gpm, 340.8-1t HEAD, 140°F
' ¢ 3373 /pm, 430 gpm, 347,311 HEAD, 140°F
13 1 - == =;
006 008 QG o2 014 0,16 C18

VALLE QF CAVITATION PARAMETER

Fig, 1.4.15. Cavitation Data for Auxiliary MaK Pump
with an Inconel lmpeller Operating with Warer Compared
with Dato Obtained with o Brass Impeller.
and Inconel vones. On the basis of these tests
the covitotion perfermance of the Inconel impeller
is considered to be entirely satisfactory for oper-
otion with NaK over the ART design range. Some
additicnal date will be tcken, however, to assure
adequcte information for comporison with dato to

be taken with NaoK,

Bosed on earlier experience with the primary
NoK pumps, 1*a % _-in.-dia hole was drilled through
the top of the discharge pipe inside the pump tank
to permit gas trapped in the discharge line during
filling to be vented. This vent permitted satis-
foctory pump priming. In order to further determine
the effectiveness of this vent, a section of piping
was instalied in the discharge Jeg of the test loop
to deliberately trop any quontity of gos up to o
valve of 0.45 t3. A series of pump priming tests
was conducted wherein the loop was filled ot
0 psig with 0.45 #13 of gas tropped in the loop.
The pump tank gas pressure was then raised to
15 psig before ottempting to stur‘tlfhe pump, A
foter series of tests was alse conducted wherein
the gas pressure was raised to 15 psig before
the loop was filled so thot twice the previous
amount of gas was trapped. In ofl coses, it was
possible to start and fully prime the pump by
following the starting and stopping technique
described below. The pump is brought to an
operating speed of 1800 rpm (this speed represents
the first control step speed on the wound-rotor
drive motor) ond operated ot thot speed until the
discharge pressure drops slightly, which indicates
that gas has entered the pump, and then the pump
is stopped. This procedure is repeated until the
dischorge pressure no longer drops. The average
number of startup attempts required ranged from a
minimum of seven for the first series of tests to
17 for the later series of tests with twice os much
trrapped gos. |t appears that with this technique
the pump can be started with troepped gas in the
system, ond it will not be necessary to vecuoum-fill
the loop.

 

15y, C. Young and J. G. Tague, ANP Quar. Frog.
Rep., Sept. 10, 1956, ORNL-2157, p 43.

PERIOD ENDING MARCH 31,1957

REACTOR COMPONENT DEVELOPMENT TESTS

D. B, Trauger
Heat Exchanger and Rodiater Development Tfists
J. C. Amos 2
L. H. Devlin'é R. L. Sean
D. R. Ward

A summary of fuel-to-NaK heat exchanger and
NoK-to-oir radiotor test operations during this
quarter is presented in Toble 1.4.2. Process
Engineering heot exchanger No. 2 (type SHE-7)
and Black, Sivalls & Bryson radiator No. 1 com-
pleted 1456 hr of nonisothermal operation, including
552 hr at o heat exchanger fuel inlet temperature of
1600°F, and 29 thermal cycles, in SHE test stand
B."7 The test was finally interrupted by o NaK
leak in the circulating cold-trap economizer ofter
o total of 1845 hr of operation. At the time of the
interruption, the test was opercting on the tenth
step of a 15step progrom '8 for obtaining data on
mass transfer in the NoK circuit. The cenditions
imposed on the heat exchonger hod been suf-
ficiently severe, however, that it wos decided to
terminate the test rather than tc complete the
mass-transfer study progrom. Dota covering the
range of interest hod been obtained and further
operation would only have provided more inter-
medicte and check points. Anolyses of the data
obtained during this test ore being made. The heat
exchanger is being examined metallurgically, and
the economizer, which had operated in this stand
tor o total of 6490 hr, is also being examined to
determine the cause of failure.

The test stand (SHE-B) is being modified to
accommodate a 20-tube, heat ex-
chonger, designated *'small heat exchanger type

SHE-9'" (Fig. 1.4.16). The dimensions of this

semicircular

 

]'601'1 assignment from Prott & Whitney Aircraft.

7 The design of haat exchanger type SHE-7 is illus-
trated in Fig. 1.4.5, p 50, of ANP Quar. Prog. Rep.
Sept. 10, 1958, ORNL.2157; the design of radiotor
Ne. 1 is the same os thot of rediator Mo, 7 shown in
Fig: 1.4.8, p 60, of ANP Quar. Prog. Rep. June 10,
1956, ORNL-2104.

18), ¢. Amos et al,, ANP Quar. Prog. Rep. Dec. 31,
1956, ORNL=2221, p 43

43
144

COMB SPACERS

MoK QUT

FUEL IN

   
  
 

 

 

 

 

 

MATERIAL . INCONEL

57 i

   
    

ORNL-LN-UWG 20886
b
-20 TUBES (Q.2315-n OD, 3025-n WALLS)
- . '_l
i Ju-.__g 3 /
3()5:5"54|n. e e -
*, I".___.-___l|r_ * Ziafa—in R
ot | '
SECTION A4 //
d HEAT EXCHANGER SHELL=" '|
/ | — Nok HEADER

| |
ifi-l Rt LW
HEAT EXCHANGER 0 | k |;._,,-E ¥
TUBES ! L

 

), NeK IN
& | L=
-*"- i J— ' l

FUEL HEADER-

l

FUEL CUT

Fig. 1.4.16. Semicirculor Heat Exchanger Designoted “'Small Heat Exchanger Type SHE-2.'"

103 SS3HH0Hd LITFFCUd dNY
PERIOD ENDING MARCH 31,1957

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 1.4.2. Summory of Heat Exchonger and Rodlator Operntlons
Hours of Total Mumber
Tast Unit Tast Naonisothermal Hours of of Thermal Status of Test
shand Operation Operation Cycles
Process Engineering Corp.  SHE-B 1456 1845 29 Terminated because of foilure
heat exchanger Na. 2 of circulating cold-trap
{type SHE-7) economizer
Black, Sivalls & Bryson IHE-B 547 911 5 Test completed
heat exchonger Mo, 1
{type \HE-B)
Black, Sivalls & Bryson tHE-B 547 211 5 Test continuing
heat exchanger MNe, 2
(type IHE-8)
York Corps radiator Mo, 11 iHE-B 547 211 5 Test continuing
York Cerp. radiator No. 12 IHE-B 547 o11 5 Test continuing
Black, Sivalls & Bryson SHE-RB 1456 1845 29 Terminoted becouse of failure
radiator No, | of circulating cold-trap
econemizer {operated in con-
junction with Process Engi-
neering Corp. heat exchanger
Na, 2, abeve)
heat exchanger, which is being fabricated of sl
ORNL-LR-OWG 20887
Inconel, are given below, ond the proposed oper- 325k
ating conditions are presented in Fig. 1.4.17, 1500 °F MoK~ TG0 IR 1070 °F
SZpva RADIATOR 15 pig
Tube material Inconel
Number of tubes 20 BFy 5T 32510 | Vapsig
Tube cutside diameter, in. 0.2315 pr-\:::p o
Tube inside diameter, in. 0.1815 Re o = 109,000 :
Tube woll thickness, in 0.025 e
NMeK flow length, ft 6.138 D Py = 150310
NaoK free flow area, f12 0.00358 BT ek = 430 °F
NaK side heat transfer area, f? 5.74 L1 W o TN 7 . e  1070°F
Spacer dimensions, in. 0.018 x 0.040 2 erg FUEL-TO-NoK 3a.2 0810
Spuocer arrangement 45-deg stagger 16O0°F HEfliE_KETEG_EE___ 1250 °F
0-deg incline 36 95”9_-_ t3psin
Number of spocers 14 arF”El‘2350fF
8Ppyp * 21 psig
Fuel flow length, ft 5.05
Fuel free flow areaq, £2 0.Q036 ¥ fopsia
Fus!l side heat transfer oreq, fr? 7.32 Regye = 2790 FUEL | 7840m
Fuel equivolent diameter, ft 0.0090 Bl
- €l y
This heat exchanger was designed to simulate i o
the tube geometries of the ART main heat ex- e Rfilgfig{x st
‘ i 3B osig 40 psag
changers, The heat exchanger will be subjected i
to a thermal-cycling program in order to verify
the stress onalysis of the ART meoin heot ex- Flg, 1.4.17. Operoting Conditions for Semicirculor

changers.

Smecll Heat Exchanger Type SHE-9,

45
ANP PROJECT PROGRESS REPORT

York Corp. radiator No. 15 was installed in SHE
test stand C for operation in conjunction with
Process Engineering Corp. heot exchanger No. 3,
type SHE-7. When test operations were storted,
proper NaK flow could not be established. The
loop was drained, ond examination revecled that
the cold-trap economizer inlet and all but four of
the radiator tubes were plugged with oxide. The
oxide had apparenily been deposited in the main
NaK piping at the end of the preceding test,'8
which, as described above, was terminated because
of a failure of the cold-trap economizer. When
operations were resumed, the oxide was trensferred
to the new rodiator and o the economizer inltet
York radiator No. 15 has been replaced with
York radiater No. 16, and all NoK piping con-
taining oxide hes either been replaced or steom

line.

cleaned. Operation wiil be resumed to obtain
information on the heat transfer and corrosion
characteristics of o typical system when operating
with NaF-ZrF ~UF, (56-39-5 mole %, fuel 70).
York radiotor ldc:-. 15 has been thoroughly cleaned
and will be installed in SHE test stand B as the
heat dump for the test of the semicircular heat
exchanger,

The first part '8 of the endurance test of Black,
Sivalls & Brysen heat exchangers Nos. 1 and 2
{type |HE-8) was completed in IHE test stand
B. The test included 503 hr of operation of heat
exchonger No. 1 with a NaK inlet temperature of
1700°F and the other conditions of temperoture,
pressure, and flow rate shown in Fig. 1.4,18.
The NaK pressure drop through heat exchanger
No. 1, calculated on the basis of constant flow,
increosed to approximately 5.4 times the initial
level, and the NaK pressure drop through the
radiator increased to approximately 1.3 times
the initial level. These pressure drop increases
in the NaK circuit ore attributed, as before, ¥
to the byuildup of mass-tronsferred materigl in
the small-diometer tubing of the test units. Black,
Sivalls & Bryson heat exchanger No, 1 is being
replaced by Black, Sivalls & Bryson heat ex-
changer No. 3, which is of the some type os No, 1.
Endurance testing of heot exchonger No. 2 will
continue during the second phase of the test.
However, detailed progroming for the second phase

 

‘IQJ‘ W. Cooke ond H. C. Hapkins, ANP Quar. Prog.
Rep. Sept. 10, 1956, ORNL-2157, p 46.

46

ORNL-LR-0WG 20820

toK-TO=AIR
RADIATORS
YORK NOS. 11 AN 12

1265 °F

 
     
 

 

5.2 pagliNTiaL]
12,831l FIMAL)

 

 

 

e o
= ad
z@ ¥ FuE 474 psig LIMITIAL) w|E 2
S =|R VEL 46,1 psigIFINAL) g
2 T8 PuMe e < e s
a - - in
o é_' 1200%F sa
T o b
T T

 

 

 

vo 26.4 perg{(INTIAL)
250 palg [FINALY

 

 

 

FUEL FtOW

mWaK FLoW
8,500 B e HINITIALY
20800 oo (FINALY
FUE:=TO=-Mok HEAT E¥CHANGER
BLACK, SIVALLS AND BRVSON NO 2
(TYPE {HE-E}
33,000 Bsnr (INITIALY
14 OO0 Wrne ITINALY
Mok - TO-FUEL HEAT EXCHANGER
BLACK, SIVALLS AND BRYSON NC Y
[TYPE {HE-3)

 

 

 

 

 

 

 

1600 °F

 

369 pmg UNITIALY
358 psiglFINAL)

1500 °F N HEATER 1700 F

[GAS-FIRED
42 5pug(INITIALY FURNACE ] 38 palg[ATEAL)
44 5 psig{FINAL] AGBSg{FINALY

Fig. 1.4.18,
Endurance Test of Black, Sivalls & Brysan Heat Ex-
chongets Nes. 1 and 2 (Type IHE-B).

 

 

 

 

Operoting Conditions During 503-hr

of the test will not be established until the resulits
af the metallogrophic examination of heot ex-
chonger No. 1 are known.

ART-prototype test rodiater No. 1 was received
from York Corp. during the quarter. It is currently
being modified by slifting the fins as shown in
Fig. 1.4.19. These meodifications were recom-
mended as the result of recent detailed stress
analyses of this radiater design., Tests of this
unit will be started soon.

Water Flow Tests of Aluminum North-Head Mockup

R. Curry 20 J. W. Cooke??

Additional tests were run of the ART twin fuel
pump system as port of the water flow tests of the
aluminum north-head mockup. For these tests the
pumps incorporated the improved model 32 impellers

 

209, aossignment from Prott & Whithey Aircraft,
PERIOD ENDING MARCH 31, 1957

_’; UMCLASIFIED
| PHOTO 17484
T— o

-

-

HRE -."

I nl.hi.-'-{:----"'"." -
o

|
f bl

RN ”'rs-.-‘-“""

oy

 

Fige 1.4.19. ART-~Prototype Test Rodiator Ma. 1 Showing Lines Where Fin Matrix Has Been 5lit to Reduce

Stress Concentrations. [Confidential with caption)
described previously.?!  With the improved im-
pellers the liquid spraying and sloshing noted
previously in the expansion tank and the resuliont
weting of the tank ceiling were greatly reduced
because of the reduction in the gross inflow to the
tank. Ingassing of the main circuits with one pump
stopped and the other still ruaning was not
appreciably different with the improved impeller,
and flow-rote fluctuation and pump suction pres-
sures remoined essentially unchanged. With the
bypass loop connecting the pump suction pipes
wide open, stopping one pump with the other stil]
running cauysed the total flow through the external
main loops to be reduced by a factor of . How-
ever, the flow continued in the normal direction in

 

2y, E. lockey, G. Samvels, and J. 1. W. Simon,
ANP Quar. Prog. Rep. Dec. 31, 1936, ORNL-2221, p 33.

each external main loop with no reversal, and the
Hlow rote in the external main loop of the stopped
pump was 7'.:"]0 that in the external main loop of
the operating pump. One significance of this
result with regard to reactor operation is that any
tendency for local overheating in an inrermediate
heat exchanger tube bundle wouyld be minimized
following stoppage of one-half the fuel, sodium,
and NakK pumps.

The model 32 impeller has greatly reduced the
problem of liquid rising in the shoft eanulus.
With the criginal impelier, liquid rose inte the
shoft annulus of a pump which was stopped if the
other pump was still running. In tests of the new
impeller over o wide range of pump speeds and
gos flow rates down the shoft annulus, the only
instance of liquid rising up the shaft annulus
upon stoppage of one pump occurred when the

47
AMP PROJIECT PROGRESI REPORT

gas inflow was cut off completely; and even then
more thon 30 sec elopsed following pump stoppage
belore water wos detected 0.4 in. wp the annulus,
the lowest sensing point. In operation of the
reactor, the reactor dump cycle would be under way
within this period if o pump stopped.

During the quoster the entire moin loop circuit
was sant fo the Hydroulics Laboratory of the
University of Tennessee for calibration of the
external moin loop orifices. This calibration
showed arifice coefficients 2 to 3% less than the
accepted values, and also showed no variation of
the cosfficients with external main loop valve
settings in the range of interest, A special test
was colso performed in which water was pumped
inte the common purp discharge pipe at o rate of
1300 gpm, and ony Huctuctions of mein loog erilice
differentiol pressures were observed.  Sewvercl
repetitions of this test yielded ditferent reswlts,

In abaut hall the runs no  fluctuctions were
ohserved. In the other runs fluctustions of on
orifice diffarentiol pressure, similar to those

observed with the complete oluminum test rig,
ware observed. In these University of Tennessee

tosts, woter was supphied by two dissimilor pumps

vperating in parallel, und one of the pumps weos
known te be subject to some instebility in the
operating region of interest,

The purap wvolutes in the twin pumps ware
carefully exomined while thaloop wos disassemblad
for the colibrotions described ahove. Differences
in the design of the volutes in comparison with
the voiute design specified for the reacior were
noted, ond the cluminum volutes were modified to
more closely resemble the reactor pump wvolutes.
Both the orifice calibration and the rewarking of
the wveolutes were undertcken in the hope of
eliminating an apparent loss of 10% of pump sutpu?
How rate at design speed and head compured with
the output ocbserved for the same impellers in the
single pump high-temperature test,

Dump Yalve Development Tests

l. T. Dudley M. H, Cooper??

Prototype Vaolves: ~ A summary of the tests
conducted this quarter on ART prototype dump
valves is presented in Table 1.4.3. One prototype

 

MOH aszigoment from Proty & Whithay Alrcrals,

Table 3.4,3, Opercting Conditions and Results of Prototype ART Dump Valve Tasts
veith N::F-"Z:'FJUF“ {50-48-4 mole %, Fuel 30)

Volve temporature: 1 iVt s

 

 

Fressors difforential across seot: 50 psi
Yolve 40C Volve 5B Valve 5C

Seat moteriol K18 1A Copper Copper
Flug material K-1ETA Malvbdenum Molybdenum
Operoting time, hr 1500 4 30 1100
Mumbar of eycles 50 3 36
Stem closing facen, 1b 750 754 1000
Stem apening force [max}, Ib féd 1850 2520
Leckage rore, c:nE,f ke

During first 25 eycles {av} 1.4 i .5

During 500-hr closure {av) 1.2 0

Curing logt cycling pariod {ov) 2.4 2,5

Maximum during tasy 506 0.6 &7

 
valve was received during the latter part ol the
quarter from Block, Sivalls & Bryson, designated
valve No. 6, which incorporates a swiveled piug.
Since the swiveled plug design is now considered
to be unsatisfoctory, leakage rote tests will be
made on this volve with water only. The two
available test stands will be used to test only
valves which have rigidly mounted plugs ond
which incorporate other refinements now con-
sidered to be essential for the suvccessful oper-
ation of this volve design.

Valve 5B, which had @ rigidly mounted, spherical
molybdenum plug and a conical copper seuat,
exhibited negligible leakoge during the first 23
operoting cycles. The test of this valve wos
terminated after 430 hr because the bellows
ruptured while the valve was closed. The bellows
is composed of stamped Inconel disks welded
together at their inner and outer peripheries, and
foilure occurred at an peripneral
Preliminary metallographic inspection showed that
the failure was located at the point of fusion
between the weld bead and the disk. Some
scattered impurities were observed in this orea
which may have contributed to the failure. The
pressure differential
approximately 3 psi and therefore not high enough

inner weld.,

across the bellows was
to couse direct stress failure,

Valve 5C was assembled by using refinished
parts from valve 5B. The conical surface of the
seat was remachined and the spherical surface of
the plug was polished, The high initial leokage
rates that occurred when tests of volve 5C were
started made it necessary to use ¢ stem closing
force of 1000 Ib. The average leakage rate was
0.5 em3/hr during the first 25 operating cycles
with this stem force., No measurable leckage
occurced during the 500-hr closure test, and teakage
rates during eleven oaperating cycles since the
E00-hr closure test hove been erratic, with on
average of 2.5 e¢m3/hr. The leakage rate has
increased raopidly since the thirty-third operating
cycle. This valve hos now been operated for
1100 hr, and the test will continue until an oper-
ating time of 1500 hr has been reached.

Valve 4C, which had a rigidly mounted, spherical
Kentanium cermet {K-131A) plug and @ conical
Kentanium cermet {K-151A) seat, was tested for
1500 hr ot 1300°F. The average leckage rate
during the first 25 operoting cycles wos 1.4
em3/he, 1.2 em=/hr during the 500-hr closure
test, and 2.4 cm3/hr during the last 25 operoting
cycles. Relatively low stem forces, 500 to 864 Ib,

PERIOD ENDING MARCH 37, 1957

were required for opening this valve. This valve
is to be tested again with the diametrical ¢clearance
of the inner stem guide increcsed from 0.008 to
0.080 in. This valve has given the best over-all
performance of any valve tested to date.

A valve of o modified design is being fabricated
for testing. The new design incorporates o backup
bellows to supplement the seal bellows and im-
proved stem-alignment control. Also, a valve that
will have o spider guide mounted in the end con-
directly obove the seor ring is being
assembled to determine whether volve alignment

nection

can be improved by guiding the plug at a point as
close to the seot as prociical. Kentonium 162B
inserts are being used for the bearing surfaces of
the guide.

"“InsLine” Valve Operaters — A simple rig has
been set up for tests of the Philadelphia Gear
Works “‘in-line’’ actuater that is to be used on the
ART NaK valves. The test assembly consists of
a stmuloted valve on which the actuator is mounted
and on opposing actuator that is vsed to determine
the “'in-line’" stem thrust. Provisions are made for
heating of the simuloted valve body to determine
the effect of temperature on the actuator operating
characteristics,

Seat Materials, ~ Twelve seot and plug materials
combinations have been tested?® thus for in
NaF-ZrF -UF , (50-46-4 mole %, fuel 30), and three
other combinations are being tested. The tests
compteted during the quarter and those now under
way are summarized in Table 1.4.4. One of two
test stands was modified for use with NaK, ond
tests with NaK are under woy. The seat and poppet
cembinations tested are described in Table 1.4.4,
The 1500-hr test of each combination consists of
25 opening and reseoting cycles, a 500-hr closure
pericd, ond, firolly, 25 opening ond reseating
cycles,

Two cermet combinations, K-151A vs K-152B
and KM vs 162B, have given the mest satisfactory

results. The compositions of these cermets are

{in wt %):
M 20 WTiC2—10 Co
K-151A 70 TiC-10 NbTuTiC3-..2(1 Mi
k-152B 64 TiC=5 NbTaTiCa—EU M3
K-162B 64 TVC-6 NbTuTiC3—25 Ni—5 Mo

 

23p0; results of previous tests see Toble 1.d4.6, p 50,
ANP Quar. Prog. Rep. Dec. 31, 1956, ORNL-2221.

49
ANP PROJECT PROGRESY REPORT

Takle 1.4.4. Conditlons ond Results of Tests of ¥alve Seat ond Plug Materiols

Al motsrials tasted in NoFquFéflUFd [50-A6=4 maole %, fuel 30) sxcept us Indicoted

 

 

 

. ol Operating Volve Mo E
3eat Ve Time Tamperature Bmwer @ {omments
Moterial Meoterial (hrl {z; Fy Cycles
e W 1875 1306--1500 7 Terminoted for exomination; leokage high
elter cyching
K-54 K94 L5460 1300 hd Tarminoted for pxomination; leokoge high
KMallory Maliney tan 1300 21 Terminated for sxamination; teakage high
TR0 1006 after cycling
Cu Ma 1538 1300 41 Cperatisg
K-1628 KM 1001 1300 31 Operating; no leokage at oay time
K-15%8 K~162B 172 1300 2 Tarminatad becouse of misclignment
fricangl Stellite & 694 0040 11 Oprrating; lzokage high; test Hluid s NoK
Ka1528 K-1514 41.5 1300-- 1500 3 Operating
K-162E KM Being installed
The leokage of the K-1514 vs K-152B combi-  Februory. The fest piece is o quarter-scole mode!

notion, tested previous!y,” wos less thoan 1
am3/hr during the test; the KM vs K-162B combi-
nation, which is still in cperation, has shown neo
The wcpening force after the 500-hr
closure of 1300°F wos 700 lbh for the K-151A
vs K-152B combination ond 900 b for the KM
vs K-1628 combinstion. A phetogroph of the
KI5YA vs K-1528 combinotion ofter 1549 hr of
operation  in NoF-ZeF -UF  (50-46-4 mole %,
fuel 30) of 1300°F is presented in Fig. 1.4.20,
Subsequent tests will be performed o determine

leakoge.

the choracteristics of these materials ot higher
temperatures.

A Stellite plug ond lnconel seat combination is
being tested with Mol ot 1000°F. Mo self-welding
of the materials hos been experienced, but the
leakage haos for example, 40
cm3/he ot o S50-psi diffsrenticl pressvre ocross
the seat. Two KM vs K.15628B combinations ave
being instolled for testing in Mal(,

been excessive,

Ouser Core Shell Thermat Stability Tes?
4. . Amos L. H. Devlin

The secend of the two proposed thermal-siability

tests of the outer core shell was started in

50

of the bottom half of the ART outer core shell,
The test program censisis of three hundred cycles,
with each cycle of mognitude comparable to those
generated by an ART power cycle between zero
and full power.

The test thermal cyole comprises 1 hr under o
temperature gradient and 1 hr of isothermal oper-
ation at 1200°F.  The fransient time 15 [0 min,
Internct thermocouples iadicate the wall temper-
atyre at both ends of the shell. During the temper-
oture gradient phase of a cycle the ftemparatures

Qre
lrner sodivm antronce 15707 F
lrwner sodivm axit 1250°F
{Ivfer sodivm exit 1075°F
Guter sadium antrance 1010°F

The ogperoting conditions are thus approximately
the same as those for the first test, 24 in which
The
twe sodium streoms, which flow countercureently

the chell suzcessfully endurad 300 cycles.

 

Mg, R, Dvtko et al., ANP Quar. Frog. Rep. Dec 31,
956, GRMNL~2221, p 55
LCL Agsal LD
T

(5]
brat
M
<
0
I

 

Fig. 1.4.20. Kentanium Cermet X-151A Plug and
K-152B S5ear After 1549 hr in NuF-Ier-UF‘l (50-46-4
Mole %, Fuel 30) ot 1300°F.

over the shell, serve as the heat transfer mediums.
Bursting pressures of 26 and 4 psi exist across
the shell at the bottom and top, respectively.

Prior to the final assembly of the test housing,
the dimensions of the inside surfoce of the shell
were carefully measured. The measyring procedure
wos the same as that used on the first shell,?®
except that a bead of weld material (to serve as a
reference point) wes placed on the large end of the
shell rather than on the housing, as in the first
test, Relecation of the reference point resulted in
more gcfurate measurements.

Sodium Circuit Water Flow Tests

S. Kress 2 R. D. Peak 26

Many regions in the ART sodium circyit which
are difficult to onalyze are being individually

 

25, Curey and A, M. Smith, ANP Quar. Prog. Rep.
June 18, 1956, ORMNL-2106, p 65.

280n as signment from Pratt & Whitney Alrcraft.

PERIOD ENDING MARCH 31,1957

tested in mockups?? in order to determine the
pressure drops.  Three of these regions were
tested this quarter. Since the heot loads, lengths,
ond entrance conditions of the five rows of cooling
holes through the beryllium reflector are different,
the sodium flow through these cooling holes must
be controlled to achieve a uniform temperature
pattern throughout the beryllium. A conveniens
method of contrel is to erifice individual holes ot
the beryllium equator joint with ¥-in.-thick orifice
plugs. Such plugs with orifice holes of 0.0595
to 0.1875 in. in inside diometer were tested,

The results of the tests are shown on Fig. 1.4.2]
as pressure loss coefficients, which include the
total loss for the entronce orifice and 4.56 in. of
hole length, vs the orifice diameter. A range of
orifice sizes was tested, with the smallest orifice
in the reflector cooling hele reducing the flow to
obout 20% of the flow with the lorgest orifice.
As the result of an error in machining the first
sovthern hemisphere of the beryllium reflector,
holes 29 to 71 in the secend row from the center

s, Kress and R. D. Paok, ANP Quar, Prog. Rep.
Dec. 31, 1956, ORNL-2221, p 58,

 

UNCLASSIFIED

ORNL—LR—DWG 20889

200 T 'y T T T
| ' ‘ | I \

PRESIURE TAP

e )
11 WATER IN
4~ OORS

 

 
  
   
   
  
   

 
  
 
  
 
 
   

| ]

o
b EQUATOR |

 

0.38-in-00, ; JOINT
Sg-in -THICK - |
] ORIFICE PLUG
: QIBT5m 1D

i— PRESSURE TAP

WATER GUT

| SET SCREW FOR
=™ PRESSURE CONTROL

DiaGRam OF TEST ASSEMELY
[ ; .

20 —

PRESSURE LOSS COEFFECIENT (dimensionless)

 

 

 

) L
r e WITH O 24Bin == —
= CMISALIGNMENT =1 Ik

s | 1
= Equp,r()n -DIGMETER OF 11— —
JOINT COOLING HOLES,
_ avigNEp—" AT GBS In —{ | )
| | | | 1
B el 1YL I i b
o005 o4 0.2 0.5 f 2
QRIFICE DIGMETER (in)
Fig. 1.4.21. Pressure Loss Coefficients vs Orifice

Digmeter ot Retlector Equater Joint.

51
ANP PROJECT PROGRESS REPORT

were drilled 0.218 in. circumferentially off align-
ment. This misalignment will introduce an extra
heod loss ot the equotor joint, but it can be
compensated by proper selection of the orifice
size. A second series of orifice tests was run to
determine the effect of this misalignment on the
less coetiicient. The results of these tests are
also shown on Fig, 1.4.21.

Another test involved a full-scole mockup of
one-half of the symmetrical sodium circuit in the
reflector entrance regicon where the sodium supply
to the retlector flows down frem the two pump
tees, impinges on the c¢ircular boron cons, and
spreads out over the top of the reflector and the
core shell. The wooden test model of the top of
the beryllium reflector with the circular boron-
containing can in place and the spacer bars
assembled is shown in Fig. 1.4.22, and Fig. 1.4.23

shows the model with the boron-containing conicol

T . gk

« !
. = COOLING HOLES

J\‘L'
CIRCULAR BORON— CONTAINING CAN =

\"‘

®
é

-

ring and the pressure ring in ploce. This model
was fitted inte aon aluminum-inconel-Piexiglas
piece that simulated the underside of the Inconel
north-head shells just obove the reflector,

The head loss in this reflector entrance region
at the design flow rate of 1.28 fi3/sec was found
to be 3.0 ft compared with an estimated 7.5 i1
The relotive tlow distribution through the beryllium
cooling holes was obicined by measuvring the
height of the water jet from each hole with the
complete test assembly turpned upside down, and
loss coefficients (Fig. 1.4.24) were calculoted for
each hole based on the jet height ond the hole
orientation shown in Figs. 1.4,22 ond 1.4.23. The
measurements showed that the cooling holes on
the shoulder around the pis-shaped depression in
the reflector carry much less flow (2 to 3 times)
than the rest of the holes.

URCLASSIFIED
PHOTO 27347

: -
NS
SPACER BaRS I &

—F’_,_-,'fll""""
—

 

Figi ].4.22-
and One-Half of the Cooling Heles,

52

 

Wooden Model of Top of Reflector Showing the Boron-Centaining Ciccular Can, the Spacer Bars,
 

 

F igl 1 14-231
Assembly.

A water test was also conducted on a proposed
bypass slot in the north-head sedium circuit. This
slot is to be cut between the sodium-to-NaK heat
exchanger shell and the external pressure shell of
the reactor to allow some sodium flow to byposs
the present slots between the pressure shell ond
the upper-deck outer shell and ta give o more
direct entrance for flow into the heat exchanger
inlet. The slot tested, shown in Fig. 1.4.25,
was approximately 5 in. wide, 0.8 in. long, and
0.1875 in. high. The Plexiglas top represented
the pressure shell. The lcss equation for this
slot, os determined experimentally, is

V'I.'i)?
AH = 1.4 ——,
2g

PERIQD ENDING MARCH 31, 71957

UNCLASSIFIED
PHOTOQ 275%%

Wooden Model of Top of Reflector Showing Boren<Containing Conical Ring end Pressure Ring

where Aff is the heod foss, V is the velocity of the
flow through the slot, and g is gravity.

AUXILIARY COMPONENT DEVELOPMENT

TESTS
J. J. Keves
High-Frequency Thermal-Cycling Water Tests
J. J, Keyes J. E. Mott28

The time coastont of a 36-goge Chromel-alumel
thermocouple wos determined to be less than 3.5
msec, and thus such ¢ thermocouple moy be used
to meosure, quite accurately, sinusoidel fluid
temperature fluctuations of up to aboui 30 cps.

 

285, assignment from Pratt & Whitney Aircraft.

53
Ps

LOSS COEFFICIENTS FOR ROWS & AND D
WERE NOT MEASURED

Fig. 1.4.24.

 

Loss Ceefficients for Cooling Holes in Reflector.

NCL ASSIF S &
DRRL - LR-{WGE 20890

 

L¥0d3d S53800dd LD237MCHd JdNVY
CONICAL BORON - CONTAINING RING

 

PERIQD ENDING MARCH 31, 1957

UMCL ASSIFIED
PHOTO #7519

Fig. 1.4.23. Wooden Model of Top of Reflector Showing Beron-Containing Conicol Ring and Pressure Ring

Assembly.

A water test wos also conducted on a proposed
byposs slot in the north-head sodium circuit. This
slot is to be cut between the sodium-to-NoK heot
exchanger shell and the external pressure shell of
the reactor to allow some sodium flow to bypass
the present slots between the pressure shell and
the wupper-deck outer shell and 1o give o more
direct entrance for flow into the heat exchonger
inlet. The slot tested, shown in Fig. 1.4.25,
was approximately 5 in. wide, 0.8 in. long, ond
0.1875 in. high. The Plexiglos top represented
the pressure shell. The loss eguation for this
slot, as determined experimentally, is

1.97
AH = 1.14 ———,
2g

where AH is the head loss, V is the velocity of the
flow through the slot, and g is gravity.

AUXILIARY COMPONENT DEVELOPMENT
TESTS

J. J. Keyes
High-Frequency Thermal-Cycling Water Tests
1. J. Keyes J. E. Mott?8
The time constont of a 36-gage Chromel-alumel
thermocouple was determined to be less than 3.5
msec, and thus such a thermocouple moy be used

to measure, quite occurately, sinusoidel fluid
temperature fluctuations of vp to obout 30 cps.

 

8o, ossignment from Prott & Whitney Aircraft.

33
7S

 

 

LOSS COEFFICIENTS FOR ROWS A AND O
WERE NOT MEASURED

Fig. 1.4.24,

Loss Coefficlents for Cocling Holes in Reflactor,

UG ASSIFED
ORML- LR - DwE 208590

 

LAd0d3Y §SSIHD08d LD3IF0US INV
UNCLASSIFIED
ORNL —LR ~DW5 20834

PRESSURE TAP\
™ B~

PLEXIGLAS TOP

SLOT ———er g e
e = BlsSs il
— o o T S |
\‘:I \ / f‘\\r ~ t.'\ » _
u? = > ¥
*\::\ : II TfiR 0
A, wp

PRESSURE TAF
Fig. 1.4.25. Diagram of Nosth-Head-5lct Test Piece.

The Inconel-nickel ‘‘gun-barrel’’ thermocouple,
described previously,?? may be used to measure
sinusoidal wall-temperature omplitudes in ¢ ]f2-in.-
dio, Y-in.-woll, Inconel tube for several surface
heat transfer coefficients, The results of experi-
ments with a film heat transfer coefficient of
3140 Btu/hr-ft2.°F cre compared with the Jakob
equation®? in Fig. 1.4.26. The deviation of the
experimental results from theoretical moy be
coused by perturbations of the boundary layer as o
result of slight thermocouple protrusion (~0.002
in.) or by entrance region effects. The thermo-
couple was located about 30 diometers downstream,
which was not far enough awoy from the entrance
for uniform turbulent flow 1o be definitely es-
tabli shed.

Measurements of the oxial fluid temperature
amplitude ettenuaticn in turbulent flow through o
straight tube as a result of heat exchange with the
tube wall were made for comparison with theo-
reticel predictions for metal and plastic wolls.
In addition, the results enabled an estimate of the
wall attenuotion ratic te be made, independent
of direct wall measurements, which is in reosonable
agreement with the Jokob equation. This work is
being continued.

 

29, E. Mott and A. G. Smith, Jr., ANP Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 54.

304, Jakob, Heat Tramsfer, vol 1, p 298, Wilsy,
MNew York, 1949,

PERIOD ENDING MARCH 31,1957

UK LASSHIED
ORBL-LR—-0wS 20892

 

0.8

Q.7

 

05

 

W
a3 - 4
oz !
Q

FREQUERLCY [cps}

Flgl ]|4-26|
the Fluid Temperature Amplitude, ., os a Function of

Raotic of Woll Temperature Amplitude so

the Frequency, /, of the Thermal Cycle Imposed on the
Fluid for a Thick-Walled lnconel Tube with a Film Heat
Transfer Coefficient of 3140 Brux’hr-fiz-oF.

High-Frequency Thermal-Cycling Test Loop

A. G. Smith, Jr.3!

The high-frequency thermal-cycling test loop,
described previously,?? was completed, and
Fig. 1.4.27 shows the pulse generator, pulse pump,
ond test piece. The pulse generotor consists of
two pistons which produce helivm pressure pulses
180-deg out of phose. These pressure pulses are
transmitted to the pulsers which force olternate
hot and cold slugs of fluid through the test piece,
A helium bubbler type of level device is used to
record and contrel the fluid level in the pulse
pump. The maximum aond minimum fluid temper-
atures will be 1650 ond 1150°F, respectively,

 

3o, assignment from Prott & Whitney Aircraft,

55
 

ANP PROJECT PROGRESS REPORT

 

Fig. 1.4.27. High-Frequency Thermol-Cycling Test Loogp,

56

 
to give o maximum pulse-amplitude fluid temper-
oture of 1250°F. The frequency of these pulses
can be varied from 0.5 to 10 cps. The maximum
flow onticipated is 12 gpm, which will permit o
calculated heat transfer coeflicient. of 5500
Btu/ hrf12.°F in the test section, @ :'}é-in. sched-40
pipe.

Most of the contrel panel is shown in Fig. 1.4.28,
including the gas panels for both the high- ond
low-temperature loops and sumps, clarm and motor
control panels, and two of the four Yariac cabinets.
There are two 14-point Brown temperature recorders,
and each contains an integral high- or low-temper-
aturg alorm os a safety measure. These recoerders
will menitor the 150 thermocouples mounted on the
loop. The four Varioc cabinets contein 42 seporate
control circuits for preheating and for normal
The 250-kva tronsformer for
main {cop resistonce heot ond the auxiliory gas

secondary hecting.

panel are not shown.

The loop has operated sotisfoctorily at design
specifications with woter, and dota are being taken
on temperature .amplitudes at the test section as a
function of frequency and flow rate. This infor-
mation will form a bosis for high-temperature
operation. Some control difficulties have occurred
during startup and shutdown, but these are being
eliminated, The woter tests will be continued
vntil therough fomiliorization with the operation
is achieved.

Cold Traps ond Plugging Indicators
R. D. Peak D. A. Gardiner

Plugging indicators are required in each of the
eight NaK cold-trap circuits of the ART to monitor
the sodium oxide level. The plugging indicater
proposed for this serviced? is essentially the
same as the type which has been used successfully
on all the heat exchanger test stonds during the
past two years. Since the plugging indicators
used in the heat exchonger tests were never
calibrated, calibration tests of an ART prototype
plugging indicotor were conducted. A cold-trap
test stand®? was used for these calibration tests
with the ART plugging indicetor mounted vertically
in the stond,

 

32R. D. Peok, ANFP Quar. Prog. Rep. Sept. 10, 1950,
ORNL-21587, p 57,

33R, D. Paak, ANP Quar Prog. Rep. June 10, 1956,
ORML-2104, p 63

PERIOD ENDING MARCH 31,1957

The results of o typical test with the ART
plugging indicator are shown in Fig. 1.4.29. The
test is started by turning the Calrod hearer off ond
the cooling air on so that the NaK flowing through
the plugging plete gradually cools. At a certain
tempercture the NaK flow rate storts decreasing
abruptly because precipitated oxide has begun to
plug holes in the plugging plate. This temperoture,
called the bregk temperature becaouse of the change
in tlow rate, is on indication of the oxide content
of the NaK. After the breck temperature is de-
termined, the air is turned off and the Calrod is
turned on in order to disselve the precipitated
oxide,

All plugging indicator tests were made at o
nominal NoK flow rate of 0.5 gom. Experience on
the heat exchanger test stands has shown that the
breok temperoture is insensitive to this flow rate.
Since the flow rates in ART applicotion may vary,
dafa on pressure drop vs flow rate were obtained
and are shown in Fig. 1.4.30. Air flow rates were
also veoried, and it was found that the break temper-
atuwre was duplicated to within 5°F at o 400°F
breck temperature with air flow rates varying from
30 to B0 scfm, and to within 20°F at a 800°F breok
temperature with air flow rates varying from 8 to 40
scfm, The air pressure drop is shown in Fig.
1.4.31.

The plugging indicator was colibrated ageinst
chemical onolyses for sodium oxide obtained by
the Argonne sampler distillation methed.?4 The
chemical onolyses showed much greater inherent
variations thon did the breck temperatures. In an
ottempt to determine the causes for these vari-
ations, the Argonne sampler data were subjected
to the standord technigques of anclysis of veriance,
There appeared te be no significont differences
in instrumentation or sampler units from test to
test, but somples taken hy one technician were
significantly higher in sodium oxide then thase
obtained by the other nine technicians. The
standard deviction of the analyses was very large
and increosed as the number of the analyses
increased. Although 86 somples were taken, only
46 were analyzed successfully and used for the
calibration. The resulis of these analyses are
shown in Fig. 1.4.32 along with the corresponding

 

3, ¢, White, Procedure for the Derermination of
Ox{fen in Sodium and NaK by the Distillation Meibod,
ORNL CF-56-4-31 (April 5, 1956).

57
 

 

 

 

 

UNMCLASSIFIED '
PHOTD 28130

 

L
S HOT LOOP

 

L 8 2
-
-fl-#-fl».

a
= FROSE CIRCLITE o
 Eormlc e o

 

 

 

 

 

 

 

 

L¥0dTY $SIHO0YL LDIFO¥d ANV
The feast-
squares correlation of breck temperature based on
oxide analyses is also shown. The calibration line

plugging-indicator breck temperatures.

has good validity, os indicoted by o correlation
coefficient of (.89, even though the date appear
to show wide scatter. The calibration curve
includes most of the practical renge of plugging-
indicator application, and it should be used as a

PERIOD ENDING MARCH 31, 1957

criterion for the oxide content of NaK rather than
the solubility line.

At low break temperctures, the correlation line
is to the right of the oxide solubility line.3%
A requirement that o given amount of oxide is

 

g, . Rodgers and J. W. Mausteller, Solubility of
Sodium Monoxide in NgK, MSA Tech, Rep. 47, Tig. 2
(March 15, 1958},

UNCLASSIFIED
ORNL~LR~—DWG 20893

 

1_
|
1
—
|

 

T

 

    

TEMPERATURE ——

1

 

 

   

ki

Le
T8 |
= = |
a
g
=
ol o | ' '
z = ‘ | |
8 |Fg | I | l
g 1 L el e - 1 -
= O |
= == ' | |
= = | |
- wl 'LL".E i el — i e g S0 — S 4 — = —
|7 |
E e | | | | |
B e o 1 - — J ! = = o e T = iz H
= O | .
| ul
L | s | |
= : — s b e e e
| | i’ |
| | :
.
| CI‘: |
L e - | 1 e R | R |
o
| | | | &
wd
3= g|
o = b — 4 = — — + A S — — —
‘ ‘ | i i
c Z
- % k- W mEl _
t C | L
PLUGGING INDICATOR ‘ £ o w
TEMPERATURE |£ o o |
- ~ ==l G e 1 — o
‘ S 2 i z [
= - =
=1 T o
L =
e .ol ]
= |
| & 8 g
= = ';'é-'
1= e
b — — b D1 B ..~ S
- = | o
; '&" a | | v
m a =
Lot

 

 

 

   

. i
L | | e i | 1

TIME ——=—

 

 

Fig. 1.4.29. Resuvlts of a Typical Test with an ART Plugging Indicoter. {Confidential with caption)
ANFP PROJECT PROGRESS REPORT

UnCE ASSIFIED
QRHL-LR- m_aq Bii_

  

LHTOR
G @
o

o
w

o
o

Lol
L

<
~

 

 

NeK FLOW THROUGH ART PLLGGING INDICAT

o

 

 

|| |
- 4 = B T 89w 20 30 40 50 80
MoK PRESSURE QROP (psl)

Fig. 1.4,30. HNaK Pressure Drop Through oa ART
Plugging Indicoter. {Confidential with caption)

UHCLASSIFIED
ORKNL-LR-DWSE 20855

alk FLOW THROUGH ART PLUGGING INDICATOR
fecfm a7 POPF 14 7 psiol

 

| 2 5 10 £0 a0
4R PRESSURL DROP (ps)

Fig. 1.4.31, Air Pressure Drop Through an ART
Pluggirg Indicotor. (Confidential with coption)

necessary to plug the indicetor holes and thus
to show the flow break would move the correlation
line to the left of the solubility line, since, as the
oxide level fells, the NaK has to be cooled to @
lorger temperature intarvol below the satyration
temperature before the required amount of oxide is
deposited. The correlotion line is to the right,
however, and the above phenomenon maoy not occur
or the solubility dota moy be in error at these low
temperatures.

UNMCLASSIFIED
DANL—LA—DWE 20898

 

 

500 4 | 1

200 L - el

& ARGONME SAM?LER RESULTS

   
   
 

 

 

 

 

C)N
E
2
=
&
T
T
o
=
&&
U
i
act | | |
o
£ |
= 100 | = S £t b
— |
5 f - | — |
s | [ [ L
= | |
5 ‘ SOLUBILITY OF | /[ |
o OXIDE M Nok {REF 35)- I
£ 5] = Sl B .‘? - — s
S /
4 — . , 4!| L i
g | 2t bl |
% | 1 i 4 .i { . | | =
2 | o "‘ ¢
55 | LW o
T/ A L0 UE
7~ ) La— LEAST-SQUARES
/ CORRELATION LINE
// | I i |
/s | ‘ ‘
10 ! ' Rt by
100 200 500 0G0 1500

ART PLUGGING INDICATOR BREAK TEMPERRTURE {°F}

Flg. 1.4.32. ART Plugglng Indlcator Calibration and
Results of Chemicol Anolyses Obtained with on Argonne
Sampler. {Confidential with caption)

Liquid-Metal-Vapor Condensers
M. H. Cooper

Tests of the modified liquid-metcl-vapor con-
densers3® were completed. The NaK condenser
functioned satisfoctorily, but the sodium condenser
became partially blocked. The results of these
tests are summarized in Table 1.4.5. The re-
duction of the gus flow through sodium condenser
Il wes coused by the condenscotion and freezing
of sodium ot the first boffle, as shown in Fig.
1:4,33;

 

aé
M. H. Cooper, ANP Quar. Prog. Rep. Dec. 31, 1956
ORNL-2221, p 54 ¢

e
PERIOD ENDING MARCH 31,1957

Table 1.4+5. Summary of Results of LiquldsMetel-¥opor Condenser Tests

Condenser infet tamperature:  1200°F

Condenser ouilet temperature:  FOO°F

 

Condenser Designation Helium Flow Reote Comments

 

Sodium |1} 1.0=0.5 liter/min Condenser became partially blocked after 230 hr ond
caused the helium flew ro foll to 0.5 liter/min. Me

further decrease in How occurred during 1500-hr test

Mak IV 1.7 efm Satisfactery operction through 10 high-flow cycles of
2]-’2 te 3 min each

MNal 1V 0.5 liter/min Sotlsfactory performance for 300 hr: hewever, entroined
MaK corried into downstreom line by occidentol, con-

tinuous high flow

 

Sodium condenser IV, which is to be tested,
consists of a 7-ft length of ¥-in. sched-4D nconel
pipe that will be maintained at 300°F by electrical
heoters ond will be sloped toward the sump. [t is
expected thot the sodium will condense and drain
back into the sump, since the entire condenser will
be above the freezing point of sodium (208°F).
The dimensions of the NaK condenser [V have
been modified, os shown in Fig. 1.4.34. The

cooling section has been reduced from 24 to 12 in,

UL ED
KO

   

and the separating section has been changed from
a 24-in. section of 4-in. sched-40 pipe to & 12-in.
section of 6-in, sched-4C pipe. The NaK condenser
wos shortened because of space limitations above
the NoK sumps in the ART, ond the diameter of the
separating section was increased to compensate,

    

PARTIAL SODIUM PLUG
AT FIRST BAFELE 1

B Zirconium Fluoride Vapor Traps

. 4 M. H. Cooper

' The detailed design of an experimental prototype
of the ART zirconium fluoride thermeol-condensation
vapor trap, describad previously,®? was completed
and a unit is being fabricated. A diagram of the

experimental trop, which wos designed to be
easily disossembled for examination, is shown

 

Fig. 1.4.33. Inlet Section of Sodium Condenser I S?M H. Cooper ond F. A. Anderson, ANP Quar.
Afrer 1500-he Test, Prog. Rep. Dec. 31, 1956, ORNL-2221, p 51.

61
ANP PROJECT PROGRESS REFPORT

UNCLASSIFIED
ORHL =LA~ 0WG 20597

GAS OUTLET

  

S-u1 SCH- 40
PiPE

   
  
   

S T C

 

——— IMCONEL, DEMISTER
PaCKING

L

 

Py ar¥eeaietByn s et

~——DEFLECTOR PLATE

 

 

 

2, — INCONEL DEMISTER
] PACKING
|
|
'| 2-1n SCH-40 PPE
-~y

Yy ~in. TUBING —- l

GAS INLET

Fig. 1.4.34. Modified NoK Condenser IV,

62

in Fig. 1.4.35.
electricolty heoted and therefore does not include
the sedium jocket that will surround it in ART use.

Three tests of AIEO traps,
described previously, *8 were conducted in order to

The experimental trap will be

reaction also
determine the effect of temperoture on the Al 203-
Zer reaction. A section through the reaction
zone of each test assembly is shown in Fig. 1.4.36.
The reaction products fused in the tests made at
1600 ond 1700°F, but they did not fuse in the
test at 1500°F. It is believed that the fusion
was caused either by the formation of complex
reaction products or by impurities in the Al O,
pocking material,

An Al,O, trop performed satisfoctorily for 2000
br in an in-pile loop test. Also, o fuel pump wos
purged ot ART design conditions for 1400 hr
through on Al O, trop without difficulty.  The
A|203 trop used in the Pratt & Whitney high-
temperature critical experiment did not completely
remove the ZrF from the purge gos during rapid
fuel dumps. Some ZrF , was detected in dewnstream
gos contro! valves. However, the ZrF  removal
wos more complete than in the ART hig;h-remper-
ature critical experiment for which a thermai-
condensation trap wos used.

lnsulation Tests
. D. Peak

Tests to measure the heat losses
through three brands of insulation on piping carrying
high-temperature NaK in order to determine the
type of insulation best suited for the ETU and
ART NoK piping.

The insviating materiols tested were: {1) Fiber-
frax, type XSW, monufoctured by The Carborundum
Company, which is o I/z-in.-thick blacket of syn-
thetic alumina-silico ceramic fiber; (2} Thermoflex,
type RF-600, manufactured by Johns-Manvilie,
which is a T.fz-in.-'rhick brown blanket of synthetic
elumina-silica ceramic fiber; (3] Hy-Temp, manu-
foctured by the Keasbey & Mattison Co., which is
dictomaceous earth preformed for 2-in.-thick pipe.
The Hy-Temp insulation wos cut to length, placed
on the piping, ond tie-wired on. The Fiberfrax

and Thermoflex insulations were placed on erdinary

Were run

galvanized window fly screen and cut te size.

 

I8¢, A, Andersan ond M. H, Cooper, ANP Quar Prog.
Rep. Sept. 10, 1956, ORNL-2157, p &1,
£G

 

AHCLRSSIPIED
ORML~ LR- 0w 20856

= - = = — -—-'Tn———- A5,

Wt IRGET

    
  
 
    

 

 

ZrFy ~VAPOR CONDENSIT 18 THIS SECTION LT ca: i ZrF; —VAPOR CONDENSES N THIS SECTION k ! fifib&,\ . "I
CLRING LOW-FPOWETE (95 RATION It‘» B Ej,‘, DURING HIGH - POWER OPERATION s H::_ B _
., i o ] h B
I e e T Bl A TR Er T T e ‘-‘-—e—t:.. : ‘ e — - . ST 3 " HED1M—ET i
I 1 : i M R —— - : e

 

M —in DIA

 

 

i R g S— L s0ouner
#-I: - e 'L'Fl'l =t i A - i ’ } HQE___

T EEEECE PO EEErTFIA OF TR E R ERELT TS fi% i;%lrflr”m”flr";fi =

L

 

Ho0 QUTLET

REAR WIEW 4—4 OF TUBE CHaMBER

 

Fig- 1.4.35. Experimental Zirconivin Fluoride Yapor Trap.

2561 °1f HOHYW 9ONIONI aoid3d

 
ANP PROJECT PROGRESS REPORT

The blanket and screen were then wropped two full
turns around the pipe to give o l-in. thickness ond
tie-wired on.

The thermal conductivities and the heot losses
of the three brands of insulation are summarized
in Toble 1.4.6. The sime required to apply the

  

 

Thermoflex insulations indicated that it could be
applied ot on average rate of 2 ft/mon-hour, os
compared with a rate of 1]{‘ ft/man-hour for Fiber-
frax. The rate for applying the Fiberfrax insuletion
is lower because it is brittle and requires greater
care in handling.

URCL ASSIFIED
. PHOTO 38001

/700°FY/1600° FY1500° F

v ——

Fig. 1.4.36. Packed Sections of Three AIEDE Traps for ZrF4I Yopor Showing the Effect of Temperature on the

Reoction Products.

Toble 1.4.6. Heat Losses Through Insulation

 

Type of Insvlation Tested

 

 

 

Mean thermal conductivity ot 700°F, Btu/ heft2.0F
Thickness, in.

Power input required to hold at 1200°F with 90°F

ambient temperoture, w/ft
On 3,:’8- and l’z-in. sched-40 pipe

On ]!-"2-in. tubing

 

 

 

Thermeflex Fiberfrax Hy-Temp
0.050 0.049 0.053
1 1 2
130 107 Bé
154 143 149

 

64

“
PERIOD ENDING MARCH 31,1957

1.5. REACTOR AND FACILITY CONSTRUCTION

W. i. Boudreau

ART-ETU REACTOR COMNSTRUCTION
AND ASSEMBLY

G. D. Whitman
H. M. Abele C. K. McGlothlan
G. W, Peach

The schedule for ossembly of the ETU ond ART
reactors was revised on the basis of delayed
delivery of components being produced by outside
vendors. |t is planned to increase the local shop
load by transferring the fabricotion of all pres-
sure shells, additional strut-load ring ossemblies
and filter plates from outside vendors’ plonts to
the shops at the Y-12 plant, The lecal shop is
fabricating the ETU north head and will assemble
the reactors.

The detailed design of the recctor is essentially
complete, with some work remaining to be done on
the design of the sodium annuli surrounding the
fuel overflow, off-gas, and fill-ond-droin lines.
Mumerous design changes have been made bosed
on further design studies, development tests,
component fabrication difficulties, and weld shrink-
age tests. The status of the major reactor sub-
assemblies and components is described befow.

North-Head Assembly

Work has been stopped on assembly of the north
head pending the arrival of the sodium-fo-NaK
heat exchangers. The lower deck assembly has
been fabricated; the lower deck skirt has been hot
formed and rough mochined; and numerous other
small subassemblies have been completed. Ap-
proximately 30% of the assembly werk is completed,

The work on the sodium-to-NaK heat exchengers
is now progressing at a more satisfactory rate
than previously. All 30 of the offset-bend-ferming
dies hove been completed, The Griscom-Russell
Co. has completed assembling the tubes for the
first unif, and assembly of the tubes to the headers
of the second unit hos been storted. Successful
brozes are now being made, and it is felt that no
further brazing development work will be needed.

Reflector-Moderater Assembly

A lower beryllium hemisphere for the reflector-
moderator assembty hos been fobricoted, and the
spacer ferrules have been installed in the inner

M. Bender

surfoce. An upper beryllium hemisphere is com-
plete, except for contouring the top surface im-
mediately below the two sodium coolant inlet
pipes. This beryllium component is shown in
Fig. 1.5.1 during the machining operations on the
top surfoce. A second setof beryllium hemispheres
for the ART has been received, ond hole-driltling
operations are nearly complete in the lower half,

Work is progressing on o strut-load ring as-
sembly, ond the struts have been welded 1o the
upper ring. The load ring forging has been re-
ceived from the outside fabricator, and machining
operations have sforted on this piece.

The majority of the boron carbide tiles, and
spares, required for two reactors have been re-
ceived, and no difficulty is onticipated in obtaining
the remainder. The fabricatorof the stoinless-steel-
clad boren-copper shielding has hod difficulty in
rofling the material to desired tolerances, and o
relaxation in the toleronce on ratic of cladding to
cermet in the matrix has been authorized. Dit-
ficulty hos been encountered in rolling e product
thot is consistently acceptable metollurgically,
and more development work remains to be done on
this aperation.

The fabrication of Inconel shells, notably the
islond ond outer fuel core shells, continues to be
a major problem. The effort has been concentrated
on the shear spinning process, ond to dote no
completely satisfactory component has been pro-
duvced. The development problem has turned out
to be more difficult then originally estimoted,
Raw material quality requirements have had to be
improved in that the surface finish must be befter
than thot normolly obtained in the stock sizes
used. The physical properties of the material,
particulerly hardness and grain size, must be
carefully controlled throughout the forming process
to prevent cracking. The tool forces on the stock
during final forming of the inner and outer fuel
coce shells produce resuitant forces that tend 1o
pull the shells away from the mondrel. This
occurs because of the reverse curve in these
shells, and several light sizing passes have had
to be employed to acquire satisfactory dimensional
control.

Despite these problems a dimensionally ae-
ceptable outer shell has been produced, but over-

65
ANP PROJECT PROGRESS REPORT

- N -
-

s

 

UnCL ASSIFIED
PHOTO 8142

Fig, 1.5.1. Upper Baryllium Hemisphere for ETU Reoctor Assembly.

pickling after sne of the heot-treating operotions
has resulted in excessive intergranular attack. |t
is possible thot the surface of the shell moy be
cleaned without going below the tolerance on wall
thickness.

A small parailel progrom has been initiated in an
effort to fabricate the thin core shells by hot
forming oversized parts and then finish machining
them to final thickness. This seems to be o more
positive approach to the problem; however, the
metallurgical risks are theoretically greater since
thicker stock, in which there is greater chance for
flows, must be vsed, and there is none of the cold
working of the row material which is beneficiol to
the metallyrgical structure if properly controlled,

Island and Pressure Shell Liner

One set of beryllium islond parts is complete,
including the insteliation of the spacer ferrules.

66

Ne acceptable Inconel shell parts have been
produced, however, as discussed above.

The lower section of the pressure shell liner is
now to he fabricated by hot pressing and machining
inlieu of combination cold forming, shear spinning,
and machining. It is felt that the present plon is
o safer approach, since there were so meny un-
knowns in the celd forming and shear spinning
aspects of the original plon. Bids for the fabrice-
tion of this component are being evaluated.

No fabrication work has been done 10 date on the
]flé-in.-rhick shells to cover the shielding in the
reactor. It is planned to fobricate the spherical
sections of these poarts by spinning and to fabri-
cote the other sections by cold pressing.

Fuel-to-NaoK Heat Exchangers

The dual effort on fuel-to-NaK heat exchanger
fabrication was evoluated, and Black, Sivells &
Bryson was selected as the fobricator, All jigs,
tools, and fixtures ordered and made by York Corp.
are being transferred to Black, Sivalls & Bryson,

In the course of machining the first channel at
Black, Sivalls & Brysen an error was discovered
in the |IBM doto which wos used to construct the
cam.  Corrections were mode ond the second
channel machined showed the mochining method to
be satisfactory but demonstrated the need for
changes in the method of forming the rough chon-
nel. The third chonnel is now being mochined,
and the second chennel will be incorporated into
rough prototype assembly. The resistonce-heating
method of setting the tubes appears to be satis-
factory, and enough tubes for the prototype have
been formed.,

Assembly Methods Development

Weld shrinkage tests have continued, with a
total of 24 tests being completed for the welds
required for the '/{3-, ]’fi'f and 1-in.-thick reactor
shells. These tests have been performed in order
to determine the magnitude of transverse and
rodial shell shrinkage as a result of the welding
and to develop welding fixtures and procedures
that will enoble reasonably accurate prediction of
shrinkage, distortion, and quality of the reactor
core shell welds.

Test results have indicated that transverse
shrinkage can be predicted to within £0.010 in.
for 'fé-in.-fhick shells. Insufficient tests have
been run with the heavier shell sectiens to esti-
mate shrinkage voriation. Very litile success hos
been obtained in controlling the radial shrinkage.
The concentricity of the parts has been held to
within 0.010 in. totol indicoted reading, but the
amount of radial shrinkage has varied from negligi-
ble values to 0.050 in. in the Y%-in.-thick shells.
Many of the tests have been performed on very
poor quality rejected spun shells which have been
very difficult to weld. The tests on shells that
were closer to the required metallurgical and
dimensional specifications hove given much more
uscble welding data. An inner fuel-annulus core
shell and welding fixture are shown in Fig. 1.5.2,
This fixtute is used in a lathe fo align and support
the shell during the welding operations.

The first test on the l-in.-thick, 54-in.-dia hoops
that simulated the reactor pressure shells showed
the necessity of changing the reactor design to
accommadate the radial shrinkage. These hoops

PERIOD ENDING MARCH 31,1957

pulled in almost ]}:’ in. on the radius. The pressure
shelt design was changed from 1- to !'5_-in. thick-
ness at the weld by topering the cross section
above the equator on the inner surface so that
radial shrinkage would not restrict the flow pos-
sage or interfere with the shell liner. Additional
tests must be performed on the thicker shelils, and
preparctions are under way now to perform these
tests,

Tests hove been completed on spot welding
0.100-in.-thick stainless-steel-cled Cu-B,C cermet
shielding material to 0.250-in.~thick lnconel, and
satisfactory results were obtoined. Satisfactory
results have also been obtoined in spot welding
the 0,005-in.-thick stainless steel cons of the
copper-coated B, C tiles to the stoinless steel
cladding of the Cv-B,C cermets.

The reactor assembly procedures are being re-
viewed continually, and revisions are being in-
corporated as design changes and better assembly
techniques are conceived. The reactor assembly
area is complete except for one main section of
floor reinforcement. The security enclosure around
the assembly, inspection, cleaning, and sterage
areos is complete, ond isolefien from the remainder
of the building will be effected when major reocter
assembly operatiens begin.

ETU FACILITY
G. . Whitman
P. A. Gnadt A, M, Smith

All installation work on the ETU facility has
been tentatively stopped until July 1, 1957, and
the present status of the facility is described
below:

1. bench fobrication of the moin and auxiliary
NoK piping 90% complete;

2. installation of the main furnoce circuits ap-
proximately 70% complete; no installation werk
started on the isothermal or auxiliary circyits;

3. all cold-trap circuit components available for
assembly;

4. all floor supports installed;

5. main air duct installation approximately 50%
complete; all parts available for installation;

&, control room complete, except for a few miscel-
lanecus items;

7. intermediote runs of the auxiliary piping in the
basement approximotely 80% complete; no
odditional woerk scheduled until components

delivered;

67
ANP PROJECT PROGRESS REFORT

UNCLASSIFIED
PHOTO 27458 |

 

 

Fig. 1.5.2.

8. instellation of the cable trays for the bosement
aree 80% complete; other normel electrical
system installation work not sterted;

9. emergency electrical system 95% complete.

The following equipment is now on hond and
ready for installation:

1. mein and auxiliary NaK pump drive motors,

2. main NaK pump drive motor controller and

resistor units,

oll tubular and clamshell heaters,

4. oll pipe and components for fabrication and
installation of cold-trap systems,

5. NaK dump tanks for the isothermal and auxiliory
circuits,

6. induction regulotors ond Powerstots for voltage
odjustment of heaters,

7. motor control centers for hydraulic and lube oil
pUMP motors.,

A2

 

: - L = a’;:—-
f':t“%xrj'

Inner Fuel Annulus Core Shell ond Welding Fixture.

Fabrication of long-term delivery components
needed to complete the facility is under way on
the following items:
ail NaK pump stationary ossemblies,
furnace drain tanks,

NaK-to-air rediators,

NaK system flowmeters,
isothermal-system throttling valves,
control ponels,

NG A N

. hydroulic drive and lube oil packages.
Mu|-::-r efforts will be expended on these items
since they offect the faocility completion date.
Some of these items are still in the development
aond testing stage, ond therefore delivery dates
connot be quoted. Flowmeters are being coli-
brated in NoK pump component development tests.
The delivery of these flowmeters is contingent

upon the calibration test resuits, which are, in
turn, contingent upon the success of the NaK
pump tests.

The NaK-to-air radiators are being fabricated by
the York Corp., ond several development problems
in brazing techniques must be resolved before a
delivery date can be estoblished. Acceptable
brazed fin sections hove been produced, but a
recent design chunge mokes the fin bank more
flexible, and sagging may occur during the brazing
cycles.,

The lube oil systems are nearly completed, but
the units have not been tested. Delivery dates
for these items are scheduled early encugh to fit
in with other piping work; however, leak testing
and operational testing resvits may indicate un-
foreseen difficulties.

Major difficulties hove been encountered in
testing hydraulic drive units similar to the drives
originally specified for use in the ETU. These
difficulties have made it secessory to hold the
purchase order for ETU wunits in cbeyance until
revised units can be proved. The specifications
are being rewritten.

Detailed design of the NaK systems piping ond
supports is approximately 75% complete, and the
detoiled electrical distribution design is approxi-
mately 75% complete. The design of the auxiliary
service piping external to the test cell is approxi-
mately 25% complete. Preliminary estimotes show
an expected ETU facility completion dote, ex-
clusive of the reoctor and i$s associated equip-
ment, of September 1, 1958,

TEST CELL COMPONENTS
G. D. Whitman A. M. Smith

A re-evaluation of the ETU progrom has brought
about several chonges in connection with items to
be placed within the test cell thot simulates the
ART cell. The equatorial ond south-heod lead
shielding on the ETU has been eliminated, and
this port of the shield contract with Knapp Mills
was canceled. Design ond fobrication of the
equotorial section of the ART lead shielding was
stopped until a review of the shielding require-
ments can be completed. This review and neces-
sary engineering work is to be completed by
June 1, 1957. The design of the south-head shield
is being continued by the vendor.

Contracts for detoil design and fabrication of
the ART fuel {fill-and-droin tank and reacter sup-
port structures were canceled. The detailed lay-

PERIOD ENDING MARCH 31, 1957

outs necessary for fabrication are now being made
by the locol design group. Fabrication of both
components is scheduled to be done in the local
shops when the necessary design is complete,

Preliminary design of the ART fill-and-drain
tank lead shield was completed, but the final de-
sign myst await establishment of a firm design for
the tank.

Detail design of the nuclear instrument pod,
which was scheduled os port of the design work
to be done by outside sources, has beenreassigned
to the ORNL Instrument and Controls Deportment.

The ETU fuel fill-and-drain tank supports that
haove been fobricated will hove to be reworked
because design changes hove increcsed the length
of these supports.

The present statys of procurement and fubrica-
tion of other components is given below:

1. The ETU fuel fill-and-drain tank has been
received,

2. Prototype fuel dump volves have been re-
ceived and are being tested and evaluoted.

3. Design of the sodium expansion tonk was
approved, and an order was placed for the neces-
sary Inconel. Fabrication of the tanks for the
ETU and ART will be done by the local shaps.

4, A contract for fabricotica of the ETU and
ART NaK manifolds has been let to the Midwest
Fiping Company, and they are now in the process
of doing development work on bends and nozzle-
to-header welds and of qualifying welders.

5. Design of the hydraulic oil piping inside the
cell has been completed, and purchose orders were
issued for procurement of the materials required
for both the ETU and the ART cells.

6. The sodium system plugging indicator for use
en the ETYU and the two ART auxiliary piping
junction panels were completed.

7. Fabrication of the ART enricher and enricher
line assembly has continued without appreciable
difficulty.

8. Welding procedures and designs for the joints
between the bellows and nipple on the twelve
junction ponel expansion joints have been ap-
proved and the vendor is preceeding with fobrico-
tion.

Design work on other components inside the cell
is continuing. Preliminory design loyouts for the
fuel-recovery tank, fuel overflow ond off-gas
piping, lube oil piping, and the north-head lead
shielding are now being made. In addition, design

69
ANF PROJECT PROGRESS REPORT

layouts showing equioment locations ot various
slevations throughout both cells were started.
ART FACILITY
F. R, McQuitkin
Test Focility

G, . Hobinsen
R. [, Shotting

W. F. Ferguson
J. M. Mills, Jr.

Contract work on the ART facility hos been
contracted in four srages established by the design
program.  The Ffirst two stoages, packages | and
A, which included building oiterations, building
additions, cell instollotion, ond installation of
auxiliary services piping, hove been completed.
The third stoge, package 2, which included the
instaliotion of diesel generators and facility, efec-
trical control centers, and spectrometer-raom elec-
trical ond air conditioning eguipment, wos com-
pleted on Februory 20, 1957, The fourth stage,
package 34, which includes o distribution system
for supplying electrical power to the pipe heaters
and equipment heaters, a dry-air plant, and o build-
ing t house the air plant and wound-retor-motor
controllers, is 28,.2% complete.

As reported previously, the contract for package
3A work was awarded to the Rentenbock Engineer-
ing Company, effective December 7, 1258, ot o
contract price of $114,434, with o contract com-
pletion date of Aprid 25, 1937,

Work perlormed during the quarter on package 2
included the compietion of installation of anchor
bolts and support plates in the penthouse walls,
wiring of diesel-generotor control ponels, tests of
the air compressor and fuel oil piping for the
diesel wunits, and diesel-generator equipment op-
erational and performance fests.

During the testing of the diesel-generator equip-
ment the Ceoterpillar unit generator bearings were
damaged because of improper lubrication and had
to be replaced.  The control equipment required
minor design ond wiring changes. The performance
testing <ensisted in operoting each generotor in
poratlel with a common generator and loading with
twe 230-hp main blowers, which is approximately
0% of full capacity of the two units. NMecessary
governor, veltege-regulater, aiid  cross-current-
compensotor odjustments were made to give sofis-
factory load and current division., The five gen-
erator units were paralleled and satisfoctorily
loaded with the available load. Additional per-

/0

formance tests will be made when o lorger tess
lood is available,

Work performed during the quarter on package 3A
included excavotion, instaliation of underground
piping, pouring of the footings, und erection of
steel and siding for the compressor house, re-
arrangement of the electrical panels in the center
bosement, and installotion of conduit and caoble
trays from the switch house to the heoter distribu-
rion panel area.

Paclege 3 work on the design of the process
piping ond process equipment has cenfinued. Pro-
curement was initiated for the abserher pit piping,
material for heat barrier doors, and steel for the
fuel f#ill-ond-drain  tank coolont
supparts,

systemn  pump

Disassembly Factlity

A. A, Abbatielle V. J. Kelleghon
J. H. Dowdy! L. W. Love?
M. G, Willey

Plons for removing and sectioning the ART cofter
vperation were studied further, with one objective
being the determination of the important dimen-
sions which must be token during essembly in
order to hove basic information availoble with
which to compere the chonges ofter the resctor has
been operated. [t is apporent that much work will
be required fo take the meosurements and to set
up the sequence of disossembly cuts required to
chtain meaningful dora,  This work is urgent be-
cavse fixed reference points must be established
in the sssembly and ‘a3 built" dimensions must
be obtained,

The hydrauvlic shear which hos been tested ap-
pears to be an excellent severing device, ond it
can be remotely operared to cut and partially seal
off lines. This wos demonstrated with the use of
a tempeorary closed-circuit felevision system. A
turther adventage of this device is that shearing
disperses the minimum of activity into the air os
compored with the sowing and grinding methods
also considered.” Bids have been invited for a
shear copable of cutting the Elfz'-in. sched-40
fncone!l lines in the NoK system and operable from

 

Yon assignment frem Convoir.
20 osgignment from Fratt & Whirney Ajrcraft,

35, C. Bolger, §. F. D’Urso, and M. H. Fantano, ANF
Guar. Prog. Rep. Dec, 31, 19856, ORNL.2221, p 88,
This would be the main
cutting device used to remove the ART from the
test cell. The locations of lines within the test
cell have been studied to assure their accessi-
bility if they must be severed with a remotely
controlled device. The detailed operations to be
undertaken in the hot cell are also being studied.
ln general, welds will be cut away and the reactor
compenents will be removed in the reverse order of
their assembly.

a remote suspension.

Conceptual designs based on the interlocking
drilled or milled hole method of disassembly have
been mode for the principal items of heavy equip-
ment needed in the hot cell. The interlacking
hole technique was selected because it permits
the greatest flexibility in the cut location with the
minimum spread of activity. it is planned currently
that the two main hot-cell tools will be a turntable
mounted on a hydroulic lift, which will support
the recctor, and o Bridgeport type of universal
milling head carried on a cantilever floor mount,
which will drive the cutting tool.

Tentative design criteria were prepared for the
proposed disassembly facility. A large hot cell,
20 £t wide, 76 ft long, and 36 ft high, is to be

PERICD ENDING MARCH 31, 1957

provided that will be divided into two rooms by o
pair of shielding doors constructed of high-density
concrete 4]@ tt thick. The cell equipment layout
ond tooling criteria will be more completely deter-
mined before more firm facility criteric are ad-
vanced.

A Hz-scule model of the hot cell is now ovail-
able. It is anticipoted that the model will be a
substantial oid in visvolizing and anolyzing the
complex hot-cell techniques. The model is shown
in Fig. 1.5.3 with the approximately 57-in.-dia
reactor mounted on a tumtoble after the water and
lead shields have been removed. Twe floor-
mounted milling machines are shown forward of the
turntable, and preposed carriers for two overhead
monipulators and a 30-ton crane are shown over-
head, The longitudinal openings in the cell wall
will provide means for determining optimum locao-
tions for viewing windows. Beyond the reactor can
be seen o tank for decontamination of parts, as
well os several ports and tools [ying on fhe floor.
The openings in the floor of each room provide
access to the proposed underground tunnel for
transfer of hot materials to storage or to smaller
hot cells for subsequent component parts dis-
assembly and close examinaticn work.

71
72

ANP PROJECT PROGRESS REFPORT

Fig. 1.5.3. S5tudy Model of Hat Cell for Disassembly of the ART After Operotion.

UNCL ASSIFIED |
PHOTO 28280

 
PERIOD ENDING MARCH 31, 1957

1.4. ART, ETU, AND IN-PILE OPERATIONS

H. W. Savage

ART QPERATIONM PLANNING
W. B. Cottrell
C. ¥. Cunningham R. A. Dreisbach'

A critical review of the reoctor system oper-
ability is being made, and conflicts among the
desigr, construction, and instrumeniation phases
of the project are being resolved.
made of the manpower required during the various
phases of operation of the ETU and the ART ond
the disossembly of the ART. The peck work lood
during operation of the ART will require 40
engineers and 18 technicians. Although the ETU
operating crew will be chout one third this size,
all ART operating perscanel will receive ETU
operating

A study was

experience. Furthermore, since the
majority of project engineers have no experience
in nuclear operotions, arrangements are being
made for troining in the thearetical aspects of
nucleor radiation, shielding, reactor theory, and
and for on-the-job experience at other

operating reactors.

control

As a consequence of an extensive design review,
supplementory provisions have been mode for the
containment of activity at the facility in the event
of concurrent heat exchanger ond radiator leaks, a
number of design and instrument changes have
been incorporated into the system, ond additional
component development tests have been scheduled
to demenstrote the adequacy of the system for the
prescribed test program. The final specifications
of the test progrom will be based on the results of
detailed onalysis work and the component de-
velopment test program (see Chaps. 1.1 ond 1.4,
“Ajrcraft Reactor Test Design’' and ‘"Component
Development and Testing,'' respectively). In order
to provide for the contsinment of octivity in the
event of concurrent heot exchanger and rodiatoer
leaks, the NaK system pressure will be maintained
above that of the fuel ot all points where o leak
between the two systems could occur. In the
event of a leak or leoks, this pressure will be
maintained until either the activity has decayed 1o
below seriously hazordous levels, or the NaK
lines hove been pinched off with remotely operated

 

IOn ossignment from Pratt & Whitney Aiccraft.

D. B. Trauger

clamps in order to seol the activity in the reactor
away from the radiator leaks.

Further studies of the off-gos system have indi-
cated the advisability of directing the reactor off-
gos through the ZrF,-vapor trap on the reactor
without the option of using the vapor trap on the
fill-and-drcin tank if the reactor trap became re-
stricted.  Thus, the vaopor trop on the fill-ond-
drain tank will never receive vapor from the reac-
tor and should be in good condition when needed
during o dump. Thys, it is expected thafi normel
operation could continve without increased hazard
even though the reactor vapor trap became re-
stricted. Provisions are being made to bypass
both traps. The equalizer line around the reactor
vapor trap will be used during every dump, but the
byposs around the wvapor trap of the fill-ond-drain
tank will be for one-time emergency use only.
Appropriate development tests that will simulate
the different conditions to which the two traps will
be subjected are to be performed. Unfortunately,
it wifl not be possible to include fission gas heat-
ing, 20 w/cma, in these tests. The postpower
fuel-sompling system and the helium supply vent
and bypass provisions to the fuel expansion tank
are to be tested also.

The effect of xenon solubility in the ART fuel
on the ability of the helium purge system toremove
the xenon hos been studied,z and recent meas-
urements have established o value for the solu-
bility of xenon in the proposed reactor fuel (see
Chap. 2.2, ""Chemical Reactiens in Molten Salts''},
Therefore it was possible to predict the amount of
xenon that will remain in the fuel for any given
purge condition.® Measurements will be taken in
the ART of the xenon poisoning, which when com-
pared with the calculated xenon poison effects
should give a measure of the purging efficiency of
the ART xenon removel system.

Procedures were prepared for filling the various
systems with helium by pressurization and venting

and for filling the NaK system with NaK. Although

 

2M. T. Robinson, A Theoretical Study of Xel?® Poi-
soning Kinetics in Filuid-Fueled, Gas-Sparged Nuclear
Reactors, ORNL-1924 (Feb. 6, 1958).

3W. Lowen, Ap in the ART, ORNL CF-57-2-31 (Feb. B,
1957).

73
ANP PROJECT PROGRESS REPORT

bEleed lines will have to be installed in the special
NeK systems in order to fill the fill-and-drain tank
tubes and annulus with NaK, recent pump tests
indicate that the trapped gos in the cuxiliary NaK
system may be removed by the pump. [f this is not
so, bleed lines will have to be added to both
auxiliory NaK systems. A drain line will be in-
stalled in each of the two special NaK systems in
order to remove the NaK from the fili-and-drain
tank. There will be some holdup of NaK in all
NaK systems, but they will be flushed with
kerosene and alcohol prior to disassembly.

Two independent analyses are fo be made of the
ART off-gas. First, o source from a branch of the
off-gas line will be analyzed with o gamma-ray
specirometer. Becouse of the complexity of this
off-gas and because of background, Compton scat-
tering, shielding, ond other effects, use of o 5-in.
crystal is plonned. The output of the crystal will
be analyzed with o 256-chonnel analyzer. For the
second analysis o suitable gas chromatography
method will be used.

As a result of the various design studies, sev-
eral significent changes hove been made in the
ART instrumentation and contrels. The pressure
regulator bypass around the fill-and-drain tank
vent valves has been removed, and a pressure
regulator hos been added to the helium supply to
provide bhetter control and to limit the maximum
pressure available for the filling operation. As
stated aboVe, an equalizer line has been provided
from a peoint downstream of the vaper trap on the
fill-and-drain tank to cuach of the fuel pump helium
inlets o permit o fuel dump ir the preseribed time
even in the event that the reactor vapor trap is
plugged. The equalizer line contains two solenaid
valves in parallel, and check valves are located
both in these bypass lines ond in the helium supply
tines where the two join ¢head of both fuel pumps.
The solenocid volves will be interlocked to open
when the fuel drain valves are opened, ond they
will be manually operable from the contral room at
other times.

A line from the helium supply will be provided
to admit gas into the fill-and-drain tonk to periodi-
cally test the vepor trap for plugs. This line will
have o fixed arifice and will include o solenoid
valve, a pressure transmitter, and a check vclve.
The solencid valves will be operated from the
control room where the pressure will be continu-
cusly recorded.

74

The frangible-disk-operated valve for bypassing
the vapor trap on the fill-and-droin tank will be en-
ergized on an emergency basis in the event of a
plug in this vopor trap, A separate nitrogen supply
has been provided for enecgizing the valve. The
downstream taps of the two fuel expansion tonk
level indicators were moved to improve the reli-
ability of the instruments in the event of a plug of
the recctor vapor trop.

Praovisions were made for venting the gas from
the NaK pumps and dump tanks through @ common
vent header to the cell off-gas line. Two vacuum
pumps in parallel with a bypass line and an isola-
tion volve were provided in the cell off-gas line
upstream of the charcoal bed and sample station.
With these provisions it will be possible to main-
tain the cell at o pressure of 5 psia se that in the
event of concurrent radictor ond heat exchanger
leaks, the cell may act as o sink for activity by
venting the reactor fuel system to the cell. A vent
connection was provided from the reactor off-gas
line to the cell atmosphere that consists of a
frangible-disk volve e¢nd o solencid valve in a tee
in the off-gos line downstream from the pressure
control valve which maintains the resctor expan-
sion tank pressure.

ETU CPERATION PLANNING
W. B. Cottrell

C. W, Cunningham R. A. Dreisbhach
5. M. DeCamp

The operation of the ETU is prereguisite to that
of the ART, ond the mointenonce of ETU flow
sheets, instrument lists, ond valve tabulations is
being emphosized.  An instrument application
tabulation was prepared which gives maximum,
minimum, and normal velues of the quantities to
be measured, as well as the type of pickup instru-
ment, location, and number required and the icca-
tion, type, accuracy, and range for the reading
instruments by which the various quantities are
presented. A valve tabulation is currently being
prepared.

A preliminory control block diagrom wos used as
a basis for preparing a list of “emergency condi-
tions for ETU operations.'' There will be no em-
ergency conditions in the nuclear sense for the
ETU reactor, but for training purposes it is pro-
posed to set up annunciations of emergencies
similar to those provided for in ART operation

18
fa

planning, and insofor as possible the same pro-
cedures will be followed as would be required by
such an emergency during ART operation.

ART DISASSEMBLY PROCEDURES
W. B. Cottrell
R. P. Shields

A number of studies have been made of proce-
dures for disassembling the ART, including dis-
pesal of rodicactive materials.
spray the inside of the test cell after operation of
the ART will be highly contaminated. The Health
Physics Division has recommended that this water
be taken by truck to the tonk farm at X-10. The
Operations Division has a tank truck with o
shielded cab and hos indicated thot it can be used
for this purpose. One stipulation is that the water
must be slightly alkaline. The problem of sodium
and NoK disposal is still being investigated.

The method for effecting the remote cutting op-
erotions required within the test cel!l for removal
of equipment is being studied, and several dis-
adventages of o proposed underwater operation
have developed, The following factors are perti-
nent to this study.

1. Iron oxide and hydroxide are good scavengers

¥. E. Browning

Water used to

for many types of radicoctivity, especially stron-
tium, ond can be expected to concentrate activity.
2, The following radiation leveis would exist
above the surfoce of woter containing the indi-
cated concentrations of rodicactive strontium:

Dose 10 in, Abpve
Sucface {me/he)

Concentration of Activity in
the Water (curies/ml)

0.35 18
1.5 85
1.9 110

3. Leod oxide metal paint adsorbs activity and
is not easily decontaminated.

4. Buoyancy and solubility effects in underwater
cutting tend to remove contamingted materials
from partially restricted areas where they might
remain undisturbed in a gaseous atmosphere.
Therefore, greater precautions muyst be taken to
prevent gscope of activity,

PERIOD ENDING MARCH 37,1957

{N-PILE EXPERIMENTATION
J. A. Conlin
In-Pile Loops

A shift of emphasis in ART supporting research
has resulted in the cancellation of further in-pile
loop tests of zirconium-fluoride-base fuels in
lnconel systems. The informotion thus far ob-
tained from the in-pile loop program has wverified
the results obtoined through irradiation of capsules
ond through out-of-pite corrosion tests. The
irradiation tests indicate that no substantiol
change in fuel properties or corrosion rate is to
be expected because of radiation. Further in-pile
loop tests will be initiated when suitable nickel-
molybdenum alloys and lithium-base fuels are
available.

In-pile Yoop No. 6, the operation of which was
described previously,? was cut up at NRTS ond
returned to the ORNL hot-cell facilities where it
is owaiting disassembly and metollogrophic ex-
amination.

Prototype In-Pile Pump

Futere in-pile loop work will require o pump
capable of operoting for longer times and at higher
temperatures than those used thus for., Therefore
a development program hos been initiated to ex-
tend the established operating range of the in-pile
pump.® To provide a margin of safety for future
loop operotion, the pump should be capable of
operating ot 1500°F at about 150% of the present
salt flow requirement and for 2000 hr. The shofts
of eoch of the first two test pumps seized after
approximately 1000 hr of operation as a result of
zirconiuvm fluoride accumulation around the shaft
slinger. A third test, previously described,® was
set up for which a zirconium fluoride vapor trap of
cluming was used ond the circulated fuel mixture
was NaF-ZrF -UF, (56-39-5 mole %, fuel 70).
This pump satisfactorily met the required c¢ondi-
tions by completing the 2000-hr test ot 1500°F and

 

4J. A. Conlin et al., ANP Quar. Prog. Rep. Dec. 31.
1956, ORNL-2221, p 89,

*J. A. Conlin ond D. M. Hoines, ANP Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 90.

75
ANP PROJECT PROGRESS REPORT

4500 rpm, which corresponds to o flow rote of
.75 gpm ot a 19-ft head. Postrun disassembly
showed all pump parts to be in the as-built condi-
tion, There waos some zirconium fluoride bhuildup
on the pump housing surrounding the slinger but
not enough to contact the shaft. The alumina in
the vapor trap, which had cperated for its design
life, appeared to be fused and the trap was almost
plugged; no difficulty had been ex-
perienced with the purge flow through the trop.

The results of the test justify the use of this
pump in in-pile loops with maximum temperotures
in the range of 1600 to 1650°F, which correspond
te o pump temperature of 1500°F. This conclusion
carties the limitation for zirconium-fluoride-base

however,

fuels, however, that the zirconium fluoride vaper
pressure ot pump temperatures must not exceed
that of this test, 3.6 mm Hg. Locop flow and test
duration may also be increased safely over the
previous Jimits.

In-Pile Loop Heat Exchanger Tests
C. C. Bolta

Future in-pile loops are expected to be operated
ot higher power densities for which the capacity,
30 kw, of the present in-pile lcop heat exchanger
is insufficient. A program wos therefore initiated
to develop o heat exchonger with a capocity of
45 kw or more. A test unit similor to o forced-
convection loop with a resistonce-heated section
and a 5-gpm centrifugal pump was set up to ac-
commodate varioys heat exchangers. The first
series of tests were run on the standard in-pile
loop heot exchonger to obtein dote for comparison
and io verify loop power generation data. Curves
of heat removal vs air flow were obtained, ond
the results verified that with inlet oir ot a pres-
sure of 90 psic o maximum of about 30 kw of heat
can be removed.

In order to increase the capacity of the heat ex-
changer, o water injection system was devised and

76

tested. This consisted of @ small, high-pressure
pump for injecting 0.12 gpm of water into the inlet
air stream through a small shower-head nozzle.
Preliminary calculations indicated that about 15kw
of heat should be absorbed by the heat of vaporiza-
tion of the water alone. The test results showed
an increase of only 2 to 3 kw, however, ot all air
It was apparent that cdmplete vaporization
Further develop-

flows.
of the water was not cbtained.
ment of this system was dropped because of the
complications foreseen in developing an efficient
water dispersion system for in-pile use. A second
test heat exchanger with machined nickel fins hos
been designed and fobrication has been started.

In-Pile Test of Moderator Material
C. C. Bolta

The first in-pile test of moderator materials for
use at high temperatures was completed. In this
test four samples of beryllium oxide were irradiated
in the HB-3 beam hole during two MTR cycles with
gamma heat estimated ot between 2 and 5 w/g and
o fast neutron flux of about 1073 neutrons/cm?.sec.
The high-temperature end of the capsule was
normeily ot cbout 1500°F. The capsule was
thermally cycled 32 times as a result of reactor
power fluctuations, with temperature differentials
ranging from 200 to 1400°F. The results of ex-
amination of the irradiated specimens are pre-
sented in Chap. 4.1, “‘Radiation Damage."

A second test assembly hos been prepared for
irradiation in the MTR which includes three cap-
sules mounted side by side in a vertical position.
A l-in.-dia capsule contains three beryllium oxide
encased in a 0.020-in.-wall Incorel
sheath; o 1.25-in.-dia capsvle contains graophite
encased in a 0.020-in.-wall nickel sheath; ond a
%-in.-dio capsule contains zirconium hydride en-
cased in @ 0.025-in.-wall molybdenum sheath.

specimens
PERIOD ENDING MARCH 31,1957

1.7. HEAT TRANSFER STUDIES

1 8

ART FUEL-TO-NaK HEAT EXCHANGER
A. L. Wantlond S. |. Cohen

Heat tronsfer studies of the delta-array heot
exchonger] were continved. Meaosurements were
obtained in the laminar flow region for Reynolds
moduli as low as 200. 1t may be seen from Fig.
1.7.1 that the data for Reynolds numbers of 500

 

'J. L. Wantland ond S. 1. Cochan, ANP Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 277.

H W. Hoffman

to 1000 are correlated by the equation
N

My

0.
= 0.47NR°3‘5

 

0.4

NPr
Below N, = 500 the dato foll progressively
farther below the curve representing this cor-
relating equotion. However, the significonce of

these data is limited by experimental uncertainties
reloting to flow control and heat generation in

UNCLASSIFIED
ORANML -LR-DWG (9447

 

 
 

 

#
”

 

 

  

 

 

/]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o | ___] | | | I | L b e e oo 0 b
400 1000 2000 3000 10,000
” Nre
Fig. 1.7.1. Heat Transfer Characteristics of the Delta-Array Heat Excharger,

77
ANP PROJECT PROGRESS REPORT

this very low Reynolds modulus region. At these
low Reynolds moduli, the effects of natural con-
vection ossume grecter importonce in the totol
heat transfer. In order to evaluate the natural-
convection contribution, the system was operated
with flow in both directions through the heat
exchanger. If the natural-convection mechanism
s involved, the heat transfer for flow in the
reverse direction (downward through the exchanger)
should be less than that for flow in the normal
direction. The date presented in Fig. 1.7.1 show
no effect of natural convection over the Reynelds

modulus range examined. From considerations of

the geometry ond orientation of the heat exchanger,
this result was to be expected.

Additionol fluid friction date were obtoined for
the fuel side of o straight-tube model of the ART
fuel-to-Nak heat exchanger. The data now avaifable
cover the Reynolds modulus range of 150 to 8000.
The new datg are shown in Fig. 1.7.2, aleng with
the previously reported results.
the predictions based on experimental measurements

For comparison,

with a 100-tube bundle having o similar spacing
configuration are given also in Fig. 1.7.2, In the
renge from N = 1000 to N = 500, the dcta fall

along the extrapolation of the earlier data. Over

UHCLASSIFIED
ORML-LR-DWG 19418

 

  

 

 

       

 

 

 

 

 

 

 

 

 

 

 

 

 

T T |— ¥ | | ‘ | | = | |
- ||
s ot ; . . = == =il = 1 e
i | 64 | ] ] ‘
CIRCULAR DUCTS, /= =— ‘ | |
| | Re | |
' |
| | |
. | | |
02 —_— —— . \ _ —_— —_— = — ,I —i —
| ]
L | |
| o
| : de |
i SQUARE CHANMELS, f= —— — | . \ | '
5 &
% 0_! — - —! — I' — _‘h‘-fi'_ — o ——fe — + —r--—!-__.
e — /[ Mgy /¥s Mg CURVE |
= — — —- — > — ——— PREDICTED BY WANTLAND +—
\ e g = )
s — — = = oy 1 :
o - [
5 L e e | |
= | | |
'S
0o — — T - . — 1
|
|
e — e _I na wang - =
|
|
-
002 — - — — ‘— |
|
| |
| |
|
| | ‘ ‘
| | o n
| | | |
0. ] | I I | i | | [ |
100 200 500 1000 2000 SO00 10,000
REYNOLDS MODULUS, A,
Flg, 1.7.2. Fluid Friction Characteristics on the Fuel Side of a Straight-Tube Mode! of the ART Fuel-te-NoK

Heat Exchanger. (Secret with caption}

78
the range N, = 500 1o N~ = 3000, the doto
can be specified by the egquation

oL
Ng.:'}fi
In the laminar region the data lie between the
curves { = 56/N,_, for square channels, and
f = 6d/NRa' for circular ducts. The early fron-
sition, in the vicinity of N, = 400, results from
the turbulent wakes behind the heat exchaonger
spacers.

Preliminory designs were made of a system for
studying the heot transfer and velocity structures
in o single fuel-side channel of on ART type of
heat exchonger for both the square- ond the delro-
tube arrangements shown in Fig. 1.7.3. Fer the
velocity profile determinotions, the opperatus will
be constructed of Plexiglas so that the profiles
can be visualized and photographed by using the
phospharescent particle technique.2 The heat
transfer experiments will be conducted in chonnels
with flow cross sections identical to those used
in the flow visualization studies. The individual
tube sections will be constructed of stainless
stee! ond will be electrically insulated from each
other, Heat will be generated electricolly in the
tube sectiors and will be transferred to the water
flowing through the test unis. It will be possible
in this system to obfoin the periphera! voriation
of the heat transfer coefficient.

ART HYDRODYNAMICS

L. D. Palmer F. E. Lynch
G. L. Muller®

Core Hydrodynamics

Earlier experimental studies of flow stobilization®
in reactor cores demonstrated the value of screens
placed in the diverging portion of the flow channel.
These experiments were made with on axial-flow
entrance sysfem. Megsuremeats are now planned

 

2, D. Polmer, £. E. Lynch, and G. L. Muller, ANP
Quar. Prog. Rep. Dec. 31, 1956, ORNL-2221, p 279;
see alsg '"‘Instantonecus Yelocity Prefile Macsurae-
ments,'" this chapter.

*0n assignment from Pratt & Whitney Aircroft.

e N Copenhaver, F. E. Lynch, and G, L, Mullar,
ANP Quar. Prog. Rep. Sept. 10, 1956, ORNL-2157,
p 225; L. D. Polmer, F. E, Lynch, ond G. L. Mulle,
A!;;-’B Quar. Prog. Rep. Dec. 31, 1956, ORNL-222],
F +

PERIOD ENDING MARCH 31,1957

UNCLASSIFIED
ORNL-LR~- WG 18419

 

    

i

{o) SGUARE ARRAY

 

 

1&) DELTA ARRAY

Fig, 1.7.3. Typlcal Flow Channels for Velocity
Profile Visvolizotion 5Studies of Parolle!-Tube Hea:?

Exchongers.

with @ swirl entrance system similor to that of the
ART core heoder to determine the effect of screens
on the hydrodynamic pattern of this system. The
experimental system will again involve the 10/44-
scole modet of the ART core, olong with provision
for the introduction of various mesh screens at
a number of planes in the diverging half of the
chonnel. Preliminary celculations indicote that
the pressure unbalances in the present ART header
are of the some order of megnitude as the friction
losses through the core. Under these conditions
peripheral flow asymmetries in the core could
develon. Such flow asymmetries were seen in
velocity measurements in the full-scale Plexiglas

79
ANP PROQJECT PROGRESS REFPORT

model of the ART core.® Becaouse of the velocity-
modifying power of the screens ond the cross-
mixing between parallel streams flowing through
the core, the extent to which these periphercl
flow varigtions are "‘ironed-out’’ cannot be pre-
dicted. Thus, experimental
the core models are necessary.

ohservations with

Some calculations were maode of the temperatures
to be expected in such flow-stabilizing screens
during reactor operation. The cnalysis indicated
that the screens would be essentially ot the fluid
envircnment temperoture during normal operation.
During a fuel dump the screens would be subjected
to gammo-ray heating without the cooling normally
afforded by the flowing fuel,
heat conduction was assumed as the only heat-
removal mechanism and the internal heat source
was tecken to be B w/cma,. ¢ moximum screen-to-
wall temperature difference of 1100°F was obtained,
Thus, for a wall temperature of 1100°F, the max-
imum screen temperoture would be 2200°F. This
calculation provides o coaservative upper limit

If one-dimensional

in that notural-convection heot tronsfer to the
surroundirg helium and radiation heat tronsfer to
the cooler walls were not considered.

Instentanecus Yeloclty Profile Measurements

The program of photogrophic recording of in-
stontaneous velocity profiles by using the phos-
phorescent dye technique has included studies
with circular tubes of severol diameters, a two-
dimensional diffuser with o half ongle of 8 deg,
ond o l-in. spiral. Examples of lominar flow in
a circular tube and turbulent flow in the spiral
were presented previously.? Further illustrations
under other flow conditions and different geometries

 

w. . Stelzman, G. D. Whitman, ond W, T. Furgerson,
private communicetion to G. L. Muller.

 

Re= 8750

aore given in Figs. 1.7.4 ond 1.7.5. Figure 1.7.4
is a photograph of a turbulent-Hlow profile (N
= 27,000} in a 3-in. glass pipe. Figure 1.7.5
shows a group of three photogrophs of flow profiles
in the diffuser over a Reynolds modulus range of
7000 to 8750. In port {a} of Fig. 1.7.5 the maximum
velocity occurs close to the left wall of the
channel. tn part () the position of the maximum
haos shifted closer to the channel center, and the
beginning of a reverse flow region on the right
wall can be seen. In part {c} well-developed
reverse flow is apparent. This shifting of the
velocity pattern is not rapid in the diffuser studied,
the photographs being taken during an entire
afterncon of intermittent system operation. After
each photograph the velocity in the system was
increased sufficiently to ensure removal of all

UNCLASSIFIED
PHOTO 40650

b—- '

Re= 27,000

 

Fig. 1.7.4, Turbulent-Flow ¥elocity Profile in o

3-in, Gloss Pipe,

UNCLASSEFIED
PHOTO 406851

Re=8750

 

Fig. 1.7.5. Velocity Profiles in o 1{; % & in. Diffuser with an 8-deg Half Angle,

80
the phosphorescent dye from the tubing wells.
The flow was then reset to the desired test value
and the system was allowed to reoch equilibrium
conditions before the next photograph was taken.

To obtain quantitative information on the fluid
velocities, the profile photograph must be super-
imposed on a grid photogroph. The grid is ploced
within the test section so that it lies in the plane
of the collimated light beam used to obtcin the
The system is filled with water
and o photograph is taken. Thus, the same dis-
tortion will exist in both the velocity profiles
and the grid phetographs. During this aperation
the camera remains rigidly positioned.

The velocity profile in the composite photograph
of Fig. 1.7.6 wos token at a Reynolds moedulus
of 10,000 in a 2!’?-in. plastic tube. The slight
dip in the profile neor the tube center probably

velocity profile.

from the particular entrance coaditions
of the test unit. The time interval for this picture
{the time that elopsed between the flash of the

results

light beam which excites the phos-
phorescent porticles and the opening of the camera
shutter) was 0,2373 sec.

collimated

The Reynolds modulus

 

Fig. 1.7.6. Composite of Grid and Velocity Profile
Photographs for Flow Through «a ?%-In. Tube at o
Reynolds Number of 10,000,

PERICD ENDING MARCH 31, 1957

based on the mean velocity evcluated from the
grid-photograph  dato checked with
Reynolds

the system
modulus as determined by calibrated
rotameters to within 10%. Further studies over
o wide Reynolds modulus ronge are planned.
The results of these studies will be compared
with datc obtoined by uwsing more conventicnal
techniques, such as pitot tubes ond hot wires.

HEAT TRANSFER IN REFLECTOR-MODERATED
REACTOR CORES
N. D. Greene W. R, Gombill

ART Core Heat Transfer

The hydrodynomic experiments in which screens
were used in the 10/44-scole plostic core model
indicoted the need for further volume-heot-source
studies of a similar geometry. A heat-generation
analysis hes shown thot @ number of fine-mesh
plastic screens can be placed in the northern
hemisphere of the core without drastic disturbance
of the flux field. In the pores of the screens (the
region in which the flux density is highest) the
heat generated will be 2]’2 times that in the free
fluid. The effect of these regions of high tem-
percture on the over-cll temperature field is made
negligible by two factors: (1) the volume of fluid
affected is only a very small percentage of the
totel fiuid volume, ond (2) the stringers of high-
temperature fluid issuing from the screen pores
are rapidly dissipoted in the highly turbulent
wokes behind the screen wires. Such local in-
creases in heat generotion would be limited to
resistance-heated systems, and thus similar screens
in the ART core would not be subjected to the
higher temperatures. Wall-temperature meas-
urements will be made in the regions between the
screens, as well as in the undisturbed scuthern
hemisphere of the core. Fluid temperatures will
be obtained with the odjusteble thermocouple
probes located ot the core outlet. These probes
will be used both to survey the fluid radially in
a fixed oxial plane and to investigate temperature-
time correlations befween the four thermocouple
probes. These latter studies will be made for
beth the swirl entrance system with the core
screens and for the swirl entrance alone. |t is
hoped that the results of these experiments will
shed odditional light on the origins of the tem-
perature fluctuations which hove been observed
in the ART type of core geometry.

81
ANP PROJECT PROGRESS REFORT

Vortex Tube Heat Transfer

A volume-heot-source experiment was designed
that utilizes the vortex flow principle.® For these
measurements @ uniform volume heat source will
be generated electrically within a sulfuric acid
solution flowing through on insulated cylindrical
pipe. The desired fluid velocity profile will ke
created by vanes in the entrance section of the
tube. Both steady-stote and iransient temperature
measurements will be obtained by using surfoce,
as well os fluid, thermocouples. The fluid ther-
mocouples will be located at the tube exit and
will provide a radial traverse of the tube.

A similar experiment will involve the measurement
of heat transfer coefficients in o heated-tube
system contoining a fiquid vortex. It is estimated
that the heat transfer in this system will be
greater than that for cenventional flow of flyid
throygh a tube. In such a system it may be
possible to maintain satisfactorily Sow tube-woll
temperatures without secondary cooling.  The
effect of centrifugol forces on the temperatyre
structure will also be investigated. Hydrodynamic
studies of the stobility of the fluid vortex ond
other phenomena at the interface between the
vortex and the main fluid flow are currently in
progress.

THERMAL-CYCLING EXPERIMENTS
H. W. Hoffman D. P. Gregory®

Studies with the system for thermally eycling
Incenel tubing were continued in order to determine
the effects of fluctuating temperatures and stresses
on Inconel exposed to flowing i"'hflF-ZrF“--UFJ1
(50-46-4 mole %, fuel 30). The results obfained
to date are summorized in Tobles 1.7.1, 1.7.2,
and 1.7.3. Figures 1.7.7 through 1.7.14 show the
maximum penetration observed in the test sections
used for the experiments described in Tables
1.7.1, 1.7.2, and 1.7.3.

In order to investigate the pessible effect on
the depth of attack in Inconel of the interaction
between the heoter section ond the test section
and of the heater length, an experiment was per-
formed (ET-S) in which the entire 12-in. test unit
served as a heoter. The heater wos subjected to

 

H. . Foppendiek, N. D, Greens, ond L. 0. Palmer,
ANP  Quar. Prog. Rep. Dec. 31, 1956, ORMNL-2221,
» 281,

82

temperature cycling under the some wall-temper-
ature-difference conditions as existed in the test
sections used in runs ET-H ond ET-0. The test
unit for run ET-H was divided into a 4-in. heoter
and on B8-in. test section, while the test unit for
run ET-0 weos on 8-in. heater and ¢ 4-in. test
section. Frem the dota in Tcble 1.7.1, it may be
seen that the general subsurfoce-void formation
was approximately the same in oll three test
sections. However, the ET-H test section showed
deeper intergranular stringers than the ET-0 test
section. The specimen taken from the test unit
for run ET-S showed a tendency toward inter-
granular attack in the regions of deepest pen-
etrotion, The exposure times for runs ET-H and
ET-O were approximotely the same, while run
ET-S lested 16 hr less thon ET-H. Although it
is difficult to draw specific conclusions on the
basis of these three runs alone, it appears thot
the heoter-test section interaction and the heater
length may not have significant effects on attack
in this thermal-cycling system.  This result
possibly could have been predicted on the basis
of the surface area-to-velume ratios in the system.

The resulss of run ET-5 probebly connot be
caompared directly with those from runs ET-H and
ET-0 from the stress stondpoint. The temperature
difference across the velume-heated metal wall
of the ET-S test section was less thon that of
the test sections used for runs ET-H ond ET-0O.
Since the thermally induced stress is a function
of the temperoture grodient in the wall, it is to
be expected thot the stresses in run ET-5 were
lower than those in runs ET-H ond ET-0. There-
fore, run ET-T was made under temperature ond
geometry conditions identical to those of run ET-S
with, however, continuous thermal cycling. Since
run ET-T lasted for 97 hr, while run ET-S lasted
for only 50 hr, the test unit vsed for run ET-T
was ot the system maximum temperature opproxi-
motely four times as long s was the test unit
used for run ET-5. A comparison of Figs. 1.7.12
and 1.7.13 shows thot somewhot deeper attack
occurred in run ET-T than in run ET-S. Further,
while the aumber of voids in the specimen from
run ET-T is less than in the specimen from run
ET-S, the ET-T voids are much larger. Additionat
dato must be obtained before hypotheses can be
based on these results.

The series of runs ET-O, ET-P, and ET-Q was
designed to obtain a comparison of the effects of

low-frequency, high-frequency, and isothermal
£8

Toble 1.7.1, Resuits of High-Frequency Thermel-Cycling Tests of Inconel Tubing Exposed to
Flowing NaF-ZrF -UF, (50-46-4 Mole %, Fuel 30}

Cycling rete: 0.4 cps
Tubing size: !‘:1 in. 0D, 0.035-in. wall

 

 

 

 

Heater Section Test Section
Ouration Interface Moximum Depth of Artack Interface Moximum Depth of Attock lnlet Mixad Mean
Run of Run Length Temperature {mils} ond Type of Temperature {mils) and Type of and Tank
{hr) (in.) (°F) Subsurface Voids °F) Subsurface Voids Temperature (°F)
Averoge  Fluctvation  General Intergranuler  Average  Fluctuation  General Intergranular
ET-A 14 4 1455 180 & 6 1273 8 o 1 1250
14 1546 225 1273 12
-8 240 4 1415 170 2 None 1330 7 2 None 1250
-C 4 4 1755 189 Nane 2.5 1613 37 2 None
-D 14 4 1845 240 None g 1590 83 4 None 1580
-E 5 4 1905 235 None B 1635 43 3 None 1600
-G 23 4 17560 190 None 4 1583 13 4 None 1545
-H &b 4 1730 125 None 13 1600 32 6.5 o 1570
-1 106 4 1850 125 None 9 1600 22 7.5 Mone 1375
-N 13 8 1740 110 3 4,5 1640 16 5 None 1585
-0 59 8 1730 120 Neone 10 1600 20 5.5 6.5* 1565
-R 100 8 1750 100 Mene 2.5 1615 30 3.5 5.5* 1565
-5 50 12 1500 22 5.5 None 1570
-T 97 12 1600 25 None 7.5 1570
-3 100 8 1785 100 1.5 13.5 1625 20 4.5 6* 1560

 

*The intergrenulor ottack vppeared as stringers of voids thot were more localized than the general void formation.

2861718 HDHVYW ONIONT gO1¥3d
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

Table 1.7.2. Results of Low-Frequency Thermal-Cycling Tests of Inconel Tublng Exposed to
Flewing NaF-ZrF -UF, (50-46-4 Mole %, Fuel 30)
Cycling rate: 0,01 cps
Tubing size: !i in. QD, 0.0353-in. wall
Heater Saction Tost Section
Maoximum Depth Maximum Depth Inlet Mixad Mean
Duration Interface of Attack (mils) lnterfocs of Artock {mils} and Tork Temperature
Run  of Run Lo Temporature ond Typa of T;mperarure and Type of (°F) and Tamparature
thr) enaf {°F) During Subsyrface {"F) During Subsurface During Sscond Hall
(el Firgt Halt Volds First Helf Volds of Cyele
i Cycel of Cycle
o e General  lntergranular Ganeral  Intrergronulor
ET-J 23 4 1800 5 5 1622 2 None 1585
-K 100 4 1850 2 13.5 1640 2 7 1565
A 59 8 1850 None g 1640 None 5.5 1600
Table 1.7.3. Results of Isothermal Tests of Inconel Tubing Exposed to
Flowing M-:uF-ZrFJ{-UF‘i (50-46-4 Mole %, Fuel 30)
Tubing size: ]/4 in. 00, 0.035%-in. wal!
Maximum Depth of Attack {mils) and Type of
R Duration of Run Temporoture Subsurfoce-Void Formation
un (hr) (°F}
General Intergranular
ET-F 5 1435 None None
- M 100 1580 5 None
-P 59 1600 i 4
operotion.  The heater sections used for runs served under isothermal conditions in thermal

ET-O and ET-Q were 8 in. long. A fresh fluoride
salt mixture was chorged to o new system for run
ET-0, and the same salt mixture ond tonks were
then used successively in runs ET-P and ET-Q.
The resuvlts of these runs are shown in Figs.
1.7.8, 1.7.9, ond 1.7.10. The specimen from high-
frequency run ET-O shows o higher frequency of
attack the the
frequency run ET-Q. This comparison is in rec-
sonable ogreement with the previously reported
comparisons’ of run ET-G with run ET-J ond of

run ET.l with run ET-K. The attack in the iso-
thermal run ET-P, although higher than thot ob-

than does specimen from low-

 

?H. W. Hoffman and D. P. Grogory, ANP Quar. Preg.
Rep. Dec. 31, 1956, ORNL-2221, p 282,

84

convection or forced-circulation loops, is less than
that for runs ET-O or ET-Q. The high isothermal
attack may be the result of the surfoce oreo-io-
volume ratio present in this system. However,
since even in the isothermal runs there was some
low-temperature-ditference low-frequency cycling,
the explanation for the high isothermal ottack is
not yet clear. :

Runs ET-R ond ET-U {Figs. 1.7.11 and 1.7.14)
were made under essentially identical temperoture
conditions. However, fresh salt ond a new system
were used for run ET-R, while the solt and system
for run ET-U hod been used successively through
ET-R, ET-S, end ET-T {(approximotely 250

The specimen from run ET-U shows a some-

runs
fe).
what finer void size than does the specimen from

run ET-R, but both specimens show intergranuiar
 

 

 

[ UNCLASSIFED

T-1142 !

. "

£aT

o

ki

- T

o

=

R e -

(B

g =
0 T
i PN
Fapsem T Loas,
. = L "] e
. ¥ W = 2|

Crpsie) K, . B
i I:"

! "1k ]

& - - M v J
L~ W - 5 213
- > i 1 g L L =, >
¥ T i & i e r Al

 

Fig. 1.7.7, Run ET-M. Inconel Tubing Afrer Exposure
to Flowing NaF-Zer-UFA (50-46-4 Mole %, Fuef 30}
at 1580°F (fsothermal) for 100 hr. Etchant: modified
oque ragia. 250X, Reduced 32%. (Secret with caption)

UHCLASSIFIED
2

 

T-1
o’
i
4 o ‘—-\_
. ks i )
- i."_'-
s T
4
:m - L2
i
# |
- } 1—".
-3 ; il
e ;G |:—-.
Al i ™ : & ! o e =
8 = bl ;-! h B
: LY
Fig. 1.7.8. Run ET-0, Specimea from Test Section

of lnconel Tubing After Thermal Cycling While Exposed
to Flowing NnF-ZrF4-UFd (50.46-4 Mole %, Fuel 30).
Cycle temperatures:
Dutation of test 59 hr.  Etchant:
250%. Reduced 32%. (Secret

Cycle frequency: D.4 <ps.
1600 * 20°F.
modified aque rtegia,

with coption)

 

PERIQOD ENDING MARCH 31, 1957

   

 

 

UHCLASSIFIED
T-11735 g
- - al el

» T - 2. W e
e oy p | f T gy gt a3

' T ' ! P
1' -.- e [ i _'_". " - St T""-_.. a4 '_:. : ?.' i - g
N A e AR 6"
& » T i l'.ou a _E_
2 =3 =%g A
& - i . . goz
P i e

e e | =5 | Lo

- l. - k] - = |

p o e :
T g . \ s o Lags

- = =z
" -3 o
i ) o
‘.. ' - E
o i

 

Fig. 1.7.9. Run ET-P, Inconel Tubing After Ex-
posure to Flowing NnF-sz4-UF4 (50464 Mole %,
Euel 30) at 1600°F {lsothermol) for 59 hr. Etchant:
250X, Reduced 32%.

modified oquo regia. {(Secret

with cupiion]

UHCLASSIEIED
~ T-1ins

EEE]

1
mc:lEs

 

 

EEEERERRE

“ % o

Fig. 1.7.10. Run ET-Q,
of Inconel Tubing After Thermol Cycling While Exposed
te Flowing NoF-Zer-UFJi (50-444 Mole %, Fuel 30).
Cycle temperotures:

Speclmen from Test Section

Cycle frequency: 0.01 cps,
1660°F during holf of ecycle and 1600°F during other
Duration of testt 59 hr. Etchont:
250%. Reduced 32%.

hall of eveie.
modified aqua regioa. {Secrat

with caption)

85
ANP PROJECT PROGRESS REFPORT

UNCLASSIFIED
T-11742

FF

T
THCHES |3
' L=

J

i
=
=

E i |

 

 

Fig. 1.7.11. Run ET-R. Specimen from Test Section
of Incenel Tubing After Thermal Cycling While Exposed
te Flowing NDF-Zer-UF‘i {50-446-4 Mole %, Fuel 30)
Cycle freguency: 0.4 eps, Cycle temperotures!
1615 + 30°F. Duration of test 100 hr. Etchont:
modified ogquo regie. 250X. Reduced 32%. (Secret
with coptian}

UMCLASSIFIED

EFE]

T
IN‘I:J:E‘S

—
1

 

Bk G

ko
=

e

 

= |2
w

Fig. 1.7.12, Run ET-§, Specimen from Heater Section
of Incone| Tubing After Thermai Cycling While Exposed
te Flowing NuF-ZrF‘-UFd (50-46-4 Mole %, Fuel 30).
Cycle freguency: 0.4 cps. Cycle temperatures:
1600 * 22°F. Duration of test: 50 hr. Etchant:
modified aguo regia. 250X, Reduced 22%. (Secret with

caption)

UNCLASSIFIED
T-11758

EEE]

 

 

w1 1

L
4 = =

o

: z
G - Y |
- - - o f‘ = ¢ ’. !- Laar,

- s g — | . Y,
. ot
; ekt W

i = L ! E.: " e o= | Lamn

= ffifi " ‘-fi ?‘-d:.':.u Vi :

& - —_Ey c - | e
% - Lo

SR i & ;i ar

Ty 5] Leaz

gt - " 'fl"" ‘-f'“ Sa
i i ETR

 

Fig. 1.7.13, Run ET-T. S5pecimen from Heater Section
of Inconel Tubing After Thermel Cycling While Exposed
to Flowing NuF-ZrFA-UF_,i {50-46-4 Mole %, Fuel 30).
Cyele freguency: 0.4 cps. Cycle temperaturas!
1600 t 25°F. Durotion of test: 97 hr. Etchant:
modified aquo regia. 250X. Reduced 32%. (Secret with

caption)

UNCLASSIFIED
T-11

Bi T

Bk waeB R

k

 

i

 

 

 

Fig. 1,7.14. Run ET-U. Specimen from Test Section
of Inconel Tubing After Thermal Cycling While Exposed
to Flowlng NuF-ZrF‘-UFd (50-46-4 Mole %, Fuel 20).
Cycle frequency: 0.4 cps. Cycle temperotures:
1625 £ 20°F, Durotion of test: 100 hr. Eichant:
modifiad aque regia, 250X. Reduced 32%. (Secret with

captien)
stringers and general comosion to opproximately
the same depths. Since the specimens from runs
ET-R ond ET-U ore reasonable duplicates, it
appears on the basis of these two runs that the
thermel cycling results are independent (within
limits) of the prior use of the salt mixture ond the
system,

PHYSICAL PROPERTIES

S. 1. Cehen N. . Greene
W. R. Gambil] ¥. D. Powers

Enthalpy and Heat Cupacity

The enthalpy and heot copacity for the mixture
NaoF-Z:F -UF (50-25-25 mole %) were redeter-
mined with the use of the copper-block calorimeter.
In the previous measurements made with the use
of Bunsen ice calorimeters, the heat capacity of
this mixture wos found to be 10.90 cal/g-atom,
which wos considerobly higher than the averoge of
8.11 cal/g-atom found for all the other zirconium
fluoride ~containing salts studied.® It wos felt
that this difference resulted in part from the high
melting temperature of the salt mixiure ond the
subsquent small iemperature range over which
the liquid was tested. The more precise measure-
ment capable in the copper-block colorimeter
resulted in a new value of 9.16 col/g-atom for the
NaF-ZrF -UF {50-25-25 mole %) mixture. The
results ore listed below:

NaF-Zer-UF4 (50-25-25 mole %)
Selid {120 to 458°C)

Hp - H

T 300{: —620 + 0115]T

0.151

0

c
P

 

Sw. D. Powers and G. C. Blaleck, Enthalpies and
Heat Capaciites of Solid and Molten Fluoride Mixtures,

ORNL-1956 {Jan. 11, 19356).

PERIOD ENDING MARCH 31, 1957

Liquid (648 to 914°C)

Hp — Hygop = ~40.5 + 0.3237 ~

- (6.87 x 10~%)1?
0.323 - (13.74 x 10-%)T
0,227 at 700°C

Heot of fusion {610°PC}
H. {lig) - Ho {solid) = 45

In these expressions

2]
1

Il

H = enthalpy ip cal/g,
¢, = heot capocity in cal/g-°C,
T = temperature in °C.

Measurements are being made for obtaining the
heat of fusion of the compound UF4. To date only
the enthalpy and the heat copocity in the solid
state have been determined. The results ore given
below:

UF, (solid)
HT - Hau"—’c = —2.25 + 0.085T +
+ (1.52 x 10-5)12

C

, = 0.085 + (3.04 x 10-3)T

After several meosurements of the liquid state
(melting point of UF, = 1036°C) were made, the
capsules had weckened and one broke. A stronger
capsule was therefore designed and is now being
fobricated. The enthalpy data from both the solid
ond the liquid stote measurements will be used to
establish the heat of fusion of UF .

Viscosity

Viscosity measurements were made for three
mixtures in the NoF-RbF-ZeF -UF, system. The
results, presented in Table 1.7.4, can be described

Toble 1.7.4. Summary of Yiscosity Meosurements of Mixtures in the NnF-RbF-ZrFJ{-UF‘ System

 

 

 

 

 

Mixture Compesition {meole %) Liquidus Viscosity (centipoises)
Temperature o
NaF Ry F ZrF, UF, (°C) Ay 600°C At 700°C At 800°C
5.0 54,5 34.5 6.0 545 8.0 4.8 3.3
44.5 15.0 34.5 &.0 540 7.5 4.75 3.5
2.5 5.2 3.5

37.5 0.5 47.0 6.0 540

 

87
by the several schemes for viscosity correlation
previously reported.”

Thermal Conductivity

The wvariable-gap device for the measurement of
thermal conductivity was redesigned to eliminate
the effects of varying the heating rate. Erratic
results were obtained from measurements with two
samples of the mixture NaF-ZrF ,-UF, (56-39-5
mole %) as the rote of heat flow was voried. The
first modification of the system involved provisions
for the measurement of the heat flow above ond
below the sample, and the modified system is
currently being tested with liquids of known thermal
conductivity., The design of o second device that
incorporates o mejor modificotion of the varichle-
gap device wos completed. The new design in-
corporates three sets of individually controlloble

 

PSummorized in 5. |. Cohen and T. N. Jones, Visocity
Measurements on Molten Fluoride Mixtures, ORNL-2278
{to ba issued).

heaters, both above and below the test sample,
in addition to o number of external guard heaters.
Provisions were also mode for many times the
present number of thermocouples. With the many
more heating circvits ond thermocouples, it should
be possible to ensure that one-dimensional heat
flow will exist in the salt somple regardless of
the rate of heat input.

Electrical Conductivity

An Inconel conductivity cell for use in determining
the specific efectrical conductivity of liquid sodium
and sodium-potassium alloy was designed and
fabricated. The purpose of this investigation is
to obtain reliable, absolute values of the con-
ductivities of these metals in the temperature
ronge of 200 to 1400PF for use in calibrating the
ART electromagnetic fHlowmeters. The measure-
ments of sodium-potassium alloys will cover those
compositions of interest to the ANP program.
Part 2

CHEMISTRY

W. R. Grimes

 
2.1. PHASE EQUILIBRIUM STUDIES'

C. J. Barton

R. E. Moore

R. E. Thoma

H. Insley, Consultant

MODIFICATICN OF APPARATUS FOR
THERMAL ANALYSIS

H, A, Friedman

Thermal analysis has been the most important
exploratory technique used in phase-equilibrium
studies at ORNL., For such analyses relatively
large charges of melt {50 to 200 g) are maintained
under a flowing cotmosphere of helium and ore
stired to assure homogeneity and to minimize
temperature gradients in the liquid. Since several
thermal excursions through the solidus ond liquidus
are used in each experiment, mixtures containing
high <concentrations of a volatile constituent
chonge appreciobly in composition during the ex-
periment.

In order to minimize the losses by vaporization,
a simple modificofion was mode to the apparatus.
The thermal anolysis crucible is fitted with a disk
of grophite that is drilled to gccommodate the
stirrer shaft omd the thermocouple well. The
diameter of the graphite disk is 0.1 in, less then
the inside diameter of the crucible. When the
charge of salt is melted, the graphite disk floats
on the molten mixture, whose free surface is thus
reduced nearly to zero. By using this technique
it has been possible to subject samples contaoining
80 mole % ZrF , to thermal analysis with negligible
loss in weight of the sample.

YISUAL OBSERYATION AND DIFFERENTIAL
THERMAL AMALYSIS OF SMALL
QUANTITIES OF FUSED SALTS

R. J. Sheil

The Leitz vacuum hot-stage microscope, which
it was hoped could be vsed for visuval observation
of small gquantities of fused salts in an inert
atmosphere ot high temperatures, was found to be
unsatisfactory for this purpose. The design of
the equipment permitted lorge thermal gradients

 

]The petragraphic exominctions reported here were
performed by H. Insley and T. N. Mc¥ay, Consultonts,
G. D. White, Metallurgy Division, end B, S, Landou,
Chemistry Division, The x-rey exominotions were
performed by R. E. Thomo, Chemistry Division.

and made measurements of melt temperotures ex-
tremely unrelioble, Since equipment that permits
thermal dato to be obtained with less than 1 g
quantities of fused salts would be highly desirable
for work with radicactive or toxic materials, the
experience obtained with the commercial equipment
was utilized in designing a new vacuum fumace
that is adaptable to microscopic observation. This
appdratus centains two concentric tubes of Alundum
or other syitable insulating material. The inner
tube is wound with a spiral of platinum-rhodium
wire to provide resistance heating. A cylindricol
nickel block 1 in. in diometer and 1 in, in
thickness haos ¢ ]x_i-in.-dia tube drilled in it to
provide ¢ sample compartment which can be lined
with thin metal foil to facilitote removal of fused-
salt samples. The furnace is enclesed in «
vacuum-fight container equipped with a glass
window through which the samples con be observed
by means of o low-power microscope. Melt temper-
atvres are measured with a thermocouple junction
immersed directly in the meft,

This apparotus wos tested initially for visual
observation of liquidus and solidus temperatures
by using approximately 0,15 g of the NaF-KF-LiF
eutectic as the test material. When the sample
was heated at a rate of obout 17°C/min, volues
ranging from 444 to 457°C were obtoined for the
first appearance of liquid and from 464 te 475°C
for complete melting. More consistent results were
obtained upon cooling at approximotely the same
rate. Values ronging from 435 to 460°C for the
first appearance of ¢rystals and from 451 to 460°C
for the solidus temperature were observed. A
value of 459.5 £ 1.5°C for the freezing point of
this mixture was obtaired by a different method
by another ORNL investigator.? The use of sfower
heating and cooling rates would uadoubtedly
improve the reproducibility of the data, but these
preliminary visual observations indicated that
reasonably accurate melt temperatures <¢an be
meosured with oaly 150 mg of material.

Liquidus determinations by visuval observation
are subject to some uncertainty when the melts

 

2'M. Blander, unpublished dota.

91
ANP PROJECT PROGRESS REPORT

ore strongly colored, and therefore it seemed
desiroble 1o adopt this apparctus to the use of
the differential thermal anglysis technique. Only
simple modifications to the apparatus were re.
guired. Two holes were drilled in the nickel
block, symmetrically arranged, and two opposed
thermocouple junctions, one for the sample and
the other for the reference moterial (alumina),
were iniroduced. The instrumentation is the same
as that previously used for similar determinations.
A Leeds & Northrup d-¢ amplifier is used to
amplify the difference in emf between the opposed
junciions before it is fed into an X-Y recorder.
The temperature of the sample is recorded on the
Y.axis of the recorder. Sharp breaks in the differ-
ential troce were obtained with the NaF-KF-LiF
eufectic in both the heoting ond the cooling cycles.
The heating breaks in three experiments varied
from 457 to 459°C, and the cooling breaks occurred
at 461 and 459°C,

The principal difficulty observed with this oppo-
ratus has been foilure to keep NaF-ZrF, mixtures
in the somple compartment regardless of the
composition of the liner material. Similar behavior
was observed with an N::II:-LH:4 mixture, in which
case the liquid was seen to be leaving the com-
partment by way of the thermocouple wires. It is
expected thot wider spacing of the thermocouple
wires and the use of taller containers will help
to minimize and possibly eliminate the creep
probfem.

THE SYSTEM LiF-UF4
H, A. Friedman R. E. Moore

Recent quenching work with mixtures in the
LiF-UF, systems that contain 20 10 32 mole % UF
has confirmed the liquidus and solidus relation-
ships previously reported.? It was found, however,
that the compound 4LiF.UF, undergoes very slow
subsolidus decomposition to LiF and 7LiF.6UF,
at about 470°C., A series of samples of o mixture
containing 28 mole % UF,, which originally con-
tained no free LiF, was heated overnight and
quenched from temperatures both above and below
470°C. Just below 470°C, LiF and ?LiF--fiUFd
were found, but, at still lower temperatures,
4LiF.UF, was found in addition to the other two

 

R, E. Moore, ANP Quar. Prog. Rep. Dec. 10, 1953,
ORNL-2012, p 77.

92

phases. Complete decomposition was apparently
not achieved at the lower temperatures.

Thermal analysis studies showed that liquids
containing 20 to 32 mole % UF, undercool readily
if the melt is first heated to over 800°C. Two
undercooling halts occur in the cooling curves ot
cbout 460 to 480°C, and the resulting mixtures
contain 4Lif.UF,, LiF, and 7LiF.6UF,. If the
liquid is first heoted eonly to 600°C, the under-
cooling effect is absent, and a eutectic type of
halt is obtained at ohout 490°C. The soclidified
mixtures from this type of experiment contain only
LiF ond 7LiF-6UF,, which are the phases to be
expected if 4LiF.UF, decomposes during the
cooling; in this event the cooling curve should
indicate o thermal effect at 470°C, as well as
the eutectic halt at obout 490°C. Becouse of
the absence of the 470°C effect and the sluggish
nature of the decomposition of 4Lif.UF,, it did
not seem 1o be likely that the absence of 4LiF.UF,
could be explained by its decomposition., These
mixtures therefore had apparently cooled to o
metastable eutectic between LiF and 7LiF.SUF ,
which must freeze only o few degrees below the
freezing point ot 490°C of the stable eutectic
between 4LiF.UF  and 7L; F.6UF,.

In view of the existence of the peculiar thermal
effects in the range 20 to 372 mcle % UFA, a
seeding experiment on a mixture containing 24
mole % UF, was conducted to be sure that the
quenching results previously obtained truly charac-
terize equilibrium conditions. The mixture was
heated to over 800°C and seeded upon cooling
with small omounts of a mixture of the same
composition that contained oll the phases found
in that region [LiF, flLiF-UF4, ?LiF-éUF4, and
VSLIF-UF," (ret 4)]. The seed material was
odded at intervals of about 10°C ofter the mixture
had cooled to about 560°C. The cooling curve
hod halts ot 485 and 480°C, with nc evidence of
undercooling.  Petrographic examination of the
mixture after it hod cooled to room temperature
showed that it consisted of 4LiF.UF,, 7LiF.6UF,,
and a small amount of LiF. These results are in
accordance with the results of gquenching studies,
except that the temperatures of the thermal effects
are about 10°C lower.

 

‘B. A. Scderberg, R. E. Moore, ond H. Dovis, ANP
Quar. Prog. Rep. Sepr. 10, 1956, ORNL-2157, p 87.
THE SYSTEM NuF-HfF_i

B. 5, Londauy H, A. Friedman
R, E. Thoma

The Hf** cation radius {0.86 3\) is so nearly
the same as that of Zr** (0.87 E} that the systems
NaF-HfF , and NoF-ZrF, are probably quite similar.
A preliminary study of the simiiarities was in-
itiated recently with thermal anolysis ond thermal-
gradient quenching experiments with compositions
in the NoF-HfF, system. In these experiments
three NaF-HfF, compounds, analegs of NaF-ZrF,
compounds, were discovered, The melting charac-
teristics of the analogous compounds are given
in Table 2.1,1, ond the optical and x-ray dote
are listed in the section of this chapter on
“Optical Properties and X-Ray Diffraction Patterns
for Compounds in Fluoride Systems.' The x-ray
Table 2,1.1. Meijting Charocteristics of Analogous
Compounds in the System NuF-ZrF‘ and NaF-HfF

 

 

4
Malting
Compound Point Type of
Formulo (°C) Compound
INaFZr F4 850 Cangruent
3NaFHIF, 850 Congruent
NaFZrF, 635 Incongruent; primary phase,
3No F'-ZrF‘1
2NaF - HiF 555 Incongruent; primary phase,
3NeF-HIF,
?NuF-éZer 520 Congruent

?NaF-éHde 540 Probobly congruent

 

diffraction patterns for the onalogous pairs
7NcF-62rF4-?NuF-6Hde and [, 2NaF.Z¢F,-
2NaF HfF, ore so nearly identical os to be
interchangeable for ?NGF-'fiZrFJ-?NGF-éHde and
almost interchangeoble for B, 2NaF.ZrF,-
MNoF.HEE,. Although 2NaF.Z;F, displays five
polymorphic forms, the 2NaF.HfF, observed in
the thermal-gradient quenching experiments oc-
curred only in one fom, the analog of B,
NaF.ZrF,, which is the lowest tempercture poly-
morph of 2NaF.ZrF,,  The x-ray patterns of
3NoF-ZrF, and 3NaF-HfF, strongly resemble each

PERIOD ENDING MARCH 31, 1957

other, ond, although the spocings of the principol
planes are
isomorphous, os s

are not
optical

the compounds
shown by their

identical,

properties,

THE SYSTEM LiFwCEFj
K. J, Sheil L.. M, Bratcher

The high-temperature furnace previcusly used in
the study of the RbF-CaF, system® woas used for
visual observations of liquidus temperatures of
mixtures in the LiF-CeF, system. A eutectic
containing 19 mole % CeF, and melting ot
755 + 5°C was previously reported® to exist in
this system. The eoarlier efforts to complete the
diogram were unsuccessful becouse no reliable
liquidus values could be obtained with mixtures
containing more than 19 mole % CeFa,

It appears that the thermel anslysis apparatus
in use at that time was not sufficiently sensitive
to detect the small heat effect associaoted with
the separation of CeF, frem the melt, Thig
difficulty was probably associated with heat con-
duction by the thermocauple well used to protect
the thermocouples, With ¢ Pt—Pt + 10% Rh thermo-
couple immersed directly in the melt, it was
possible to detect thermally the crystallization
of CeF, at the some time the crystals were
observed visually to be forming in the melt. The
diagram shown in Fig. 2.1.1 is o composite of
the thermal analysis data obtcined earlier {solid
circles}) and the recent data {cpen circles). The
melting point of CeF, used in Fig. 2.1.1 (1460°C)
is the literature value.” Examination of several
slowly cooled melts showed only LiF and Cef,
to be present, and this confirmed the thermal
indications of the absence of complex compounds
in this system.

THE SYSTEM KF-BGFZ
R, E. Meadows R. J. Sheil

The compound KF-BeF, crystallizes too rapidly
from the melts in the system 'KF-BeF, to permit
determination of its melting relatienships by the

 

L. M. Brajcher ond R. J. Sheil, ANP Quar. Prog. Rep.
Dee. 31, 1956, ORNL-2221, p» 109.

SL. M. Bratcher, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 89.

“H. von Wartenberg, 2. anorg. allgem. Chem. 244, 37
{ 1940).

@3
ANP PROJECT PROGRESS REFPORT

UNCL ASSIFITD
GRNL-LR-DWG 13365

 

TEMPERATURE (*C)

g00 ‘ . =

 

LiF o 20 30 A

Fig. 2.]-1. THE

use of the quenching method.® When this material
was subjected to thermol analysis in an apparatus
in which its behovior could be observed visuvally,
the data, presented in Table 2.1.2, indicoted an
incongruent melting point of 390°C and ¢ peritectic
compos ition with about 52.5 mole % BeF,.

X-ray examinations of equilibrated and quenched
materials have shown the inversion temperature of
a ond 3 KF.2BeF, to be 334°C., The high-
temperature modification, & KF-?Ber, is the one
found in slowly cooled preparations, presumably
because the inversion is too slow to permit for-
mation of the 3 form of normal cooling rates.

 

e M. Breteher end R, E. Meadows, ANP Quar. Prog.
Rep. Deg. 31, 1956, ORNL-2221, p 115,

94

 

(Fr PR —_

 

_#__L_W.

 

 

 

 

540 &0 70
Ce F} {mofe %)

System LiF-Ce F3-

8o

20 CeFy

Table 2,1.2. Visval-Observation Dara for
KF-Ber Mixtures

 

 

Obsarved Ohserved
Composition Liquidus Solidus
{mole % B&Fz) (OC) (OC)
dé 557 (385)*
48 E42 {385y
50 447 390
59 413 3590
54 388 330

 

*Yalves in parentheses ware

certainty than the others,

established with lass
Recent studies hove shown no evidence of the
previously reported® possibility of subsolidus de-
composition of KF-?BeF2 into KF.BeF, and BeF,.
The solidus temperoture between KF.Bef, and
KF-2Bef, is 330°C; the eutectic mixture conteins
59 mole % BeF,.

The compound 2KF-Bef, was found as crystals
with rounded corners and edges in semples con-
taining 33.3 to 47 mole % BeF, that were equili-
brated above 370°C and quenched. Although the
raw materials for these preparations were weil-
crystallized 2KF-BeF, and KF.BeF,, the rounded
grains were obtained regordless of whether the
equilibrium temperature was approached from above
or below.

Rounded crystels are often considered evidence
for the existence of immiscible iiquid phases.
This possibility can be eliminated becayse the
rounded

single-phase

crystais  consisted of

PERIOD ENDING MARCH 31, 1957

2KF.BeF,. Mo explanation for the rounded grains
can be offered at present.

Thermal analysis showed an inversion of
2KF-BeF, ot 690°C, The inversion was so ropid,
however, that only one modification (the B or
low-temperature  form) found in samples
quenched after equilibration at temperatures above
and below §90°C.

was

The solidus temperature between 3KF.Bef, and
2KF.BeF, (730°C), which was previously de-
termined by thermal anclysis, was confirmed by
quenches of compositions with 27 and 30 mole %
BeF,. The liguidus and solidys temperotures for
the mixture containing 27 mole % BeF, are nearly
the same, and thus it appears that this camposition
is near the eutectic composition. The phase-
equilibrium diagram for this system is presented
os Fig. 2.1.2.

HNCEASSIFIED
ORMNL-LE-DWG 19366

 

900

800

 

 

700

 

600

2G0

TEMPERATURE (°C)

400

a-KF 2BeF, |

 

i ——

B - 2KF-Bef,

L

e

JKF - BeF;l

 

 

 

 

200 i

 

KF - Befs

 

 

 

 

 

B-KF-2Bef,

 

 

 

 

l L L

 

L
L]

KF 10 20 40

50 &0 7O B8O <0 Bef,

BeF, tmote %)

Fig. 2.1.2. The System KF.BeF,.

3
ANP PROJECT PROGRESS REPORT

THE SYSTEM NoF -RbF-BeF
R. E. Meadows

Quenching studies of the NaF-RbF-Bef, system
were started recently,
pound previously reported in the low-BeF, part
of the system was tentatively identified® as
2NaF-RbF.Bef, on the bosis of petrographic and
x-ray studies of slowly cooled melts. Exami-
nations of quenched compositions on the join
NaF-RbF-BeF, iadicate that the compound probably
conteins closer to 45 than 50 mole % NaF, but
further study is needed to establish the identity
of this phose conclusively. Previous exominations
of slowly cooled compositions estoblished the join

The only termmary com-

NeF.RbF.BeF, as a binary system. Quenches of
mixtutes on the join containing 33.3 and 40 mele %
NaF have shown that the ternory compound is
the primary phose, that the solidus temperature
is 452°C, ond that RbF-BeF, is the other phase
below the solidus. The phase relationships be-
tween the liquidus and solidus are rot clear, as
yet, and more work is required in this region.

THE SYSTEM NGF-LIF-UF"
H. A, Friedman B. S. Landav

A lorge number of thermal-gradient quenching
experimenis completed during the quorter have
shown the tentative phase-equilibriuvm diagram
previcusly reported!® for the system NaF-LiF-UF,
to be in error in several respects, A revised
diagram is presented os Fig. 2.1.3. The un-
identified moterial reported previously to be a
ternary compound was found'! to be the compound
NaF.-2UF,, which has been shown to have o
primary phase field in the ternary system. This
compound thus became an apex of the compatibility
triangle NuF-?UF4-?LiF-éUF‘{-muF-{SUFd.

The <compound 3NoF.UF, is unstable below
497°C, ond it is accordingly impossible for the
primary phose field of LiF to contact that of
3NaF-UF,. The latter compound is not therefore
an opex of a compatibility triangle. The compound
4LiF-UF4 is unstable below 470°C, and the

 

LM Bratcher, ANFP Quar. Frog. Rep. Sept IO,
1956, ORNL-2157, p 94,

mH. A. Friedmon et al.,, ANP Quar. Prog. Rep. Dec.
10, 1955, ORNL-2012, p 78-81, esp. Fig. 4.3,

”R. E. Moore, ANP Quar. Prag Rep. March 10, 1956,
ORNL-2061, p 72.

96

primary phase field of this material does not
contact that of o sodium compound. Therefore
4LiF-UF, is not an apex of a compatibility
triangle.

The compound 5NaF.3UF, was shown to exist
ot the liquidus in the ferary system, and the
extent of its primary phase field was established.
Boundary paths separating the phase fields of
LiF-4UF, ond NaF.2UF,, LiF.-4UF, and 7LiF.UF,,
and 7NaF.6UF, and 7LiF.6UF, were also es-
tablished. The boundary paths between the
7NaF.6UF, ond 7LiF.6UF, phase fields were
difficult to establish because of the extensive
solid solution which occurs between these two
compounds.

NoF-ZrF,-STRUCTURAL METAL FLUCRIDE
SYSTE MS

H. A. Friedman 8. 5. Landau

Phase equilibrium studies of the systems NaF-
CrF,-ZiF, NaF-FeF,-ZrF,, ond NaF-NiFE-ZrF4
were made during the quarter in order to determine
the formulo and phase relationships of the ternary
compound in each of these systems. Only one
ternary compound was observed in each system,
The phases identified in several slowly cooled
melts [50-g quantities) indicote, as shown in
Table 2.1.3, that the NOF-GFE-ZrF4 ternary
compound probably has one of the formulas
NaF.CrF,-ZrF,, 3NoF.3CtF,-dZrF,, NoF.CrF,.
2ZrF,, or NaF.CrF,.3ZtF,. The compound is
isomorphous with its analeg in each of the three
systems, ond therefore it is likely to have the
some stoichiometric formula in the three cases.

Thermal analysis ond thermal-gradient quenching
experiments were made along the join NeF-CiF,-
ZrF,, which includes all the possible ternary
compound formulas mentioned above., Egquilibrium
conditions were not opproached closely enough for
a determination of the formulo by the use of phase-
equilibrivm experimentel techniques. The compo-
sition that produced the most nearly single-phase
material was NuF-CrFI-EZer. Single crystals
of the NGF-FEFE-Zer ternary compound were
obtained from a portion of a slowly cocled melt
containing 33]6 mole % of each component of the
mixture. Chemical analysis of the sample showed
that the composition of the compound is NaF-FeF,-
QZer. The isomorphism displayed by the analegs
of this compound ia the NaF-CrF,-ZrF, ond

4
PERIOD ENDING MARCH 31, 1957

o

ORNL-LR-DWG tE9%28A

ALL TEMPERATURES ARE IN °C
(COMPOSITIONS ARE N mole %)

HHHH IMDICATES S0LID SOLUTION

 

TNoF - 6 UF, &

w“ “"“\‘ -BUFq ‘
S5MoF - 3yF, ‘ 600 500 £ m
875 i ’

ENOFJ?U ‘ - <) </
3mur-‘u: :35

 

 

 

2 T — ] 480 4 Lif - UF

1o R o (oo <N

800

900
o
Nty

VNV VY v/ v v/ \/ \/ \
NoF 950 800 850 800 750 700 630 LiF
$80 850

Flg. 2.71.3. The System NuF-LiF-UFd.

NaF-NiF,-ZrF, systems indicates thot the for-
mulas for the ternary compounds in those systems
are almost surely NaF.CrF,.27¢F, and NaF.NiF,.
2ZrF,. A tentotive diogram for o portion of the
join NaF-Cri,-ZrF, is shown in Fig. 2.1.4.

The very large thermal range of the Crf,-ZrF,
primary phase field {see Fig. 2.1.4) suggests that

it may be the scturating phase for solutions of
CrF, in NaF-ZrF, (53-47 mole %), Further ex-
periments to outline the extent of the Cer-ZrF4
and the analogous FeF,-ZrF, and Nif,.ZrF,
primary phase fields will be initiated. Optical
and xsray data for the compounds NaF.CrF . 2Z¢F ,
NaF FeF,-2Z¢F,, and NaF-NiF,-2ZrF, are pre-

sented in the following section.
ANP PROJECT PROGRESS REPORT

P e

Teble 2.1.3. Phases Observed in Slowly Cooled Melts in the System NaF-CrF,-ZcF,

 

Composition {mole %)

 

MaF C':rF2

ZrF

Fhasez Obsarvad™

 

4
50 15 35 2NaF+ZrF ,, ternary compound
46 15 39 2NuF-ZrF4, ternary compound
43 15 42 2NoF+ZsF ,, ternary compound, Cr Fz'z'F.s
45 10 45 WMNoFZrF ,, TNaF-6ZrF,, ternary compound, CrF -ZrF,
48 3 47 MNaF-ZrF ,, TNoF+6ZcF ,, ternary compound, CrF2~ZrF‘
51 5 44 ?NoFtéZrF4, ternory compound

 

TEMPERATURE (*C)

98

*ldentification made by petrographic ond x-ray diffroction exomination.

 

UNCLASSIFIED
ORNL-LR-DWG 19367

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000 | | T i (iR I 1 |
|
. | |
| |
S e sl =
\ |
|
N
BOQ —— 11 = S
% | |
N ]
| & '
700 e p—— = — .
| | | 1
Mo - CefFp + LIGUID I
l‘I—NUF'CfFE EZFF4 |
| | + LIGuUrD
GO0 Em— : 3 y — = - == — — e
1 a—NaF-CrFy- 22rFy + ZrFy
MoF - CrFy + a-MaF - CrFy - 22rF), ; |
. | | | . | ‘ | | |
[ I [ ! | | * ! |
200 | e = —t—— F T = —t — I
. [ | i
NaF-CrFp + B-NoF - Crfy- 2Z¢F, | B‘NOF-IC:F2'22rF4 + ZrF,
400 L. . | | — —— r |
MNaF - CrFp 10 20 30 40 S0 60 70 80 890 ZrF,

ZrF, (mole %)

Fig- 2.1.4. The Join NnF-Cer-ZrF4 (Tentative),
OCPTICAL PROPERTIES AND X-RAY
DIFFRACTION PATTERMNS FOR COMPQUNDS
IN FLUGRIDE SYSTEMS

H. Insiey T. N. MeVay
B. S. Landou R. E. Thoma

The identifying characteristics of some new com-
pounds encountered in pohase studies are listed
below. The symbol d{A) denotes the distance
between reflecting planes meosured in ongstroms,
The term J'/J’I refers to the relative intensity of
the lines as compared with an arbitrary value
of 100 for the strongest line.
properties, N, and N,

Under optical
refer to the lowest ond
highest indices of refroction, respectively, for
bioxial crystals, and 2V refers to the acute angle
between the optic axes in biaxial crystals; N |
and N, refer to the ordinary and extraordinory

indices of refraction of uniaxial crystols.

;'3 KF+25&F2

Optical Data

Biaxial; sign net determined

Mean refroctive indax, 1.319

X-Ray Data
d(g} LH.I
8.20 20
7.25 10
6.56 10
6.19 5
5.40 15
5.20 5
4.93 30
4,48 25
4.29 15
4.15 20
3.88 100
3.74 40
3.34 20
3.25 100
3,14 35
3.01 100
2.644 40
2,501 20
2.455 10
2.270 25
2,253 S
2.164 BO
2.074 49

PERICD ENDING MARCH 31, 1957

1.987 15
1.972 20
1.848 20

Opticol and x-ray dato were reported pmviouflyu
for I{F-ZBeF2 (printed incorrectly as KF-QBeFd). The
discovery of the low-temperoture form, fi KF°2BeFZ,

requires that the previously reported form be desig-
nated & KFQB&FE.
2KCI'FECI2

Opticol Dato

Uniaxial pesitive

N, = 1.600
N, = 1.636
X-Ray Data
AR 11,
6.02 1G0
5.69 15
4.06 25
3.47 75
3.10 15
2.86 30
2.675 100
2.615 50
2,590 100
2.518 25
2.489 15
2.467 50
2,203 15
2.108 15
2.022 15
1.994 25
1.884 30
1.749 25
KCL-FeCl,

Optical Dara

Biaxial positive

N, = 1.700
N, = 1.740
X-Ray Data
diR i1,
7.5 30
7.2 30

 

uR. E. Thome et al.,, ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 101.
ANP PROJECT PROGRESS REPORT

d(A) 11 INaF-HfF
1.17 10 Optical Data
3.43 36 Biaxial positive, 2V = 40 to 60 deg
3.1 25 Mean refractive index, slightly lass
2.89 20 than 1,384
2.83 100 Low birefringence, ™~ 0,006
2.75 30
2.695 35 X-Roy Data
2.675 100 4R 11,
2.495 15 8.01 30
2.276 3G 6.39 5
2.243 15 554 30
2.124 15 513 100
1.960 1 4.80 90
1.957 10 449 20
1.940 35 4.33 40
o o
1:351 10 382 %
) 3.55 5
1.832 25 3.24 90
1.78%9 15 .08 20
3.02 B0
2.88 10
2.822 5
3NGFHIF 2,735 40
2,704 5
Optical Data 2.667 20
Mean refractive index, slightly lass 2,564 20
than 1.386 2-529 10
Biaxial negative, 2V = 40 deg 2.405 25
2,363 10
X.Ray Dota 2,327 10
%) i, 7202 "
5.26 80 2.194 5
4.74 100 2,154 10
3.75 70 2.034 25
.06 100 2.008 10
2.93 40 1.97¢ 0
2,373 15 1.935 20
2.321 35 1.898 40
2.154 20 1.840 15
2.120 5 1.819 20
1.967 25 1.792 5
1.883 20 1.754 10
1.872 S0 1,742 25
1.770 10 1.5%9 25
1.685 5 1.567 15
1.604 25 1.535 5

100
ru

MNaF -Cer 2Zr F4

Optical Data

Biaxicl negeotive

Exhibits palysynthetic twinning

Mean refroctive index, ™~1.456

X-Roy Doto
o
d{A)
5.21
4.50
4.06
3.86
3.48
3.43
3.3
2.622
2.194
2.034
2.01%
1.951
1.770
1.73¢%

i/t

49
15
100
10
50
30
45
15
23
25
15
15
15
20

N-:IF-Fc;-.F2-2ZrF4

Cptical Date

Bioxial negative, 2V = 75 deg

Ny = 1.452, N,

Exhibits polysynthetic twinning

Compound is probably monoclinic

X-Ray Data
R

5.20
4.50
4.06
3.79
3.69
3.36
2.607
2.521
2.034
1,943
1.812
1757
1.733
1.642
1.511

= 1.459

I,/.!',I

40
10
1G0

100

PERIOD ENDING MARCH 31, 1257

NuF-Nin-ZZrF‘

Qptical Data
Bioxial positive
Exhibits polysynthetic twinning
Maan refractive index, ™1.460

X-Ray Data
(A i1
5.42 14
515 s
4.48 25
4.05 100
3.34 100
2.500 10
2,022 25
1.932 15
1.670 15

SUBCONTRACT WORK AT MCUMND
LABORATORY AND AT BATTELLE
MEMORIAL INSTITUTE

During the period covered by this report the
Mound Loboratory continued investigations of
beryllium-containing fluoride mixtures of interest
to the ANP program. This work included phose
equilibrium  studies of the following systems:
LiF-BeF,-UF,, NoF-BeF,.UF,, LiF-RbF-BeF,.
In addition density and viscosify values were
obtuined for various compositions of these mixtures
ot 600, 700, and 800°C. This work is reported in
detail in Mound Loboratory memoranda numbered
57-1-28 ond 57-2.24,

Studies of interest to ANP were clso continued
ot the Battelle Memorial Institute under subcontract
Ne., 982, Determinations of the vapor pressure
of three compositions of the RbF-ZrF, system by
the tronspiration methed were carried out, and
fobrication and assembly are under way on appo-
rotus for the stafic method of determiniag vapor
pressures. The tronsport of chromium from Inconel
exposed to the NaF-ZrF UF  sait system is being
studied by using Cr®'-lobeled CrF,. Apparatus
is being built for determining the obserption
spectra of fused mixtures of the fluorides of
sodium, uronivm, ond zircenium. Studies of the
stability of uranium-bearing liquid-metal slurries
are continuing, This work is reported in detail

in repott BMI-1171, issved February 20, 1957,

101
ANP PROJECT PROGRESS REPORT

2.2. CHEMICAL REACTIONS IN MOLTEN SALTS

F. F. Blankenship
R. F. Newton

EQUILIBRIUM REDUCTION OF Nle BY l"l2
IN NaF-ZrF4

C. M. Bloed
Investigation of the equilibrium

NiF, (2) + H, (g) &= Ni(s) + 2HF ()

in Naf-ZrF, (53-47 mole %) at 550°C was con-
cluded during the quarter, and the odditional data
obtoined are summarized in Table 2.2.1. The new
data were combined with the dote previously
obtained’ at 350°C, ond the values are plotted
in Fig. 2.2.1, A straight line drawn from the
origin with a slope of 0.76 x 10% atm, numerically
equcl to K _, moy be seen to agree quite well with
the experimental resulis. This line has been
extended to the value of the solubility of NiF,
in this solvent ot 530°C.  Accordingly, it is
observed that, within the precision of the ex-
periment, K _ is independent of the NiF, concen-
tration over the concentration range studied, which

 

1C. M. Blood ond G. M. Palmer, ANP Quar. Prog. Rep.
Dec. 31, 1956, ORNL-2221, p 120.

L. G. Overholser
G. M. Watson

includes values not for from the concentration of
Ni F:2 in the saturated solution of this temperature.
The most dilute solution studied contained 85 ppm
Ni as NiF,, and the most concentrated solution
had 313 ppm. The solubility of NiF, in this
solvent at 550°C was determined to be 352 £ 11
ppm, as described below.

UNCLASSFIED
ORML=LR=-DWG 20235

  

 

o0, — - et
[ | i I
8.0 -+ EXPERIMENTAL POINTS FOR EQUILIBRIUM AT 550°C..
20 (o SOLUBILITY = 352 % i{ppm Ni AS NiF, 1N
MNoF - ZrF, [53-47 mole %} |
20 ——g— g2 = —t
HE
iy N . |
qi 20 | ¥ '("“Fz W | -
N‘f 40— = =4 _'0'5 ‘\_’_QQ.
5 g
s e

 

C
i | LN } 1. — {
T | [ ]
0 10 20 30 aQ 50 &0 (%1077
*N.FQ("""'E froction )

 

Fig. 2.2.1. Equilibrium Reductiorn of NiF, by H
NeF-ZrF ;.

2in

Toble 2.2.1. Equilibrium Ratios for the Reaction NiF2 {4y + H2 {g) = Ni (5} + 2ZHF (g)
ot 550°C in Nc-F-ZrF4 (53-47 Molie %)

 

 

 

NI in Mely Pressurs of Hy Pressure of HF -

(pom) {atm) (otm) *

x 1ot
85 0.0530 0,239 0.74
@5 0.0517 0,257 0.79
120 0.0494 Q.287 0.81
120 0.0493 0.292 0.85
135 0.0491 0.295 0.77
220 0.0448 0.356 0.75
313 0.0413 0.406 0.75

Average value of 12 dererminations previously reportad 0,76 F (.04

 

Total average 0.76 & 0.04

 

*Kx e PI%IF’,XNIFZPHQ' whetre XN&FQ it mole froction of NiFQ found in solution and P is pressure in atmospheres,

102
SOLUBILITY OF NIF2 IN NoF-ZrF4
C. M. Bloed

Considerable attention has been given to
determining the solubility of Nif, in NaF-Z:F,
{53-47 mole %). |t was shown |:.~r45t".-'i~'.:pusI';.r2 that
the solubility of NiF, in this solvent wos o
function of the total amount of NiF, odded in the
range 0.5 to 10 wt %, but no determinations were
made with only traces of the soturating (nickel-
containing) phase present. The results obtained
electrochemically? were of considerable interest
in this regard, but unfortunately these results

2-4

appear to be high by cpproximately o factor of 2.
Consequently, on accurate determination of the
solubility ot 550°C was undertaken with only
troces of the saturating phusa present,

In order to approach the saturation concentration
gradually, o mixture of HF, hydrogen, and helivm
was bubbled through NaF-ZrF , contained in aickel
apparatus at 550°C. The apparatus was the same
as that used for equilibrium studies (see preceding
section on "'Equilibrium Reduction of NiF, by
H2 in NaF-ZrF, "), The HF-H, ratio of the
gaseous mixture was set at appropriate values so
that nickel metal from the cantainer was converted
slowly to NiF,. The changes in NiF, concen-
tration of the molten NaF-ZcF, were followed by
chemically analyzing liquid somples drown through
filtering tubes. The filtrates were token in pairs —
one at 550°C ond another from the apporatus
tempororily heated to 700°C. For the unsaturated
solution the concentrations of NiF, were identical,
within the analyticol precision, in both filtrates.
As the solution became saturated, the 550°C
filtrate indicated the solubility of Nif,, ond the
700°C filtrate gave a measure of the total amount
of NiF, in contact with the liquid. The results
of these meosurements are presented in Fig. 2.2.2.

The data obtained for o saturated solution gave
o solubility of 352 £ 11 ppm Ni as NiF, in moiten
NcF—Zer (53-47 male %) at 5350°C. The data olso
indicate thot the solubility remains constant when
the amount of soturating phose is chonged from o
trace to a total of 1210 ppin Ni*T. The volue of

 

2, D. Redman, ANP Quar. Prog. Rep. june 10, 1956,
ORNL-2106, p 101.

3. E. Topol, ANP Quar. Prog. Rep. june 10, 1956,
ORNL-2106, p 103.

4C. M. Blood, ANP Quar. Prog. Rep. June 10, 1956,
ORNL-2106, p 99.

PERICD ENDING MARCH 31, 1957

WHCLASSiFED
ODRNL=LA-DWG 20236

{400 T I i
| ; |

 

 

 

 

 

1200 — L = — i
| |
|

%1000 - — e — ]
:: | o SOLUBILITY =352 X4t ppm Mi
4 | &S NiF, IN NoF — 2rF, {53~ 47
2 male%) AT 550°C
§ 800 - L ‘ ]
b
g |
o
1.
5
& 600 - — _ | =
£ |
©
&
o
Z 400 — - — e

200 — b = -

|
0 | l
o 260 400 &00 800 1000

Wi lppm} 1M FILTRATE TAKEN 4T 530°C

Fig. 2.2.2. Solubility of NiF, in NaF-ZrF, at 550°C,

352 ppm is in substantial agreement with a velue
of 400 ppm previously reported? for the solubilily
obtained with o total addition of 1700 ppm Ni ™ .

STABILITY OF CHROMIUM FLUOQRIDES 1N
LiF-ThF, ARD LiF-BoF,-ThF, MIXTURES

J. D. Redman

The stability of CrF, and CrF3 in two THF"-
bearing melts at 800°C wes examined by application
of the equilibration-filtration technique described
in previous reports in this series.® (This work,
while of interest to the ANP project, is more
directly applicable to the Fused-Salt Power Re-
actor Project.) These experiments were performed
by charging pure LiF-ThF, (72-28 mole %) or
LiF-BeF,-ThF , (67-23-10 mole %) along with pure
CrF, or CrF 4 to hydrogen-tired nickel apparatus,
equilibrating for 5 hr at 800°C under helium,
filtering ot the equilibrium temperoture, and anp-
alyzing the cooled filtrate for Cr'' and total
chromium ion,

 

5J. D. Redmaon, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 120,

163
ANP PROJECT PROGRESS REPORT

The dota obtained, os given in Table 2.2.2,
indicate that Cer is stable in both these solvents.
The dispreportionation regoction

3CrF, == 2CF, + Cr°

seems to occur to a very limited extent, if at all.
It appears, however, that CiF, reacts with the
nickel contoiner, when added 1o either solvent,
presumably by the reaction

2CF, + Ni° == Nif, + 20/F,

The reaction is manifestly incomplete under the
conditions used. [t is likely that the heating
period was teoo short and that the surfoce areo
of the nickel container wos too small for the
heterogeneous equilibrium to be attained.

RECUCTION OF UF4 BY STRUCTURAL METALS
J. D. Redman
Reduction in Molten RbF-Zer Mixtures

Dota were reported previously®:? for experiments
in which Cr® and Fe® were equilibrated with UF
(~7 wt % U) in RbF-ZrF, (60-40 mole %) as the
solvent medium. Since the volues reported for
the reaction with Fe® were somewhst uncertain,
the experiments were repeated with longer equili-
bration periods being ollowed. The results of

 

). D. Redman, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 123.

7). D. Redman, ANP Quar. Prog. Rep. Sept. 10, 1956,
ORNL-2157, p 106.

these experiments, presented in Table 2.2.3, show
that equilibrium is estoblished within 5 hr and
that the equilibrium concentration of Fe'' is
significantly higher at 600 than at BO0°C. The
concentrations obtained in this reaction medium
fell in the range thot has been found in the other
alkali fluoride—ZrF, binory mixtures. No sotis-
factory explenation can be given for the fact thot
the Fe®-UF, system is not morkedly offected when
different alkali fluoride=ZrF, mixtures are used
os the reoction medium. In contrast the studies
of the Cr°-UF , reaction showed thot the chromium
concentration increases in the order of use of
RbF, KF, NaF, eor LiF, in the reaction medium,
which is the order of decreasing ion size.

Data for the reaction of UF, with Cr® at 600
and 800°C when RbF-ZrF4 (52-48 mole %) is used
as the recction medivm are given in Table 2.2.4.
Approximately 2 g of Cr® wos reacted with UF,
(8.9 wt % UF,, 4.0 mole % UF,) dissolved in
approximately 40 g of the RbF-ZrF, mixture ina
nickel container.

The chromiom values given in Table 2.2.4 show
that the UF -Cr® reaction has a very small tem-
perature coefficient over the temperature rtonge
studied. These equilibrium chromium concentrations
are slightly higher than those found® in RbF-ZrF
(60-40 mole %) ot both temperatures. The small
differences indicate that the fluoride ion activity,
as measured by the Cr®-UF  reaction, increases
only slightly when the RbF concentration is in-
creased from 50 to 58 mole %, the RbBF con-
centrations of the ternary mixtures containing

4 mole % UF,.

Toble 2.2.2, Stohility ot 800°C of Cer and CrF3 in Molten LiF-ThF“ ond LEF-Ber-ThFd

 

Conditions of Equilibration

 

 

Present in Filtrate

 

 

Salvent Additive et {wt %} Total Cr {wt %) Ni (ppm)
LiF-ThFd {72-28 mole %) CrF2 4.0 4.1 100
Cr F2 4.0 4.0 50
CrFB 0.5 3.9 4400
CrF3 0.9 3.9 5500
LiF-BeF,-ThF, (67-23-10 mole %) CrF? 3.6 4.0 10
'f:'lrvl"'2 4.0 4.1 10
CrFa 0.5 4.1 3200
CrF3 0.8 3.9 4100

 

104

R
a}

Table 2.2.3. Data for the Reaction of UF4 with Fe°
in Molten RbF-ZrF‘ {60-40 Mofe %) ot 600 and 800°C

 

Canditions of Prasent in Filtrate

 

 

 

__Equil ibrotion Total Totel Total
Temperature Time B Fe* M
°C) (hr) (wt %) (ppm)  (ppm)
&00 5 7.2 500
5 0 510
16 7] 460
16 7.3 490
800 5 7.4 170 50
5 7.4 220 &0
5 -2 120
5 7.1 130
16 7.5 180 100
14 el 140
14 i 230 80
15 7.3 170 110

 

*Blank of 70 ppm of Fe at 800°C,

Table 2.2.4. Duta for the Reaction of UF, with Cr®
in Molfen RbF-ZrF‘ (52-48 Mole %) at 660 and 800°C

 

Conditions of

Eguilibration

Present in Filtrote

 

 

Total Total Total

Temperature Time u Cr* Mi
(C} (hr) (wt %) (ppm)  (ppm)
600 3 4.8 &80 40

3 4.8 &60 20

5 6.8 710 <1

5 6.9 720 <

300 5.8 740 <1

3 6.8 780 20

5 6.8 760 120

*Blank of 170 ppm of Cr ot 800°C.

Reduction in ThF -Bearing Melts

Data were reported previouslyfl for the reaction
of Cr® with UF, with low uranium cmcentrations
(0.5 wt % U) in molten LiF-BeF, (48-52 mole %)
as the reaction medium. These experiments hove
heen extended to include dilute solutions of UF4
in LiF-ThF, (72-28 mole %) ond LiF-BeF,-ThF,

(67-23-10 mole %} mixtures os solvents,

FERIGD ENDING MARCH 31, 1957

Data for the reaction of UF , with Cr® a1 650 and
800°C in LiF-BeF,-ThF, (67-23-10 mole %) are
given in Toble 2.2.5. Approximately 2 g of Cr°
was reacted with UF4 {0.12 mole % UFJ, (.5
wt % U) dissclved in approximately 40 g of the
LiF-BeF ,-ThF , mixture confained innickel. Similor
dato obtained with LiF-ThF4 (72-28 mole %) os
the reaction medivm and ¢ UF, concentration of
0.21 mole % {0.5 wt % U) are presented in Table
2.2.6.

The equilibrium chromium values presented in
Tables 2.2.5 end 2.2.6 approximate the blank ob-
tained for chromium. Therefore, the reduction of
UF, by Ct® in these reaction media when low UF,
concentrations are used is negligible, as wos
the cose when LiF-BeF, (48-52 mole %} wos
used as the reaction medium.” While the method
is not sufficiently sensitive to determine slight
chonges in chromium concentrations thot occurred
in this series of experiments, the data do indicate
that no extensive attack of Cr® by UF, occurs
vader these conditions.

MEASUREMENTS QF SOLUBILITY OF RARE-
EARTH FLUORIDES IN MOLTEN NoFf-ZrF,-UF,
MIXTURES

W. T. Word

Raore-earth fluorides labeled with Ce
153

141 140
; Lag ™

or Sm and the techniques previcusly described®
were used fo study the solubilities of CeF, ond
L-:F:3 alone and in combination ond of SrnF3 alone

 

BW, T. Word, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 125.

Taeble 2.2.5. Deoto far the Reoetion of UF4 with Cr°
in Molten LiF-Ber-ThF‘ (67-23-10 Mole %)
at 850 and 800°C

 

 

Conditions of Present in Filtrate

 

 

Equilibration Total Total Tetal
Temparoture Time u Cr* Ni
(°C) (hr) (wt %) {ppm) (ppm}
650 5 0.38 210 80
5 0.37 240 20
800 s 0.44 270 g0
5 0,48 270 80

 

*Blank of 260 ppm of Cr at 800°C,

105
ANP PROIJECT PRQGRESY REFORT

Tuhle 2.2.6. Doata for the Reoction of 'UF'4 with Cr°
in Malten LiF-Th F‘i (72-28 Mole %) or 650 ond angic

 

Prosant in Fillrate

Conditions of

Equilibration

 

_____________________ Toral Total Taotal
Temperoture Tirne U Cr* i
(%) (h) {wt %} {ppm)  (ppm)
450 5 g.51 280 40
5 0.49 240 S0
800 5 0.1 260 54
B 0,52 280 G

Tekle 2.2.%. Solubility of 'f.':rr'F3 in M-:::|"'-Za‘1""&~l.l'l=|i
(50-46-4 Mole %:} Determined with the Use of

Ce“” us 3 Trager

 

Cerium in Filtrote {wr %)

 

Temperature Experiment Mumber

 

 

Q) - Avegrage
W 2 23 24

OO+ 5.3 534 537 5,29

700 360 375 230 3.43 3.70

500 2.80 2,83 285 .47 279

559 2,47 253 2.5 233 2.46

 

“Blonk of 280 apm of Cr at 330°C.

ond in combination with CeF, in NoF-IrF ,-UF,
{50-46-4 rmole %). In oddition, the effect of E’ron
and 5rF, os the solubility of CeF, in this melt
was excmingd.  Solubility voluves shown ot precise
temperatures in the following tables were gbtained
from smooth cuorves of the experimental dota taken
at approximately the temperatures shown,

Sclubility of CeF,
The solubility of CeF, in NoF-ZrF -UF  (50-46-

4 role %) was redetermined in three seporate
experiments with the use of Ce' as o tracer
by the technique described previously.® The
sotubilities  thus  obtained, together with those

previously reported, are given in Table 2.2.7.

Solubility of Laf,
The solubility of LaF 4 in MoF-Z:F -UF, {50-46-

4 mole %) wos redetermined with La'?% as « frracer,
angd the tempercture ronge of the determinations
was extended. The solubility values thus measured
ore helisved 1o be the most occurgte obtoined 1o
data; they are given in Table 2.2.8.

Solubilities of CeF, and LaF,

in Combination

An experiment was completed 1o determing the
solubilities of mixed CeF, and LaF; in NoF-
Zrf -UF, (50-46-4 male %) with the vse of C-:-:F:,_
confaining Ceté!

tracer and LaF. coniaining
140 3
La tracer,

The filtrates and standurds were
counted several times over a period of o week,
and the ditferences in decay rotes (Ce'*!. 32.5.day
holf life; La'?, 40-hy helf lifel were used to
caleculate the cerivm and lanthanum concentrations.

126

 

“Freviously reported.

*4in this and liowing tobles in which the precise
temperatures of B00, 700, 600, ond 550 are listed, the
solobilitiss were recd frem the curve drown threogh the
asztual values obtoined,

Toble 2.2,8. Solubility of LnFq in NaF-Ier-UF&
{50464 Mo\e %) Determined with the Use of

 

 

Lo 4% os & Traces
Teamparature Lanthonum in Filtrate
(°C) (wy %)
By 7.59
315 4.81
7h4 4,19
77 3.51
b4 2.0
415 2.68
582 2.5
539 2.22

 

As o check on the accurocy of these caloulaticas,
a set of five control somples wos made up that
contoined different weighed ameunts of the radio-
active CeF, and LaF,. These were counted
along with the Ffiltrates obtoined during the ex-
veriment aond were calouloted in the some way.
The wecuracy of the concentrations obtained for
these control somples is shown in Tobie 2.2.9.
The average error was 1.9%,
error wos 5.6%.

and the woaximum

For the gsolubility determinations sutficient tobeled
CeF, wos added 1o the Nof-ZiF -UF , mixture to
give o cerium concentration of a little over & wt %.
¥R

The mixture was then heated to BBPC for 1 hr,
and filtrates were taken successively ot opproxi-
mately 800, 700, 600, and 550°C after equilibration
for 1 hr at each tempercture. Labeled L0F3 was
then added to give 1.9 wt % lonthanum, and the
procedure was repeated. A second LaF, addition
was then made to bring the lanthanum concentration
up to 8.4 wt %, and the procedure wos again re-
peated. The cerium and lanthanum concentrations

calculated from the octivity dota obtained are
given in Table 2.2.10.
Solubility of SmF,
The solubility of SmF, in NaF-ZrF -UF , {50-46-
4 mole %) was determined with the use of Sm'23

PERIOD ENDING MARCH 31, 1937

as a tracer over the temperature range of 800 to
550°C.  After the fiitrates were obtained, CeF,
was oadded, and a second set of filtrates was
obtained. A second CBF3 addition was then made,
ond a third set of filtrates wos obtained, The
activities of all these filtrates were counted to
determine the and the
results are given in Table 2.2.11. The amount by
which the SmF, solubility decreased wupon the
addition of CeF, is shown graphically in Fig. 2.2.3.
It moy be noted that the decrease in the solubility
of SmF, coused by the odditien of CeF, is very
similar to the decrease in solubility of CeF, upon
the addition of L0F3, as shown above,

samarium congcentration,

Table 2.2.9, Accuracy of CeF3 and L\:|F3 Determinations in Synthetic Control Somples

 

 

Co F3 Found {mg)

 

Le F3 Found (mg)

 

 

Sample Error Erreor
No. By Synthesis By Analysis (%) By Synthesis By Analysis (%)
1 6.9 7.0 1.4 27.4 27.5 0.4
2 13,9 14,0 0.7 14.1 14.0 0.7
3 19.2 18.8 2.1 39.2 37.8 .6
4 3.2 30.5 2.2 32.3 2.6 0.9
5 60.4 1.0 28.8 27.2 5.6

61.0

 

Table 2.2.50. Solubilities of CeF3 and Lch3 in
Combinotion in Molten NuF-ZrF“-UF‘ (50-46-4 Mole %)

 

Approximate Found in Filttrate

Initial Temparatyre

 

fwet T)
Raore-Earth °C)
Content Ce La
6.2 wt % Ce, B20 5,72
na La 729 4,12
4638 3.20
548 2.47
6.1 wt % Ce, gl4 4.08 1.25
1.9 wt % Lo 726 3.06 0.%946
636 2.27 0.71
523 1.79 0.51
5.7 w1 % Ce, Bié 1.89 3.1
B.4wt% Lo 723 1.52 2.1%
632 .16 1.62
553 0.95 1.36

Takble 2.2.11. E#ect of CeFS on 3olubility of
':'rmF3 in NaF-Z¢ F“-UF4 {50-46-4 Maole %)

Somarium in

tniticl Rare- Temperature

 

Filtrote

Earth Content (°C) {wt %)
9.4 wt % S, BOO &.48
no Ce 700 4,45
600 3.38
550 2.99
7.3 wt % S5m, 800 2.88
6.7 wt % Co 700 2.10
600 1.64
550 1.48
9.8 wt % Sm, 8OO 1.73
12.1 wt % Ce 700 1.41
400 1.14
550 1.03

107
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORHL-LR-DWG 20237
TEMPERATURE (7]
800 BOO 700 &0 500
8 T_I— T 1 |
1 1 !

£

  
   

SAMARIUM | FILTRATE [wt %)

! I
93wl %Sm, BT wt % Ce
|

3 |
[ S8 wl T Sm, 12 1 wi % Ce

g _ e I o]
8 9 Wy i 17 130010~

¥riek)

Figs 2.2,3. Effect of CeF4 on Solubility of SmF4 in
Ma F-Zer-UF_‘ {50-46-4 Mole %). (Secret with captien)

Effect of BaF, or SrF, on Solubility
of CeF,

It has been determined that the addition of BaF,
or 5rF, to a melt contoining CeF, dissolved in
NaF-ZrF ,-UF, (50-46-4 mole %) reduces the solu-
bility of the CeF ;. In one experiment approximately
7 wt % Ce os labeled CeF, was equilibrated with
NoF-ZrF ,-UF, ot 900°C, ond then filtrates were
obtained ot approximately 800, 700, 400, and
550°C, os described above. After these filtrotes
were obtained, about 1.0 wt % B¢ woas odded as
8aF, and equilibrated ot 900°C, and another set
of four filtrotes wos token, More BGF? was then
odded to bring the borium content to 7 wt %, and

108

the procedure was repeated. The activities of the
filtrotes were then counted to determine the cerium
cong¢entration,

A similar experiment was performed in which
StF., was added to the extent of 4.6 wt % 3rto o
previously equilibroted CEFE-NQF-Zer-UF_‘ mixture
containing 6 wt % Ce. Filtrotes were obtained
by using the procedure desceiibed for the BoF,
additions, and the octivities were counted to
determine the cerium concentration. The results
of these experiments are given in Tobles 2.2.12
and 2.2,13 ond are shown graphically in Fig. 2.2.4.

Table 2.2.12. Effect of BuF2 on Selubility of
CeF3 in NuF-Zer-UFd {(50-46-4 Mole %)

 

 

Cerium in

Initial Rare- Temperature Eiliiais

Earth Content °c) (wt %)
7.0 wt % Ce, BOC 5,29
ne Bo 700 3.70
600 2.79
550 2,46
7.0 wt % Ce, 800 4.93
1O wt % Be 700 3.38
600 2.52
350 2.20
6.9 wt % Ce, goo0 3.23
7.0 wt % Ba 700 1.98
{6 male % Ba) 600 1.32
550 1.10

 

Table 2.2.13. Effect of Ser on Solubility of
CeFa in NnF-ZrF“-UFd (50-46-4 Mcle %)

 

Cearium in

Initial Rora- Tamperature

 

Filtrate

Earth Cantent (°C) (wt %)
6.0 wt % Ce, 800 5.29
no Se 700 3.70
600 2,79
550 2.44
5.9 wt % Coe, 800 4.57
4.6 wt % S¢ 700 2.82
(6 mele % 5¢) 600 1.86
350 1.5¢%

 
UNCLAZSIFIED
QRML—LF—DwG 20238

TEMPERATURE (°C})

 

    
   
 

 

 

 

200 800 Y00 a00 500
5
}L_.;_l_.I_F 1
|
- 1 — -
|_ ] | l N _I
5 - =i — - -
—
L 59wl % Ce,
46wl % 5¢ |
|
=
i |
3 :
i ] . ]
q
{r
Z |
o
= T i
=
= |
=
L
5 |
g e 70wl % Ce,
{.0wt % Ba
| 1
69'#17..Fe,?.0w1%3a
i
|
; | | | _'
5 a ‘0 " iz 1307
Vriek)
Fig. 2.2.4, Effect of Bl:||=2 or Ser on Solubiltty of

CeFa in NuF-ZrFJ-UF“ {50-46-4 Mole %). (Secret with

cupfiun)

PERIOD ENDING MARCH 31, 1957

SOLUBILITY RELATIONSHIPS AMONG
RARE-EARTH FLUORIDES IN
MOLTEN NaF-ZrF,-UF

R. A. Strehlow W. T. Ward
Solubility of Individual Rare Eorths

As shown in the preceding sections, the solubility
of LGFS, CeF3, and Sn'-.F3 in the molten solvent
NaF-Z¢F ,-UF, (50-46-4 mole %) has been carefully
examined os a function of temperature. The dato
reveal considerable differences in selubility for
materials whose chemicol properties are so similar.
The c¢ompound SmF, is, as indicated in Table
2.2.14, nearly 25% more soluble than LaF, ot
each temperature stydied, and the solubility of
CeF, is intermediate in each case.

The product of the solubility of sach compound
multiplied by the third power of the cotion rodius
is the some within 2% for each of the three
rare-garth tluorides studied (Table 2.2.14). Mo
explanation for the success of this correlation
has been offered, and it is possible that the
agreement is fortuitous. Subsequent studies of
solubility of these materials as a function of
NaF-ZrF, composition need to be completed before
a discussion s fo the generol validity of this
correlation can be made. In any event it is evident
that the solubility of the rare-earth fluorides is
sufficiently high to permit quite long-time operation
of o reactor before precipitation of these materials
becomes ¢ problem.

Solubility of Binary Rare-Earth Mixtures

When binary mixtures of CeF, and LoF, in NaF-
ZrF,-UF, were studied, as described above, the

Tabie 2,2.14. Solobility Relationships Ameng Rare-Earth Fluorides in I*nl:}l':-Zs-I':"-UF4 (50-46-4 Mole 72)

 

Product of Solubility and Third Power

 

 

 

Hampurdions Solubility {mols %) S w6
& SmFy  CeFy  LaF, Mean SmFt CeFl LaFy Mean
550 2.23 2.02 .80 2.02 +0.23 3.2 3.3 3.27 3.26 + 0.05
400 2.52 2.27 2.02 2,27 +0.25 3,43 3,72 3.68 3.67 * 0.05
700 3,38 3.00 2.68 3.02 ¥ 0,36 4.87 4,92 4.88 4,89 * 0.03
800 4.97 4.30 4.03 443 +0,54 714 7.06 7.33 7.18 +0.14

*Catien radii: 5m+++, 1.13 2._: Ce+++, 1.18 3; Lt

+

 

 

1.22 A.

109
ANP PROJECT PROGRESS REPORT

dota shown in Toble 2.2.15 were obtained. The
solubility of CeF, ond LaF, in combination is
intermediate between that of pure L{:F3 and that
of pure Cef,. The binary solubility data are
plotted in Fig. 2.2.5, along with the solubility
data for pure 'i:ne,-F3 and pure LcFa, as a three-
component sytsem, with the pure solvent (NaF-
ZrF,-UF,, 50-46-4 mole %) as one component.
The liquidus lines established by the four ex-
perimentel points are nearly straight.

In a preceding repor?q equilibrium quotients were
defined for the partition of the rare earths between
the liquid and their solid solution:

® N

NL0F3 {solid) C°F3 {solution)

K =

 

N

Can {salid] NLOF3 {salution)

where N is the concentration (in moie fraction)
of the designated material in the designated phase.
It was also established® that if Henry’s law were

cbeyed by both the liquid ond the solid solution,

K should be the ratio of the solubilities of the
pure component trifluorides. The foct that the
experimental points for the mixtures of Lc:F3 and
CeF, foll on the straight lines connecting the
original composition and the pure solvent, as
shown in Fig. 2.2.5, is equivelent to the statement
that K, as defined above, is unity rather than the
rofic of solubilities (obout 1.1). The ohserved

 

W, T. Ward, ANP Quar. Prog. Rep. Dec. 31, 1936,
ORNL-2221, p 126.

Table 2.2,15. Concentrations of L::F3 and E«aF3 in o Saturated Solution in HuF-ZrF4-UF

values are, accordingly, unexpected, and they indi-
cate some deviotion from Henry's law in the liquid
or in the solid solution or in both,

The preliminary studies mode to date for the
CrF,-SmF,-solvent system have yielded the X
values shown in Table 2.2.16. These values,
which were obtained with the use of a single-tracer
technique and with the assumption that the liquidus
lines are similar to those of the LaF;-CrF,-
solvent system, deviate significontly from unity.

UNCLASSIFIED
ORML ~LR—-0WG 20239

SCLVENT

 

RELATIVE COMPOSITION AFTER
ADDITION OF MORE LaFy
L — T 8

 

Flg. 2,2,5. Pseudoternary System Lc|F3-'C'|E'.-F3 and the
Solvent Mixture NaF-ZrF‘t-UFJ (50464 Male %). {Secrot
with coption)

4

(50-456-4 Mole %) for Different Relative Amounts of CeF3 and LC:FB in the 5olid Phase

 

 

Cemposition of Selid Phose (mole %)

 

 

 

 

Temperaturs Concentration in Solution (mole %)
CeF, LaoF, {°C £1%) CeF, LaF,
76.5 23.5 550 1.43 0.41
633 1.82 0.57
720 2.47 0.78
820 3.33 1.03
40.1 59.9 550 0.76 1.10
635 0,93 1.31
720 1.23 1.78
820 1.54 2.55

 

110

"1
Table 2,2.16. Equilibrium Quotient for Portition of
.."lrflF3 and Ca F3 Between Liguid aad 5olid Solutions

 

Compoesition of Solid

 

 

Tamperurure

Phase (mole %) cc; K
Sm F3 CEFS

7.6 8.8 550 1.2

&35 1.4

720 1.5

820 1.7

8.4 11,33 550 1.1

635 1.4

720 1.7

820 2.3

 

NSm FS{snIId] X NCu stsoluf ion}

K =

 

NCaFa(solid] x Nsm F3(|o|uflon)

Further studies with the use of the dual-tracer
technique on the SmF,-CeF,-solvent system will
be required before it can be established that the
two systems are reclly dissimilar. |f the data in
Table 2.2.16 con be verified, SmF, can be removed
from solution by equilibration with CeF, ond
filtration more readily than would be expected
from Henry's law estimates.

SOLUBILITY OF XENON GAS IN NaoF-ZrF ,-UF
N. V. Smith

The determination of the solubility of xenon
gas in molten NaF-ZrF -UF, (50-46-4 mole %) as
a function of pressure and temperoture was suc-
cessfully completed. As o conseguence of the
high cost of xenon, a recovery system for the xenon
used in the determination was added to the apparotus
previously described? for the determination of the
solubility of heliuvm in NaF-ZrF, -UF . In this gas-
dispensing and recovery system three xenon cylin-
ders were connected through their conventional
pressure-regulating valves to a manifold which served
to introduce the gos at the desired pressure into
the solubility equipment. Between each cylinder and

 

N, v Smith, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 130.

PERIOD ENDING MARCH 31, 1957

its pressure regqulator an odditional connection,
which could be closed with o needle valve, was
made to a second manifold which received the
xenon gos emerging from the sclubility apparatus.
With the use of a liquid-nitrogen cooling bath the
xenon could be recovered in one of the cylinders
while the experiment was fed from a second
cylinder.  The third cylinder wos maintained as
a standby in cose the source cylinder become
depleted during an experimeni. This assembly is
shown as Figs. 2.2.6 and 2.2.7.

The xenon was obtained in c¢ylinders from Linde
Air Products Co. Mass spectrometic analysis
showed its purity to be better than 99.99%, as
received. The high purity grodually decrecsed to
§5.9% after eight complete experiments.

The limited supply of xenon ovailable necessitated
@ constant wotch during the experiment in order
to aveid exhausting the soyrce cylinder below
o sofe preset pressure. Accordingly, the soturating
time wos lowered to approximaotely 5 hr. The
results appeared to correlate satisfuctorily re-
gardless of whether the time was 4 or 5 hr. In a
single experiment in which the saturotion was
attempted in 1 hr, the amouat of xenon found in
solution was only about one-half as much as that
obtained with 4- or 5-br saturation times. From
the previous experience with helium, it appeared
that 3 hr of saturation time was sufficient.

The results shown in Table 2.2.17 and Fig.
2.2.8 appear to be consistent to within approxi-
mately +10%.  All the experimental wariables
appear to have been controlled to within con-
siderably closer vuriations than £ 10%. The mass
spectrometer ancalyses agree, on an acverage, with
the synthetic compositions of the stondard gas
mixtures to within approximately 10%.  As o
consequence, it does not appeor unregsonoble to
assume thot the resulfs presented are correct to
within t20%, if it is assumed that the saturating
time used was sufficiently long.  Appropriate
experiments are under way fo demonstrate quan-
titatively the time required for saturction.

It may be seen that, as in the case of helium,
the solubility of xenon in NeF-ZrF,-UF, (50-46-
4 mole %) increases with increasing temperature.
This effect is discussed in more detail in the
following section,

111
ANFP PROJECT PROGRESS REFPORT

L

 

LHCL AMSIFIET
FHOTG 2T

Fig. 2.2.6. Xenon Recovery Assembly.

COMPARISON OF THE SOLUBILITIES OF
HELJUM AND XENON GASES IN
NOF-ZFF4-UF4

N. V. Smith

Thermodynamic comparisons of the solubilities
of inert gases in NaF-ZrF,-UF, (50-46-4 mole %)
ore more readily performed by recalculating the
selubilities on o mole froction scele. The mole
fraction, N, of o gos dissolved in this solvent is

defined by

"
E

~ (1/110.48) + ",

where n_ is moles of dissolved gas per gram of
solvent and 110.48 is the average moleculor weight

of the solvent. Since the solutions used were

112

very dilute, the value of n_  in the denominater
could be omitted, and the valves of the saolubilities
previously given in moles of gas per cubic centi-
meter of melt were chonged directly to mole
fractions by dividing by the number of males of
melt per cubic ceatimeter. The density values!!
were taken to be 3.37, 3.28, and 3.19 g,f’c::m3 at
600, 700, ond BOOPC, and thus the number of moles
of melt per cubic centimeter were 3.05 x 10-2,
2.97 x 10°2 ond 2.88 x 102 ot the three tem-
peratures. [he experimenta! volues of the solu-
bilities of helium and xenon obtained on this
bosis are shown in Fig. 2.2.9. The average valves

 

Ve 1 Cohen, W. D. Powers, and N, D, Greene, A
Physical Property Summary for ANP Fluoride Mixtures,
ORNL-2150, p 41 (Aug. 23, 1956).
VNCLASSIFIED
ORNL-LR= DWG 20272

£l

LIguID-LEVEL

 

 

 

 

 

 

 

 

 

 

 

 

 

   
 
  
   
 
     
   

   

 

 

 
 

 

 

 

  

 

 

 

 

 

 

 

 

GAaGE
T VACUUM ’F
SYSTEM :
—-— T0
MANOME TER
HIGH-PRESSURE vaLvES{3}
1
rl
TWO - WAY MERCURY H
BUBBLER r:
L
I|
= e ME Va Ly |
|r_T - EOLE LVES (3] :|
|| Lo I
|
I Lo DIFFERENTI &L '
P || PRESSURE 1
| L TRANSWMITTER L __,_1
I L bl - s —
1 I
V& | L— DEWAR FLASK | o |
b o | 1 FOXBORO b o \
e | e - RECORDE R | ;
l[' 1. : 1 | :l I|: 1
| | ' | |
L l } BYPASS VALVE il [
1 E I
F s . iy |
L el | | i |
[ r W I
| 11 1
L i |
r k
S L e - :
— | "y !a 1
MORMALLY GPEN SOLENOLD viliff// ' I 2 |
OPERATED BY MANOMETER 'r “ 2 |
I I 3 }
| 11 |
| |
I L I'
_ ] ! 8
2000- omp FURNACE '. J
' '
1 E

 

 

 

 

 

 

 

 

 

FREEZE VALVE

TO RECEIVER =—a——o

Fig. 2.2.7. Dicgram of Xenon Recovery System.

£S61 ‘1Tt HOUVYW ONIGN3 doiy3d
ANP PROJECT PROGRESS REPORT

Toble 2.2.17. Solubility of Xenen in NuF-ZrF4-UF4
{50.46-4 Mole %)

 

 

Xanon Salubility
Temperoture Saturating (moles of
°C) Pressure xenon per K
{atm} s:m3 melt)
x 108
00 0.49 0.245 1.3
1.50 3.28 2.18
1.92 3.75 1,96
Av 2.0 £ 03
700 1.00 3.67 3.67
1.96 8.60 4,38
Av 4.0 % 0.4
800 1.00 &.18 6,18
1.48 }1.32 6.74
Av 6.5 0.4

 

moles of xenen X 103
iK -_

 

&
P lfa.':rrl3 af melt} X (otm)

of Henry’s law constants ot the three temperctures
are listed in Table 2.2.18, ond the temperature
dependence of the constants is illustrated in Fig.
2.2.10,

For the cose of a noble gas, X, in equilibrium
with its saturated solution in o nonvolatile liquid,
the chemical equation for the process is

(1} Xidy = x(g) ,

where (d} and (g) refer to dissolved and gaseous
states, The equilibrium constant
for the process is given by

respectively,

@y fg)

 

(2) K, = ,
“ T e td)
where a refers to activity of the indicated species.
At the high temperctures ond low pressures used,
the helium ond xenon goses may be assumed to
behave ideclly and the pressures may be ossumed
to be equal to the fugacities. Accordingly, if the
standard state of the gas is taken at a fugacity
of 1 atm, the octivity of the gas is aumerically
equal 1o the pressure. The activity of the gos
in solution may be defined by

{3) gy = Yxta) Vx (ay

114

UNCLASSIFIED
QANL-LR- OwE 20240

 

 

 

o
o

&
o

XENON SOLUBILITY [meles OF Xefcm3 OF MELT)
Lo o
O o

I
L=l

20

 

 

 

 

 

 

O 050 100 1,50 200 2.30
SATURATING XEMON PRESSURE (atm]

Fig. 2.2.8. Solubility of Xenon in MuF-ZrFJ‘-UF4
{50464 mcle %) as & Function of Temperoture,
with caption)

{Sacret

INCLASSIFIED
ORNL-LR-DWG 20241

 

 

 

 

SATURATING PRESSURE {omm]

 

0 a 8 12 '8 20 24 28 {x1075)
MOLE FRACTION OF DISSCLVED GAS

 

Flg. 2.2.9. Selubilities of Xenon end Helivm in NaF-
Zr Fd-UF“ (50-48-4 Mole %)}, {(Secret with coption)
 

Teble 2,2,18, Henry's Law Constants of Heliym ond
Xenon Dissolved in NaF.ZrF -UF, (50-46-4 Mole %)

 

Herry's Low Constant, P/N (atm)

 

 

Temperufure
(GC} For Heliom For Xenon
x 10° x 108
400 1.54 1.51
70% 1.09* 0.75
800 0.71 0.45

 

 

*Based on only one experiment. The results of twe
other experiments at 700°C were discorded becouse of
mechonical faitures of the apporatus during the sxpari-
ments.

UHCLASSIFIED
ORNL-LR-OWG 20242

TEMPERATURE [}
2040 L0 e0a

I T

 

 

HENRY'S LAW CONSTANT, 2% {(atm)

 

A 1 | i |

 

10

900 950 000 1050 4100 W50 1200 (xi0™H

{
1)

Fig. 2.2.10. Temperature Dependence of Henry's Law
Ceonstants for Sclubilities of Xenon ond Helium in
NoF-ZrF ,-UF , (50-46-4 Mole %]). (Secret with caption)

PERIOD ENDING MARCH 31, 1957

and Eq. 3 reduces io

P K
X{g) X
W K, = - -

Yo i) Nx ) Y

 

Ower the region where the gas obeys Henry's law,
Yx (4) (S Unity, and consequently
{5) K =K |

& X

This is the equivolent of making the activity
reduce to the mole fraction over the concentration
renge mentioned. The fugocity of the standard
state implied by this postulation is numerically

equal to Heary's law constant for the particular
gas ot the given temperature.
Standard free energy changes, AF®, moy be

calculated for Eg. 1 from the vatues of the equi-
librium constants. The slopes of the curves in

Fig. 2.2.10 readily give the numerical volue of

AH® of Eq. 2. The standard entropy change,
AS®, is obtoined by the relotion
AH® - AF®
(&) AS® = ——— ——
T

Some calculoted values of AH®, AF®, ond AS® ot
1000°K are listed in Table 2.2.19.

The results oppear to be significant in that the
standard entropy changes are the some even if
the solubilities of the goses are widely different.
In oddition, some thermodynemic changes con be
calculated for the isothermal expansion at 1000°K
ot on ideal gas from a concentration of ¢ to ¢,
moles per liter.
numerical velue of the Hemev's low constant ex-
pressed in (moles of gas)/(liter melt):(otm) ond ¢,
is ossumed to be the ideal gas conceatration ot
1 atm, the values listed in Toble 2.2.20 are ob-
tained. As may be observed the vaolues of g, w,

I cq is allowed to assume the

Table 2.2.19, Stondard Enthalpy, Free Energy, and
Entropy Changes of X () = X {g)} at 10600°K

 

 

AL AF® As?

{cal) {cal) fe. v.)
Healium -7,400 —22,800 +15.4
Xanon -11,400 =26, 600 +15.2

 

 

115
ANP PROJECT PROGRESS REPORT

Table 2.2.20. Mumerical Yalues of Heat {g), Work (), Enthalpy Changes (AH), ond Entropy Changes {AS)

Involved in the Reversible Isethermal Expansion of on ldeal Gos at 1000°K

 

 

 

Gas ¢ (moles/liter) ¢, {moles/liter) g = w (col) AH AS {e. u.) TAS (col)
Helium 3.0 x 1p~4 1.22 % 10™2 —7.300 0 _7.3 _7,300
Xenon 4.85 x 10°° 1,22 x 10~2 _11,000 0 ~11.0 ~11,000

 

and TAS for both gases arc equol, within experi-
mental uncertainties, to the values'? of AH® given
in the Table 2.2.19. Further discussion of the sig-
nificance of the agreement between the values
colculated for both processes must await oddi-
tional experimental data. It is te be noted, however,
that this ogreement, if it is general for monatomic
goses in fluoride solvents, would encble the
investigator to calculate, from a single experimental

measuvrement, as many isotherms os would be
desired.
SOLUBILITY OF HF IN NnF-ZrF‘
J. H. Shaffer C. M. Blood

Knowledge of the solubility of HF in molten
fluoride mixtures would be wvaluable in studies
of methods of solt purification, some aspects of cor-
rosion problems, ond processes for disselution of
solid fuel elements for uranium recovery, No record
of direct measurement of the solubility of HF in
molten flucrides has been found. As ¢ byproduct
of studies of salt purification, attempts'® have
been made ot times during the past severol veors
to estimate the sclubility of HF in molten salis
from rotes of stripping of the HE by inert gas
carriers. The results obtained have not, however,
been conclusive.

In recent months it has been possible to measure,
by a direct equilibration method, the sclubility of
HF in on NoF-ZrF4 (53-47 mote %) melt at 600°C,
When data are available ot other temperatures,
comparisons of these values with those estimated
from meosurements
discussed.

of stripping rates will be

The direci equilibration procedure consists, in
brief, of sotyroting the melt with HF at constant

 

"2 The agreement of the calculated valves of TAS®

with the experimental AP values was peinied oui by
M. Blander.

13G. M. Watson, C. M. Blood, ond F. F. Blankenship,
ANP Quar, Prog. Rep. March 10, 1954, ORNL-1692, p 62,

116

pressure, transferring (by melting a frozen plug)
a portion of the saturated melt into o closed re-
ceiver containing on initial pressure of helium
equal to one-half the HF saturotion pressure,
re-establishing the frozen plug in the transfer line,
and stripping the HF from the melt in the receiver
by sparging with helium gas. The HF is absorbed
in ¢ stondard KQH solution, and the amount of
HF evalved is determined by titration. The pro-
cedures for soturotion ond transfer of the meflt
and the apperatus for these operations are very
similar to those uvsed for studying helium ond
xenon solubility.’® A simplified dicgrom of the
assembly used in these studies is shown as Fig.

2.2.11.

Whan the empty receiver is stripped ot 600°C
with helium, about 4 x 10=% mole of HF is re-
covered per liter of exit gos, ond after the passage
of 50 liters of helium through a transferred melt
the HF concentration in the exit gas is ogain
4 »x 107° mole of HF per liter of exit gas. This
“residusl'’ HF is probsbly o consequence of
desorption from the receiver and lines, but these
concentraiions ore completely negligible in com-
parison with the omounts soluble in the melt.
When molten salt saturated with helium was
transferred ond stripped at §00°C with 350 liters
of helivm, 0.32 mmocle of HF was
This *'blank™ is small but not negligible, and
it hos been used to correct the gross quontities
of HF (11 to 27 mmcles) recovered from the
solubility experiments.

recovered.

Summaries of the experimental results obtained
to date ore presented in Table 2.2.2) and Fig.
2.2,12. From these data, which will be extended
to 700 and 800°C and to pressures in the ronge
of 0.25 to 2.5 otm, it oppears that Henry's law

 

HN. Y. Smith, ANP Quar, Prog. Rep. Dec. 3I, 1956,
ORNL-2221, p 130.
PERIOD ENDING MARCH 31, 1557

UNCEL ASSIFIED
QRML-LR-DWE 20243

® ®

 

   
 
  
 

  
 
    

THERMOSTATED BaATH

FLOWRATER
OFF - GAS LINE

 

W

-

[

]

i
SATURATOR |

]

!

{

 

 

RECEIVER

   
 

 

| 3

 

 

 

VACUUM SYSTEM ‘ \_~ | - \_“e
ey
| "
SAMPLING LINE | TRANSFER LINE
i HFE

Fig, 2.2.11. Apparaotus for Determining the Solubility of HF in Molten Fluoride Mixtures.

Table 2.2.21. Solubitity of HF in Malten NuF-ZrF‘ {53-47 Mole %) at 600°C

 

 

 

Saturoting Quantity of Melt HF HF

Prossure Transtarred Stripped Salvbiliry Ly
{atm} (l:rn3} (mmeles)* {mnlefcma of melt)

x 103 x 103

2.53 210 27.73 3.1 1.23
2.52 8§72 26.29 3.02 1.20
2.05 33 22.80 2,45 1.1¢
2.05 950 26.14 2.72 1.33
1.55 927 17.86 1.92 1.24
1.50 034 16.38 1.75 .17
1.05 B49 11.27 1.33 1.27
1.05 ¢10 12.63 1.3% 1.32

Mean 1.25

 

*Corrected for 0.32 mmole blank.

“*K expressed as moles of HF par cubic centimeter of melt times pressure in otmosphares.

117
ANP PROIECT PROGRESS REPORT

UNCLSSEIFIED
QRHL- L R-CWG 20244

 

 

DISSOLVED HF (moles OF HF/cm® OF MELT)

 

 

O 05 0 1.5 2.0 2.5 30
HF FRESSURE fafm]

Fig. 2.2.12. Selubility of HF in Nt:ll‘:-Z.rF4 (53-47
Mole %) at 600°C.

is obeyed ond that the solubility is 1.25 x 1077
mole of HF per cubic centimeter of melt at 1 otm.

EME MEASUREMENTS IN MOLTEN FLUORIDES
L. E. Topol
Concentration Cells with Ze®|ZrF, Electrodes

The solubility of a metal fluoride in a molien
fluoride salt can be determined from the dis-
continuity of the liguidus femperature in a plot
of emf vs temperature'® for a concentration cell
of the type

MCPIMF

2
AF(CE)HAF(C;), ME_ (M0,

n(c]

rlc, ¥

where M is o meto! ond AF is a molten fluoride
salt.  Attempts to use this technique, however,
to establish the liquidus temperafure for the NoF-
ZrF o (39-61 mole %) mixture with the use of Zr°
or Zirceloy electrodes have been generally un-
successful. The potentiols of the cells exhibited
irregular  fluctuations with temperoture, and in

 

13 Topol

L. ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL- 2221 b 139,

18

some cases the tempercture dependence of emf
was in the direction oppeosite to that expected.

Concentration Cells with Ni°| NiF, Electrodes

Preliminary experiments with concentration cells
of the type

Ni°\NiF2(C]}HNiFE(CE)]Ni“

with KF-LiF {50-50 mole %) as the solvent indicate
the solubility of Nin in this solveat to he about
1Q times thot in the solvent NaF-ZrF  (53-47 mole
%). This difference can probably be attributed
to the hosic nature of the KF-LiF solvent in which
the weckly ocid NiF, would be more soluble than
in the more acid NoF-ZrF, mixture. (See section
of Chap. 2.3 on "'A Structure Model and on Acid-
Bose System for Fuel Mixtures,”” this report.)

An illustration of the acid character of ZrF,
was obtoined in measurements of the cell

     

Ni|NiF sle,)” NiFucl}mi

in KF-Lif, where ¢; = ¢ = 1.00 wt %, ond
= 1.5 wt %. The emf of this cefl, 7 to B mv

at 630°C with the ZrF, melt positive, is equal

to

 

RT af ag L
- — In [— S
L) fi'.l aa
oy a, 2¢,
= -0.09 log - ,
.84 23

where x, a, y, and t, denote the mole fraction,
activity, activity coefficient of a species, and
the average transference number of the olkali
cations, respectively; the subseript 1" refers
to the NiF, and the subscript 3" 10 the KF-LiF;
the primes signify the Zer-free melt. Since the
mole fractions x, ond x| were approximately equel,
it was found that

121 ay \*+
— iff [ = 1.20
3 a3

and hence the NiF, activity in the ZrF  -free melt
is, as an opproximation, at leost 20% greater then
in the ZrF ,-containing melt.

ok
In both half cells, as well as in the nickel half
cells described in Toble 2.2.22 in the foilowing
section, the NiF, solutions were reddish, as they
were found to be in NoF-ZrFi (ref 16). Dilute
NiF, solutions in KF-LiF-NaF eutectic were
previously'® found to be yellow. The red color
is probably due te the larger concentration of
NiF, used in the emf experiments described here.

Daniell Cells

In order to obtain more information regarding the
change of solute octivity coefficients with solvent
composition, efforts are under way to develop
more relioble types of Daniell cells and to measure
a series of these cells, as wos done previously,!”?
in different media with varying but dilute con-
centrations of the metal fluorides, Nin, Fef,,
CrF,, and possibly other solutes. Some results
are given in Toble 2.2.22 for measurements made
with Daniell cells of the type

FelFef 5o yIINIF 5 INi

 

6L, E. Topol, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 148.

]?L. E. Topol, ANP Quar. Prog. Rep. Dec. 10, 1855,
ORNL-2012, p 97.

PERIOD ENDING MARCH 31, 1957

in NaF-ZrF, (53-47 mole %) ond KF-LiF (50-5C
mole %) contained in iron oad nickel crucibles,
respectively, The measurements were corrected
for thermoelectric effects, but junction potentials
were neglected.

The results agree well with the results of
previous measurements in the NaF-ZrF solvent,

which gave about 500 for yNinf?Ferf even

though most of the cells studied exhibited a de-
crease in emf with time, oand none were reproducible
to within 5 mv for more than one day. The FeF,
melts showed o greoter tendency to creep than
the NiF, melts in both solvents,

It oppears that the activity ratic of NiF, to
FeF, does not change much with solute con-
in the roange meosured but that fhis
ratio vories from on averoge of about 600 in
NoF-ZrF, to 200 in KF-LiF. The two values of
around 300 (Table 2.2.22) obtained in the NaF-
ZrF‘{ solvent should probably be disregarded,
since any failure of the cells seems to result in
a decreasing emf, which in turn produces a smaller
calevlated activity coefficient ratio.
value of about 400 found for VNin/}’FeFQ in the

centration

The averoge

NaF-ZrF‘i solvent may be compared with a value

of 700 at 625°C taken from date obtained in

Toble 2.2.22. Activity Coefficient Rotios of Nil‘:2 to Fer in Molten Fluoride Solvents as
Obtained from Daniell Call Meosurements at §50°C

 

Concaentration

Concentration

 

Solven of NiF, of FoF, Yie,/YFoF,
{mole %) {mole %)

NaF-ZrF, 0.212 0.744 500*
0.214 0.224 310
0.202 0.205 301
0.190 0.192 497
0.210 0.258 650
0.213 0.706 510
0.258 0,700 640
0.247 0.659 730

KF-LiF 0.545 0.62) 193
0.530 0.617 218

 

*This valus is based on data obrained previously, ref 17,

119
ANP PROJECT PROGRESS REPORT

previous studies of the eguilibrium reduction of
meta| fluorides by hydrogen in NaF-ZrF, {ref 18).

The observed change in the volue of the activity
coefficient ratic with the chonge in selvent was
expected, since a strong fluoride donor such as
KF-LiF would tend to *'level”

between two

the differences
“amphoteric"’ It should be
noted that the large values found for the activity
coefficient ratios of iron and nickel fluorides are
probobly due to on error in the free energy of
formation of Nin, which enters the colculation
of activities from Daniell cells. Thus by de-
creasing the present volue of AF° for NiF, by
about 10 keal (from =126 te =116 keal ot 1000°K),
the experimental results can be brought into better
accord with theoretical postulates.

solutes.

ACTIVITIES FROM EMF MEASUREMENTS QF
MOLTEN AND SOLID SOLUTIONS OF SALTS

M. B. Ponish

It has been noted os o result of continving
investigations directed toward understanding the
corrosive behavior of salt melts that mony of
the distinctive features of the molten fluorides
of interest as reactor fuets can be explained with
models based on the premise that the ions hehave
as hard, cherged balls. For example, the negative
devictions from Raoult’s low, which ore charac-
teristic of the fuel mixtures, can be correloted,
os for as relative magnitude is c¢oncerned, with
the calculated electrostatic forces between charged
spheres hoving the same charge-toredius ratio
os the ions in the fuel mixtures.'® {See also
section in Chap. 2.3 on ""Some Aspects of Solution
Behavior in Fluoride Meits,’’ this report.) Com-
porison with the behavior of other salt mixtures
which cannot be explained by such a simple model
shows that the difference is due to the presence
of polarizable ions, thet is, ions whose charge
distribution con be distorted inte nonsymmetrical
configurations by neighboring ioens.  Estimates
based on refractivities?? indicate that the F-
ion hes a polarizability of 0.81, compored with

 

T8¢, M. Blood ond G. A, Paimer, ANP Quar. Prog. Rep.
Deg, 31, 1956, ORML-2221, p 120, ond previous reports

in this serjies,

%M. Blander, ANP Quar. Prog. Rep. Dec. 31, 1956,

ORNL-2221, p 140.

ZQJ. A. A, Ketelour, Chemical Constitution, Table ¥,
o 90, El sevier, Houston, 1953

120

2.98 for CI™ and 4.24 for Br=. The Ag’ ion has
a polarizobility of 1.9 compared with 0.03 for
Li” and 0.24 for Na”,

The effects of composition on activities in the
systems NaCl-AgCl (ref 21} and LiCl-AgCl were
measured to illustrate positive deviations from
Raoult's law in centrast te the negative deviotions
predicted by the hard sphere model. Positive devi-
ations were found experimentolly thet were smaller
with Na* than with Li%, as shown in Fig. 2.2.13.
Some of the stability of pure AgCl can be atiributed
to the partially covalent nature of the bonds be-
tween Ag' and CI™ ions, which in turn is due to the
polorizability of these ions. The effect of neigh-
boring alkeli cotions on the loss of the covalent
character (“‘depolarization’ of the chloride jon)
ond the consequent decrease in the stebility of
AgCl is reflected in the trend toward positive
deviation,  The increasing negative trend with
increasing alkali cotion size is in accord with the
expected smaller electrostotic attraction between
chloride 1ons ond the larger alkoli cations,

The correlotions described above apply to several

other systems which contain polarizable ions

 

24, B. Panish, ANP Quar. Prog. Rep. Sept. 10, 1956,
ORNL-2157, ¢ 115,

UNCLASSIFIED
ORNL=|R=0QWwG 20245

 

 

 

 

7 | | ]
<
2 I
s
:
E 05 (— T == T —]
e
=
2
b_
<
g |
| e S -
0 AqCI-NaCl
® AgCi-LiCl
0 I
U 0.9 i
AgCl {moie froction)
Fig. 2.2.13, The Activity of AgCl in the AgCl-NaC|

ond AgCIl-LiCl Systems ot 300°C.
ond which exhibit o tendency toward positive
deviations.?? By contrast, the fluoride fuel sol-
vents give little evidence of behavior resembling
that ottributed to polarizable ions. This contros?
strengthens the conclusion that fluoride
can be considered as relatively unaffected by
polarizotion and indicotes that flueride melts can
be adequately represented by simpler models than
those required for molten solts in general.

In the course of the electrochemical studies
of the LiCl-AgCl and NaCl-AgCl systems, ot-
tention was also given to the solid selutions
found below the solidus. Measurements2® af
400°C, for example, give the activities shown in
Fig. 2.2.14, The significance of these meas-
urements is twofold. First, they provide activities
in o solution whose structyre is known, since

melts

the solid-solution crystals hove a simple cubic

lattice. Second, they indicate the existence of

 

227, Fonl ond, On the Properties of Some Mixtures of
Fused Saits, Penn. State Univ. Tech. Rep. No. 4%
{June 1958).

Bpsigningl aF ik zall Ag®/AgCl (solid solution)/ClZ°
was obtained as a function of temperature and composi-
tion. There is apparentty sufficient electrolytic con-
ductivity of the crystal to gllow potential measurements
reproducibie to X1 mv. These potentials were considared
to be a close opproximotion to the reversible values.

UNCLASSIFIED
ORML~LR-DWG 20246

 

a5

  

ACTIVITY (orbitrary units)

o AgCI—LiCl
& AgCI—NaCl

 

 

G 0.5 {
aqCi{male fraclion)

Fig. 2,2,14, The Activity of AgCl in AgCl-NaCt and
AgCi-LiCl Solid Solutiens at 400°C,

PERIOD ENDING MARCH 31, 1957

concentration ranges in which the activity of AgCl
decreases with increasing concentration.  Such
unusual activity curves apply only te metastable
conditions '‘frozen’ by the slowness of diffusion
in the solid; they could never persist in @ melt,
but it is of interest to note that they can have a
real existence if the most obvious interpretation
of the date is correct.

The results on the LiCl-AgCl system lend support
to those previously found for the NaCl-AgCl

21 and favor a revision of the shope of

System
the activities curve described in the earlier report
to conform to Fig. 2.2.14. The main deterrent
to the application of similar techniques to fluoride
fuels is the lock of a svitoble substitute for o
fluoring electrode. As a consequence, it is neces-

sary to moke comparisons with chloride systems.

ELECTROMIGRATION IN MOLTEM SALTS AS A
POSSIBLE SEPARATIOR TECHNIQUE FOR
PROCESSING FUEL

M. B. Panish

Electromigration in molten salts has been

shown to be a successful means for separating
is.cmc:;:utaesa,M and therefore it is of interest to de-

termine the usefulness of this technique for the

presumably eosier problem of seporating ions
of different elements in fuel mixtures. Electro-
migration accomplishes separation by taking

advantage of the differences in mobility of the
ions. The electrochemical cell used is contrived
so that the electrolytic product at one electrode
is consumed ot the other, and ne net chemical
transformation is accomplished. After operation
of such a cell for o pericd of time, the fastest
gccumulate near the cathode, ond the
ln a molten salt

cations
fastest onions near the anode.
mixture containing a common anion, such as moalien
fluorides or chlorides, a separation of the cations
can be occomplished if remixing s prevented.
Remixing can be suppressed by the use of a smoll
cross-section cell and packing and by the avoidance
of temperature grodients which would result in
convection ¢urrents.

The separation of Li* ond Ni*" ions in a chloride
melt wos studied ot temperatures between 650

 

sz. Klemm, in 2. Eiektrochem 58, 60% (1954), has
demonstrauted the feasibility of seporating the isotopes
of lithium by electrolyzing molten LiCl wirh chlorine
olectrodes as both anode and cathode.

121
ANP PROJECT PROGRESS REPORT

and 750°C. Chlorides were chosen in order to
allow the use of & quartz-glass container ond
chlorine electrodes. The Li’ ion was chosen
both becouse it is of practical interest and because
it has high mobility. The Ni*" ion was chosen
becouse its blue color served as @ good indicator
and allowed the movement of the ion tc be fellowed
visuolly, even for Ni** concentrations as low as
0.001 mole fraction. A diagram of the cell is
shown in Fig. 2.2.15. Substitution of HCI| for
Cl, os the gas supply to the electrodes eliminated
much of tne difficuity from bubble formation in
the pocking, This effect could be explained by
a smaller temperature coefficient of solubility
for HCl than for Cl,, coupled with higher tem-
peratures in the pocking because of the higher
resistance.

The NiCl, was added by breoking o frengible
seol at the bottom of the U-tube. A 0.5-amp
current, which corresponded to o nominal current
density ol approximately 20 amps/cm® in the
electrolyte in the packed column, was maintained.
The cell was viewed throuvgh o slit in a clamshell
furnace; the boundory changes
with o cathetometer.

were followed

The experimental results obtained at 650 te
750PC  showed that the boundary between the
colorless solution and the solution colored blue
by the Ni*" ions moved toward the anode, and
thus the Li* ions moved faster than the Ni'’
ions. If Li" ions had corried all the current, the
boundary would have moved at a calculated rafe
of 22 mm/hr under the conditions of the experiment.
The measured rate was 20 2 mm/hr, If the experi-
mentally observed value of 20 mm/hr is assumed to
be correct, and if the chleride lattice is considered
to be stationary, the Li* ions moved about 20
times as fast as the Ni'* jons. As a result of
the uncertainties in the assumptions, however,
this conclusion must be considered as repre-
sentotive aond qualitative rather than quantitative.
Within the ogccuracy of the measurements, no
temperature effect was found between 650 and
750°C.  |n principle, electromigration should be
a very efficient means of recovering Li* from a
spent fuel, but the engineering arrangements for

122

working in fluorides rother than chlorides represent
some extremely formidable problems. Even the
chloride systems seem likely to be very awkward
from an engineering standpoint.

UNCLASSIFIED
DRAML-LR-DWG 20247

GAS SUPPLY

GRAPHITE TUBE ELECTRODES

 

 

 

MOLFEN LiCl

 

 

     

 

o CRUSHED, FUSED SILICA PACKING
L [30-50 MESH)

 

s B=-mm 1B

| o ’———— ROD T BREAK SEAL

 

 

FRANGIBLE OISk SEAL

MiCly

Fig. 2.2,15, Apparatus for Studying Electromigration,

=
PERIOD ENDING MARCH 31,1957

2.3. THEORY OF MOLTEN FLUORIDE BEHAYIOR

F. F. Blankenship

CORRELATIONS OF MOLTEN FLUORIDE
BEHAVIOR BASED ON CHARGE-TO-10ON
RADIUS RATIOS
M. Blander

An attempt has been made to correlate the be-
havior of molten fluorides from the standpoint of
cation-gnion attroction. Emphosis on cation-onion
offinity has, however, largely obscured the effects
of repulsion of like charges in the melt, which
ore  probobly more important than cation-anion
attraction effects in causing negative deviations
from ideal solution behavier in liquids in which
the components mix with little or ne structural
change. To the extent that emphasis on cation-
anion attraction implies the existence of discrete
complex iens in all fueride melts, the concept is
misleading.

A mixture of solts with slightly ditferent Z/R
(charge-to-ion radius) ratios, in which no structural
change from that of the melten components results,
will be energetically more stable than the unmixed
salts, olthough the anion-cation attroctive forces
are not aitered by the mixing. An example of this
is shown in Fig. 2.3.1. Configuretion A for the
mixture 15 more stoble than the average of config-
vrations 8 aond C, which represent the pure com-
ponents, in thai the larger the difference in 1/r for
Li* and K* or in Z/r at equol r for the twocations,
the more negative becomes the energy of mixing,
AE {mixing}. f the Z/R ratios of the two salts
differ only slightly, the siructure of the mixture
will differ only slightly from the structures of
the <¢omponents, and the extra stobility of the
mixture will be small.

I the Z/R rotios differ considerobly ond es-
pecially if the salts are of widely different valence
states, the extra stability of the mixture will be
large, ond the structure will be considerably dis-
torted in comparison with the structures of the
component salts, that is, the larger the difference
in Z/R ratios, the more stable the mixture relative
fo its components. For components which mix
without o change in coordinotion number and with
little distortion, the negative deviation from ideal
solution behavior is occounted for, in the main,
by a decrease in cation repulsive energy upon
mixing.

UNCLASSIFIED
ORNL-LR-DWG 20248

o b L e

L
CONFIGURATION A

o]

CONFIGURATION B

_et et + __2e? ~ AF (MIXING) < O
2R 2R ¢ R * e

 

Fig. 2,3.1. Diagram of Catlon Repuision Energies for
Vorious lon Conflgurations Showling the Gain im Srability
upon Mixing,

SOME ASPECTS OF SOLUTION BEHAVIOR
OF FLUORIDE MELTS
R. F. Newton F. F. Blankenship

The polarizability of ions, which can be inter-
preted in terms of portially covalent bonds between
ions, is a factor in the solution behavior of molten
Huoride melts thot is usuvoily secondary but which
moy be important in many melten electrolytes.
Fortunaiely, the fluoride melts of interest as
reactor fuels cepresent a rather extreme cose in
which the effects of polarization of the ions appear
to be of little consequence.

Foremost among the factors controlling the
solution behavior of molten electrolytes is the
effect of ion charge ond size on the coulombic
force between neighboring fons. Progressively
smatler contributions to the energy of a solutian

123
ANP PROJECT PROGRESS REPORT

configuration result from electrostatic inferoctions
between next nearest neighbors, and so on, until
the effect becomes negligible with distence. This
electrostatic behovior has long been recognized
as a controtling factor in the lattice energy of
ioni¢c crystals.

All the moiten fluorides which have been con-
sidered as possible fuels in the large develepment
program contain cationic componenis of widely
different valence state (U'”, Zrt', Be?", Na™h
which may be expected, from apy standpoint, to
show large negative deviations from ideal solution
behavior and large distortions from the structures
of the pure molien components. Accordingly, it
)5 convenient 10 correlate the observed behavior
of these complex liquid systems in terms of the
attraction between the fluoride ions and the neigh-
boring cations.

The attractive force is given, for the case in
which the ions are in direct contact by the relation

Zez
F = —— .

{ry + ?’_)2

where F is the attractive force, Z is the charge
on the cation, e is the charge on theoeleciron, T,
is the cation radivs, and r_ is 1.33 A, the radius
of an F~ ion. It is convenient to vse r in ang-
stroms, the charge of the electron as the unit of
chorge, ond Z as the volence of the cation. In
these wunits the attractive energy (E) between ¢

flouride ion and a cation it is touching is

2
Ze
- r
Ty + r_

which is numerically equal teo —-Z/R where R
=7, br_ .

The ions in ¢ selution tend to arronge them-
selves in positions of lowest energy. Accordingly,
the mutual energy velue -Z/R determines the
position of the fluoride ions with respect to alter-
native cations, and fluoride ifons will prefer as
nearest neighbors those cations which yield the
lorgest Z/R rotios.! The Z/R values derived

 

]This viewpoint ossumes that altermotive positions
exist. Such an assumption can be avoided and the some
gualitative correlation can be reached by cttributing the
negotive deviotions to the decreose in cation-cotlon
repul sien that cecurs when salts of different Z/R volues
are mixed.

124

from estimates of ionic radii in crystals are given
in Table 2.3.1.

An increase in the difference in the Z/R ratios
of the cations of a flueride solutien will result in
an increase in the negative devistions from ideal
solution behavior shown by the mixture. Exomples
of conformity with this Z/R rule were given
previously.?

In order to predict the behavior of selutes in
binary solvents it is necessary to know the effective
Z2/R for the binary solvent.  Presumcply the
effective Z/R for the solvent (as for ony mixture)
varies with composition between the volues for
the pure components; however, the precise shape
of the curve (or surface) for Z/R vs composition
If the Z/R ratio for a
simple solute is identical with the effective Z2/R
ratio for the complex selvent, the solution com-

is not readily ascertained.

position at this point can be termed the neutral
composition. The activity coefficient of the solute
15 ot a maximum ot the neutrol composition, and
the further the composition of the solvent from
this composition the smaller will be the activity
coefficient of the solute. This principle provides
a basis for evoluotion of effective Z/R ratios for
the NaF-ZrF, system, for exomple, as a function
of composition,
coefficients of soiutes as a function of solvent
composition, neutral compositions ond effective
Z/R ratios for the salvents can be established.
This process will be relatively labericus, and
completely quentitetive data will be unovailable
for some time. An attempt is moade in the next

From measurements of activity

section of this chapter to extend this concept
with the available data. The principle is gquite
useful, however, even in qualitotive form, in that
it can be used to exploir the opparently anomalous
differences in vapor pressures of the NaF-ZrfF,
system when UF, is added or substituted.

Mixtures of NaF (Z/R = 0.43) and ZrF, (Z/R
= 1.82) corresponding to Na,ZrF, show large
negative devictions from ideality, and the vaopor
pressure of Zer from such a mixture is much less
than would be expected. The addition of UF,
(Z/R = 1.68) to such a melt serves to increase
the average Z/R for the mixture and thus to lower
the activity eoefficient of NeF ond to raise the

activity coefficient of ZrF,.  Accordingly, the

 

M. Blonder, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 140.
PER!OD ENDING MARCH 31,1957

 

 

 

Toble 2.3.1. Valence-to-Radius Roties for Flueride Salts Bosed on Goldschmidt's® Volues of lon Radii
Cation Cation
Cation Radius Zhry + 1) Cation Radius Zhr, 4 1)
(r,} (ry)d
ee! 1.65 0.34 Lo " 1.22 1.18
rb* }.49 0.36 Ead* 1.18 1.19
i« 1.33 0,38 Pt 1.16 1.20
Y 0.98 0.43 ngdt 1.15 1.21
G’ 0.96% (0.44) 3t 1.14¢ (1.21) 2
i 6.70 0.49 B 1.13 1.22
i Eu?’ 112 1,22
Bd ].43 D.?3 3+
. Gd 111 1.23
5¢ 1,27 0,77 o
- Bi 1.1 1.23
o 1.06 0.84 5
% 5 ¥ 1.06 1.24
Cr 0193 (D-;BE} 3+
- P 0.67 1.50
Fe 0.83 0.93 34
. e 0.64 1.52
CO 0.32 0.93 3+
. Al 0.57 1.58
Cy 0.8 0.94
Nt 0.78 0.95 Thit 1,10 1,45
Mgt 0.78 1,00 gt 1.05 1.68
g b 0.34 1.20 i 0.87 1.82
et 0.64 2.03

 

 

Y, M, Goldschmidt ted., A. Muir), Geachemistry, Clatendon Prass, Oxford, 1954.

L. Pauling, j. Am Chem Soc. 49, 765 (1927).

“From deta on UCL 73 given in ). J. Katz and E. Rabinowitch, The Chemistry of Uranium, NNES, p 453, McGraow-

Hill, New York, 1951.

dme J. H. Hildebrond and R. E. Powell, Principles of Chemistry, dth ed., p 267, Macmillan, New York, 1940.

addition of UF, to Na,ZrF should increase the
vapor pressure of ZrF4, and it has been cbserved®
that, when sufficient UF, is added to Na, ZrF,
to ottain NaF-ZrF -UF, (50-25-25 mole %), the
vapor pressure of the melt is increcsed from 3 to
10 mm Hg ot 912°C. This remarkable increase,
which occurs even though the ZrF, concentration
of the melt is reduced from 33.3 to 25 mole %,
indicates that smoll Z/R changes
noticeable effects.

hove very

 

jFrom equation given by R. E. Moore in ANP (lhuas
Prog. Rep, March 10, 1953, ORNL-1515, p 159.

The effective Z/R for the mixture is reduced
slightly, however, when 7NaF.6UF, is added to
7NaF.6ZrF | (replacement of ZrF4 by UFA). This
shift of Z/R in the direction oway from the high
Z/R value for ZrF, reduces the activity coeffi-
cient for ZrF ,, and accordingly the vapor pressure
of ZrF  should decrease more rapidly than would
be expected from the decreased concentration of
ZrF,.  This expected more pronounced decrease
was found experimentally, as reported previously,?

 

45- D. Christian and 5. Contor, ANP Quar. Prog. Raep.
Dec. 31, 19536, ORNL-2221, p 144.

125
ANP PROIJECT PROGRESS REPORT

A STRUCTURE MODEL AND AN AC|D-BASE
SYSTEM FOR FUEL MIXTURES

0. G, Hill F. F. Blonkenship

A useful qualitative model for describing the
liquid structure of fuel mixtures based on the
NaF-ZrF, system was devised in which liguid
Zrf , is considered to be ¢ somewhat disorganized
lattice of flyoride bridges connecting zirconium
ions.”> At the other end of the system, molten
NaF contains sodium ions and fluoride ions. The
addition of NeF to ZrF4 results in the breaking
of the bridges to give zirconium ions which cre
surrounded partially by bridging fluorides and
partially by nonbridging flucrides.

According to the proposed model, the addition
of ZrF, to a large excess of NaF results in the
formation of ZrF,~77 iens, and thus for the com-
position range of up to 25 mole % Zrf the liquid
is composed of Nu+, F=, and ZrF,” ™" ions. At
25 mole % ZrF | most of the free fluoride ions have
been incorporated in ZrF,” 77, since the stoi-
chiometry of the meit conforms to Nu3ZrF? (ref &).
The fluorides attoched to the ZrF =77 ions are
examples of nonkridging fluorides, as distinguished
from free fluorides and bridging fluorides. If
more ZrF, is odded, it connot form additional
ZrF,=~~ lons but can only share some of the
olready complexed fluorides. The shared fluorides
are referred to as bridging fluorides.

With increased Zer concentrations, there is
a further decrease in the numbar of nonbridging
fluorides and an increase in the number of bridging
fluorides to eight per zirconium in pure ZrF

(ref 7).

4
If on assumption is made regarding the

 

5R. D. Burbank in The Crystal Structure of Uranium
Tetrafluoride, K-769 {June 8§, 1951), describes the strue-
ture of UI:4 crystals, which are isemorphic with Zer,
as an fonic lattica in which each uranlum ion is sur
rounded by eight fluorine ions,

6T|'|e K+ ions and ZrF;“_ ions are the structural units
in solig KEZrF?, which is isomorphic with Naa?_'rF?,
according to W. M. Zachariasen, Acta Cryse 7, 792
(1954).

correspending to this fomlly of compounds indicates

The shorpness of the phose diagrom maxime

that they melt with lirtle dissociation,

7s, Langer, in Chap. 2.4 of this report, has observed
thot UF4 melts without opprecioble expansion. The
model assymes that ZPF‘ behaves analogously und that

the coordinotion number remains unchoanged.

126

coordination number of the Zr*' ions, o more
definite picture of the liquid structure of fuel
compesitions near 50 mole % ZrF4 is available.
Because of the rodius reltaionships, no more
fhan seven fluorides can touch a Zr#* ion at one
time, |f seven, which is near the upper limit,
is used as the coordingtion number for the 50
mole % ZrF, region, about 40% of the fluorides
are shared, or bridging, and 60% are unshared,
or nonbridging, as in ZrF?‘“"'. There are no
discrete groups of ions that form a structural
entity or complex, such as ZrF_." ions. Rather
there is an interlocking network of Zr*' ions
joined by an average of four bridges to other
ions. The U4" ions substitute for Zr*" ions
without altering the picture very much, and Na”*
ions, also present, maintain electroneutrality in
the vicinity of nonbridging fluorides. The situation
is not static, since there is on extremely mobhile
equilibrium between the kinds of fluorides, and
the bridges are easily broken ond reformed.

For simplicity the model has been described
as though free fluoride ions did not exist at higher
concentrations than 25 mole % Zer. Actuolly
they are present throughout the composition range
but their effective concentration diminishes rapidly
with increasing ZrF, concentrotion, Since the
complexing potentialities of a fluaride melt depend
on the fluoride ion activity, it is very important
to know how this activity vories with compesition,
Unfortunately, it connot be measured.

The fluoride ion activity can, however, be
conveniently correlated in o qualitative manner
by an acid-bose scheme in which cations are
rated according to their tendency to behave as
fluoride donors or occeptors, |If FHuoride denors
are defined as bases and tluoride acceptors as
acids, the cations foll into the same general
pattern as thot used for the classificotion of
oxides as acidic or basic, The NaF compound
is on example of a strong base and ZrF, of o
strong acid., The stronger the acid ond the bose
that ore combined, the stronger they complex ecch
other on mixing, or stated in thermodynamic terms,
the greatsr are the negative deviations from ideal
solution behavior shown by the mixtue. The
resulting mixture con olso be clossified as acidic
or basic. Fuels based on the LiF-NoF-KF sutectic
are strongly basic, and they highly complex cor-
rosion products, since the latter are usually

LT

T
amphoteric. o contrast, fuels based on the NaF-
ZrF ,-UF, system are nearly neutral with respect
to corrosion products.

In order to map the relative acidity of a fuel
sclvent on a numerical scale, it is convenient
to use a plot having Z/R ratios for the ordinate
and solvent composition {in mole %) for the
abcissa, os shown in Fig. 2.3.2, The coulombic

UNCLASSIFIEDR
DRHL-LR-DWG 19240

 

COULOMBAC ATTRACTION ENERGY OF A0JOINING IGNS, /7 = 7. o8+ r_i]
ACIMC FLUORDES

——— BASIC FLUOQRIDES

 

g oo 1

o 20 G el 80 100

 

—_ —. BASIC SOLUTIONS ACIDNIG SOLUTIONS
Zri, 1M SOLVENT (mole 7a)

Flg. 2.3.2. An Acid-Baose Scale for Predicting Neutral
Points of Solutes in NuF-Zer Melts (Tentative),

PERIOD ENDING MARCH 31, ) 957

attraction energy of odjoining ions (Z/R ratios),
obtained by dividing the valence of a cation by
the center-to-center distance between the cafion
and on odjoining fluoride ion, are, os explained
in the preceding section, energy terms which con
be used as a measure of the fluoride affinity of
A line joining the Z/R ratios for the
extremes of the compesition ronge is a very rough

a cation.

first approximation for evaluating the effective
Z/R values at
the neutral compositions, described in the preceding
section, the flyoride offinity of o salt is the same
os that of the solvent. Thus the corrected curve
for matching compositions to Z/R values is found

intermediate compositions. At

by plotting the neutral compositions for various
solutes.

Currently only cone intermediate neutral point
can be approximated. The factors which influenced
the choice of NGF-Zer (53-47 mole %) as a fuel
solvent served to give ¢ melt in which FeF,
behaves idte;e::lfw;.r.3 This point has been plotted
in Fig. 2.3.2 to give three points through which
to draw a curve. Since the liquid structure probably
changes around 25 mole % ZrF,, and since the
position of the curve is uncertain, the maximum
curvoture was located ot 25 mole % ZrF,. The
simplest application of the plot involves the im-
plicit assumptions that only negative deviations
and thot marked positive deviations are
artifacts. This assumption is valid to the extent
thot effects due to polarization of the ions are
absent and thot real positive deviations can be
regarded os evidence of polarization.

The purpese of such plots is to aid in predicting
the relotive acidity or basicity of a solute toword
This permits estimates of the relotive
negative deviotions of solutes and, hence, also
relotive solubilities in case the pure solute is
the saturating phase. An important use is the
selection which will
reduce cocrosion by complexing the corrosive
ageats rather than the corrosion products. Because
of uncertainties in theory, ion radii, and neutrcl
the plot

accur

a sclvent.
compositions

of solvent

compositions, is subject to revision.

 

B¢ W Blssd and G. M. Watsen, ANP Quar. Prog. Rep.
March 10, 1956, ORNL-2061, p 84.

127
ANP PROJECT PROGRESS REPORT

2.4. PHYSICAL PROPERTIES OF MOLTEN MATERIALS
F. F. Blankenship

YAPOR PRESSURES OF MOLTEN FLUCRIDES
5. Cantor

The importance of detailed knowledge of solt
vapor behovior in both engineering considerations
and in the solution theory of fuels hos focysed
attention on the need for determinations of the
vapor pressures of the less volotile components
For a theoretical understanding
of solutions of molten electrolytes, vapor pressures
provide in principle a direct means of obtaining
the thermodynamic activities of the constituents.
In practice, the presence of associated molecules
in the vapor phase greatly complicates this method
of arriving at activities. It was formerly thought

in fue! mixtures.

that temperatures high enough to melt the olkali
halides should prevent the existence of all but
the simplest, or monomeric, in the
vapor; however, this is not the cose. As was
recently shown convincingly by Miller and Kusch,'

molecules

there are appreciable quantities of dimeric and
trimeric molecules in the vapor of alkali holides.

In order to proceed with a determination of
vapor species, the absolute vopor pressures of
molten olkali fluorides were meosured by the quasi-
static methed? used in this laboratory for obteining
vapor A sufficiently wide
temperature range was covered to give curvature

pressures of fuels.

 

]R. C. Millar and P. Kusch, J. Chem. Phys. 25, 860
{1958).

2W. H. Rodobush and A. L. Dixon, Phys. Rev. 26, 851
(1925).

on a log p vs 1/T plot. Hence the data, sum-
marized in Table 2.4.1, cen be more occurately
represented by means of the Kirchhoff equotion,
A
IngmmHg sl BlogT + C

The constants were evaoluated by following ¢
procedure resembling that used by Kelley.® Com-
parison with previously reported absolute pressure
meaf:us.-..nremems't'5
the uncertainties in the extropolations (as much
as 350°C) required to reach the ronges shown in
Table 2.4.1.

Transpiration or corrier gas methods were used
at lower temperatures for LiF,% NaF,” and RbF®
at Baitelle Memorial Institute, and for MNaF and
KF by MNaryshkin.” In these methods, however,
the weight of vapor required te saturate o known

is scarcely feasible because of

3. K. Keliey, Bulletin 383, Bursou of Mines, 1935,

JiH. von Wartenberg ond H. Schulz, Z Elektrochem.
27, 568 (1921},

50. Ruff, G. Schmidt, end 5. Mugdan, Z. anorg. aligem
Chem. 123, 83 {1921).

8E. M. Simons ond R. . Carlson, Progress Relating
fo ANP Applications Augusi~October, 1956, BMI-1148,
p 19 (Nov. 27, 1956).

7K, A. Sense el al., Vapor Pressures of the Sodium

Fluoride-Zirconium Fluoride System and Derived Informa-
fion, BMI-1064 {Jon. 9, 1956).

8R. W. Dayton and C. R. Tipten, Progress Relating to
Military Applicairons During Ocrober, 1936, 8MI-1145,
p 85 {Now. 1, 1954).

9|. I. Naryshkin, J. Pbys. Chem USSR 13, 528 (1939).

Toble 2.4.1. Yopor Pressures of Pure Alkoli Fluorides

 

Experimental Ronges

 

 

Compound Equation Temperature Pressure
K} {mm Hg)

LiF Log p = —(13,890/T} — 4.021 log T + 23.165 13281575 1.3-30.8
NoF Log p = —(14,60%/T) — 3.518 log T + 21.863 1381-1592 726,06
KF Log p = —(11,910/7) — 3.518log T + 21.002 11991505 1,8-81.4
RbF Log p = —(11,373/T) — 3.518 log T + 20,995 1154~1453 2.2-106
CsF Log p = —{9,845/7) — 3.518 log T + 20.556 1009-1379 1.7-242

 

 
volume of carrier gas is measured, and absclute
vapor pressures cannot be obtained without an
independent determination of the molecular weight
of the vapor. The molecular weights that must be
used to bring tronspirotion data into agreement
with the obsolute pressures givea in Table 2.4.]
have hkeen calculoted, in some cases, where
svitable data were avoiloble.  The molecular
weight calculations were based on the ossumption
that the vapor consisted only of monomeric and
dimeric molecules, and the results, shown in Table
2.4.2, will undoubtedly be subject to considerable
revision as better measuremenis are made. The
increasing proportion of dimeric molecules with
increasing temperature in the case of LiF is not
as imprebable as it might seem at first glance.
The wapor is in equilibrium with the liquid, and,
if the pressure effect is grester than the tem-
perature effect (heat of wvaporization is greater

Table 2,4.2, Tentative Estimates of the Percentoge of
Monomerle Molecules In the Scturated Yapor of Yarious
Molten Afkali Fluorides

 

Menomeric Molecules in Yapor (%)

 

 

Temperature
(CK) LiF NaF KF RbF
1400 52.9 48
1200 69.2 45,2 76.5
1200 82.1 {(Solid) 8.7 76.3
1100 {Solid) {Solid) {Solid) 70.5

 

PERIOD ENDING MARCH 37,1957

than the heat of dissociation}, the proportion of
dimeric molecules should increase.

SURFACE TENSION AND DENSITY
OF MOLTEN UF,

S. l.anger

A preliminory value for the density of molten
UF, was obtained for use with previously reported
sessile-drop dota ' to give the surface tension of
molten UF,. At 1075°C, a density of 6.80 + 0.09
g/cn® was found by using grophite pycnometers.
There were some difficulties in calibration of the
pycnometer which prevent this velue from being
regorded as final. |t leods to o surface tension
of UF, of about 180 dynes compared with the
value of chout 140 dynes found from the pre-
liminory estimotes given in the previcus repon.w
Further experiments are currently being made in
order to obtain a firm value for the density.

The tentotive experimental velue given obove
for the density of liquid UF, is, surprisingly,
greater than the density of solid UF, {6.63 q:.],»"'t:m3
based on x-ray dato).!! The greater density of
the molten material and the behavior of the salt
surfoce leve! upon freezing in the pycnometer
appear to indicate that UF, either controcts or
expands only very slightly upon melting.

 

10 Longer, ANP Quar. Prog. Rep, Dec. 31, 1936,
ORNL-2221, p 152,

PR, D. Burbonk, The Crystal Structure o Uranium
Tetrafluoride, K-769, p 14 (June 6, 1951}
ANP PROJECT PROGRESS REPORT

2,5. PRODUCTION OF PURIFIED FLUQRIDE MIXTURES
G, J. Nessle

PREPARATION OF VARIOUS PURE FLUODRIDES
B. J, Sturm L. G, Overholser

Preparation of YF3

The continving interest in moderators for use at
high temperotures encouraged considerable study
of the hydride of yttrium. Since the hydride is
prepared from metallic yttrium, which is made by
reduction of YFS, it has been necessary to prepare
considerable quantities of YF, for eveluation in
the metal manufacturing step.

The method of preparation used ta date has been
to dissolve Y,0, in the minimum quantity of
aoqueous HC! solution and to precipitate the
fluoride by the oddition of aqueows HF sclution.
The solid product is washed by decantation,
recovered by centrifugation, and given a pre-
liminary drying at 150°C in air.

The c¢rystalline product is dried for 4 to & hr
at 250°C under flowing helium in nicke! equipment
end then treated, in this equipment, with anhydrous
HF at 700 to 750°C for 12 hr. About 1400 ¢ of

YF, has been prepared in this fashion,

Preparation of Fluorides of Molybdenum

A previous report in this series! described the
preparation of what appeared to be a lower fluoride
of molybdenum by reduction of MoF, with FeF,.
tn odditional experiments this reduction product
has been collected in a condenser maintained at
110°C and subsequently redistilled in glass appa-
ratus. The yellow product, which is very unsiable
toward traces of moisture, melts at 78°C. This
is considerably higher than the 60°C previously
reported! as the melting point of the material
before distillation,

The unstability of the material to air and moisture
mokes it ditficult to obtain a clear x-ray pattern.
The best data available to date indicate the solid
to be tetragona) with a = 11.62 A ond ¢ = 9.26 A.
The available chemical analyses do not serve to
demonstrate conclusively that the moterial is
MoF,.  Additional studies will be required to

 

'8, J. Sturm and L. G. Overholser, ANP Quar. Prog.
Rep. Dac. 31, 1956, QRNL-2221, p 153.

130

identify the compound or compounds in this yellow
material.

Preparation of Other Materials

The method of preparation of FeF, from
FeCl,-4H,0 has been improved considerably by
partial dehydration of the commercial materici
before the hydroflucrination step. The substitution
of anhydrous CrCl, for bydrated CrF, as the raow
materiol for CrFa preparation results in a pure
product; this improvement is entirely due to the
higher purity of the commercial chleride,

Hydrofluorination of CrCl, ot 700°C yields a
product coniaining o smali quantity of Cng. Most
of this contominant can be removed by treatment
with H, at 700°C. When this reduction is per-
formed in silver vessels, less Cr% contaminates
the product than when nicke! vessels are used.

FILOT-SCALE PURIFICATION OPERATIONS
C. R. Croft J. Truitt

During the quarter the pilot-scale equipment (5-,
10-, and 50-1b capacity) processed 42 batches
totaling 1575 |b ot the wvarious fluoride compo-
sitions needed in phase equilibrium studies,
physical property measurement, small-scale cor-
rosion tfests, and other experiments. A large
fraction of this material was MNaF-KF.-LiF-UF,
{11.2-41-45.3-2.5 mole %). Usage of this material
seems to be fairly constant at about 500 Ib pes
month,

Nickel storage receivers have been shown to be
inadequate for the NoF-KF-LiF-UF, material. The
lorge expansion of the moterial upon melting
results in failure of 75% of the 50-Ib storage
containers upon melting of the charge. Inconel
cans have, however, been shown to be satisfactory
for the 50-Ib botches. No failures have occurred
in the ten tests made to date; it is not certfoin,
however, that Inconel cans will be satisfactery
for 250-1% or lorger charges. When inadverteny
exposure of the molten salt mixture to air or water
securs in Inconel cons, extensive contamination
of the melt with Crf, ond Fef, results. Since
these contaminants cannoi be removed economi-
cally by retreatment of the batch, the material must
be discarded.
PRODUCTION-SCALE OPERATIONS
J. E. Eorgan

During the quarter 46 batches totaling approxi-
mately 11,500 |b were processed in the production-
scole focility. The transition to a seven-day
continuvous production schedule was begun on
November 5, 1956 ond full capocity of 4000 Ib per
month was attained in December 1956, No failures
of the cepper-lined stainless-steel reactor cans
occurred during the quarter. The reactors now in
service have each processed neorly 10,000 b of
material since their installation. If the supply
of storage cons can be maintained and no major
breckdowns occur, the present rate of production
con apparently he maintained indefinitely.

This acceleroted program was adopted ot the
request of the AEC to meet the estimated re-
quirements for processed fluorides to be used in
the ANP program by Pratt & Whitney Aircraft and
by ORNL. During the last month of this quarter,
however, the Pratt & Whitney requirements were
greatly reduced. The projected total requirements
for the last half of fiscal year 1957 dropped from
37,000 1o opproximately 12,000 |b. Because of
the firm commitments made for the large omounts
of row materials, the personnel, and the operating
equipment oaeeded for the highly accelerated
program initiated in MNMovember 1956 and becouse
of the economies achieved by continuous operation,
it was deemed advisoble to maintain the high
oroduction rate until the end of this fiscal year.
However, the main objective has changed from
meeting immediate requirements to stockpiling
sufficient quantities of processed fluorides for
maintaining the ORNL effort in the ANP program
in fiscal year 1958. If this objective is attained,
the production focility will not be operated during
fiscal yeor 1958, except on a limited basis as
dictated by speciol large-scale demands of nen-
standard fluoride compositions, At present the
estimated ORNL-ANP requirements are foreseen
as some 14,000 |b of processed fluorides for fiscal
year 1958,

Sore 18,000 of the 30,000 Ib of NaZrF, ordered
has been received. Difficulties encountered by
the vendor in eliminating oxide contamination and
adjusting the NoF-to-ZrF, rotios have been re-
solved by a visit to the vendor's plant. Subsequent
shipments have had the proper ratios, and the
previous high purity levels have been meintained.

PERIOD ENDING MARCH 17,1957

New NaF has been obtoined to replace the high-
woter-content NaF previously received ond used
as described previously.? The new maferial meets
specifications, in genercl, but some of the drums
have a prohibitively large quantity of sulfur
With the availability of good
quality NeZrF, and the resultant low vsage of
NaF, it is becoming more feasible to restrict the
use of NaF to cp-grade moterial.

The design and construction of the Y-12 fluoride
processing focility, which would hove permitted
much greater throughput, has postponed
indefinitety.

(2000 ppm or more).

been

Reprocessing of a low-hafnium barren fluoride
mi xture (NaF-Zer, 50-50 mole %) to remove
excessive oxides and oxyfluorides wos successful.
The matericl wos then certified for use in the
Pratt & Whitney high-temperoture critical experi-
ment,

BATCHING AND DISPENSING OPERATIONS
F. A. Doss D. C. Wood

A total of 15,700 b of processed fluorides was
di spensed during the quarter in botch sizes ranging
from 1 t0 250 |b. At the beginning of the quarter,
with the onticipoted accelerated production ac-
tivity, the stock inventory was completely modified
ond redefined. Cleaning moterial previously in-
ventoried was removed from the inventory, and
compositions were listed so that
“‘Special Compositions'
sitions besides those specifically numbered. As
a result, the present inventory figures differ con-
siderably from those of the previous guarter, but
the totals differ only some 900 |b, or the amount

the heading
contained all compo-

of cleaning fluoride mixture previously carried in
the inventory. Also not listed in the inventory is
some 2500 |b of stored salvoge material being
stockpifed for use by the Chemicsl Technology
Division for testing purposes.

A material balance for the past quarter is given
in Table 2.5.1. As can be seen from the figures
presented, a very sharp increase in production and
dispensing occurred during the quarter as a result
of the accelerated pregram initiated in November
19546. The lorge shipments to Pratt & Whitney
Aircraft and some increose in local consumption

y ———— - ————

 

2| E. Eorgon, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 153.

131
ANP PROJECT PROGRESS REFORT

Table 2.5.1. Material Balanee for Fluoride Mixture Production and Use

 

 

 

 

 

 

 

 

 

Material {Ib)
304 3? 45¢ 1084 107¢ Special Tatal
On hand ot beginning of quarter 1452 1439 1787 994 564 1684 7,920¢
Produced during quarter 3703 690 2343 5279 1386 82 13,483
Total 5155 2129 4130 65273 1950 1766 21,403
Dispansed during quarter 3147 1346 3573 5519 1327 771 15,703
On hand ot end of quarter 2008 763 557 754 623 995 5,700

 

“*Composition 30: NnF-Zer-U Fd {50-46-4 mole %).
bComposition 31: NaF-ZrF, (50-50 mole %).
“Composition 45! NaF-ZrF4 {53-47 mole %),

d(:ornpcsifion 108: NflF-:".'r|"'4-|..l|=_‘1 (56-37.5-6.5 mole %Y.

Compasition 107: NaF-K F-LiF-UF, (11.2-41-45.3-2.5 mole %).

”Revised figure obtained by subtracting 907 |b of cleaning fluoride mixture from last quarter’s bolance.

resulted in an appreciable drop in stock inventory
by the end of the quarter,

The main consumers of processed fluoride mix-
tures and their alletments during the quorter are
given below:

 

Consumer Amount (Ib)
ORMNL-AMP groups
For chamical and physicel property 284
studies
For experimental engtneering tests Z,444
For metallurgical studies and fuel 1,948
reprocessing develiopment
Pratt & Whitney Aircraft
For experimental use at Hartford 7,500
For high-temperoture criticel 1,366
experiments at OQRNL
Other contractors, including BMY, NRL, 187
and WADC
Salvoge ond reprocessing 1,974
Taral 15,703

132

SPECIAL SERVYICES

F. A, Doss J. Truiit
D. C. Wood

Filling, Draining, and Sampling Operations

Approximately 4000 (b of processed fluorides and
2500 |Ib of liquid metals were charged into test
The charge sizes ranged from 1 to 500 Ib,
and some 170 operations were performed.

units,

Fitling of Pratt & Whitney High-Temperature
Critical Assembly

The Prott & Whitney high-temperature critical
ossembly wos charged with low-hafium-content
NaF-Z¢F, (50-50¢ mole %) and the enriched con-
centrate NuF-UFd 66.7-33.3 mole %). The entire
filling operation proceeded very smoathly, but the
first charge of barren material hod to be withdrawn
from the sump shortly after it wos loaded because
of a leak in the bottom of the sump. It was
possible to withdraw this material in a cendition
suitoble for re-use, and, after the sump was
repoired, the barren material was again charged
inte the sump and enriching was started. The
loading operation begen on Janvaery 9, 1957, and
was compteted on February 4, 1957, at which time
the responsibility for loading of the assembly wos
transferred to the Operations Group of the Criti-
cality Section for final additions of enriched
uranivm with the ART enricher apparatus.
PERIOD ENDING MARCH 31, 1957

2.6, COMPATIBILITY OF MATERIALS AT HIGH TEMPERATURE

F. Kertesz

PFPENETRATION OF GRAPHITE BY MOLTEN
FLUORIDES

H. J. Buttrom

The study of the behovior of graphite exposed
to molten fluorides was continued.! Tests were
conducted to evaluate a commercial graphite
specimen that was impregnated with NoF-ZrF,
(53-47 mole %) before being exposed to a fluoride
mixture containing uranium. The impregnation of
commercial APC graphite specimens was ac-
complished by immersing them for 24 hr in the
fluoride mixture melt kept ot BOOPC. After this
treatment, the impregnated specimen was heated
in vacuum for 24 hr at S00PC in order to remove
part of the ZrF, and thus to obtain a residual
mixture with o higher melting temperature. [t
would be expected thot on impregnating material
that was solid when exposed to the uronium-
bearing melt would resist penetration more suc-
cessfully than one thot was liquid and could thus
diffuse from the pores.
and a control sample that had not been heated in
vacuum wete submitted for analysis. The results
of the analyses are presented in Table 2.6.1, along
with the calculated Naf-to-ZrF | rotios.

The analytical results indicote that the NaF
penetrated the graphite to o greater extent than
the ZrF, did, by a facter of 12 (4.16/0.34). The
sublimation step effected a further separation of
the ZrF,, as wos expected, and ofter sublimation
the Naf-to-ZrF, ratio was further increased about

Both the treated specimen

 

]H. J. Buttram, ANFP Quar. Prog. Rep. Dec. 31, 1954,
ORMNL-2221, p 158.

fourfold, with the final rotio of NaF to ZrF  being
about 16:1,

Two other types of graphite, code Nos. R-28
and M-80, obtained from the National Carbon
Company were also tested by impregnation, os
described above. The samples impregnoted with
NaF-ZrF, (53-47 mole %} were subsequently im-
mersed into NaF-ZrF,.UF, (53.5-40-6.5 mole %).
The olpha counts obtained from these specimens
were compared with the counts obtained with
specimens immersed directly into the uranium-
bearing mixture without preliminary impregnation.
The data obtained are presented in Table 2.6.2.
The results indicate that the grophite specimens
were not penetrated to any consideraoble degree by
either of the salt mixtures, and thus these speci-
mens compare favorably in this respect with the
sample obtoined from the Degussa Company of
Germany and, in oddition, appecr to have greater
strength.

ELECTROLYSIS OF THE MeOH-MNi SYSTEM
F. A, Knox

In the course of electrolytic studies of the
NaOH-Nj system it has been established that, in
a nenisothermal system, the hot nickel electrode
is usually onedic and the cold electrode is
cathodic. It has been found, however, that this
situation can be reversed by the use of a hydrogen
atmosphere,? Since NaOH is a potential moderator-

coolant material, attempts are being made to

 

2.l'-'lL_ Nichols and D, R, Cuneon, private communication
to F. A, Knox.

Toable 2.6.1. Analyses of the Fluoride Mixture NuF-ZrF“ Before and After Penetration into Graphite

 

 

 

 

NoF Zrf, NaF-to-ZrF
{wt %o} {wt %) Rotie
Original fluoride mixture 25.3 74.7 0.34
Flucride mixture n grophite specimen B0.&6 19.4 4,16
Fluoride mixture in graphite specimen ofter sublimation 24.0 6.0 15.80

 

133
ANP PROIECT PROGRESS REPORT

Table 2.6.2. Penetration of Graphite Specimens by l"lnt:lF-ZrF'_,L-UF4 {53.5-40-5.5 Mole %} in 24 hr ot 800°C

 

Alphe Activity {cpm)

 

 

Initial Weight Weight After
Graphite Specimen Aftar Exposure to 1’, in 'I/ -
Specimen Weiah Treatmant . . . B 4
gight Impregnation Uronium-Bearing At from i
Code Mo, ( . g ° rom
9) (o) Mixture veface Surfoce  Surfoce
R.28 093117 Placed in Nu::|F-Zrl:d-UF‘1 0.9328 >1 1 Q
{53.5-40-6,5 mole %) as
recaived
R-28 0.9394 Impregnated with NuF-Zer {.9585 0.9544 >3 0 0
(53-47 moie %) before being
placed in NcF-ZerfiU Fq
{53.5-40-6,5 mole %)
M-80 0.9474 Ploced in NoF-Zer-U F‘1 0.9518 1 0
{53.5-40-4,5 mole %)} as
received
M-80 0.9574  lmpregnated with MaF-Zr Fd (0.94680 0.94664 0 > >1

{53-47 mole %) before being
placed in NoF-Z¢F -UF .
{53.5-40-£.5 maole %)

 

balance the current in an effort to find a way to
eliminate the mass-transfer effects observed in
sodium hydroxide circylated under a temperature
gradient.

The experimental arrangement being used con-
sists of three closed concentric nickel crucibles,
which serve os electrodes, as shown in Fig. 2.6.1.
The outer vessel is 6 in. long and 21{1 in. in
diemeter. The middle container is ]]é in, in di-
ameter and 8 in. long, and a small hole was drilled
in the bottom to atiow the molten sodium hydroxide
to rise to the same level as in the outer container,
The inner electrode consists of a ]é-in.-dio nickel
tube, The bottem of the tube was c¢losed ond o
thermocouple was welded inside. For tests, the
three electrodes were placed concentrically in
the furnace in o controiled atmosphere. About
150 g of sodium hydroxide was used as the charge.
in this orrangement the outer electrode had hy-
drogen gas on one side and molten sodium
hydroxide on the other side. The middle electrode
had melten sodium hydroxide on both sides. The
ettect of voricus arrangements could be studied
by connecting the electrodes to the positive or
negative poles of batteries,

In one experiment a relatively low test temper-
ature (550°C} was used; the middle electrode was

134

connected to the negative peole; and the inner
electrode was connected fo the positive pole of
the outside bottery. Current and voltage measure-
ments were made by slowly raising the applied
voltage and meesuring the corresponding current.
Two sets of measurements were made with the
system under o hydrogen otmosphere ond also with
the system under vacuum.

The results, plotted in Fig. 2.6.2, show that
under a hydrogen atmosphere the current increosed
much foster than it did under wvacuum, which
indicates that hydrogen participated in the current-
carrying mechenism,

Measurements were aiso made at 800°C, with o
hydrogen atmosphere, for which the curment flowed
either from the inner electrode to the outer elec-
trode or in the reverse direction. |t may be seen
in Fig. 2.6.3 thot, when the current was flowing
from the inner electrode toward the middle elec-
trode, the current increased continuously with the
voltage. In the reverse case o definite ploteau
was reached, that is, as the voltoge was increased,
the current remained constant up to a value some-
what below 2 v. A theoreticol interpretation of
these results will require ¢ more systematic study
of the effect of current density, The polarization
of hydrogen electrodes is probably invelved.
UNCLASSIFIED
ORNL- LA~ DWGE 20255

THERMOCQUPLE LEADS

|| INNER ELECTRODE, Yo -in - 0D
I GAS AVAILABLE TO INSIDE QF
ELECTRODE I CONTACT wiTH

|
ARaRATuS IN !‘ | MoOH ON QUTSIDE.

CONTROLLED
LTHOSPHERE

 
   
   
    
 
    

(Il
MIDDLE ELECTRODE, 1'%, -0n - OD
| NoCH D BOTH $I0ES OF Wall
BELOW LIGUID LEVEL .
QUTER ELECTRDNE, 2% -in-00
[ NaOH ON WSIDE, GAS ON OUTSIDE,

 

 

  

[ MOLTEN NaOH I il
V27 aND OUTER CHAMBERS

.

e l=d
- _f ~ L2 THERMOCOUPLE WELDED TO
|| ¥ BOTTOM OF INMER ELECTRODE

Fig:s 2.6.1. NaOH-NI Electrolysis Apparatus.

 

 

 

 

 

PHYSICAL PROPERTIES OF ELASTOMERS
EXPOSED TO ATTACK BY LIQUID METALS

D. Zucker

The first in-service test of valve seot materiols
exposed to helium saturated with NaK veopor was
completed in the apparotus described previousty.”
The system and valves were opened and washed
with water, reassembled, and tested with helium
at 30 psi. The appearance of the valve seats and
their performance after reassembly indicate that
General Electric SE-550 and SE-450, Dow-Corning
Silastic 80, and Ou Pont 5570 withstoed NaK

SD. Zucker, ANFP Quar. Prog. Rep. Dec. 31, 193¢,
ORNL -2221, p 151.

PERIOD ENDING MARCH 31,1957

UNCLASSIFIED
ORML-LR—-DWG 20256

 — g

 

  
 

 

a
£ f
o
: - HYOROGEM ATMO$PHERE§ I
S . o el
z = ——— = =
i . — —
i, | e - - |
e
u i 4 S —
605 = _ ]
) 1 |
0.02 ! L !
]
,O 01 i . —— . = —_— ——— -
o .4 08 |2 1.6 2.0 2.4
WOLTAGE (v)
Fig. 2.6.2. HNoOH-Mi Voltage-Current Relationships

Under Hydragen or Yacuum at 550°C.

WHNCLASSIFIED
DRNL—LR—DWG 20257

 

 

    
  
 

— 1
“CURRENT FLOWING —

FROM INNER TO —
_MIOOLE ELECTRODE o8

A 1)

T CURREMNT FLOWING FROM
MIDDLE TO INNER ELECTRODE

 

 

 

a ol e S0
s |
£
o |
= = = =
z . —1 = =
Lt = = | {
&
= i
5 = —
o
0.5 — — =
| - = 4
|
02 e e =
i b S e
o Q4 o8 b2 16 20 2.4
VOLTAGE (v}
Fig. 2.6.2. MNaOH-Ni ¥oltage-Courrent Relationships

Under Hydrogen ot 804°C,

135
ANP PROJECT PROGRESS REPORT

attack fairly well, with SE-5350 possibly being very
slightly better thon the other moterials. These
materials connot be used directly in the Skinner
electric valves currently being supplied for use
in the ART, but they might be acceptable if the
edge of the volve orifice were rounded,

136

A cursory test also indicated that curing by
heating (200°C) and pressurizing of these materials
for 24 hr increased their resistonce to cottock.
The reason for this is not obvious, since these
elastomers were probably heat treated previously
at higher temperatyres although not under pressure.
n

PERIOD ENDING MARCH 31,1957

2,7. ANALYTICAL CHEMISTRY

J. C. White

DETECTION OF TRACES OF NoK IN AIR
A, 5, Meyer, Jr. J. P. Young

The apparatus for the detection of submicrogram
quantities of sodium (NaK) in air wos assembled and
subjected to preliminary testing.
made with this instrument, which was discussed

Measurements

previously,! are based on the absorbance of sodium
resonance radiation by sodium atoms, The instru.
ment is illustrated in Fig. 2.7.1, In the compartment
on the right, the light from o sodium vepor lomp
(not shown) is collimated by a simple lens. The
collimated light posses through a mechanical
chopper, which consisfs of o disk perforated to
pass light through three equally spaced 60-deg
sectors. The disk is rotated ot 30 ¢ps by a synchro-

nous motor. The interrupted light passes through

 

‘A. S, Meyer, Ir., et al., ANP Quar. Prog, Rep. March
10, 1956, ORNL-2061, p 207.

J. R, Sites

limiting circular apertures 3’2 in. in diometer to
produce the two beams of light shown in the central
porticn of the photograph, The copertures con be
rotated about axes through their centers that are
perpendicular te the radius of the chopper in order

Under

conditions of exoct alignment and balance the light

to balance the energy in the two beams.

ts collimated into two beoms of equal intensity, and
eoch beom is interrupted ot a frequency of 90 cps
in a manner which maintains the total ensrgy of the
two beams constant.

In the
through a 20-mp bandspass interference filter, and
the sodium-sodium doublet at 5890 A is isoloted.
Both beams of light are focused by a converging
lens on the cothode of a photomultiplier tube. The
output of the photomultiplier tube is fed to a $0-cycle
tuned amplitier 1o yield o signat that is proportionasl
te the difference in intensity of the two beoms of

light.

second comparimen! the light passes

UNCL ASSIFIED
FHOTS 19304

AMPLIFIER

 

il FOWER SUPPLY
~

 

 

A P

Fig. 2.7.1. Apporatus for the Detection of Submicrogram Quantities of Sodium in Air.,

137
ANP PROJECT PROGRESS REPORT

In the tests which were carried out, a 5-in.-1D fube
furnace was placed between the two optical com-
ponents of the instrument. Absorption celis, con-
structed of 24-in. lengths of :}:‘-in.-diu Inconel
tubing and fitted with end windows of optical
quartz, were ploced in the furnace so that each
light beam traoversed an chsorption cell, After the
furnocce was heated, the optical components were
aligned and the instrument was balanced by rotating
the spertures until a minimum signal wos abserved
A sample of
sodium-containing air, which wos prepaored by
admitting meosured quantities of helium thot had
been saturcted with sodium vapor to a stream of air,
was introduced into one of the absorption tubes, and
the deflection of the cutput meter wos observed.

When o sample calculated to contain 25¢ ppb of
sodium was introduced into the absorption tube,
the observed signal produced o deflection equal to
twice that of the original balance of the meter. On
the basis ofthe observed stability of the instrument,
it was found that 100 ppb of sodium could be easily
detected. As was reported,? however, the technique

on the ocutput meter of the amplifier,

used for the preparation of somples of sodium-
containing air did not yield quantitative fransfer of
sodium to the air, ond thus these preliminary
measurements may not truly reflect the sensitivity
of this instrument, It wos found that no evidence of
sodium could be observed when the effluent gases
containing the sodium from the ocbsorption tube were
pessed through an air-propane flame, It oppears
that the stipulated limit of detection of 10 ppb of
sodium can be achieved by oppropriate sampling
techniques and refinements in the alignment of the
instrument.

Tests were carried out for selecting o suitable
material tc use in the fabrication of orifices with
which to inject small quantities of NaK into air,
Zirconivm was the only satisfactory material found
among the materials tested. A 0,003-in, orifice in
crystal-bar zirconium was entarged less than 20% by
the ejection of NaK for o period of 30 min at o
temperature of 450°C., Under these conditions
nickel, type 316 stainless steel, and
molybdenum are severely corroded by the oxides
of the alkali metals. Measurements of the flow
rates of NoK through small orifices indicate that
orifices approximately 0,002 in.

Inconel,

in diamerer will
be required in order to mcintain a sufficiently

 

QA. 3. Meyer, Jr., ond J. P, Young, ANFP Quar. FProg.
Sept, 10, 1956, ORNL-2157, p 129.

138

small How of NaK to provide fest samples for the
leok detectors. Since holes of this diameter connot
be readily mochined, electralytic methods of drilling®
ore being investigoted,

DETERMINATION OF OXYGEN IN METALLIC
LITHIUM

A. S, Meyer, Jr.
G. Goldberg R. E. Feathers

A method is being tested for the determination of
traces of oxygen in metallic lithium that is based
on the following reactions

2Li + 2CH,0H ——> 2Li0CH, + H,
ti,0 + 2CH,OH ——— 2Li0CH, + H,0
LiOCH, + C,H O, (solicylic acid) ———>

LiC,H G, +CH,OH
and the subsequent measurement of the water by the
Karl Fischer titration.

developed by Nuclear Development Corporation of
America,® offers a unique advantage in that the

The procedure, which was

sample can be washed free of surface contamination
immediately before the determination is perfarmed.

The determination is carried out in the apporatus
shown in Fig. 2.7.2. A sample {~] g) is placed in
the reaction vessel, and the opparatus is freed of
oxygen by evacuation and purging with helium,
Surface oxide is removed by transferring three 20-ml
portions of dried methanol from the titration flask to
the reactien vessel ond allowing eoch portion to
react for 1 min ot room temperature. After she final
wash solution has been withdrawn, a 60-ml volume
of methanol is transferred from the titration flask
to the reaction vessel, ond the sample is dissolved
by gentle heating. A 20-ml portion of a 50% solution
of salicylic acidin pyridine that has been pretitrated
to the Karl Fischer end point is admitted to the
titration Hosk, and, when the dissolution is com-
pleted in the reaction vessel, the sofution of lithium
methylate ond water is tronsferred to the titration
flask. The solution in the titration flask is then
retitrated to the Korl Fischer end peoint. The
titration is corrected by subsitracting the volume of
Karl Fischer reagent thot is required to titrote a
60-ml portion of methanol.

 

3A. Uktic, Jr., Rew. Sei Instr. 10, 965 (1955).

I{N. l. Sox, private communication to J. C. White.
300-mi SEPARATORY
FUMNNE L

SOLUTION OF
SALICYLIC ACID
N PYRIDINE

METHANQL IN MACHLETTE BURET

HARL FISHER REAGENT ELECTROCES

N MACHLETTE BURET

-

&

250-mi TITAATION
FLASK

 

DRAIN

 

e

T
" MAGNETIC STIRRER

PERIOD ENDING MARCH 11,1957

WNCLASSIFIED
CRNL-LR-DWG 12874

FRESSURE RELELSE

\

{___—’_‘,} HEL UM

VaCluwm

. !

WATER JACKET

REACTION VESSEL
L —— LITHIUM SAMPLE

   

3 HELIUM
VaCcuUumM

g e e ——

Flg. 2.7.2. Apporotus for the Detarmination of Oxygen in Metallle Lithium by the Methanol Method,

The salicylic acid is used to drive the second
reaction to completion and to eliminate the inter-
The
contaminants, Li,C, and Li,N, which
interfere with the determination of oxygen by the
omolgamation procedure, are decomposed according
to the following reactions:

ference of lithium methylote in the titration.
common

Li,C, + 2C,H,0, —> 2LIC,H, O, + CH,

|

LiN + 4C,H Oy ——— 3LIiC,H,0, +
NH4C1H503

Since at least 60 ml of methanol is required to
dissolve a 1-g sample of lithium metal, the methanol
must be thoroughly dried to avoid high blank fti-
trations and consequent loss of precision in the
determination of traces of oxygen. Metheds for
drying methonol were tested that inciude distillation,
possage through chrometegraphic columns packed

with molecular sieves, vopor phase drying over
molecular sieves, anddrying over oxides of alkaline-
earth metals. Theconcentration of water inmethanol
was reduced to less than 10 ppm by distittation
through a 100-plote fractionation column after
preliminary drying over barium oxide. Alkoline-
earth ond alkali metals cre being tested as drying
agents in an effort to find o less time-consuming
method of reagent preparation.

A pressurized apparatus for the dissolution of
somples of lithjum metal in ethereal solutions of
butyl iodide and iodine® was constructed, It is
anticipated that in the pressurized apparatus the
dissolution of samples can be carried out with
smaller volumes of liquid reagents than are required
for the dissolution of atmospheric pressure, and
thus the effect of contaminents in the reagents con
be reduced.

 

SA. S. Meyer, Jr., et al., ANP Quar. Prog. Rep. Sept.
10, 1956, DgNL-ZIS?, p 127.

139
ANP PROJECT PROGRESS REPORT

APPARATUS FOR SAMPLING AND ANALYZING
ALKAL) METALS

A, S. Meyer, Jr.
R. E. Feathers G, Geldberg

A convenient receiver and transfer vessel was
designed for the handling of samples of alkali
metals, The opparotus, which is shown in Fig.
2.7.3, consists of a Jamesbury volve fitted with a
vacuum-tight compaortment at one end and coupled
by a Tygon sieeve to o stondard taper joint ot the-
opposite end. When the apporatus is fitted 1o the
complementary joint of the analyticel apparotus,
the tronsfer of the somple of alkali meta! con be

UNCLASSIFIED
ORNL-LR-DOWG 19875

- f}}il

|1 [l =——29/42 STANDARD TAPER
ol PYREX JOINT

TYGON TUBING

   

 

I H—u STAINLESS STEEL PIPE

=—- TEFLON -PACKED
JAMESBURY VALVE

 

 

 

 

 

 

—STAINLESS STEEL
SAMPLE RECEIVER

 

Fig. 2.7.3. Tronsfer Receiver for Alkali-Metal Somples.

140

corried out without atmospheric contamination, The
transfer vessel can be loaded in a suitable inert-
atmosphere dry box. It is also ideally suvited for
use with the modified Mine Sofety Appliances Co.
bucket sampler,® Sample buckets filled with NaK
can be tronsferred if the apparatus is kept under
refrigeration,

A high-vacuum dry box that will have o 10-ft? work
chamber is being assembled for processing samples
of alkeli metals. It is to be equipped with an
evacuable entrance port, a high-capacity diffusion
pump, and a purification train for inert gases which
includes heated scrubbers containing titanium and
Nal(,

An dpparotus for the purification of mercury by
distillotion was constructed and is now in service.
The two-stage still used for the purification hos @
totel capacity of 30 |b of mercury per day. The
apparatus is used to remove traces of oxide ond
water from mercury in order to eliminate errors in
the determination of oxygen in sodium by the
amalgomation method and also to reprocess the
large quantities of mercury which are reguired in
the amalgamation procedure,

ANALYSESOF LITHIUM-BASE FLUORIDE 3ALTS
FOR METALLIC IMPURITIES

The development program on the application of
nickel-molybdenum atloys as container materials for
fused lithium-base flyoride salts has necessitated
o number of modifications in existing methods for
the analysis of metallic impurities in the fluoride
salts. In additien to the usuval determinations fer
iron, chromium, and nickel, the sclts are presently
being onalyzed for molybdenum, niobium, titanium,
tungsten, vonadiuvm, and aluminum, Methods for
determining all these metals were developed previ-
ously, but some modifications were necessory in
the methods used for the aluminum, vanodium, and
titanium analyses in order to eliminate interferences
introduced by the presence of other metals. The
status af the work on the modifications is given
in the following sections.

Aluminum
W. F. Vaughan

The determination of microgram quantities of

aluminum in fluoride soits with the use of the

 

 

Y Goldberg, A. S Meyer, Jr., ond J, C. White, The
Sampling of Alkali Metal Systems with the Modijied
MSA Sampler, ORNL-2147 {Aug. 21, 1956).
ammonium salt of aurin tricerboxylic acid (Aluminon)
is subject to interference from zircenium, iron,
titanium, ond vanadium, White and Meyer’ removed
these interfering metals by precipitating them as
the insoluble cupferrate salts. As a result of more
recent work by Ross® on the extraction of metal
ions with tri-n-octylphosphine oxide from ocidic

solutions, these metals are now removed by
extracting them from a chloride solution. The
separation is complete in obout 10 min. The
aqueous phase is then free of iron, titanium,

vanadium, zirconium, and uronium, and the aluminum
is essentially isoloted.

Vanadium

J. P. Young J. R. French

The method of Wise and Brandt® for the determi-

nation of microgram quontities of pentavalent
vonadium with benzohydroxamic ocid was applied
to fused fluoride salts, The method is based on
the reaction of vanadium(V} with benzohydroxamie
acid at ¢ pH of 2.0 to produce a complex which
exhibits on absorption maximum ot 450 my in a
The melar absorbance index of

this complex is 3500, A linear relationship exists

1-hexanol medium.

between cbsorbance ond the concentration of the
vanadium in the complex in the range of 1 to @
ig/ml. The coefficient of variation for this method,
based on standards, is 2%.

Most of the interfering ions, including nickel,
iron, chromium, ond molybdenum, are removed by
deposition at the mercury cathode. Although the
vanadium is reduced to the quadrivalent oxidetion
state by the mercury cothode, it remains in solution
and is accompanied by the uronium, alumiaum, and
titanium.  Argentic oxide, AgQO, is then used to
oxidize the vanadium to the pentavalent state. As
much as 50 mp of uraniy.n can be tolercted in ihe
final determination, When the ratio of either alumi-
num or titanium to vanadiom greatly exceeds five,
howewver, positive errors Miobium and
vanadivm ordinarily do not exist together in these
samples, so the possible effect of niobium on the
determination ef vanodium waos not studied.

result,

 

7), C. White and A. S, Meyer, Jr., Determination of
Traces of Aluminum in NflF—ZrF4—UF4, ORNL CF-56-310
{March 1, 1956).

8w, .. Ross, Extraction of the Elements with Trioctyl-
phosp hine Oxide from Acidic Solutions, ORNL CF-56-9-
18 (Sept. 6, 1958).

"W, M. Wise and W. W. Brondt, Anal. Chem. 27, 1392
(1955).

PERIOD ENDING MARCH 31,7957

Titanium

J. P. Young 1. R, French

The established method for the determination of
trace amounts of titanium'? in fused fluoride salts
include

titanium in the presence of niobium and molybdenum,

was extended to the determination of
Since niobiym and molybdenum interfere with the
determination of titonium by the Tiron merhod,w it
was necessary fo develop some means of removing
these interferences. The oddition of sodium hydro-
sulfite to reduce the iron{lll) also reduced the
molybdenum and thus decreased but did not coms.
pletely eliminate the interference of molybdenum.
In the determination of titanium in NaF.ZrF .UF ,
quantitotive separation of the titanium was achieved
by an amimonia precipitotion in the presence of
corbonate ion. The seporation was adequate to
eliminate the interference of molybdenum in the
determination of the titonium,

Quantitative separation of titanium from the
niebium was achieved by an ether extraction of the
The nicbhium thio-
cyanate complex extracted quantitatively into the
ether phase, whereas the titanium thiocyanate com-

two metals as thiocyonates.

plex did not extract. The titanium in the aqueous
phase is determined by the Tiron methed.

SPECTROPHOTOMETRIC DETERMINATION OF
NICKEL IN ALKAL] METALS

A, 5. Meyer, Jr. B. L. McDowell

The dioxime derivotive, 4-isopropyl-1,2-cyclo-
hexonedionedioxime, which was reported ' to be a
very sensitive recgent for nicke!, wos investigoted
further, In order to increase the sensitivity of the
determination of nicke! by extraction when the ratio
of the aqueocus volume to the orgonic veolume is
large, organic compounds were investigoted that
are less water soluble thon the chloroform used
initially, 12 Hydrocarbons, chlorinated hydrocarbons,
esters, alcohols, ond aromotic compounds were
tested as extroctonts for the nickel chelate, but

 

19, ¢ White, Determination of Microgram Amounts of
Titanium in the Presence of NdF-ZrF‘-UF‘i, ORNL CF-
366111 (June 20, 1958).

Mp, 1, Hooker and C. V. Bonks, Preparation, Proper-
tzes, and Analytical Applications of 3ome Substituted
Alicyelic Vie-Dioximes, |5C-597 {March 1955).

124 5. Meyer, Jr., B. L. McDowell, and R. E. Feathers,
.Jr.f ANP Quar. Prog. Rep. Dec, 31, 1956, ORNL-222],
p

141
ANP PROJECT PROGRESS REPORT

onfy benzene and xylene proved to be better than
chloroferm,  The melar absorbonce index for the
chelate at the absorption maximum of 385 my was
unoffected by the solvent, The partition coef-
ficients, however, were much higher for the chelate
in both benzene {1000} and xylene {5000) than in
chlaroform (300). la view of these more favorable
partition coefficients, it will be possible to de-
termine even lower concentrations of nickel thon
the previously postulated'? 0.5 ppm in a 5-g sample
of alkali metal,

The two eiements that most commonly interfere
with the determination of nickel in alkali-metcl
Methods for the
elimination of each of these elements were therefore
developed. Copper{!l) is reduced to copper(i), and
the reduced copper is complexed with thiocyencte
ion 1o prevent the extraction of copper as the
dioxime complex. The interference of iron(lIl} was
eliminated by the oddition of fluoride. Since the
Huoride ion inhibits the precipitation of zirconium
upon neutralization of the selution, this procedure
can also be applied to the determination of nickel
in solutions of zirconium- or lithium-base fuels,

samples are iron and copper.

DETERMINATION OF TRACE AMOUNTS OF
ZIRCOMNIUM IN SULFATE SOLUTIONS WiITH
PYROCATECHOL VIOLET

J. P. Young J. R. French

Experimental work on the determination of trace
amounts of zirconium '3+ in solutions of sulfuric
ocid with pyrocotechol violet wos completed, as
well as o study of the interference caused by
various cations., The data obtained in this in-
vestigotion are summarized in Tabie 2.7.1.

Microgram omounts of zirconium ia suliate so-
lutions can be determined in the presence of high
concentrations of thorium, wvranyl, cerium, ond
lanthanum ions, |t was possible 1o determine
zirconium with a coefficient of variation of 2%
even in the presence of more unfovorable rotios of
cerium and lanthanum by lowering the fina!l pH of
the test solution to 4.5.

The determination of zirconium with pyrocatechol
vialet is unaffected by the presence of anions such
as chloride, perchiorate, nitrate, or borate. Small

 

13_]. P. Young ond J. R. French, ANP Quar Prog,
Rep. Dec. 31, 1956, ORNL-222), p 165.

144, Floschko ond F. Z. Sodek, Z. anal. Chem 150,
339 {1954).

142

Table 2,7.1. Effect of Catlons on the Spectrophotometric
Determinotion of Zirconium ia Sulfate Selutions
wlth Pyrocatechol ¥Yiclet

Concentration of zirconium: 1.2 ug/ml

 

Weight Ratio Error Resulting from

Cotion of Cation te Pressure of Cation
Zirconium (%)
Th? 10 5
Lad? 30 5
Ce” 15 5
vo,?’ 30 0
TR 3 5
NI 15 0
Cr6+ 15 Q
ct? 15 0
Fod* 6 5
Fe2’ 6 5
sod? 6 5

Ti4%, v,

VAR YL 1 20

amounts of aniens such as citrote, oxalate, tartrate,
fluoride, and ethylenediominetetraccetate, however,
will destroy the color of the complex. The coef-
ficient of variation for this methed, based on the
determination of zirconium in stondords, is 2%.

CETERMINATION OF SULFUR IN
FLUORIDE FUELS

T. W. Gilbert, Jr.

The methylene-blue method'® for the determi-
nation of small omounts of sulfur in fused fluoride
tuels was modified by the substitution of a new
reducing selution for the mixture of red phosphorus,
hydriodic acid, and formic acid that was previously
used, The new solution, prepored by dissolving
stapnous chloride in dehydrated ohosphoric acid,

 

155, C. White and G. Goldberg, ANP Quar. Prog. Rep.
Dec. 10, 1954, ORNL-1816, p 131.
wds reported to reduce sulfates quantitatively to
hydrogen sulfide,'® and therefore experiments were
carried out to test the efficiency of this solution
for reducing microgram quontities of sulfote and to
test its applicability to the determination of sulfur
in fluoride salts.

Difficulty wos experienced with the red phospho-
rus reducing mixture when the sample size was
increased beyond 200 mg, and therefore the phos-
phoric acid mixture, which can conveniently handle
l-g samples, has en important advantage in that a
fivetold increose in sensitivity is achieved., A
second advantage of the phosphoric acid reducing
mixture is that the fluoride sclts are decomposed.
Uronium tetrafluoride is brought into selution, and
zirconivm  fluoride is converted to zirconium
phosphate,

The determination is carried out by adding about
20 ml of a solution of snCl, and phospheric acid
to o 1-g sample of fluoride salt. Boric anhydride
is then added to complex the fluoride and to mini-
mize etching of the gloss apparatus. The phos-
phoric acid solution is heated to 300 to 310°C to
reduce the sulfates to hydrogen sulfide, which is
passed into a solution of zinc acetate by means of
a stream of orgon. The quontity of hydrogen sulfide
is then determined by the methylene blue procedure.
The coefficient of variation of the method was found
to be approximately 10% for 5- to 20-pg quantities
of sulfur,

OXIDATION OF CHROMIUM WITH ARGENTIC
OXIDE

J. P. Yeung J. R. French

A method for the oxidation of chromivm with
argentic oxide, Ag(D, reported by Lingene and
Davis,'” was investigated for use in the spectro-
photometric determination of chromium with di-
Lingane ond Davis report that
argentic oxide serves advontagecusly as a very
powerful oxidimetric reagent in acid media. They
recommend AgQ for the oxidotien of manganese,
cerigm, or chromium prior to volumetric determi-
nations of these metals. Oxidotion with orgentic
oxide is ropid and complete for these ions in 2 to
5M acid at room temperature.

phenylcarbozide,

 

U1 Kiba er al., Buli. Chem Soc. Japan 28, 641
(1955).

”J. J. Lingane and D. G. Dawvis, Anal. Chim Acta
15, 207 {1958).

PERIOD ENDING MARCH 31,1957

A standard solution of potassium dichromate was
used in order to establish & calibration curve for
the chromium determination by the diphenylcorbazide
methed. The chromium in o portion of the standard
solutien was reduced with hydrogen peroxide to
vield o standord solution of chromium{lll). The
calibration curve obtained from the sclution of
chromiom(l1]) after oxidation of the chromium with
argentic oxide waos identical, within experimentol
error, with the ocne cbtained with chromium(VI). |t
is recommended that the solution be boiled for
several minutes to ensure the reduction of all the
excess silver(l}),

These investigations indicate that oxidation of
the chromivm with argentic oxide con reploce
oxidation with perchloric acid in the colorimetric
determination of chromium with diphenylcarbazide.
The use of argentic oxide results in @ considerable
saving of time in the oxidation step. [t is further
expected thot oxidation with argentic oxide will
produce results which are ot least as precise as
those obteined by the present method, An investi.
gation of this latter point is presently being carried
out,

SOLVENT EXTRACTION OF MOLYBDENUM FROM
ACIDIC SOLUTIONS WITH
TRI+n-QCTYLPHOSPHINE OXIDE

W. J. Ross

180 the isclotion, concentration,

Previous success
and determination of elements encountered in ART
fuels, corrosion products, and fission products
through sclvent extraction of these elements with
tri-n-octylphosphine oxide (TQPQ) has led to o
continued study of the anolytical applications of
this reagent, It wos found that hexavalent melybde-
num is amenable to separation from acidic soluticns
by extraction with tri-n-octylphosphine oxide. A
systematic study of the variables involved in the
extraction showed that hexavalent melybdenum
can be concenirated greoter thoan a hundredfold by
o single extroction with §,) M TOPO in cyclohexane
from HCl and that as much as 20 mg of Mo®* con be
quantitatively extracted by 0.5 mmole of this reagent,
Less efficient extractions are obtained from stod'
HNO,, or HCIO, systems. The extent of the ex-
troction is highly dependent on the acid concen-
tration of the aqueous phase, and therefore under

- aa

 

Yy, 1, Ross, ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 147.

143
certajn  conditions practically no extroction is
obtained from sulfate, nitrate, or perchlorate media.
Extroction with tri-n-octylphesphine oxide is thus
o rapid means of isoloting or concentrating hexo-
valent molybdenum in either the aqueous or erganic
phase.

PREPARATION OF RARE-EARTH
FLUORIDE TRACERS

A. S. Mevyer, Jr, G. Goldberg

The precipitation procedure described previausly V¥
wos used for the preparction of additional samples
of radicactive tagged tuoride saits. The salts pre-
pared were Y7 l.bearing YF,, Ce"“-beoring Cef,,
Sm %% bearing SmF,, Bal33.bearing BaF,, and
Lo '*C-bearing lLaF, in amounts that voried from
100 to 500 g. The somples contained from 1 10 3 mc
of total activity.

SERVYICE LABORATORY
W, F. Youghan

The 1428 samples analyzed during the guorter
involved 5746 determinations for an average of 4.0
per sample., A breakdown of the work according to
the group requesting the cnalyses is given below:

Number of Humber of
Somgples Reported Results
Ragector Chemistry &66 2255
Exparimental 589 2431
Enginzering
Metallurgy 28 33
Miscellaneous 145 1027
Tetal 1428 5746

 

a5 Meyer, Jr., and G. Goldberg, ANF Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 186,

144

YAPOR-PHASE POLYMERS IN
ALKALI-HALIDE SALTS

L. Q. Gilpatrick

A knowledge of the composition of the vapor phase
of the fused-salt fuel mixture is important to the
investigation of the vapor-deposition problem in
circutating-fuel reactors, It is alsc indispensable
for the measurement of activities of components of
salt mixtures when vopor-pressure methods of
Therefore o study of the
vapor phase of alkoli-halide sclts was initiated.
It has been reported??=22 that the equilibrium
vapor phase of mony simple moterials, such as
copper chloride, carbon, and petassium chloride,
contain simple polymeric species that are most
frequently dimers and trimers. When the olkali-
halide salts ore heated in on effusion cell, it is
possible to meosure the change in the vopor com-
position as o function of cell temperature. The
mass spectrometer is well suited for application
te this proeblem, since with it the species which are
derived from polymers in the vapor con be measured

measurement ore used.

independently, Onlyrelative values can be measured
with the use of this technique; however, energies or
heats of vaporization and the heats of palymerization
can be determined from the retative values, The
mass spectrometer is also o sensitive tool for the
qualitative examiration of the vapor phase of halide
salts and other moterials,

A satisfactory effusion cetl wos developed, and
potassium chloride is currentlty being examined at
temperatures from 550 to 750°C. Preliminary values
for the energy of vaporization of KCi monomer and
dimer were found to be 49.3 and 58.3 kecol/mole,
respectively, No higher polymers were detected.
The next salts to be exomined will be the olkali
fluorides, since they are of interest in the investi-
gotion of fused-salt fuel systems.

 

2041, M. Rosenstock et al., J. Chem Phys. 23, 2442
{19255).

2) | Grewer, P, W. Gilles, and F. A. Jonkins, J. Chem.
Phys. 16, 797807 {1948).

225, C. Miller ond P. Kusch, J. Chem Phys. 25, 860
876 (1956).
Part 3

METALLURGY

W. D. Manly
 

 
3.1,

DYNAMIC CORROSION STUDIES

J. H. DeVan

FORCED-CIRCULATION LOOP TESTS
J. H. DeVon R. 5. Crouse

Fuel Mixtures in Inconel and in Hastelloy B

Three forced-circulation loops fobricated of
Hastelloy B were examined that had circulated
NaF-KF-LiF-UF , (11.2-41-45.3-2,5 mole %, fuel
107) at a maximum flyid temperotyre of about
1650°F, The loops were heated by direct elec-
trical resistonce and were constructed from !’Q-in.-
0D, 0.060-in,-wail tubing, The operating con-
ditions for these three loops are given in Table
311,

Two of the loops completed 1000 hr of operation,
although interruptions that required temporary
shutdowns occurred during both tests. Loop
7641-18 developed o leak while fuel was being
admitted to the loop for o test run, The loop was
successfully repaired, and it then operated for
the scheduled 1000-hr period without further diffi-
culties. Loop 7641-2B developed a leak after
65 hr of operation and was likewise repaired and
operated for the remainder of the scheduled 1000-hr
period, Operation of the third loop, 7641-3, was
terminated by a pump tfailure ofter 407 hr of opera-
tion,

The hot legs of all three loops were found to be
badly pitted after the tests, with the pits in the
most heovily attacked areas reaching to o depth
of 4 mils, Unfortunctetly, the surfaces of the as-

received Hastefloy B tybing were rough, as shown
by the sample in Fig. 3.1.1, and thus o large part
of the pitting was probably present before the
test. The cold-leg sectiens of loops 7641-18 and
7641-3 were entirely free of metal deposits, but
they had deeper surface imperfections thaan those
found in the hot-leg sections, In some areas of
the cold legs the pits reached a depth of 7 mils;
however, as may be seen in Fig. 3.1,2, which
shows a cold-leg section of loop 7641-1B, the
surface imperfections appear to be the result of
corrosion deepening of the pits present before the
test,

The examination of loop 7641-2B likewise re-
vealed deeper pits in the cold leg than in the hot
leg. In contrast to the other two foops, joop
7641.2B, however, was found to have o few small
metal crystals in the cold leq. These crystalline
depesits occurred in clumps, opproximately 2 mils
thick, that were confined to g short section of the
loop near the cold-leg outlet, As shown in Fig.
3.1.3 the crystals were completely detached from
the loop wall., The quentity of deposited material
was insufficient for ¢ chemical analysis, ond it
s not known whether the crystals resulted from
groins becoming detached from the extremely rough
as-received surfaces or from a mass-transfer
mechanism, Loop 7641-2B operated with the
lowest Reynolds number used in the three tests,
and it is doubtiui that mass transfer would take

Table 3.1.T Conditions of Operation of Three Hastelloy B Forced-Circulation Loops with
NaF-KF-LiF-UF , (11.2-41-45.3-2,3 Mole %, Fuel 107}

Temperature grodient:

300°F

Ratio of hot-leg surface area to loop volume: 2.2 in2/in.3

 

 

Operating Conditiong

 

Moximum fuel mixture temperature, °F
Moximum tube wall temperoture, OF
Revnolds number of fyel

Yelocity of lual, fps

Qpercting time, hr

Loop Number

 

 

7641-1B 7641-28 7641-3
1635 1640 1660
1767 1710 1750
10,000 5,500 10,000
2.613 1.16 2.61
1000 1000 407

 

 

147
ANP PROJECT PROGRESS REPORT

 

Fig. 3.1.1. As-Received Hastelloy B Tubing. Etchant:

 

 

 

"
o T
G L alE

Fig- 3-]‘.2.

Area of Deep Pitting in Cold Leg of
Hastelloy B Forced-Cirevlotion Loop 7641-TB Which
Circolared I\fcnl':-'l"CF-LiF-UF4 (11.2-41-45.3-2.5 Mole %,
Fuel 107} Under the Conditions Given in Takle 3.1.1,
Riddle's reogent. 250X, Reduced 32%,
{Secret with caption)

Etchant:

148

LNCL ASSIFIED
T-1181%

 

 

 

 

 

 

; i
" - _.

i 3 b o3
E = - ;-. i
5 i 014
+ ¥ 3 o "‘_’—]

# £ i - a5

I ¥ 5 A

o »
= 3 .--- =

' . ‘,.‘ L]

b -

. Q18

Riddle's reagent. 250X,

 

 

 

Fig. 3.1.3,
Hastelley B Lcop 7641-2B Which Cireulated MoF-KF-
LiF-UFA {11.2-41-45,3.2.5 Mole %, Fuel 107) Under the

Crystols Found Near Cold-Leg Qutlet of

Conditions Given in Table 3.1.1.
Raduced 32%. (Secret with caption)

Unetched., 250X,
place in this loop and not in the others, Analyses
of the fuel mixtures used in these tests showed
no significant changes in any of the metal con-
during the tests. The resutts of the
analyses ore presented in Toble 3,1.2,

An Inconel forced-circulation loop (7425-20) was

stituents

operated at a maximum fluid temperature of cpproxi-
mately 1700°F in order to complete studies of the
effect of temperature on the corresion of Inconel
by NaF-ZrF -UF, (50-46-4 mole %, fuel 30), The

maximum wall temperature of the loop was 1840°F,

PERICGD ENDING MARCH 31,1957

and the fluid temperature drop between the hot-leg
cold-leg outlets was 300°F, Other loop
opercting conditions are given in Table 3.1.3.
After 1000 hr of operation, voids were observed in
the hot feg to a moximum depth of 9 mils. In addi-
tion to the hot-leg otteck there were a few scattered
particles of metal deposited on cold-leg surfoces,
The deposits contained chromium and, peossibly,
on the basis of etching characteristics, some
vranium or zirconium compounds. The attock on
this loop was 4 to 5 mils greater than the attack

and

Tabie 3.1.2, Anulyses of NoF-K E:-LiF-LIF4 (11.2-41-45,3-2.5 Mole %) Mixtures Circulated in
Hostelloy B Forced-Circulation Loops

 

Uraoium Cantent

Metel Constituents (ppm)

 

L.oop MNa, Sample Taken e .
{wt %) i Cr Fe Mo

7641-1B During filling 13.7 160 128 150 <2
After operation 12.4 85 330 95 <2

764)-28 During filling 13.0 195 1§D 165 <2
After operation 12.8 185 325 120 2

7641-3 During filling 13.0 20 35 55 15
After operation 12.6 205 270 75 9

 

Toble 3.2.3. Conditions of Operation of Three Inconel Forced-Circulation Loops with Fuel Mixtures

 

Loop Number

 

Operating Conditions

7425.20

Fuel mixture

fuel 30)
Maximum fue! mixture tempergture, °F 1715
Maximum tube wall temperature, °F 1840
Raynalds number of fuel 23,700
Velocity of fuel, fps 5.0
Operating time, hr 1000
Ratio of hotleg surfocs arec to 2.21

leop volume, i:1.2,"'in.:3

Temperoture gradient, °F 300

NaF-Zr F4-U Fd
(50-46~-d mole %,

7425-25

742526

Ma F-ZrFé-UF‘i
(50-46-4 mola %,

(56-39-5 mole %

fuel 30) fuel 70)
1620 1510
1710 15615
5,800 14,000
2.54 4.79
500 716.5
2.15 2,15
360 200

149
ANP PROJECT PROGRESS REPORT

of Inconel loops operated previously with a tuel
temperature of 1500°F, a wall temperature of
1600°F, and a temperature differentiol of 200°F,!
Examination of ancther Inconel forced-circulation
locp (7425.25) that operated 500 hr with fuel mix-
ture No, 30 also revealed hot-leg attack to a depth
of 5 mils, but no cold-leg deposits were found.

A third Inconel forced-circulation loop was
operated with fuel mixture 70, NaF-ZrF .UF,
(56-39-5 mole %). The results of two other forced-
circulation loop tests which compared the corrosion
properties of this fuel mixture with those of the
more commonly used mixture 30, NaF-ZrF -UF,
(50-46-4 mole %), were reported previously,?2 |In
this lotest test in the series the {oop operated with
a maximum fuel temperature of approximately
1500°F and a maximum wall temperature of about
1600°F, Other operating conditions for this loop,
7425-26, are given in Table 3.1,3. Although
scheduled for 1000 hr, operation was terminated
after only 716 hr because of a pump pulley failure.
VYoid formation during operation progressed to a
depth of 5 mils,
in the cold leg that wos similor to the layers found
in the Inconel loops operated previously with this

A thin metallic layer was noted

The layer wos too thin to provide
material for chemical analysis, and
attempts are now being made to determine its com-
position spectrographically,

The attack to a depth of 5 mils in this loop is
similar to the 7 mils of ottack in each of the two
loops operated previously with fue!l 70 for 1000 hr,
One of these latter loops operoted under tempero-
ture conditions identical fo those for loop 7425-26,
~while the other loop operated with a maximum fluid
temperature of about 1650°F. The depths and type
of attack in these loops are similar to the results
cbtained for loops that circuloted fuel 30 under
similor

fuel mixture.
sufficient

conditions, except for the presence of
thin layers in the cold legs of loops that circulated
fuel 7O. V:sible layers or deposits are not normally
observed in loops that have circulated fuel 30,

NaK in Inconel

Studies were made of two Inconel forced-circula-
tion loops that circutated NaK (56-44 wt %) and

 

1. H. DeVan end R. S. Crouse, ANP Quar. Prog. Rep.
June 10, 1956, ORNI_-2104, p 133,

2). H. DeVan, ANP Quar. Prog. Rep. Sept. 10, 1956,
ORNL-2157, p 135.

150

containad oxide cold traps operated ot 600 and
800°F. The design of the loop and cold trap is
shown in Fig. 3.1.4. The data obtained with these
loops, together with those cbtained from two loops
operated previously with cold-trap temperatures of
100 and 300°F, respectively,®*? provide a bosis
for an enalysis of the effeci of cold-trap operation
on corrosion and mass transfer in Inconel-Nak
systems.
loop wes 1500G°F and the temperature drop across
the cold-leg section was 300°F, All the loops
were operated for 1000 hr.

An investigotion of the design of the cold trap

The maximum tluid temperature of each

used in these loops was made prior to beginning
these studies to determine the effectiveness of
such a trap in controlling the oxide level. In a
spectal test loop containing a bypass cold trap
and a device known as a "'plugging meter,’’ NaK
was circuloted isothermally ot 1500°F, and meas-
vrements were made of the oxide concentration of
the NaK under vorious conditions of cold-trap
temperatures ond flow rates, The plugging meter
was wsed to indicate the soturation temperature of
sodium oxide in the NaK under each of the cold-
trap opercting conditions, The scturation tempera-

ture, together with published information on the
solubility limits of sodium oxide in NaK as a
function of temperature, was used to establish

the oxide concentration, These tests demonstrated
that the cold-trap operating temperature in the
range from 100 to 1200°F accurately reflected the
oxide concentration; that is, the oxide concentra-
tion which existed in the NaK corresponded to the
limit of solubility for sodium oxide in MaK at the
cold-trap temperature,

The effect on mass transfer of the various cold-
trap operating temperotures
Table 3.1.4. Samples token from the cold legs ot
similar lacations in each of the loops are compared
in Fig. 3.1.5. Difficulties in removing deposits
from the walls of the loops, as discussed pre-
viously,3 made it impossible to obtain weight
determinations of the deposits found in loops
7439-1 and -3, However, in the loops with higher
cold-trop temperotures (loops 7426-17 and -18) the
deposiis were comprised of larger crystals which

t5  summarized in

 

3.3. H. De¥Yan, E. A. Kovocevich, end R, 5, Crouss,
r"-fii’{‘;’? Quar. Prog. Rep. March 10, 1956, ORML.-2061,
p .

). H. DeVan and R. S, Crouse, ANP Quar. Prog. Rep.
fune 10, 1956, ORNL-2106, p 135.
151

 

WOTE  QIMENSIONS 1M INCHES

. UNCLASSIFIED
- _-\ SURGE Tank ORNL-LR- W3 15353

=

CoLD TRAP

   

ELECTROMAGNETIC

 

ELECTROMAGHETIC [T PUMP _
FLGWME TER A
S
..z'r
1 g
.‘(,l‘:.’
| “
TO SUME '
\\ |
Ry .,
™ .
M TRAF PACKRED WITH INCONEL

WIRE MESH

Fig. 3.1.4. Design of Inconel Loop, lncluding Cold Traps, for Forced Circulotlon of Alkoli Metals,

= B

LS4 °1E HONYW ONIQN3 QOl¥3d
ANP PROJECT PROCGRESS REPORT

UNCLASSIFIED
19448

 

{fgl LOOP 7439-3 { &) LOOP 7439-1

UMCLASSIFIED
UNELASSIFIED T-108%3
[-10872

 

Ll_‘_i_.\_i_u_]

ONE INCH
[} LOQOP 7426-18 (g} LOOP 7426 {7

Fig, 3.1.5, Mass.Transfer Deposits In Inconel Loops Which Circulated NoK ot o Maximum Temperature of 1500°F
and o Temperoture Differential of 300°F with Cold Traps Operated ot Vorious Temperotures. {(a) Cold trap at 100°F,
{6) 360°F. {c) 600°F. {d) 800°F. (Confidentiol with caption)

152

W
Toble 3.1.4. Effect of Cold-Trap Operating Temperature

on Moss Transfer In Inconel Loops that Circuloted NaK

 

 

Minimum Moximum Weight
Loop Cold-Trap Deposit af Metal
Mo, Temperoture Thicknass= Daposited
(°F) (mils) (g}
7439.3 100 ¥ o
74391 300 12 o
7426-17 800 20 6.29

 

*Determined by scroping deposits from loop wolls
**MNot recorded.

were more easily seporated frem the loop walls,
The weights of the deposits in these loops show
that a noticeoble increose in the quantity of ma-
terial deposited was brought about by an increase
in the cold-trap temperature from 600 to BOO°F,
This increase is also reflected in the deposit
thickness. A smaller difference in deposit thick-
ness is apparent in comparing the results for the
test with o cold-trap temperature of 600°F with
those for the two tests with lower cold-trap tems
peratures, Metallographic examinations of cold-
ieg sections from these loops also indicate that
the change in the amount of deposited material
accompanying a decrease in cold-trap temperature
below 600°F is less than the chonge from 800 to
500 P

As mentioned above, the size of the particles
comprising the deposits is opparently affected by
the cold-trap temperature and the oxide concentra-
tion. Deposits taken from similor sections of the
toops operated with cold-trap temperatures of 100
and 800°F are comparad in Figs. 3.1.6 and 3.1.7.
As con be seen, the average particle size of the
crystals found in the loop which included o cold
trap operated at BOO°F was considerably larger
than that of the crystals from the loop operated
with a cold-trap temperature of 100°F, The compe-
sitions af the deposits did not vary with cold-trap
temperafure; in all the loops, the deposits were
found to contain approximately 90% Ni and 10% Cr.

The results of these studies of the effect of
oxides in Inconel-NaK systems show trends quite
similor to those found in the results of tests of

Inconel-sodium systems., The tests of Inconel-

 

PERIOD ENDING MARCH 31,1957

UNCL ASSIFIED

T8

PEEERR

T T
1Nl'-IHE

 

HERINENES

 

%
&

 

 

 

 

Fig. 3.1.6. Mass-Tronsfer Deposits in lnconel Loop
7439.3 Which Circuloted NoK und Which lncluded a Cold
Trap Thot Was Maintoined ot 100°F, Etchant: modified
150X, Reduced 35%, (Confidantial with

URCLASSIFIED |
T-11113 _ga1 |

aquo regia.

coption}

 

 

2
o

 

LI

(1L LE]

o
n
o

o
[

 

e[ 1 1]

|

1
I150%
11

 

T

Mass-Transfer Deposits in Inconel Loop

Fig. 3.15?-
7426.17 Which Cirevlated MNaK and Which Included @

Cold Trep Thot Was Maintoined ot BOO°F. Etchant:
modified aquo regia. 150X. Reduced 9%. {Confidential
with caption)

153
ANP PROJECT PROGRESS REPORT

sodium systems, which were discussed previously®
were carried out by making oxide additions directly
to the sodium. The effects on mass transfer were
small for oxide additions of 0.05 wt % and less but
became significant at higher concentrations, Addi-
tiona! data {not presented previcusly) showed that
the deposit porticle size alsc increcsed as the
oxide level increosed,

A difference in the guantity of mass-transferred
material in sedium-lnconel systems compared with
thet in NaK-lnconel systems has now become
apparent, The weighis of the deposits in the two
recently tested NeK-Incenel systems are somewhat
less thon one-half the weights of deposits found
in sodium-inconel systems tested under similar
temperature conditions, The thicknesses of the
deposits found in both systems are comparable,
although as discussed in conjunction with the first
NaK ha-s.nr.,:3 thickness measurements do not cccu-
rately reflect the weights of the deposits because
the crystals comprising the deposits build up quite
randomty and in discontinuous patches.

The composition ond manner of deposition of the
crystals in the two systems oppear to be identical,
The average size of the crystals comprising the
deposits in both systems is also similer, Thus it
appears that the same mechanism for mass transfer
exists in both systems but that the process pro-
ceeds slower in NaK-lnconel systems than in
sodivm-lnconel systems.

Sodium in lnconel ond in Stainless Steel

Two Inconel forced-circulation loops of a design
similar 1o that shown in Fig. 3.1.4 were operated
with sodium at ditferent velocities to evaluate
the effect of flow rate on mass transfer. The
sodium temperatures ranged from a maximum of
1500°F in the hof leg to a minimum of 1200°F in
the ¢old leg. The pumps were operated at different
power levels to establish sodium flow rotes of
Both loops con-
tained bypass cold traps fo maintain low oxide
levels.

Examinations of the toops following 1000 hr of
operation revealed a substantial difference in the
quantities of deposited material recovered from

1.5 and 3.0 gpm, respectively,

the cold tegs. In the loop operated with ¢ sodium

 

 

5J. H. DeVan, ANP Quer. Prog. Rep. Dec. 13, 1955,
ORNL-2012, p 105,

154

flow rate of 1.5 gpm (loop 7426-22), the deposit
weighed 10.4 g, A doubled flow rate of 3.0 gpm
(loop 7426-23) resulted in an increase in the weight
of the deposit to 14.0 g. Hot-leg attack in both
toops was in the form of intergrenular penetration
and reached maximum depths of 2 mils in each
case. [he average thickness of the deposit alsc
varied with the flow rate, The thickness at the
lower flow rote was 12 mils, while at the higher
flow rate the thickness increased to 22 mils.

Several other sodium loops have been operated
at flow rates of 2.5 gpm, and the results can be
compared with the results of tests described above,
While the mojority of the loops operoted with o
sodium flow rote of 2.5 gpm, each loop test in-
corporated some variable which introduced o slight
difference from the condifions described above,
An averoge value of deposit weight for the tests
which most closely simulated the conditions given
above is 13 g. A plot of the data given above and
the 13-g volue is presented in Fig. 3.1.8 to show
the effect of flow rate,

A type 304 stainless steel forced-circulation
loop completed 1000 hr of operation with sodium at
o maximum temperafure of 1500°F ond a tempero-
ture drop of 300°F, The loop contained an oxide
cold trap that was maintained ot 300°F., Visval
examination of this loop revealed only slight traces
of moss-transferred particles in cold portions, A

QRNL =LA~ 0DWG Z0330

6.0 r — H———;-— —; — e _—I

LOOP OPERATING PERIOD: 1000 hr

  

150 " HOT-LEG TEMPERATURE: 1500°F ™

g | | TEMPERATURE DIFFERENTIAL: 300°F

= OXIDE COLD TRAP TEMPERATURE | 300°F

o 1440 f + =+ ' e e

a |

4 |

5 [

o t3 0 e l

h_

P |

©o

Ll

g0 b J,

-

< 1 |

i

b~ |
1€ 0 |

SODIUM FLOW RATE {qum)

Fig. 3.1.8. Effect of Flow Rate on Mass Transfer in

Inconel-Sodium Systems,

‘n
typical cold-leg section is shown in Fig. 3.1.9,
The deposits were similar in extent and oppearance
to those which were observed in fype 316 stoinless
steel loops operated under simifar conditions, 3
and were much less in amount than deposits found
in comparable lnconel loops. The hot leg of the
loop revealed slight surfoce pitting and void forma-
tion to a depth of 2 miis,

Long-Duration Tests of NaK and Sodium in Incone!

An Inconel forced-circulation loop, 7439-51, was
operated with NaK for 2760 hr to determine the
long-ronge effects of a high-thermal gradient on
mass transfer and corrosion in on Incenel-NaK
system,
Fig. 3.1.10 and operoted with a thermal gradient
of 850°F and o moximum buik fluid temperoture of

The loop wos of the design shown in

UNCLASSIFIED

7426~25""

 

HONI ANO

 

Fig. 3.1.9. Cold-Leg Section of a Type 304 Stainless
Steel Forced-Circulation Leoop Which Circulated Sedium
for 1000 hr at o Maxlmum Temperature of 1500°F and a
Grodient of 300°F. Reduced 31.5%.

{Confidantiol with caption)

Tempercture

PERIOD ENDING MARCH 31, 1957

1500°F, Fluid flow was meintained ot 1,28 gpm.
After 2480 hr of operation, a [eak developed in the
heater coil of the loop, ond the NaK had to be
dumped, The heater coil was replaced, ond the
loop was again put inte operation. Afrer an addi-
tional 280 hr, operation was terminated when a
similar leck developed in the new heater coil,

Metallographic examination of the hot leg which
wos removed from the loop after 2480 hr revealed
light-to-heavy intergranular attack to o depth of
1.5 mils, A needlelike deposit, 7 mils in depth,
was also observed in the section just beyond the
heater coil where the loop first experienced a
thermal gradient, The cooled coil, which was in
operation an cdditional 280 hr, showed only light-
to-moderate pifs, along with intergranular attack to
a depth of 1 mil. Metal deposits reached a maxi-
mum thickness of 14 mils in the section at the
entrance to the cooled coil.

A similar {oop was operated with sodium as the
circulated fluid. Operation of this loop {7426-51)
was terminated after 4000 hr, when a leak developed
near the pump. The operating conditions, with the
exception of the temperoture gradient, were similar
to those for loop 7439-51, Because of the better
heat tronsfer properties of sodium, a thermal
gradient of BDO®F was achieved., Sodivm attack in
the hot-leg section of the loop reached a maximum
depth of 5 mils. A heavy deposit of metal crystals
that reached o moximum thickness of 50 mils was
found in the section at the entronce to the cooling
coil. The depesit is shown in Fig, 3.1.11.

Carbide precipitates were found at the surfaces
of the inner walls of the tubing which hed been
exposed to sodium and, to a lesser extent that
varied with depth, in the wall of the tubing. The
diffusion of carbon therefore appears to have pro-
ceeded from the inner wall of the tubing into the
tubing, ond thus the source of the carbon must have
been the sadium, This loop contaired ¢ mechanical
pump, and ¢ slight leak through the pump seal of
the oil for lubricating the pump could have been
responsible for the carburization found,

The loops for these long-duration tests were of
a unique design, and therefore the results cannot
be compored directly with the results of other
tests of shorter duration,
observed in the loop which circulated sodium indi-
cafe that moss transfer proceeds at a rate which

The extensive deposits

is not significantty diminished with time, clthough,
based on the carburization observed in this loop,

155
ANP PROJECT PROGRESS

REPORT

UNCLASSIFIED
| DRHL—LR-DWG 1535
fre
G
LN

 

FiNnED-THLBE
AR -CO0LED CONL

A
'jtf\fi;f# 2

 

 

'.NGHI?S
Fig. 3.1.11.

Deposit Found in Cooled Coil of Incenel Loop 7426-51 Which Circulated Sodium for 4000 br at o
Moximum Temperoture of 1500°F und o Thermal Gradient of 800°F. Etchant: modified agua regio, 100X, (Cenfi-
dential with coption)

156
impurities may clso have contributed o mojor part
to the moss-transfer process,

THERMAL-CONYECTION LOOP TESTS

J. H. DeVan E. A, Kovacevich
D. A. Stoneburner

NaF-KF-LiF-UF, and Sedium in Nickel-
Melybdenum Alloys

Fuel 107, — Corrosion tests of the experimental
nickel-molybdenum alloys exposed to scdium and
to NoF-KF-LiF-UF, (11.2-41-45,3-2.5 mole %, fuel
107) in thermal-convection loops were completed.
The compositions of the alloys tested, the test
conditions, and the results of metcllographic ex-
aminations of as-tested specimens are given in
Table 3.1,5.

The daota show that a maximum hot-leg attack of
3 mils occurred in loops 1105 (16% Mo=-2% Al-—
1.5% Ti—bal Ni) and 1108 (17% Mo-2% Al-bal Ni).
These results are in good agreement with the data
regorted previously for materials with relotively
high oluminum content. The attack wos accom-
panied by high aluminum pickup in the fuel, as
shown in Table 3.1,6.

Loop 1093 {17/% Mo—7% Cr~bal Ni} showed hot-
leg attack which wos somewhaot higher than that
found previously in loops fabricated from material
containing comporable amounts of chromium.® Also
the attack changed from the moderate general sub-
surface-void formation in the loops operated pre-
viously to heavy general subsurface-void formation.
Loop 1094 (20% Mo-7% Cr-2% Nb—1% Fe~bal Ni)
revealed hot-leg attack similar to that for loop 1093,
The addition of niobium and iren thus appecrs to
have little effect on the corrosion of o nickel-
molybdenum atloy centaining 7 wt % chromium.,
Hot-leg specimens from loops 1093 ond 1094 are
pictured in Figs. 3.1.12 and 3.1.13. As may be
seen in Toble 3.1.%, the concentration of chromium
in the fuel of loop 1094 is also comparable to thot
found in the fyel in other nickel-
molybdenum alloys containing 7 wt % chromium,

circulated

Comparisons showed similor maximum depths of
attack in loops 1094 and 1099, which were fabri-
cated of alloys of similar composition except that
the otloy used for loop 1099 contained 1 wt %

®E. A. Kovecavich and D, A, Stoneburner, ANP Quar,
Prog. Rep. Dec. 31, 1956, ORNL-2221, p 179,

 

PERIOD ENDING MARCH 31, 1957

aluminum, However, as shown in Table 3.1.6,
the chromium buildup in the fuel of loop 1099 was
somewhat higher than that reported for loop 1094,
Also, the fuel from loop 1099 had a high aluminum

 

 

content, Exomination of loop 1101 (16% Mo—5% Cr—
1.5% Ti-1% Al-bal Ni} also showed the usual
high aluminum buildup found in the fuel after circu-
lation in aluminum-containing alloys. The attack
-um..mlmb )
T Tiesa et
_'E:‘a‘_"_
i-fl_-!_l-
: o
Lt
Lin

!_
o r'—:‘

 

o
|

Fig. 3.1.12,  Moximum Hot-Leg Attock Found In
Thermal-Convection Leop 1093 (17% Mo—-7% Cr—bal NI}
Which Operated for 500 hr with NaF.KF-LIF-UF, (11.2-
41-45.3-2,5 Mole %, Fuel 107) ot 1500°F. Erchans:
aqua regia. 250X, Reduced 32,5%. (Sacret with caption)

b

 

&
r
T
IHCHES

T
1

 

EREEEE

3
=

 

o
-

7

 

ko

Fig. 3.1.13. Maximum Hot-Leg Attack Found n
Thermal-Convection Loop 1094 (20% Mo—7% Cr-2%
Mb-1% Fe-baol Ni} Which Operated for 500 hr with
NoF-KF-LiF-UF4 {11.2-41-45,3-2.5 Mole %, Fuel 107)
at 1500°F. Etchaont: aqua regio. 250X. Reduced
32.5%. (Secret with caption}

157
ANP PROJECT PROGRESS REFORT

Table 3.1.5. Conditions and Results of Thermal-Convection Loop Tests of Nickel-Malybdenum Alloys with

Na F-KF-LIF-UF4 (11,2-41.45,3-2,5 Mole %, Fuel 107) and Sedium -

Maximum fluid temperoture 1500°F

 

 

 

Loop Alloy Circulated Operating Metallegropbhic Results -
N Composition Fluid Time
O 1
1023 11 Mo=2 Al-bal Mi Sadium 1000 Heavy surface pits fo a Modarate surface roughening
depth of > 1 mil
1087 20 Mo-1 Nb-1 Ti- Fuel 107 500 Haovy voids to a depth Light surface roughening
0.8 Ma—=bal M of 2.5 mils
1088 20 Mo-1MNb-2 Ti- Sodium 1000 Heavy voids to a depth Meoderate surface roughening
0.8 Mn=bal N; of 1 mil
W59 15Mo-3 Nb=0.5 Al- Sodium 1000 Heavy voids to a depth Moderote surface roughening
3 Webal Ni of 1.5 mils
1093 17 Mo =7 Cr—=bal MNi Fuel 107 500 Heavy voids to o depth Light surface roughening
of Z mils
1094 20 Mo-7 Cr=2 Nb~ Fuel 107 5S040 Heavy vaids to a depth No attack
1 Fe—bal Ni of 2 mils
1098 17 Mo-2 Ti=bol Ni Fual 107 500 Moderate surfoce pits to Mo atteck
o depth of > 1 mil
1099 20 Mo—7 Cr—1 Al- Fuel 107 500 Heovy voids to a depth  No attack -
ZNb-1 Fe-bo!l Ni of 1.5 mils
1100 17 Mo-2 ¥=bol Ni Fuel 107 500 Fow voids to a depth of Light surface roughening s
Q.5 mil -
1101 16 Mo=-5 Cr-1.5 Ti—  Fuel 107 500 Heavy voids to a depth Yoids to a depth of 1 mil )
1 Al —bal Ni of 2 mils
1102 16 Mo=5 Cr=1.5 Ti— Sodium 19000 Heavy voids to o depth Light surfoce roughening™
T Al=bal Ni of > 1 mil
1103 16 Mo=1 Al=1.5 Ti— Fuel 107 500 Heavy voids te o depth Light surface roughening
bal Ni of 1 mil
1105 16 Mo=-2 Al-1.5Ti- Fuel 107 500 Heavy voids to o depth Mo ottack
bol Ni af 3 mils
1108 V7 Mo =2 Al=bal Ni Fuel 107 500 Heavy voids to o depth Light surfoce roughening
of 3 mils
1113 17 Mo -2 W=ba! Ni Fuetl 107 s500 Intargrenular voids to o Mederate surface roughening
depth of 1.5 mils
1117 17 Mo=2 W—bal Nj Sodium 10006 Intergranular voids to o Heavy surface roughening
depth of 2 mils
1120 17 Mo =4 W=hal Nj Fuai 107 500 Heavy voids to o depth Few voids to a depth of 1 mil
of 1 mil -
1128 17 Mo ~3 Mb~ba! Ni Fuel 107 500 Few smal) pits toc o depth  Moderate surfoce roughening |
of » 0.5 mail
1131 17 Me—=5 Mb—bal Ni Fuel 107 500 Heavy surfoce roughening Light surface rovghening =

to o depth of <1 mil

 

*A thin discontinyous metallic loyer was also found.

158
Leop
Na.

PERIOD ENDING MARCH 31, 1957

Table 3,1.6. Chemical Anolyses of Fuel 107 Bafore ond After Cireulation in
Nickel-Molybdenum Alloy Thermal-Convection Loops

Alley Composition
fwt %}

Somple Tolken

 

1087

1692

1094

1098

1095

1100

1al

1103

1195

1108

13

rag

1124

1131

 

20 Mo—1 Nb=1 T-0.8 Mn—tal N,

17 Mo ~F Cr=bal b

20 Mo—7 Ce-2 Nb-1 Fa~tal N

17 bho—2 Ti=bo! M,

20 ¥o—7 Cr=1 Al-2 Nb-1 Fa-bhal M

17 Mo ~2 ¥=bal Mi

16 Mo -5 Cr=1.5 Tuw] Atcbal Ni

s Mo=1 Al=1.5 Ti-bal N,

16 Mo=2 Al=1.5 Ti—bal Ni

17 Mo -2 Adxbai N

V7 Mo=2 W-bal N

17 Mo W =bal Ni

W bo=3 Mb=bal M

17 Mo —5 Mh=bal dh

* Rawwlis ngd yat avaslabis.

During filling

Adter aperation
Mo |eg
Cold leg

During Hlling

Afler operation
Hot leg
Cold lag

During lilling

Alrar operakign
Hot leg
Cold leg

During §illing

J‘\fler opardlion
Hol =g
Cald leg

Buring filling

Alar gperalion
Hot leg
Cold leg

Dunng fliiing

Afrer aperatien
Hot teg
Cold lag

Curing filling

Aldter oparolian
Hor leg
Cold 1ag

Curing fillina

Alter gperation
Hol leg
Cald leg

Durmg irHing

Afrer aperotian
Ho! fey
Cold leg

During filhing
Adrar operation
Hat leg
Cold leg
Oyring hilling

Aftar apardtion
Hat leg
Cold &g

During lilling

Afrar gparation
Hor leg
Cald leg

During fF“lng

Ablar eperation
Hot leg
Cald log

Dwring fitling
Altar operotron

Hot leg
Cald l=g

 

 

 

 

 

 

 

Uramium Matallic Constituants [ppm)
Content ———— ———— e 5 el
{wt %) i Cr Fe Ma Al T ¥ w Mk Ma
12.0 535 i 125
12,2 190 3% 110 80 3T 20 100
12.4 Y80 55 B0 20 3680 g 1 1g
12.6 35 15 125
12.0 145 435 5 140
12.4 200 655 70 25
12.9 170 &0 165
12.3 270 665 80 115 15
12.5 G 455 &0 5 10
12,2 g5 50 210
12,9 140 30 125 135 390
12,9 85 40 5 25 I8z
12.7 3 40 170
12.7 165 755 210 30 2815 ¥
12.8 14 735 90 15 2945 7
12.5 < 25 105
12.5 85 a0 115 215 < 50
2.8 130 25 S0 40 <50
13.2 &5 55 125
12.8 45 245 130 a0 200 375
12.8 40 255 &3 15 1830 43
12.8 115 &5 125
1.9 &5 as 95 120 2535 3185
V2.4 105 30 115 55 2650 430
12.6 35 is i25
1.9 S0 20 Fiv &0 2210 445
12.2 70 20 80 20 2460 485
i2.9 195 100 135
124 <2 40 &5 43 1238
12.7 110 40 35 335 W0
2.7 0 120 L
1.8 45 50 100 280 1215
1.2 43 45 g5 340 1095
12.46 185 s 175
1.9 0 % g0 15 1465
12.2 215 59 Fiv 80 i 185
12.7 <3 30 125
12.1 70 40 105 170 D
12.5 220 50 g5 i 7
12,1 b5 40 15
. 6 - - 50 sm
¥ 30 * * 25 135

159
ANP PROJECT PROGRESS REPORT

of 2 mils in this loop (1101) is comparable to the
attacks in the above loops containing 7 wt %
chromium, However, the chromium content in the
fuel after test was noticeably lower in the case of
loop 1101 (5 wt % chromium) than in the loops
with 7 wt % chromium,

Additions of 2 1o 4 wt % tungsten to the nickel-
molybdenum alloy did not affect the depth of attock.
The hot-feg attack to a depth of 1.5 mils found in
loop 1113 (17% Mo-2% W-bal Ni) is shown in
Fig., 3.1.14. As may be seen the attack was pre-
dominontly intergranular, which is uncommoen for
nickel-molybdenum alloys exposed to fuel 107,
Another loop fabricated of material from the scme
heat alse showed intergronuler ottack after opera-
tion with sodium. The atteck by fuel 107 in loop
1120 (17% Mo~4% W-bal Ni}, fabricated from a
different heat containing 4 wt % tungsten, was of
the usual general subsurface-void-formation type
and there were surface pits to a depth of 1 mil,
The differences in the ottack of the alloys con-
taining 2 and 4 wt % tungsten moy result from
differences in the groin-boundary constituents of
the two alloys. The chemical onalyses of the fuel
mixtures circulated in the two loops show sizable
amounts of tungsten, as high as 1465 ppm,

Additicnal loops fabricoted from nickel-molyb-
denum olloys conteining vanadium, nicbivm, or
titanivm (loop 1100, 17% Mo-2% Y-bal Ni; loop
1126, 17% Mo—=3% Nb~bal Ni; loop 1131, 17% Mo-—

UNCLASSIFIED
T 14980

    
  

P

 

2

NC-:‘!ES &

3 B &

 

Fig, 3.1.14.  Maximum Hot-Leg Attock Found In
Thermal-Convection Loop 1113 {17% Mo-2% W-hol Ni)
Which Operated for 500 hr with NuF-KF-LiF-UF4
(11.2-41-45.3-2.5 Mole %, Fuel 107) ot 1500°F. Erchont:
aque ragic. 250X, Reduced 32.5%. (Secret with caption)

160

 

5% Nb-bal Ni; and loop 1098, 17% Mo~2% Ti-bal
Ni} and operated with fuel 107 showed attack of
0.5 to 1 mil. Chemicol analyses were made of the
fuel mixtures circulated in these loops, There was
no appreciable vanadium buildup in the fuel from
loop 1100, There was, however, as indicated in
Table 3.1.6, a buildup of niobium in the fuel from
loops 1126 and 1131, with the niobium content of
the fuel from the loop containing 5 wt % niobium
being greater than that of the fuel from the loop
containing 3 wt % nicbium, A slight buildup of
titanium was evident in the fuel from the loop con-
taining 2 wt % titanium (lcop 1098), although the
level of titanium in the fuel after the test was
much less than the level of aluminum in fuel from
loops Loops containing
titanium along with other alloying elements (1087,
1101, 1103, 1105} showed similar titanium con-
centrations in the fuel aofter the tests, with emounts
up to 485 ppm,

containiag aluminum,

Sadium, — Date on the nickel-molybdenum alloy
loops operated with sodium are presented in Table
3.1.5. Mass-tronster depasits were visible in
nickel-molybdenum alloy loops 1023 {11% Mo~2%
Al=bal Ni), 1088 {20% Mo=1% Nb=2% Ti=0.8% Mn-
bal Ni), 1089 (15% Mo-3% Nb-0.5% Al-3% W-baol
Ni), 1102 (16% Mo~5% Cr—1.5% Ti-1% Al-bal Ni),
and 1117 {17% Mo-2% W-bal Ni}, whick eperated
for 1000 hr with sedium at a maximum temperature
of 1500°F; however, deposits were found metallo-
graphically only in loop 1102, These deposits
differed from deposits in Inconel-sodium systems
in that they did not adhere tightly to the locp wall.
Consequently, the deposits could be mechanically
removed, and thus they were lost during preporation
of metallographic samples, The quantity of de-
posited material visible in these foops was com-
parable to the quantities found in Inconel ioops
operated with sodium under similar conditions.
Chemical cnalyses of the deposited material found
in loop 1102 showed 6350 ppm Mo, 3400 ppm Al,
94% Ni, and 3.57% Cr.

The moximum hot-leg attack in atl these loops
averaged slightly more than 1 mil. The attack
aoppeared es subsurface voids and pits, ond thus
was similar to that found in the hot legs of loops
thot circulated fuel 107, With the exception of the
otloy that contained 2 wt % tungsten, the grain
boundaries were not preferentially attacked. This
result is in confrast to the results for Inconel loops
inwhich attack by sedium is normally intergranulor,
Yoid Migration in Inconel Exposed
to Na F-ZrF‘-UF4

Three standard Inconel thermai-convection loops
were operated with NoF-ZrF -UF , {50-46-4 mole %,
fuel 30) at o bulk fluid temperature of 1500°F,
One foop {1085) was terminated ofter 277 hr of
operation; the fuel in the second loop (1086) was
repiaced with helium after 280 hr, and the loop
was held at test temperature for the remoining
scheduled operating period of 1000 hr; the third
leop (1010), included as a control locp, operated
the scheduled 1000 hr with fuel.

An analysis of the corrosion process in Inconel
systems containing fluorides indicates thot the
rate of attack (that is, void formation) becomes
constant when the test period exceeds 250 hr,
This linear relationship of ertack with time is
reached following an initial period of rapid attack
resulting from the equilibration of the fluorides
with |nconel, It was first assumed thai the linear
migration of voids into Inconel with time repre-
sented a constant rote of chromium removal with
time. However, questions were posed as to whether
even in the absence of further chromium removal
some void formation might continue as a result of
the rapid stage of atteck represented by the equili-
bration of flucrides. The purpose of these tests
was therefore to determine whether the migration
of voids into the Incenel, instituted by the leaching
of chromium during the corrosion process, would
continue if a loop were left at test temperature
without fuel,

Me tallegraphic examination of loop 1085 revecled
heavy surface roughening and heavy general sub-
sutface-void formation to o depth of 1 mil in the
hot leg, accompanied by heavy intergranuiar
subsurioce-void formation to a maximum depth of
8 mils., The cold leg showed only heavy surface
roughening, Comparison of the hot legs of loops
1085, 1086, and 1010, shown in Fig, 3.1.15, indi-
cated the attack in loops 1085 and 1086 to be
identicol in depth, extent, and type of void forma-
tion, Further metallographic exaemination through-
out both loops revealed similar corrosion of re-
spective metaliographic somples. 1n control leop
1010, which operated with a different batch of fuel,
the maximum hot-leg attack was 12 mils, which
raprasents the normal depth of corresion of Inconel
vnder these conditions,

The results of these tests provide evidence that
the migration of voids into Inconel, instituted by

PERIOD ENDING MARCH 31,1957

 

 

UNCL ASSIFIED
T 11850 s
L
ooy
4
wr L.
- -4 o
e G L - | =
1 . * e -
- { ‘

o @ - | fos

L 5 o a \ |
3 L ! | pemr
- = | 1 oan

% \ i _._-‘ : _1 |
- = g 1‘ | LODY

w I
i ? ; 1 Long
. i Y = e
el S . g 3 x - i_m_z_

 

EEE]

T
INC:‘IES

 

 

EFEREEERE

 

e
‘: | Lssz
L : ¥ r ?

£ L. & S e ™ £.13

- ® ', ¥ “

. o i & & e o i'ril—

- & f.--- o Foil e

! y OBy a i/
4 . -y b
: ’ ir ; % .\'h I‘l|l ‘;

 

=
- ,
x
P
FFEEEED

Fig. 3.1.15. Specimens from Hot Legs of Loops 1085,
1084, ond 1010 Which Operated with NuF-ZrF‘-UFd
{50-46-4 Mole %, Fuel 30} ot o Maximum Fluld Temper-
ature of 1500°F, Loop 1085 (a) operated 277 hr with
fuel; leoop 1088 {b) aperated 2BD hr with fual and 720 hr
with helium at the test temperature; loop 1010 {c} oper-
ated 1000 hr with fuel. Etchant: modified oqua regisa.
250X, Reduced 33%. (Secret with coption)

161
ANP PROJECT PROGRESS REFPORT

the leaching of c¢hromium during the corrosion
process, does not continue if the loop is left at
the fest temperature without fuel.  Thus, the
coalescence of vacancies in the metal to form
visible veoids must proceed at a rate of least
commensurate with the rote at which vacancies
are formed, |f the removal of chromium from the
metal surface is stopped, void formation is alse
stopped.

NaF-KF-LiF-U F“ and Sodium in Micbium

The compatibility of nicbium with NaF-KF-Lif-
UF, (11.,2-41.45,3-2,5 mole %, fuel 107} and with
sodium was studied in thermal-convection loops
operated at @ maximum fluid temperature of 1500°F,
The loops, which were fabricated by Battelle
Memorial
the standard loops employed for corrosion studies
of more common materials, The niobium tubing
used in the loops was fabricated from welded
sheet and had on inside diometer of 0.53 in. ond
a wall thickness of 0.020 in. In order to eliminate

Institute, were slightly smaller than

the niobium oxidation problem which would exist
at the test temperature, the niobium was encosed
in an outer sheath of type 321 stainless steel.
This sheath completely surrounded the nicbium
loop but wos not integrally bonded to if, Fabri-
cation of a sound loop of this design was difficult
in that cracks often developed in the niobium
tubing while the sheath was being assembled,
Three niobium loops (1072, 1073, ond 1022) were
operated 1664, 1000, ond 775 hr, respectively,
with fuel 107 ot a hot-leg temperature of 1500°F,
Termination of operation of loops 1072 ond 1022
was caused by o power failure and loss of flow,
respectively. An onelysis of the fuel circulated in
foop 1022 revealed a high buildup of chromium,
715 ppm, and pointed to the probability of a teak
that allowed the fuel to contact the cludding. The
examination of this foop showed an absence of
cold-leg deposits and only slight hot-leg attack in
the form of shallow pitting. Loops 1072 and 1073,
on the other hond, apparently remained sound
during operation, since no elements other thon
niobium were found in the fuel after the tests.
However, macroscopic examination of the loops
revealed metallic crystals in the fuel layer close
to the cold-leg wall. These crystols were evident
metallograghically, and are shown in Fig. 3,1.16.
Hot-leg attock in these loops occurred as shallow
pitting to a depth of less thon 1 mil. A typical

162

os-tested specimen from the hot leg of loop 1072
is shown in Fig. 3.1.17,

An odditional niobium thermal-convection loop
(1091) was operated for 1000 hr with sodium ot @
maximum temperaiuvre of 1500°F, Exomination of

UHCLASSIFIED
T 11957

 

 

Flg. 3.1.16. Metallic Crystals in Fuel from Cold-Leg
Section of MNioblum Thermal-Convection Loop 1072
Which' Opercted 1664 hr with NGF-KF-LEF-UF“ {(11.2-
41-45,3-2.5 Mole %, Fuel 107) at o Maximum Temper-
ature of 1500°F, Unetched. 50X. Reduced 32,5%.

(Sacrat with caption)

 

\'l " _.." iCII 1

¥ I__'ut

. 1y i‘vl— !m:
. » .

—J ' "\‘ ’ 'i L4

Figs 3.1.17.  Maximum Hot-Leg Attock of Nicbium

Thermol-Convection Loop 1072 Which Operated 1664 hr
with NuF-KF«LiF-UF‘ (11.2-41-45.3-2.5 Mole %, Fyel
107) ot a Maximum Temperoture of 1500°F, Erchont:
HF-HNDa-H2504-H2CI. 250X, Raduced 31%, (Secret

with caption)
the hot leg of this loop showed little evidence of
ottack other than, possibly, uniform surface re-
moval. Cold-leg deposits were, however, observed
macroscopically,  Chemical analysis of the de-

PERIOD ENDING MARCH 11,1957

posited material indicated that the sodium probably
jeoked into the clodding loop during operation,
since large emounts of iron, nickel, and chromium,
in addition to niobium, were present in the deposit,

163
ANP PROJECT PROGRESS REPORT

3.2. GENERAL CORROSION STUDIES

E. E Hoffman

TESTS OF INCONEL TUBE-TO-HEADER JOINTS
WITH RECRYSTALLIZED WELDS

D. H. Jansen

The lost of o series of lnconel tube-to-header
joints with recrystallized welds were corresion
tested in NaK (5644 wt %) and in NaF-ZrF ,-UF
{50-46:4 mole %) for 100 hr in seesow-furnoce
apparatus with o hot-zone temperature of 1500°F,
These somples were fabricaied ot the Glenn L.
Martin Co. by using a prechamfered header plote
of the type shown in Fig. 3.2.1. In comparisen

UMCLASSIFIED
¥-20540

 

Fig. 3.2.1. Schematic Drowing of Prechamfer Method
of Tube-to-Header Jaint Fabrication,

with the welds tested previously ' the prechamfered
type of weld (Fig. 3.2.2) resulis in decreased
material deformation and greater weld surface areo,
A prechamfered type of recrystatlized weld that
was exposed to I\l-::F-?_’er-UF:J1 (50-46-4 mole %,
fuel 30} for 100 hr at 1500°F is shown in
Fig. 3.2.2. The attack alorg the tube-header
interface was no greater than the ottack on the
tube ond on the header. A maximym of 0.5 mil of
attock was found on the Incone! components of
the specimens tested in NaK, and the attack on
the tube-header interface was similar,

 

B, H. lansen, ANF Quar. Prog. Rep. Dec. 3i, 1956,
ORNL-2221, p 184.

164

 

Fig- 3-2-2.

oxalic acid, 12X, Reduced 79.5%.

URCL ASEIFIED
¥- 2193

 

 

Fig. 32.3.

Recrystallized Weld Fabricoted by
Prechomfer Method After Exposure for 100 hr in MoF-
ZrFq'UF.; {(50-46-4 Mole %, Fuel 30) at 1500°F, Etchant:
agua regia. 75X, Reduced 43.5%. (Secret with caption}

TESTS OF HAYNES NO. 8244 BRAZING ALL QY
D. H. Jansen

Inconel tube-to-header joints brazed with Hoynes
No. 8244 brezing alloy {nominal composition, in
wt % 9.4 (Cr, 4.4 Fe, 3.7 5i, 2.2 8, 0.3 Mn,
0.06 C, bal Ni) were corrosion tested in Naf-
ZrF,-UF, {50-46-4 mole %, fuel 30} and in NaK

As-Received Recrystallized Tubesta-
Header Weld Fobriceted by Using o Prechamfer Method.
Mote large tube-te-header contoct area. Etchont: 10%
{56-44 wt %} in the seesaw-furnace apparotus.
The test periods were 100 hr, and the hot zones
in which the specimens were retcined were main-
tained ot 1500°PF. The brazing alloy showed good
corrosion resistance to NoK, but there was light
subsurface ottack to o depth of 3 mils on the

specimen tested in the fuel mixture, as shown
in Fig. 3.2.4.

COMPATIBH.ITY OF MOLYBDENUM AND
NICKEL-MOLYBDEMUM ALLOYS IN
NaF-KF-LiF-UF,

0. H. Jansen

In cennection with studies of cladding bydride
moderator materials for use in high-temperature cis-
culating-fuel reactors, tests were conducted in
seesaw-furnace cpporatus to investigate the com-
patibility of molybdenum and nickel-molybdenum al-
loys in NaF-KF-LiF-UF, (11.2-41-45.3-2.5 mole %,
fuef 107). The test assembly consisted of o

© UNCLASSIFIEDR
Y- 21589

p.o2 |

PERIOD ENDING MARCH 31, 1957

%-in.-0D, 0.035-in.-woll alloy capsule in which
o cold-rolled strip of molybdenum, 0.018 in. thick
and one half the length of the test capsule, wos
confined in the hot zone.

A capsule fobricated from 15% Mo—6% Cr—bal
Ni olloy tubing wds tested for 500 hr ot o hot-leg
temperature of 1650°F with a temperature grodient
of 300°F, and a 17% Mo—2% V=bal Ni capsule was
tested under similor conditions, except that the
maximum temperature was 1600°F. Spectrographic
onalyses indicated the presence of nicke) on the
surfaces of the molybdenum strips from both the
copsules as a result of dissimilor metal mass
transfer. Mo evidence of chromium or vanadium
was detected. Metollographic examination showed
that the meolybdenum specimen tested at 1650°PF
had become recrystallized, while the one tested
ot 1600PF showed the elongoted ond distorted
grains typical of rolled materials., No attack to
a depth greater than 0.5 mil was found on either
olioy copsule, and no thickness chonges or attack
of the molybdenum strips could be detected.

UNCLASSIFIED
Y-21405

  

 

  

 

';‘ :1.0 - '.
ST st
1-# ' '_.,ég r G034
a .
s & '. .
;‘; d "#‘ «‘ ':f'-j j IIJ
AT 2 L
" ._ |. x K
o R i O
o IR | e

 

 

Fig. 3.2.4, Hoynes Braxing Alloy No, 8244 on lnconel Tubo4o-Header Joints After Exposure to {a) NaF-
szd'UFd {(50-46-4 Mole %, Fuel 30) and {&) NaK (5644 wt %) in Seesow-Furnace Apparatus For 100 hr at o Hot-Zone

Temperoture of 1500°F, Light subsurfoce atteck con be

seon in {2}, Specimen nickel ploted lo preserve edge

during polishing. Etchont: none, as-polished. 100X. Raducad 6%. (Secret with caplion)

165
ANP PROJIECT PROGRESS REPORT

More severe tests of these materials are to be
conducted, since the cladding problem would be
greatly simplified if it could be established that
molybdenum and nickel-molybdenum atloys were
compatible in the fuel system. At present it is
felt that o duplex clod of molybdenum and nickel-
molybdenum clicy may be required to prevent dis-
similar metal mass transfer. The molybdenum is
required os o hydrogen diffusion barrier for the
hydride moderator material.

THERMAL-CONVECTION LQOP TEST OF
50DIUM IN HASTELLOY B

E. E. Hoffman L. R. Trotter

A thermal-convection loop fobricated of Hastelioy
B and filled with sodium wos operated to cbtain
a comparison of the moss tronster and corrosion
resistance eof Hastelloy B and Inconel in sodium
systems, The loop operoted for 1008 hr at a
hot-zone temperature of 1600PF and a cold-zone
temperoture of 990°F. No significent differences
were found belween the amount of attack and
mass transfer found in this loop and in two lnconel
loops operated under similor conditions.? In all
three tests an air blast thot impinged on the
bottom of the cold leg was used to effect o very
sharp femperature drop and thus concentrate any
mass transfer deposits. OCnly o few very small
mass-transfer crystols were detected in the Hostel-
loy B loop following the test, as shown in Fig.
3.2.5. In general the crystals found in this loop
did not seem to adhere to the walls as well os
those found in Inconel loops. The hot- and
cold-leg surfaces of the Hastelloy B loops were
ottacked to depths of from 1 te 2 mils, as shown
in Fig. 3.2.6. It may also be noted in Fig. 3.2.6
that scattered areas of the as-received tubing
had sericus surfoce defects,

SODIUM-BERYLLIUM-INCONEL COMPATIBILITY
IN A STATIC S5YSTEM (TEST NO. 3}

E. E. Hoffman L. R. Trotter

The apparotus used for tests of the compatibility
of sodium, beryllivm, and lncorel and results of
two earlier compatibility tests in this series were

 

%e. E. HoHmon, ANP Quar. Prog. Rep. Sept. 10, 1956,
ORNL-2157, » 161,

166

UNCLASSIFIED
Y-20702

_ - .'-_-‘

 

COLD ZONE -990°F

ONE INCH

Fig. 3.2.5. Sections from Hastelloy B Thermal-
Convection Loop Which Circuloted Sodium for 1008 hr,
(Confidential with caption)

described previously.* The interface conditions
tested in this third experiment gre described in
Table 3.2.1, along with results of metallographic
examipations of the interfaces after the test.
During the 1000-hr test the temperature was cycled
from 1300 to 500°F ten times to determine the
extent to which the brittle intermetallic compounds
which form the interfaces might spall.  Each
cycle required o 5-hr cooling end an 8-hr heating
period; therefore, the specimens were actually ot
1300°F for 870 hr of the 100-hr ftest. The results
of metallographic examination of the specimen
interfaces following the test are illustraoted in
Fig. 3.2.7. Several of the phases which formed
when the Inconel specimens were plated with

 

3F_- E. Hoffman, ANFP Quar. Prop, Rep. [une 10G, 1956,
ORNL-2108, p 148; E. E. Hoffman ond R. Carlander,
ANP Quar. Prog. Rep. Sepe. 10, 1956, ORNL-2157, p 160,
PERIOD ENDING MARCH 31, 1957

UNCL ASSIFIED
Y-21548

UHCL ASSIFIED
Y-21647

UNCLASSIFIED
¥-2 1666

    

0
INCHES
1
1
INCHE S

B

- " “\ -_:.1_ i S I a -
i '.' v‘l' v - n l 'h"- rl"
. E’ ey ool & 5 ' » . [oDOE
* f * nh. A ® 4
N o w F q" - - =y - "
B . : £

L * et . . "
et ) | D04 - - + = o o oo+

& _J[ -cQ', ® e - B
i " } e . - .":'

 

 

 

 

 

® g e W
| AR B L T
38 geat Mol 0 VUL LA
‘:‘ G 2 .“: -::"_4 '::‘* p
om e - o N Y
-fl..!.,f". J . T f _"; _;.. "*._ o ") Q06 |
~ . - i - 2 _r_._ *
Al S ;% i i _r_:“- ey
- .‘; T ¥ 1-— . ? gy - = - --"
o s S 7 _§ ._M‘ B2 et B ¥y By ','-§.. : _ ;
El.]‘ ".; -« AS RE{?ENED U w | s S HOT LEG UEOQQF'}'.,‘_ . |»n COLD LEG (990°F) v
R de e A e O e Fome T e f Y ' ’

Fig. 3.2.6. As-Received Hastelloy B and Sections from the Hestelloy B Thermol-Convection Loop Which
Circulated Sodlum fer 1008 hr, The hot leg wos ottacked 1o a slightly greatar axtent than the cold leg. Tha as-

received material contained a few widely scattered surface irregularities. Etchant: modified agque regia. 500X,
Reduced 12%. (Confidential with coption}

Table 3.2.1. Results of Metcollographic Exomlination of Interfaces of Specimens from
Sodium-Berylium-lnconel Compatibility Test Ne. 3
Test durotion: 1000 hr
Test temperoture: 1300°F, cycled te S00°F ten times

Contact pressure between specimens: 500 gsi

 

 

interface Results of Metollographic Examination Across the lnterface
Incone! plus T-mil nickel plote plus 4-mil lnconet plus 1.5 mils of o nickel-rich ghase plus 2.5 mils of an
chremium plete vs beryllivm unidentified phase plus 3.5 mils of Bech (see Fig. 3.2, 7a)
Inconel vs beryllium, direct contect Inconal plus 1 mil of BeMi plus 18 mils of BeflNis {sae
{standard) Fig. 3.2.7b)
Inconel plus 3-mil chromium plate Inconel plus 2 mils of BeNi ptus 1 mil of chromium plus 6 mils
vs beryllium of Be_?Cr (see Fig, 3.2,7¢)
Inconel plus 3-mil chromium plate lnconel plus 2 mils of BeMNi plus 1 mil of chromium plus & mils
vs beryllium (duplicate of above) of BBECG‘

 

 

 

 

167
ANP PROJECT PROGRESS REPORT

@  UNCLASSIFIED

   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, e : Y-21647
- * s 1854
' g 338 a
g ‘ 3
)0 INCONEL  NICKEL (D (?) | Be,Cr
: s wmeA '
2 g .- E 8 8 8 mcfaes
".: ’ e - UNCL ASSIFIED
® S Y-21548
: !
: 7438 1495 ®
144 - & | &
o - 8 P % e N
INCONEL ' "* Be Ni . e .
() o .~ :
2 g 500X S 8 8 8 3 |NCI{IE5
b _ —
- UNCLASSIEIED
N Y-21646
: 30
392 Cr .- 7439
> SN e NI INC
= 2015 s 4
= s 227@CT
_,._ 0 . L
\ /

Fig. 3.2.7. Metallographic Specimens of Interfaces Described in Table 3.2.)1. The nuymbers are diomond pyramid
hardness volues for o 25-9 load. As-polished. 500X,

168
ra

both nickel and chromium have not vet been
identified. It appears that nickel plating of the
Inconel prior to chromium plating is unnecessary
and pessibly undesirable for this application.
The results found at the chromium-plcted Inconel
vs  berytlium interfaces confirm the previous
findings which indicated that a minimuym of 5
mils of chromium plate is required to ensure that
all the chromium is not consuvmed by reaction with
the beryllium vnder the conditions of this ex-
periment. No serious spalling of the brittle
compounds Be, NI, ond Be,Cr was detected in
these tests.

EFFECT OF ZIRCONIUM ON CORROSION IN
AN INCONEL-SODIUM SYSTEM

R. Carlander?

An lnconel-sodium thermal-convection loop with
a zirconium sleeve in its hot leg was operated
for 1000 hr with the hot teg at 1500°F and the
cold leg ot 1230°F in order to determine whether
the zirconium would remove oxygen from the
sodium and thus affect the corrosion of the Inconel.
Since no mass transfer and little attack occurred,
the test indicoted that the presence of zirconium
had a beneficial effect.” In an effort to sub-
stantiote these results, two Inconel standpipe
test copsules, one containing a zirconium speci-
men, the other an |nconel specimen, were 50%
filled with sedium and tested for 400 hr with the
tep and bottom of the sodium bath at 100G and
1600°F, respectively,

Mass transfer occurred in both copsules, and,
although the weight of the crystsls found in each
capsule was opproximately the same, the com-
position of the erystals (75.5% Ni—10.4% Cr-14.1%
Fe) from the copsule containing the Inconel speci-
men differed markedly from that of the crystals
(95.6% Ni=3.5% Cie~0.9% Fe) from the capsule
which contained the zirconium specimen. The
effect of zirconium on the moss-transfer ¢rystal
composition is net understood.

Visual examingtion of the =zirconium specimen
following the test indicated that no oxide scale
had formed. Metollographic exomination revealed

 

4Dn ossignment frem Pratt & Whitney Aircraft.

SR. Carlander, ANP Quar. Prog. Rep. Dec. 31, 1956,
GRNL-2221, p 192.

PERIOD ENDING MARCH 31, 1957

that o transition from alpha to beta zirconium
occurred in the center of the specimen, as was
expected as a result of heating the zirconium
from room temperature fo 1600°F. The outer
surfaces of the specimen, however, retained the
alpha structure to a depth of 40 mils, as shown
in Fig. 3.2.8. The retention of the alpho-structure
on the outer surfaces was thought to bhe due to
oxygen absorption which raised the alpha to beta
transition temperature, which is normally 1548°F,
to above the test temperature of 1600°F and thus
stabilized the olpha phase. In order to test this
hypothesis, two sets of millings from the surface
of the tested specimen, cne to o depth of 40 mils
(alpha region) and the other from a depth of 40
to 80 mils (beta region), and two sets of similar
millings from an as-received specimen were
chemically analyzed. The onolyses, reported in
Table 3.2.2, revealed that the alpha region did
absorb oxygen during the test.
ubilities of iron, nickel, ond chromium are higher

Since the sol-

in beta zirconium than in alpho zirconium, these
metais diffused from the alpha region to the beta
region during the test, as shown by the onolytical
results, The data obtained from these tests thus
indicate that the zirconium did remove oxygen
from the sodium; however, there wos no observable
reduction in the corrosion of Inconel as a result
of the oxygen removal by the zirconium.

MOLYBDENUM-BERYLLIUM COMPATIBILITY
IN SODIUM

R. Carlander

Two tests of the compatibility of molybdenum
ond beryllium in contact in static sodium were
conducted in the apparotus illystrated schema-
tically in Fig. 3,2.9. The temperoture of the
test system was maintained at 1500°F for each
of the tests, one of which was for a period of
100 hr and the other for 500 hr. Metallographic
examination of the specimens after the test re-
vealed thot o two-phase diffusion zone hed formed
between the specimens in the 100-hr test, whereas
a three-phase diffusion zone formed in the 500-hr
fest.

The examination of the specimens from the
100-hr test showed that the diffusion layer adjacent
to the beryllium specimen was the intermetallic
compound MoBe ;. This layer, which was 1 mil

169
ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED
Y-20675

 

 

 

 

 

 

 

Fig. 3.2.8. Zirconium Specimen Exposed to Sodium in an Inconel Copsule for 400 hr ot 1600“F. Outer surface is
olpha-zirconium to depth of 40 mils. The precipitates at the grein boundories are probably intermetallic compounds
el iran, nickel, and chromiom with zircoanium. Etchant: 46% Hzo-dé% HNDS_B% HF. 50X. {Confidential with

capticon)

Tabkle 3.2.2. Chemical Anolysis of Millings from As-received Zirconium and a
Zirconium Specimen Tested in Sodium in an Inconel Capsule

 

 

Composition {ppm)
Specimen e —
Iron Nickel Chremium Oxvgen

 

 

 

As-recaived zircenium
First 40 mils 350 4 28 910
Second 40 mils 350 2 25 720

As-tested zirconium
First 40 mils (alpha region) 68 8 20 2500
Second 40 mils (beta region) 850 140 55 870

e ——— e B, ey i =

 

 

170

iy
UNCLASSIFIED
ORNL-L R-0WG 17019

TYPE 316 STAINLESS STEEL CAP

     
   

| ——MOLYSDENUM POSITION ROD

& R

%, . N 5 by

L e i e o
e i

lm+—— TYPE 216 STAINLESS STEEL CLAD-
, ’ MOLYBDENUM -CAPSULE

i

 

 

- 5Q0IUM

    

 

  
 

-MOLYBDENUM TEST SPECIMEN

T

 

BERYLLIUM TEST SPECIMEN

 
   

e
F

  

MOLYBDENUM DISK

B
\ T "' . i
e T T el

0 i 2
ey

——

SCALE iN {NCHES

Fig. 3.2.%
bility Test Capsule.

Molybdenum-Beryliium-Sodlum Compati-

thick and pink, was crocked and had a hardness
of 13598 VHN (50-g load). The layer adjacent
to the molybdenum specimen was MoBe,. It was
0.5 mil thick and white, and it had @ hardress
of 2107 VHN. The identifications of the inter-
metallic compounds MaBe,, and MoBe, were
based on previous investigations.®

The layer formed adjacent to the beryllium
the 500-hr test spalled from the
surfoce when the bond between the specimens
separated during preparation for metallographic
The moterial that spalled has not
been positively identified, but it appears to be
It may have formed as a result

specimen in

examination.

beryllium oxide,
of oxygen contamination of the sodium. The
middle layer of the three-phase diffusion zone was
MoBe,.. This loyer, which was 2.5 mils deep ond

 

65. G, Gordon, J. A, McGurty, and G. E. Klein, Inter
metaliic Compounds in the System Molybdenum Beryllium,
NEP A-1686 (Jan. 10, 1951),

PERIOD ENDING MARCH 31, 1957

pink, was cracked and had a hardness of 1598 VHN,
The loyer adjacent to the molybdenum specimen
was MoBe,. It was 2.5 mils deep and white, and it
had a hardness of 2390 VHN. The specimens
from the 500-hr test are shown in Fig. 3.2.10.

The inherent brittleness of the two intermetallic

compounds formed in the tests indicates thot
the solid-solid diffusion that occurs between
molybdenum and beryllivm ot 1500°PF is of a

deleterious nature that would seriously affect
the structural properties of the twe metals in o
high-temperatyre application.  Additiona! tests
in which a 500-psi pressure will be applied to
the specimens are to be conducted. The results
of these tests will be compared with the results
obtained in inconel-beryllium-sodium compatibility

tests.

TITANIUM CARBIDE-NICKEL CERMETS
iN MOLTEN SODIUM

¥W. H. Coock

Specimens of the four TiC-nickel cermets iisted
below were corresion tested in sodium in Inconel
capsules for 100 hr in a seesow furnace operating
ot | ¢pm with hot- ond cold-zone temperatures

of 1500 and 1120°F, respectively.

K150A: 80% TiC—10% NbTaTiC, - 10% Ni
K1S1A: 70% TiC—10% NbTaTiC,-20% Ni

K152B: 64% TiC—6% NbTaTiC,-30% Ni

K162B:  64% TiC—6% NbTaTiC,~25% Ni-5% Mo.

The specimens were retained in the hot zenes of
the capsuvles during the test, The surfoces of oll
four cermets were roughened to depth of 0.5 mil,
but the K1528 and K162B specimens appeared
tc be the most severely attacked. The ottack
in each case may have been on on Ni-TiC solid
solution along the TiC-binder interface, since
it appeared that the metal binder had been removed
and that the TiC grains were slightly reduced
in size. The typical appearance of the attack
is shown in Fig. 3.2.11. Chemical onalyses of
the sodium baths after the tests showed erratic
scatter of the results from less than 15 te approxi-
mately 60 ppm of fitanium or nickel in 20-g
samples.

171
ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED
Y-20794

 

ik
T
INC!;‘E‘S

o
b=t
A

 

 

 

 

 

¢ os

500X

DIAMOND PYRAMID HARDNESS
50-g LOAD

" § 008
- =

Fig., 3.2.10. Intermetailic Compounds Formed in Compatlbility Tests of Molybdenum and Beryllium in Stotic
Sodium After 500 he ot 1500°F. Layer with hardness of 1598 is MoBe|3. Loyer with hardness of 2390 is MDEOT
As-polished, 500X, (Confidential with coptien)

 

 

 

—-., UNCLASSIFIED ™ e . o UNCLASSIFIED
21261 fl '%‘!" Y-212462
- ‘ o
i LB q SN I S S
= % : = 1"& » _z' 1 il ¥ ‘“!. ~ 4 3

 

 

 

 

 

T
' : - V O kel
B A f"}\?-"\" L, $= 7
L {L ! ;‘-: J-T . i ow % “r; : Y.'b " - &
-.r ;; ” 1 '\/ \ f-:l- y » 3 : \ -': &' -A_’-— ;‘
e - \-_‘;_ i..' . A .‘n,.' \.- %4 > ; i e
r dob g * o2 1o . ’t . :\.' ' ".1_‘ .l'-
o - : L d u I 7
SRR ERR N T
oy A * w T Ll = L -
fllfl # k k l"i' N = ‘)‘ i \ . ‘fi ® \ I
- \ ..:l,‘; * . m‘ ‘l-‘fi e 1'?’. ?.. ll- nl * ‘*- A ‘.'
- ", F - - ;"“ A YR 4 i { X
‘2“@"’ .y = i, 3 7 ‘{-1 i_‘-_fii_‘ = -’\- . X Y.* % .:‘
= .,\."f r 4 i i 1 Bt oW “" L'J
3 " “ % 3 L 0
i l#‘x = 'fi-:' o3| ™ i% ? "& L :. Ta JC?A- i
3 @'/..- o i}" (5) *®q: b7 2 :. i L .'g. £ o
e SN ST AT e E e 2 ‘h

Fig. 3.2.11, Cermet XKI151A (4) As-Received and {b} After a 100.-hr Exposure to Molten Sodium in ¢ Seesow
Furnoce with Hot- and Cold-Zone Tempecatures of 1500 and 1120°F, Respectively. The white phase is nickel and
the large gray, angular particles ore TiC. Both the untested ond the tested specimens were nickesl plated to protect
the edges during polishing. Unetched. 1000X. (Confidential with caption)

172
In similar tesis in NaK (50-50 wt %), performed
ot Battellie. Memorial Institute, with hot- and cold-
zone temperotures of 1600 and 1200°F, respec-
tively,” there was surface roughening to o depth
of 0.5 mil on cil these cermets, except K162B,
in 110-hr exposures. Previous similar tests at
ORNL skowed that these four cermets were not
attacked by MNaF-ZeF -UF, (53.5-40-6.5 mole %,
fuel 44).3

BORON CARBIDE AND BORON NITRIDE IN
MOLTEN S50DIUM

W. H. Cook

The ceramics baron carbide ond boron nifride
are of interest as refroctory materiels for use
in reactors as thermal-neutron obsorbers, In
recent stondard screening fests of ceramics in
liquid metals, two boron corbide specimens,
with theoreticol densities of 85 and 90%, re-
spectively, had better corrosion resistance to
molten sodium than that previously observed.®

Each of the B,C specimens was exposed to
static  sodium for 100 % at 1500°F in Inconel
containers. 1he results of these tests are com-
pared with o result of an earlier test in Table
3,23, Since it is difficult to completely remove
sodium and sodivm compounds from a porous
ceramic body, 1t is surprising that the calculated

 

?5. J. Besham, J. H. 5tong, and E. M. Simens, Tilting-
Furnace Corrosion Screening of Materials Exposed o
Molten Fluoride Salt and NakK, BMI-1132, pp 46, 48
(Sept. 18, 1956).

8E, E. Hoffman, W. H. Cook, end C. F. Leitten, Jr.,
ANP Quar. Prog. Rep. March 10, 1955, ORNL- 1854, p 84.

?E. E. Hoffman, W. H. Cook, ond C. F. Leitten, ANP
Quar. Prog. Rep. Dec. 10, 1954, ORNL-1814, p 84.

PERIOD ENDING MARCH 31, 1957

and measured specimen weight losses given in
Table 3.2.3 agree so well. The results calculated
from the chemical onalyses are considered to
be the most important in determining the extent
of corrosion. The only visible changes in the
tested specimens were that they both had duller
surfaces than they had before the tests. The
8,C of 100% theoretical density, which was
tested previously,g cracked and fell aport after
exposure to sodium.

A specimen of boron nitride {nominally, ]{1 % ]a:'
% 1/2 in.) with a density of 98% of theoretical
(theoretical density of BN = 2.20 g/cm® was
severely attacked in o 100-hr exposure to malten
sodium in a seesaw furnace operating ot | cpm
with hot- ond cold-zone temperatures of 1500
ond 115(PF, respectively. The specimen container
was [nconel.

The measured weight loss of the specimen was
10.8%, ond the weight loss based on chemical
analyses of the sodium after the test wos 8.2%
if all the boron found was in the form of BN before
the test. Since the BN was extremely soft even
in the dense state, some of the weight loss could
have been due to atirition between the specimen
and its container. However, metallographic exomi-
nation with reflected light showed thot the as-
tested specimen hod develoeped crocks below
and paralle!l to its surfoces. Under polarized
light, the specimen appeared to have cltered
throughout.  There was o dense, shell-like ma-
terial approximately 0.003 in. thick surrounding
the core of the specimen. This ““shell’’ contained
the cracks observed with the reflected light. The
core of the specimen oppeared to be less dense
than it was before the test and to be covered
with black spots.

Table 3.2.3, Results of Exposure of Boron Cerbide Specimens to Static Sedium for 100 hr ot 1500°F

 

B,C Specimen Density Dimensional

Weight Loss of Specimens {%)

 

 

 

i By Weight
{% of thearatical™) Change (%) ~
5 Measurements Calculated
’\"85 "'04..' ].6 1.6
o ~0-04 0.4 0.3
100 6

 

*Theoretical density = 2,5} gz"crna.

**This represents the weight toss calculeted on the basis that oll the boron found in the soedium by chamical

onolyses ofter the test was originally Bdc in the specimen,

173
ANP PROJECT PROGRESS REPORT

S5iC-51 I MOLTEN 50DIUM
W. H. Cook

The silicon was completely removed from an
SiC-Si specimen during 100 hr of exposure to
molten sodivm in an lnconel container in o seesaw
furnace operating at 1 cpm with het+ and cold-zone
temperatures of 1500 and 1150°F, respectively,
The silicon was found on the walls of the Inconel
container.  [he specimen was submitted as a
dense SiC moterial.'® The nominal composition
of this type of material is 85 10 0% SiC plus
10 to 15% free silicon plus minor quantities of
iron, alumina, and silica, The free silicon probably
results from the fobricotion procedure, and the
minor constituents aid in the densification of
the SiC.'! As-received and tested specimens
are compared in Fig. 3,2,12.

 

Designoted Body 4107-22-7 by the Carborundum
Campaony, Perth Amboy, N. J.

TR, A. Alliegro, L. B. Coffin, and J. L. Tinklepaugh,
f. Am. Ceram. Soc., 39, 386-387 (1956).

CLASSIFIED 7
Y-21248

The resuvlts of this test were consistent with
previous resulis obtoined in o static test of o
different type of SiC-Si matericl!? ond in similar
tests of pure silicon ond silicon-containing alloys.
The nature and rapid rgte of the silicon removal
in these tests indicate thot the silicon was
ottacked ond that perhaps the ottack was assisted
by concentration-gradient moss tronsfer. These
data appear to disagree with reporfsw that sodium
does not react with silicon.

LEAD-LITHIUM ALLCY INWATER
D. H. Jansen

Lead-tithium alloy specimens were tested in
distilled water at room temperature and at the
boiling peint, and cracks and dimensienal in-
creases, such as those shown in Fig. 3.2.13,
were observed after all the tests. The lithium

 

12\'1'. M. Cook, Met. Semianm Prog. Rep, Oct, 10, 1953,
ORNL-1988, ¢ 17.

B Sittig, Sodium: Its Manufacture, Properties, and
Uses, pp 53, 243, Reinhold, New York, 1956,

   
 

o UNCLASSIFIED
' Y-21249

 

INCHES

Q.02

 

Q.03

 

 

 

 

 

Fig. 3.2.12. Silicon Corbide~3ilicon (2) As-Received and {(4) After Exposure for 100 fr to Sodium in o Seesow
Fumoce with Hot- and Cold-Zone Temperatures of 1500 and 1150°F, Respectively. In {a)} the white areas ore

silicen and the groy areas are 5iC; in (&) the silico

caption)

174

n has been removed, Unetchad. 100X, {Confidentol with

A
content of the water ofter the tests and the weight
chonges of the individuo! samples are given in
Table 3.2.4. The loss of lithium to the water
in the 100-hr room-temperature test was 16.6
mg/in.? of alloy surface, and it has been cal-
culeted thot this would be equivclent to ohout
53 g of lithium lost to the woter in the Bulk
Shieiding Facility {BSF) in 100 hr from a 36-in.-
square, 4-in.-thick slab. |f the 4-in.-thick slab
were made up of eight 1/2-in. plates with water
between them, os proposed, the amount of lithium
leoched by the woter would be about 350 g in
100 he.

UNMCL ASSIFIED
¥-21522

{a),

 

Fig. 3.2.33, Lead-Lithivm Alloy Specimens (0.65
wt % Li) [a) As-Received, (#) After Exposure to Warer
for 100 hr ot Room Temperature, ond (c} After T00 hr
in Boiling Weter. (Confidential with coption}

FERIOD ENDING MARCH 31, 1957

The lead-lithium spectrometer housing proposed
for use in the BSF is to be poured at approximately
400°PC ond ollowed 1o solidify slowly. Therefore
a test was conducted to determine whether gross
segregation would occur in the alloy if the housing
were fabricated in this manner. The test was
conducted by filling a !fi-in.-OD, 30-in.-long Armco
iron capsule with lead-lithium alloy to o height
of 24 in. The copsule was held in o verticol
position, heated for 4 hr ot 400°C, and allowed
to cool in the furnace. Samples of the alloy were
taken from various locations in the copsule
and analyzed for lithium content. The analyses
showed the lithium contents at the top and bottom
of the column to be about the some, 0.73 and
0.70 wt %, but a sample from the center showed
an abnormally high lithium content, 0.94%. An-
other fest is planned in which the capsule will
be ciimped ot varicus places along its length
in order fo isolate one section from another while
In this monner it can
be determined whether segregation takes place
while the alloy is in the liquid state.

An ottempt was made to cast o 36-in.-square,

the alloy is still molten.

!’Z-in.-Thick plate of the alloy, and it wes found
that dimensioncl tolerances could not be adhered
te and thot there was some poresity in the casting.
Therefore, it hos been decided to cost s slab
21 in. square, but thicker ('l?{fi in.), and have it
relled to the required finished dimensions (36 x 36
x 1;’2 in.).
VOLATILITY PILOT PLANT CORROSION
STUDIES
E. E. Hoffman L. R. Trotter

Investigations of the various corrosion problems
relative to operation of the volatility pilot plont
for fuel reprocessing were continved. These

Table 3.2.4. Resuhlts of Cotrosion Tests of Lead-Lithium Alloy Specimens (0,65 wi % Li) in Distilled Woter

 

 

 

Water Test Specimen Lithium Last per Unit
T emperature Duration Weight Gain of Alloy Surface
CF) (he) (%) (mg/in.?)
212 50 +0.5 67
212 1490 +0.9 Sample lost
72 50 +0.3 7
72 100 +0.6 T

 

 

175
ANP PROJECT PROGRESS REPORT

studies ore being performed in cooperation with
members of the Chemical Technelogy Division and
are reported in Chap. 4.2, ""Fuel Recovery and
Reprocessing,’' of this report.

EXPERIMENTAL STUDIES WITH MOLTEN
LITHIUM '

E. E. Hoffman T. Hikide?

An investigotion of the feasibility of ysing
tithium as o reactor coolont was begun at NDA
in July 1955 under sponsorship of the Aircraft
Reactors Bronch of the AEC, In July 1956 co-
ordination of the centract was transferred to the
Metallurgy Division of ORNL, The results of the
work in fiscal year 1956 were reported by NDA, !¢

The over-all objectives of the fiscal yeor 1957
program were to defermine, principally by the
use of thermal-convection loops, the effects of
nitrogen and oxygen impurities on mass transfer
in type 316 stainless steel-lithivm systems.
Other container moterials thot were to be in-
vestigated were type 430 stainless steel, niobium,
and possibly molybdenum.  Fundemental inves-
tigations were to include o study of the rate of

 

I‘Repnrr on subcontrect work performed at Nuclear
Development Corporation of America (NDA), White Plains,
M. Y. This subcontract is coordinated at QORNL by
E. E. Hoffmeon and T. Hikide.

'ISOn gssignment from USAF.

R. €. Ross, Quarterly Progress Report and Reactor
Development fuly 1. 1955 Through September 30, 1953,
NDA-15 (Qet. 21, 1955); Quarterly Progress Repori and
Reactor Development October I, 1955 flmugb December
31, 1955, NDA-20 {Jan. 23, 1958); Quarterly Progress
Report and Reactor Development fanuary I, ]9);6 Through
March 31, 1956, NDA-25 {Moy 1, 19568); J. M. McKee and
R. C. Ross, Quarterly Progress Report and Reactor
Development April I, 1956 Through june 30, 1956, NDA-
32 (Dec. 24, 1956}

176

solution of conteiner meterials in pure lithium,
the vultimate solubility of container materials
in fithium, and possibly the solubility of tithium
nitride in lithiom.

Considerable work on the development and
improvement of onalytical techniques for deter-
mining the nitrogen, oxygen, and corbon contents
of lithium wos performed ot NDA in fiscal year
1956 and has been continued into fiscal vyear
1957. The thermoi-convection loop design was
standardized so that ORNL and NDA test results
covld be correlated. A lithium still has been in
operation at NDA for several months, ond oll
tests are performed with the high-purity material
{less than 20 ppm of N,) produced in this still,

Among the most significant results obtained
to date are indications that titanium end zi¢conium
are benefical in increasing the time required for
type 316 stuainless steel Joops (hot leg, 1600°F;
cold leg, T1100°F) to plug with moss-transfer
crystals. Loops containing titaniuvm and zirconium
operated for approximately 500 hr, whereas standard

loops with no additions plugged in roughly 100 he.

Cold tropping of the lithium by means of g
positive-flow cold trap to reduce the impurity
content shows promise of incregsing the time
required for stainless steel loops to plug. Three
recently operated loops of type 316 stainless
steel, fabricated from an ORNMNL design {hot and
cold legs 30 in. long, horizontal connecting legs
18 in. long), plugged with mass-transfer crystals in
95 £ 25 hr. The results of the tests will be used
as base-line data for future tests in which attempts
will be made to study the effect on mass transfer
aof nitrogen, oxygen, titonium, zirconium, and
sositive-flow cold traps. No solution rate data
have been obtained to date, but tests have been
initioted.
PERIOD ENDING MARCH 371, 1957

3.3, FABRICATION RESEARCH
J. H. Coobs

NICKEL-MOLYBDENUM ALLOY DEYELOPMENT

H. Inouye T. K. Roche

General $tatus of Development Program

The progrom for developing nickel-molybdenum
alloy container materials for fluoride fuels that
have better heat transfer properties than those of
lnconel was continued during this quorter, The
mosi favorable characteristics of the new alioys in
comparison with [nconel are their superior corrosion
resistonce and strength. The corrosion resistance
of nearly every one of the new alloys tested in
NoF -KF-LiF-UF, (11.2-41-45.3-2.5 mole %, fuel
107) was equivalent to that of Hastelloy B, and the
average stress-rupture life at 1500°F and a stress
of 8000 psi was about 400 hr, which is better than
that of Inconel by a factor of 4. These properties
were attained without seriously offecting the
fabricability, ond the new alloys do not become
brittle during aging or welding.

In terms of reactor performance, it appears that
the nickel-motybdenum alioys will permit the use
of o more efficient fuel than that planned for use in
the ART and probably an increase of 150°F over
tie ART operating temperature.  Consequently,
every effort is being made to translate laboratory
dato into commercial practice. With this objective,
the effort is being concentrated on one or two
promising compositions that are to be made in
ptlot-plont quantities at the Westinghouse Electric
Corporation plant ot Blairsville,

The properties of the alloys that have been

Pennsylvenia.

prepared in the faboratery are summorized below:

Fabricability., — The fabricability of the nickel-
molybdenum alloy depends on the guantities of
other elements that must be added to provide the
necessary strength, The indications are that the
fabricability will be between that of Inconel and
that of Hastelloy B.

Corrosion Resistance, — The corrosion resistance
of the new atloys in fuel 107 and in sodium at
metal temperatures of 1650°F is equivalent to that
of Hastelloy B when the criterion is the depth of
corrosion. A more sensitive indication of corrosion
susceptibility is the chenge in the chemical compo-
sition of the fuel as a result of contact with the
indicaoted that

alloy. The corresion tests hove

chromium, aluminum, ond tungsten should be held
to a minimum in the new alloys.

Stress-Rupture Properties, — The stress-rupture
data show that, in most coses, the new alloys fall
short of the goal of o rupture life of 1000 hr at
8000 psi and o temperature of 1500°F; however,
the considerable scotter of the data prevents
accurate strength estimates, The average creep
and stress-rupture properties of the various compo-
sitions tested to date are plotted in Fig. 3.3.1.
Since dota from atl the alloys tested were gveraged,
it is expected that the composition selected for
further development will have properties somewhat
better than those shown in Fig. 3.1.1.

Joining Characteristics. — The weldability of
these olloys has not been determined as fully as
desired. Hot ductility tests on some of the olloys
indicate that titanium and oluminum are undesirable,
as ore nicbivm and fungsten when used in con-
junction with a high carbon content. Similerly, for
dry-hydrogen brazing the alloys should contain o
minimum of aluminum and titanium,

Oxidation Resistance, — The oxidation rote of
nearly every composition has been determined to be
tolerable in that over 60 thermal-convection loops
have been operated in air without failures of-
tributable to oxidation. The metal temperature
attained in these tests was approximately 1850°F
for periods of up to 1000 hr. The alloys that showed
the most oxidation contained tungsten ond niobium,

Aging Characteristics. — The stability of each
alloy wos determined by observing the changes in
tensile strength, stress-rupture properties, and
hardngss after agingat 1300°F, The agingresponse
was significant in compositions that contained 0.25%
Be or 2.0% Al + 1.5% Ti. Aslightresponse to aging
was observed for olloys containing 1.0% Al + 1.5%
Ti.

Summaries of the properties of laboratory heots
of nickel-molybdenum alloys INOR-T through INOR-
8 are presented in Tables 3.3,1 through 3.3.8. The
tuel mixtures referred to in the tobles have the
foliowing compositions:

Fuel 30 NoF-Z¢F -UE, (50-46-4 mote %)
Fuel 107 NaF-KF-LiF-UF, {11.2-41-45,3-2.5

mole %)
ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UNCLASSIFIED -
ORNL-LR-0WG 20985 =
e o e ey s e o L L LRt
E e et W W LT W Ve ) LN _ | T ' [
| | - -_¢_‘_T o - T e
=T It ] —— A e | T
[— — . —— e 1 — - P s e -~ —'— 'r———l—-— — _— —] —— i e ek . .l - ”
| | —|— | | ! |
50000 | e =l = = r__| T‘ il — 14 ] 1
T i a [ o L 1 - s e ll l_._. el el - N i !_‘, 14 '
. | .
| [ | | | | | | l | |
| | S | == -
— T T e T T
n
| | J- | | |
20,000 e ———— e .—L—.—lr i | — et L B ;
| | | | | | |
E i | | l | l ! | | ] |
= | | /~ RUPTURE | | | ] ,
2 10000 L 5% ELONGATION | ¥l e
E’ ' e = - JI/:»—'{ 1~ 2% ELONGATION | ——1 11
i ] | - A | 111
R e T % . [
— = =115 - : i =
sooo f— L 0 L1 P17 ;&;‘: -
= _z
— == — —RUPTURE LIFE OF INCONEL | 1
j.__ . — 8 : L 4 e LR - 2
' ] | t Fo ELONGATION —
| |
2000 40 gl L ._! =

 

20 50 1000 2000

TEST PERIOD (hr)

100

 

5000 {0,000

Flg. 3.3.1. Average Creep ond Stress-Rupture Propertias of Nickel-Molybdenum Alloys at 1500°F in Argon and

in MaF-K F-LiF-UF4 (11.2-41-45,3-2.5 Mole %, Fuel 107), (Secret with caption)

Table 3.3.1. Properties of Nickel-Melybdenum Alloy INOR-1

Ni

——

Bal

Mo Other

15-20

 

 

 

Composition, wt %: 0.06 C; 0.5 Al
Mumber of heats: 7

Mumber of compositions: 3

Fabricability: Equivalent to that of Inconel

Oxidation resistance: Inadequate; minimum resistance with about 20% molybdanum

Joining: Brazoble by conventional techniques; weldability between that of Inconel ond thot of Hastelloy B

Stress-ruptura: 1000-hr rupture strength ot 1500°F about 4500 psi [extrapolated)}

Corrosion: [n fuel 30 for 1000 hr at 1500°F, 1 mil
In fuel 107 for 1000 hr at 1500°F, 2 mils

Aging choracteristics: None

Remarks: All propertiss favorable except oxidation resistance and strength

 

178
|f_°£.

PERICD ENDING MARCH 131, 1957

Table 3.3.2. Proparties of Mickel-Malybdenum Alloy INOR-2

 

Mo Cr Othee Ni

Composition, wt % 16-20 32-10 0.5 Al, 0.5 #n, 5 Nb, 0.1 C (max), 0.25 Ce (max} Bal

 

Number of heats: 16

Mumbker of compositions: 8

Fabricability: More difficult te fabricate than lacanesl

Oxidation resistonce! Sotisfoctory only with 7% er more chromiuvm

Jeining: Brezable by conventional techniques; waldobility between that of Incenel and that of Hostel loy 8

Stress-rupture: 1000-hr rupturs streagth at 1500°F {extrapoloted)
With 2% Cr 4500 psi
With 5% Cr 6000 psi
With 7% Cr 6500 psi

Corrosion: In fuel 30 for 1000 hr at 1500°F, 0.5 to ¥ mil

n fuel 107 for 500 hr at 1500°F, 1 mil

n fuel 107 for 1000 hr ot 1500°F, 2 mils

n sedium for 1000 hr ot 1500°F, 0 to 1 mil, slight mass transfer

—_—

Aging characteristics: Nane
Romarks: Corrosion in fuel 107 progresses from surface pits to subsurfoce voids as chromium content of
alloy is increased; pickup of chromium in fuel olso increases with increasing chromium content

of alloy

 

 

 

 

————— ———— i —

Mo Ti Al C Mi

Composition, wt % 15-16 1.5 0.5~1.0 0.6 Bal

 

Number of heats: 3
Number of compositions: 2

Fobricability: More difficult to fabcicate thon lnconel; sensitive to changes in forging temperature ond raduc-

tion schedule
Oxidation resistance: Inodequote; oxide scole spolls on coeling
Joining: Net brezoble in dry hydrogen; weldability similar to that of lnconel X
Stress-rupture; 1000-hr rupture strength at 1500°F about 5000 psi (extrapolated)

Corrosion: [n fuet 107 for 500 hr at 1500°F, 1 mil of penetrotion consisting of heavy pits ond subsurface voids,
with metal deposits in the cold leg

Aging choracteristics; Solution-annealed samples show oging tendencies during stress-rupture tests ot 1500°F;
no response observed when anneaied at 1900°F and oged ot 1300°F

Remacks: Stress-cupture ductility improved by vocuum melting

 

 

 

179
ANP PROJECT PROGRESS REPORT

Takle 3.3.4. Properties of Nickel-Molybdenom Alloy INOR-4

 

Mo Ti Al C Ni

———

Compositien, wt %: 16 1.5 2.0 m Bal

Number of heats: 4

Fabricability: Vacuum melts similar to INOR-3; air melts more difficult
Oxidation resistance: tnodeguate; oxide scale spalls en cooling

Jeining: Not brozable in dry hydrogen; weldability equivolent to that of Inconel X

Stress-ruptura: 1000-hr rupture strength at 1500°F obout 5500 psi {extropoloted); ductility generally less thon
5%

Corrosion: In fue) 107 for 500 hr at 1500°F, 3 mils, accompanied by heavy pits and subsurface voids to o depth
of 2 mils

Aging choracteristics: Age hardens at 1300°F, but oves-ages at 1500°F; ductility obout 2% when tested
ot 1300°F

Remarks: Adverse properties in oll categories

 

 

Toble 3.3.5. Properties of Nickel-Molybdepum Alloy INOR-5

 

 

Mo Mb W C Other Mi
Composition, wt % 15~14 2-3 2-3 0.06--0.25 0.5 Al, 0.5 Mn Bal

 

 

 

Number of heats: 7
Number of compositions: 5
Fabricability: Maore difficult to fobricote thon Inconel; corbon content greoter than 0.10% net recommendad

Oxidotion rasistonce: Inodeguate; oxide scele spalis on cooling; oxidation rate higher than that of other alloys

because of Wngsten cantent
Joining: Brozable by conventicnal techniques; weldability equivalent to thet of Inconel X
Stegss-rupture: 10Q0=he rupture strength at 1500°F about 6000 psi (extrapolated)

Corrosion: In fuel 107 for 500 hr at 1500°F, no apporent corrosion
In sodium for 1000 Ar et 1500°F, 1.5 mils, with subsurfoce voids and slight mass transfer

Aging cheracteristics: None

Remarks: Stress-rupture strength influenced by heat treatment; variations in carbon content affect fabricability
and strength; 1400 ppm of tungsten picked up in fuel during corrosion test

 

180
PERIOD ENDING MARCH 31, 1957

Table 3.3.6. Properties of Nickel-Melybdenuvm Alloy INOR-6

 

Mo Cr Ti Al Qthear MNi
Compositien, wt %: 14 5 1.5 1.0 0.02-0.06 C, 0.5 Mn Bo!
Number of heots: 4

Number of compositions: 1

Fabricabifiry: Cite difficult to fobricate but sesiar than Hostelloy B

Oxidation resistance: Marginal; oxide scale slightly unstable during cycling
Joining: Not brozable in dry hydrogen; weldability equivalent to that of lnconel X

Stress-rupturer 1000-hr rupture strength ot 1500°F about 6000 psi lextrapolotad)

Corrogion: o fue! 107 for 500 hr ot ]SUOOF, heavy subsurface voids and pits to a depth of 2 milsg
In sodium for 1000 hr ot 1500°F, heovy subsurfoce voids ond pits to o depth of 1 mil, mass-transfer

deposits in cold leg
Aging characteristics: Slight age hardening at 1300°F; none observed ot 1500°F

RBemarks: Air mefts difficult to fabricate

 

Table 3.3.7. Properties of Nickel-Molybdenum Alloy INOR-7

 

Mo Mb Al Cr Othar Mi

———

Composition, wt %: 17-20 1-2 1.0 6.7 0,06-0.12 C, 0,5-0.8 Mn Bal

 

 

MNumber of heats: 4

Mumber of compositions: 3

Fabricability: Similer te thot of Hastelloy B
Oxidation resistance: Equivatent to that of inconel

Jeining: Mot &rezable in dry hydrogen; weldobility poor for compositions conteining the upper limits of the

quantities shown chove
Stress-ruptore: 1000-hr rupture strength at 1560°F about 8000 psi
Corrosion: In fuel 107 for 500 hr at 1500°F, heavy subsurface voids ond pits to o depth of 2 mils
Aging characteristies: Insufficient doto

Remarks: Proparties sensitive to small changes in malybdenum and gluminum content, high-carbon olloys not
emenable to tube reducing; 755 ppm of chromium and 4000 ppm of aluminum picked up in fusl during

corros1an fest

 

 

 

181
ANP PROJECT PROGRESS REFORT

Toble 3.3.8. Properties of Nickel-Molybdenum Alley IMOR-8B

 

Ma Cr Fe

Other Mi

 

Composition, wt %: 10-24 3-10 4-10
Mumber of heats: 25

Mumber of compositions: 18

 

0.06 C, 0.5 Al, 0.5 Mn Bl

Fabricability: Ranges from that of lnconel te that of Hostelloy B, depending on additives

Oxidation resistance: Sotisfactery when chromium content is greoter than 6%

Joining: Brozable by conventianal methads; weldability between that of Inconel and that of Hastelloy B
Stress-ruptu:e: 1000-hr rupture slranglh a¥ ]50001: about 7000 ro 3000 psi

Carrosion: In fuel 30 for 1000 he ot 1500°F, | mil of subsurface voids
in fuel 107 for 1000 hr at 1500°F, 0.5 to 2 mils of subsurface voids

In sodivm for 1000 hr ot 1500°F, 1 mil of carresion and moss-transfer deposits

Aging characteristics: Alloys that contain less thon 20% molybdanum show no response to oging

Remorks: Wide ronge of strengths possible without detracting from other favorable properties

 

Praperties of Alloy INOR-8

It appears at this time that the alloy INOR-8,
which has as its basis the Ni-Me-Cr-Fe system,
will offer the best compromise for meeting the
requirements of a container material for fused
fluorides, In order to investigate the mechanical
properties developed in alloys in this system, as
well as the mechanical property trends that resuit
from changes in the total alley content, the tensile
and stress-rupture properties of a series of alloys,
included within the composition limits described
below, are being determined:

10, 15, and 20 wt %
5 7, and 10 wt %

Molybdenum levels
Chromivm lavels

4 and 10 wt %

lran lewsls

QOther 0.5 wt % Al, 0.5 wt % Mn,
0.06 wt % C
Mickel Balance

Cighteen compositions, approximotely 315 ib
eoch, were melted vunder vacvum and cast in
graphite, Aluminum, mangonese, ond carbon were
odded to eoch chorge, primarily for deoxidation of
the melt. Each ingot was rolled to 0.065-in.-thick
strip from which twe 1l-in,-long creep specimens
were prepared. The remainder of the material was
rolled to 0.040-in.-thick strip from which die-cut
tensile specimens will be blanked.

Stress-rupture data were obtained for six of the
alloys in the solution-anneoled condition ot a

182

stress of 10,000 psi and a temperature of 1500°F
in argon. The results are presented in Tabte 3.3.9.
Although the stress-rupture properties do not follow
the anticipated trends, the alloys are stronger than
lnconel ot this stress level and approach the
strength of type 316 stainless steel,

A Torced-circulation loop is to be fabricoted from
an INOR-8 alioy centaining 17% Mo-10% Cr-7%
Fe-0.06% C-bol Ni in which NeF-KF-LiF-UF,
{(11.2-41.45.3-2,.5 mole %, fuel 107) will be circu-
lated, Therefore two 3.5-ib ingots with the same
nominal compoesition, alloys VT-51 and VT-52, were
cast for preparing stress-rupture specimens for
testing in fuel 107, One melt, VT-51, was prepared
from elemental stock, and the other, VT-52, from
Inconel diluted with Ni, Me, and Fe., Actual
onalyses of these two aloys showed the molybde-
num content to be approximately 19.5%; however,
the composition is still brocketed by the limits
described for the study of the INOR-8 alloy system,
Both these olloys had 1000-hr rupture strengths of
8000 psi ot 1500°F in fuel 107. Metollogrophic
studies of alloys VT-51, VT-43 (10% Mo-5% Cr~4%
Fe-0.06% C-bal Ni), ond VT-49 (15% Mo-7%
Cr—4% Fe-0.06% C-~bal Ni) indicated that ¥T-5]
responded to aging ot 1300 end 1500°F and thus
nad higher strength than alloys VT-43 and V7-49,
which did not show this gross aging response. It
therefore appears that higher sirengths can be
achieved in the Ni-Mo-Cr-Fe system by raising the
quantities of the elements added to the nickel.
Further evoluation studies of the alloys prepared

»
FERIOD ENDING MARCH 31, 1957

Table 3.3.9. Stress-Rupture Dato for INOR-8 Alloys Bosed en the Ni-Mo-Cr-Fe System

Stress:

Temperature:

10,000 psi

1500°F

Environment: Argon

 

 

Alley No. Nominal Composition (wt %} Rupture Life Elangation

{hr} (%)
VT-43 10 Mo—5 Cr—4 Fe-0.5 Aj—-0D.5 Mn —0.06 C—bal Ni 234 32
VT-44 10 Mo =5 Cr=10 Fe~0.5 Al-D.5 Mn—0.06 C—ba} Mi 152 21
Y T-47 10 Mo =7 Cr—4 Fe-0,5 Al—0.5 Mn —0.06 C—bal Ni 85 1
Y T-48 10 Mo—7 Cr=10 Fe—0.5 Al=D.5 Mn—0.06 C—ba} Ni 106 16
VT-45 15 Mo—5 Cr—4 Fe—0.5 Al-0.5 Mn —0.06 C—bal Ni 192 16
VT-46 15 Mo—5 Cr—10 Fa—0.5 Al—0.5 Mn—0.06 C—ba! Ni 152 24

 

in the [NOR-8 system ore to be carried out with
particular attention to the siress-rupture properties
and aging characteristics of the materials,

Properties of Alloy INOR-9

Compositions designated INOR-9 are prepared by
adding Nb, V, or Fe plus C to the besic Ni-Mo
compos ition. Certain elements, namely, Cr, Al, Ti,
and ¥, are excluded because of their corrosion
tendencies in fuel 107. Previous studies of the
oxidation choracteristics of nickel-molybdenum
clloys showed that stable oxide scales were formed
when the molybdenum content was lower than 15%.
The strengths of the solid-solution alloys of nickel
ond molybdeaum show strong dependence on the
molybdenum content, and thus support is given to
the theory that solid-solution strengthening is
dependent upon the extent of alloying.! {n order to
maintain the molybdenum level as low as possible
for oxidation resistance and yet to have the highly
olloyed composition needed for strength, niobium
additions are being made to the nickel-molybdenum
base alloy, The compositions which exhibit mexi-
mum solid solubility ore being determined, and their
oxidation characteristics in the tamperature range
of interest are being studied. For practical reasons
the alleys contain iren, which is derived from the
ferroniobium melting stock, The compesitiens and
the properties of the alloys being studied ore given

in Table 3.3.10.

 

TAs used here, "'extent of olloyin refers to the
kinds and quaoatities of elements added te the basic
nickel-molybdenum atlloy,

 

Fabrication of ORNL Experimental Alays

Three experimental compositions were extruded
to tube blanks which are presently being processed
into seamless tubing at the Superior Tube Co. for
fobrication of ferced-circulation loops which will
be operated with fuvel 107. Six tube blonks of each
composition were fobricoted, without difficulty,
at o temperature of 2150°F and an extrusion rotio
of 7:1. The compositions which have been ex-
truded are described in Taohle 3.3.11. la view of
the time delay anticipated in the fabrication of
tubing frem the INOR-8 alloys, two laboratory heats
of the composition 15% Mo-6% Cr-5% Fe-0.5%
Mn=0.5% Al-0.06% C-=hal Ni were cast, and suf-
ficient tubing is to be prepared for fabrication of
an odditional forced-circulotion loop, which will
also be operated with tuel 107,

Production of Commercial-3ize Heats of
Nickel-Molybdenum Alloys

The various products which were fobricated from
the biflet material obtained from the six 4800-Ib
heats of nickel-molybdenum base olloys designated
0s INOR-1 through -6 were received frem the
International Nickel Company.

of these alloys and @ complete moterials inventory
2

The compositions

were presented previously.” Only o small quentity
of uscble material of the compesition INOR-4
{77.83% Ni=16.9% Mo~1.96% Al=1.68% Ti + re-

siduals) remained after ingot breakdown, and it was

 

2, Inouye and T. K. Roche, ANP Quar. Prog. Rep.
Dec. 31, 1956, DRNL-2221, p 205.

183
ANP PROJECT PROGRESS REPORT

Toble 3.3.10, Scme Properties of Ni-Mo-Nb-Fe Alloys

 

 

Nominal Composition (wt %) Forgeability” Oxide Stubili’ryb Phoses®
10 Mo-~5 Nb—4 Fe-ba} Ni Foegeable Stoble Singla
10 Mo=7.5 Nb—6 Fe—-bal Ni Forgeoble Mot yet tested Not yet testad

10 Mo—10 Nb—B Fo—bal Mi
12,5 Mo~5 Nb~4 Fe—bal Nj
12.5 Mo-7.5 Nb~6 Fe—bal Ni
12.5 Me—10 Nb—8 Fe—bal Nij

Linforgeable
Forgeable
Forgeable
Unfargeable

15 Mo=5 Nb—4 Fe-bal Ni Forgeable
15 Mo=7.5 Nb=6& Fe~bal Ni Forgeable
15 Mo-10 Nb-8 Fe—-bal Ni Unforgeable
17 Mo -3 Nb-bal Mi Fargeabls
17 Mo -5 Nb—bkal Ni Forgeaoble
17 Mo~5 Nb~4 Fe-bal Nij Forgeable

Stable
Mot yet tested
Stoble
Stoble
Not yet tested
Stable
Stable
Spalls
Spolls

Not yet tested

Multiphose
Not vet tested
Multiphose
Multiphose
Mot yet tosted
Multiphase
Multiphose
Single

Single

Not yet tested

“Cast arc meits rolled ot 2100°F; cold rolied ta 0,065 in, thick.

bSpecimens exposed in air for 168 hr at 146 50°F,
“Annealed end aged 100 hr at 1300°F.

Tabie 3.3.11. ORNL Alloys Extruded to Tube Blonks

 

Allay No. Nominal Compesition {wt %)

 

30-46 and 3047 17 Mo—10 Cr-7 Fe-0.5 Al-

0.5 Mn-£.06 C~bol Ni

17 Mo—7 Fe-0.5 Al0.5 Mn—
0,06 C-bal Ni

17 Mo—10 Cr-0.5 AlI-0,5 Mn—
0.06 C—bal Ni

30.48 and 30-49

30-50 ond 30-51

 

requested that }’z-in.-thick plate and 0,065-in.-thick
strip be rolled from this stock. The plote and sheet
received were of poor quality, with numerous surfoce
cracks on the pieces. Thirty-four feet of 3- by
3-in, billet moterial of the alley INOR-5 (74% Ni-
16% Mo—=5.3% Cr-1,15% Al-1,69% Ti + residuals)
was to hove been used for fabricating plote, sheet,
wire, and bar stock, but this material wos scropped
during subsequent billet processing. The report
on this difficulty hos not yet been received. The
alloys INOR-2 (76.7% Ni-16.1% Mo-5.28% Cr +
residuals} and INOR-5 {78.6% Ni-13.1% Mo-2.69%
W-2,15% Nb-0,10% C + residuals) were fabricoted

successfully,

184

All the products received to dote ore listed in
Table 3.3.12. The plate and wire ore being used
for weldability studies, ond the 0,085-in.-thick
strip is being used to evoluote mechanical
properties,

Eight 9-in.-die, 12-in.-long billets, two esch of
INOR-1, -2, and -5 and one eoch of INOR-3 and
-4, were extruded to 315-in.-0D, 0.500-in.-wall tube
shells at the International Nickel Company in
August 1956. These shells, together with one
Hastelloy W ond three Hastelloy B tube shells,
which were extruded in Morch 1958, are presently
being reduced te 2-in.-0D, 0,187.in,-wall tube
blanks. The blanks are scheduled for delivery to
the Superior Tube Co. in March 1957, where they
witl be processed to small-diameter seamless
tubing. A summary of the fabricability of the alloys
INOR-1 through -6 is presented in Table 3.3.13.

The contract with the Westinghouse Electric
Corporation covering the mefting, casting, and
fabrication of lorge pilot heats of nickel-molybde-
num base alloys ot their Blairsville, Pennsylvania,
metals plont hes been signed, ond work is to begin
immediately. Although two compositions were
originally specified forscaling up to semiproduction
quantities, only the alloy INOR-8 will be melted
PERIOD ENDING MARCH 31, 1957

Table 3.3.12. Products Fabricered from 4800-1h Heots of INOR Alleys Produced by Internationol Nicke! Company

 

Hect No. Alley Type Quontity Size Remarks

 

].:{‘.{-in.-Thick Plote — Hot Ralled

Y8195 INOR-1 o A §ix T Eligit-coacleson suwfoes
Y.8197 INOR-2 7 pieces 3 124w
Y8196 INOR-3 8 pieces 6 x 12 in.
Y-8198 INOR-4 8 pieces 3 x 12in, Surface ond edge cracks
Y8200 INOR-5 8 pieces 3 x 12 in,

Weld Wire — Cold Drown

Y-8195 INOR-1 6 Ib '{3 i dia

Y8197 INOR-2 9 Ib 1;3 in. dia

Y-8194 INOR-3 8 Ib % in. dia Seam along length
¥.8200 INOR-5 8 b !fi in. dia

Y-8195 INOR-1 6 Ib 3:52 e

Y.8197 INOR-2 10 b 3;52 v ifiaa

Y-8196 INOR-3 6 Ib 3@2 i di

Y.8200 INOR-5 9 1b 362 i i

2?}’3-in.-d|'u Fargings

Y-8B195 INOR] ] picce 54 in. leng Crack at one end
Y-8197 INOR-2 1 piece 34 in. long

Y-8196 INOR-3 1 piece 48 in. long

Y-8200 INOR-5 1 picce 44 in, long

0,065-ia.-Thick Strip - Cold Rolled

Y8195 INOR-1 25 pieces 6 % 12in,
Y-3197 INOR-2 33 pieces 3 x 12 in,
Y-8196 INOR-3 17 pieces & %X 12 in.
Y3198 INOR-4 21 pieces 3 x 12 i0n. Surface crocks
Y -3200 INOR-5 36 pieces 3 x 12in.

 

185
AMNP PROJECT FPROGRESS REPORT

Toble 3.3.13. Fobricability of 4800-1b Heots of
Nickel-Molybdenum Alloys

 

Fabricability Characteristics

 

Alloy

 

Designotion Forging Rolling Extrusion Dr\::?ng
INOR-1 Fair Fair Good Good
INOR-2 Good Good Good Good
INOR-3 Fair Fair Good Fair
INOR-4* Poor Poor
INOR-5 Good Good Good Good
INOR-6** Fair Goad

 

*Allay INCR-4d became ?5% scrap during initial forging
af the ingot.

=*The billet material fabricated from allay |INOR-6 be-
came scrop during subsequent processing,

af present. The composition of alloy INOR-8 has
been altered slightly from that given in the contract,
and the nemincl composition (in wt %) is now the
fol lowing:

Molybdenum 1517
Iron 4-5
Chromivm 68
Carbon 0.04-0.08
Maonganese 0.8 (max)
Silicon 0.5 {max)
Micke! Balance

The average molybdenum content wos raised from
13.5 to 16%, and the average chromium level was
increased from 6 to 7%. Before the alloy is pre-
pared, details of the melting proctices used by the
Internationat Nickel Compony on the 4800-1b heaots
of alloys INOR-1 through -6 end by Battelle Memoriol
Institute on similar nickel-molybdenum ¢lloys will
be made available fo Westinghouse. These melting
practices witl be compared with those used ot
Westinghouse in order to determine the best melting
procedure,

Initially, one 35600-lb heat will be prepared for
casting two ingots of approximately 1100 Ib and
four ingots of approximately 250 Ib. Preliminary
forging studies will be catried out on the smaller
ingots in order to establish the hot-working temper-
oture range of the olloy. Priority will be given to
the fabrication of a 3-in,-dia, 6-ft forging which

186

can be further processed, on a loboratery scale, to
tube blanks and mechanical property test specimen
blonks.,

Melting of the second composition specified in
the contrect, INOR-7, has been postponed because
of the detrimental aluminum content
would have on its corrosion resistance
fivorides. 1t is believed thet on alloy based upon
the Ni-Mo-Nb system, the Ni-Mo-Cr-Nb system, or
the Ni-Mo-Cr-Fe system with higher total olloy
content than that of the INOR-8 composition can
be specified in the near future to replace the INOR-

effect its
in fused

7 composition,

Phase Diagram Studies at the University of
Tennessee

The present status of the phase diagram studies
being performed ot the University of Tennessee
Subcontract 582 is summorized here. A
previous report® described the studies of nickel-
base alloys containing 20 to 30% molybdenum, In
this work, metallogrophic and x-ray examinations,
hardness tests, ond other techniques are being
used to establish phase relationships and charac-
teristic phase transformations which relate to the
mechanical properties of nickel-rich alioys con-

under

tairing molybdenum ond molybdenum plus chromium,

[n the binary system, Ni-Mo, the three secondary
phases previously reported in the literature were
identified, although only the crystal structure of
the beta phase was confirmed. The temperctures
ot which the beta and goamma phases decompose on
heating were found to be higher than those reported
in recent literature and ore at present being con-
sidered to be 868 and opproximotely 925°C, re-
spectively. A simple ordering mechanism for the
formation of the beta phase from the alpho phase
wos worked out that yielded o structure from which
the observed x-roy dota could be derived., Pre-
cipitation of beta ond gamme material from the
alpha phase increased the hardness of the alloy to
450 DPH, with little if ony decreose in hardness
ofter aging for periods of up to 2000 hr.

Nickel-rich nickel-molybdenur alfloys with ad-
ditiens of chromium in the range of 4 to 14% were
studied, A chromium content of up to 5%

does not appear to alter the total selubility of the

alse

0. W, Stoffel ond E. E. Stansbury, A Metallograpbic
and X-Ray Study of Nickel Alioys of 20--30 Per Cent
Molybdenum, AECU-3105 {Dec. 1955).
alloying elements in nickel, whereas larger amounts
of chromium seem to permit an increase in the total
solubility. Additions ot chromium quickly restrict
the formation of beta material, and systems that
precipitate gomma ond then delto phases develop.
There is evidence of an additional phase in the
higher-chromium-content alleys which has not vyet
been identifiad.

Alloy Development at Battelle Memoriel Institute

Work on the nickel-molybdenum alloy development
program being conducted by Battelle Memorial
Institute under Subcontract 979 continued along the
lines described previously.*® ORNL has assisted
Battelle in their program by determining the stress-
rupture properties of selected olloys in the fuel
mixture NaF-KF-LiF-UF, {11,2-41.45,3-2.5 mole
%, fuel 107) and by extruding numerous tube blonks
of their alloys for corrosion evaluation, From the
tubing successfully fabricated to dote, ORNL has
operated thermal-convection loops with both fuel
107 and sodium. Emphosis is now being ploced on
the fabricotion of tube blanks of alloys selected
for forced-circulation loop tests with fuel 107 to be
performed at Battelle,

Four tube blanks of each of the Rattelle olloys
B-3404, B-3407, B-3412, and B-3418, described
previously,® were extruded and sent to the Superior
Tube Co. for redrawing to small-diameter tubing.
However, all the blanks failed on the tube reducer
and thus substantiated eorlier evidence that atloys
with 0.12% carbon are not amenable to tube-reducing
operations. At the recommendotion of Superior Tube
Co., Battelle hos discorded the high-corbon heats
and has melted alloys with a lower carbon level,
The compositions of the new olloys are listed in
Table 3.3.14. Bottelie has prepared o sufficient
number of extrusion billets of these alloys for
fabricating the necessory tubing to operate a total
of 16 forced-circulotion loops with fuel 107, but
only alloys B-3870 and B-3871 are to be used ot

 

4E. M. Simons, Agendaéor ANP Meeting at Baitelle,
April 16=17, 1956, ORNL CF-56-4-156 {April 16, 1956).

5\E. M. Simons, Semi-Annual ANF Meoting, Battelle
Memorial Institute — Ocilober 3l-November 1, 1936,
ORMNL CF-554-11-15 {Nav. 2, 1956).

'SH. Inouye and T. K. Roche, ANP Quar. Prog. Rep.
Dec. 31, 1956, ORNL.-2221, p 203.

PERICD ENDING MARCH 31, 1957

this time. ORNL hos received 12 billets of eoch
of these two compositions which will be extruded
to tube blanks and further precessed to 0,380-in,-
0D, 0.065-in,-wall tubing ot the Superior Tube
Co. This should provide sufficient tubing for twe
forced-circulotion loops of each of the two alloys.

In addition to the alloys which have been pre-
pared for loop studies, Bottelle has cast 100-1b
heats of the six compositions shown in Teble
3.3.15, which will be evaluated for fabricability,
heat treatment, weldability, and stress-rupture
properties, An evaluation of these alloys, os well
as the alloys scheduled for forced-circulation loop
studies, shovld provide additional data for de-
termining the effect of aluminum and chromium on
the properties of nickel-molybdenym alloys.

Table 3.3.14, Bottelle Alloys Scheduled for Forced-
Circulation Loop Studies with Fuel 107

 

 

Alloy Nori L
No. ominal Composition (wt %)
B-3870 19 Mo -7 Cr—-2 Al-0.02 C—bol Mi
B-3871 19 Mo=7 Cr-1 Al-2 Nb=1.4 Fe 0,02 C-
bal Ni
B.3a72 19 Mo—-2 Al-7 Mb-0.07 Fe-0.02 C—-bal Ni
B-3873 19 Me~2 Nb=1.4 Fe=0.02 C—bal Ni

 

Table 3.3.15, Alloys Prepared by Battelle for
Evaluation of Fabricability, Heat Treatment
Weldahility, ond Stress-Rupture Properties

 

 

Alloy
Maminal Compeosition {wt %)
Mo,
B.3874 19 Mo-=1.5 Al=1 Nb-0.7 Fe-0.04 C—bal Ni
B.3875 19 Mo =7 Cr=1.5 Al =1 Nb 0.7 Fe-D.06 C~
bgl N
B-3876 19 mo—=2 Al=1 Nb-0.7 Fe-0.06 C-bal Mi
B-3877 19 Mo~7 Cr=2 Al=1 Nb-{L.7 Fe -0.0& C-
bol Ni
B-3878 19 Mo—1.5 Al=1 Nb—4 Fe-0.06 C—-hal Ni
B-3B79 19 Mo=7 Cr=1.5 Al=1 Nb—4 Fe-0(.06 C-

bal Nj

 

187
ANP PROJECT PROGRESS REFPORT

FABRICATION OF COMPOSITE TUBING AT
SUPERIOR TUBE co.’

T. K. Roche

The corrosion resistance of the nickel-molybde-
num alloys to attack by sodium is about equal to
that of Inconel. Thus it appears that temperoture-
gradient mass transfer is an inherent characteristic

H. Inouye

of nickel-base alloys and thot it increcses in
severity with increasing temperatures. Since the
nickel-molybdenum olloys cre being developed for
use at higher temperatures than those contemplated
for luconel, the use of composite tubing may be
desirable or even mandatory.

The fabrication of composite tubing by coextrusion
has been shown experimentally to be entirely
feasible.® On the bosis of this experimentol work
a considerable quantity of duplex tubing was pre-
pared by commercial practice ot the Internctional
Nickel Compony ond the Superior Tube Co. without
apparent difficulty, In this production run, 9-in.-dia
compos ite billets were coextruded 1o produce 3-in.-
0D tuke sheils having o wall thickness composed
of 60% Inconel over 40% type 316 stainless steel,
The tube shells were tube-reduced to 2-in.-QD,
0.187-in.-wall tubing by the Intemational Nickel
Company and then tube-reduced again ond redrawn
to small-diameter tubing by the Superior Tube Co.
An evalustion of tuhing thus produced is presented
in Tables 3.3.16 ond 3.3.17. The results clearly
indicate that layer thicknesses can be held to
within 2% of those intended, and that, in general,
the defects found by inspection are those normolly
associated with conveaticnal tubing., The trans-
verse sections of this tubing shown in Figs. 3.3.2
and 3.3.3 illustrate the uniformity of the layers and
the metallurgical bond at the interface of the
finished tube.

Terminal Report No. 1633, Subcontroct 531, which
was submitted by the Superior Tube Co., summorizes
this work with the following conclusions: (1) The
condition of the os-received 2-in.-0D, 0.187-in.-
wall composite tube shells was excellent and equol
to that which has been furnished to the CX-900
Inconel specifications. (2) The composite tubing
was processed without difficulty by using con-
ventional tubing precesses. {3) The ratio of wall

 

TSubcontrocl 531,

84 R. D'Amore and H. Inocuye, The Extrusion of
Composite Tubes, ORNL CF-54-4-123 {April 18, 1956).

188

thicknesses was maintained essentially as it
existed in the as-received tuvbe shells. (4) Over-
all yields were excellent. {5} Additional orders for
this type of composite tubing mey properly be
processed by production schedules.

SHIELD PLUGS FOR ART PUMPS
J. P. Page

Two full-sized gamma-ray shielding plugs for
ART f{uel pumps were prepared. The plugs are
composed of 75 wt % tungsten carbide and 25 wt %
Hastelloy C. The first plug was hot pressed in o
large horizontal press, and the die material was
CS-312 graphite rather than the AGHT-grade previ-
ously used, The die fractured ot about 1800 psi
and coused loteral expansion of the pressed plug
and the die parts.

A verticel hot-pressing facility was adapted for
pressing the second plug, and o die identical to
the first one wos used. Even though the pressure
was limited to 1600 psi, this die olso failed at the
end of the pressing cycle, and some lateral ex-
pansion occurred. However, the pressed plug had @
vniform density of 12,1 g/cm®,

These plugs will be surface ground on one end
face and used for brazing and thermal-shock tests.
A new graphite die has been designed and ordered
for which the dimensions ore such thot the hot-
pressed plugwill require lirtle or no grinding on the
inner or outer surfaces. This die was made some-
whot stronger by increasing the outside diameter
and eliminating the sight hole.

Meutron shielding disks hove been cut from
Ludlum
sheet material. These ore ready for insertion into
the plug assembly,

Allegheny stoinless-steel-clad Cu-8,C

NEUTROMKN SHIELD MATERIAL
M. R. D'Amore’

The Aliegheny Ludlum Steel Corp. continued to
experience difficulty in preparing the stainless-
steel-clod Cu-B,C shield material. Of a total of
17 experimental composite packs rolled during
Qctober ond November, none yielded more than
40% usable stock, At the suggestion of ORNL, the
sintering temperature for the Cu-|34C cores was
increased, and the frames for the composite packs
were redesigned,  Since then on odditional 15

 

9On cssignment from Prott & Whitnay Aircroft.
PERIOD ENDING MARCH 31, 1957

Toble 3.3.186. Metallogrophic Evaluotion of Inconel-clad Type 316 Stainless Steel Tubing

 

Nominal Tube

 

Wall Thickness {mils)

 

Dimensions

 

 

Specimen  Specimen Type 316 . Incone]
Outside Wail No. Pasitian Stainless Steel Inconel Composite (Average P&r?ent:l;ge
Diometer  Thickness A NMox Mo A Wex Min A Men MO comeositel
(3a.} {in.)
0,500 0.047 1 Tronsverse 17.5 18 17 29 A 4.5 47 46 62.1
2 Transverse 17.5 18 17 8.5 0 ¥ 44 48 44 51.9
3 Tronsverse 17.6 18 17 221 30 28 46.7 47 46 42.3
4 Transverse 7.9 19.5 17 279 X% 458 48 45 60.9
5 Transverse 7.9 18,5 17 229 30 3.5 438 47 45 60.9
b Transverse 17.3 18 17 26.3 30 X 46.5 48 44 62.9
7 Longitudinal 18 18 18 27.5 29 2% 455 d4 47 41
8 Longitedinal  18.1 1§ 175 274 28 2% 45.3 47 44 60
0.250 0.025 9 Transverse 10 0.5 %5 163 W 155 w2 % 217 62.3
10 Transverse 10.5 11 9 156 16 15 258 26 25 58.2
11 Transverse 10.2 1N 190 155 16 15 258 26 25 601

 

“Averogn of meosuramant: a1 lour locotiong.

**Spacilication, 60% loconal; volves given are averoges of values at four locations.

Table 3.3.17. Nondestructive Inspection of Inconel-clod Type 316 Stainless Steel and

incenel-clad Type 310 Stainless Steel Tubing

 

Inspection Method

Results

 

Resononce-type wltrascund

o mojer unbonded oreas disclosed in any of the tubing; nenbonded

areas 0.0&0 in. thick and ],fz in, long can be detected by this method

Immersed pulse-ache vltrasound
Encircling-coil eddy currant
Postemulsification flucrescent penatrant

Radiagraphy

A few cracklike indications
Mo detrimental conditions within the tubing
A few small pin-hole or pitlike indications

Several pin holes ond lengitudine! indications, believed to be shellow

 

packs have been relled with considerable success;
the last six composites rolled yielded virtually
100% vsable materiel,

TUBULAR CONTROL RODS
M., R. D'Amore

The feasibility study of extrusion of contrel rods
Previous studies indicated that
cracking of the cermet core (30 wt % Lindsoy oxide—
70 wt % Ni) in extruded three-ply tubes may have
been coused by thermal stresses between the core
and Hastelloy X cledding. Therefore the cladding
material for the tube-extrusion studies has been
changed 1o [nconel, which hos a thermal expansion

was continued.

necrly equal to that of the cermet core material.
The dispersion of fine particles of Lindsoy oxide,
which form a nearly continucus network in fobricated
cores, may also contribute to the poor ductilities
observed in tensile specimens and in extruded
three-ply tubes, 1?

The major effort has been directed toward the
praduction of strong, high-density 44- 1o 105-u
porticles of Lindsay oxide. The procedure de-
veloped for obtaining the desired particles con-

sists of the following: first, the os-received

 

V4 R D’ Amore, ANP Quar, Prog., Rep. Dec. 31,
1956, ORNL-2221, ¢ 210.

189
ANP PROJECT PROGRESS REPORT

100
i _qulnnhmhu i

L

 

        

 

Fig. 3.3.2. Tronsverse Section of Composite Tubing,
Etchant;
50% HC! + 10% solution of Cr03. 7K. Reduced 32%.

Inconel=clod Type 3146 Stainless Steel.

 

 

F;u.gl 3-3.3-
Etchanit glycerio regia. 75X.

190

e

powders ore compacted in a steel die ot 4 tsi;
second, the compacis are fired ot 1900 to 2000°C
in a hydrogen atmosphere; and, finally, the fired
compacts are crushed and ground in a micropulver-
izer to obtain the desired patticle size.

A core that was to be clad ond extruded was hot
pressed to an apparent density of 91% of theo-
reticol by using oxide of the 44. 10 105-p particle
size ronge. This core was assembled into a five-
ply biliet, with |nconel cladding oend mild steel
cons both inside ond out, This billet is now ready
to be extruded. Two additional billets have been
prepared for extrusion of rods for preparing tensile
specimens of the cermet,

EVALUATION OF ARC-CAST AND
POWDER-METALLURGY NIOBIUM

J. P. Page

The collection of wsable mechanical property
dota for arc-cast ond powder-metallurgy niobium
at both low andhigh temperatures requires knowledge
and control of the grain structure of oll test speci-
mens, and therefore the recrystailization process

| LINCL ASSIFIED
Y 20786

.0l

P v

w

T

U

=

0.0

 

e
.

 

Metollurgical Bond ot Interfoce of Composite Tubing, Inconel-clad Type 316 Stalaless Steel,
is being studied, A knowledge of recovery and
recrystallizotion kinetics con be obtoined from a
small amount of materiol by metallographic and
x-ray techniques and thus sufficient material from
the specimen remains for further comprehensive
festing.

The chemical
powder-metaliurgy richium ore quite similor. The
microstructures of the two materials, however,
differ greotly, as shown in Figs. 3.3.4 and 3.3.5.
The inclusions shown are probably niobium car-
bides. The effect of the prior fabrication histery on

compositions of arc-cest and

the recrystallizotion charocteristics can be seen
in Figs. 3.3.6 ond 3.3.7. These photomicrographs
indicate thot, as would be expected, the arc-cost
material has a tendency to coalesce and form
large-grained structures,

The first phase of the recrystallization study is
essentially complete. The objectives of this phase
were to form controlled, fine-grained microstructures
in both materials ond thereby minimize the effects
size on subsequent recrystallization
behavior, to determine the range of recrystallization
femperatures for the moterials, and to determine
the effects of various annealing atmospheres on
niobium. Quontitotive recrystallization studies will
be carried cut on the fine-grained material over o
wide range of temperatures and test periods. The
present informotion indicates that the recrystal-
lization ronge for annealing times of wp to 1 hr

of grain

UNCL ASSIFIED
Y-21327

: &

Fig. 3.3.4, Arc-Cost Niobium, As-polished. 100X,

PERIOD ENDING MARCH 31, 1957

5 . : ] UNCLASSIFIED .
. = Y-21320
s -
t - . +
-y e i
e
-
* -
— o =
e s i -
- % .
- 3 —-
i s e
. a
a1 . - '
e
Sk 2 7o L
e ' - -
" i -
: . ~y -
¥ # . C . : .- -
» .. . “-i 5
o _J.‘ " - *
e SR ot
- x o
= >4
- ok ; ¥ £
* "’f . = i
- . L *
- 7 |
-
o,
. 1 'y
] b @
';.: » - g "

Filg. 3.3.5. Powder-Metallurgy Miobium. As polished,
100X,

, potia UNCL ASSIFIED
I 4 - ¥-21328
| | ’ | . -
.ff .J:I 4 }
v \1 f.;. l\ I
c dYy) \
} -’Fn'. llf J’\_I

 

Arc-Cast Niebium Anneoled 10 min ot
50 cme HZO' 14 e H2504a
HNOa. 100X,

Fig. 3-3;6-
1450°¢. Etchunt;
25 cm® HF, 5 em®

191
ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED

iy S
¥ 4
F"'I : = e -f-i ‘f-2|323
. J = A T
=y e Ly R "
B v ..I‘\ "
b e e A g i
- o
“ " ek {.. 5 5 -_.f_ ¥ J'-..- -
G i e e
= A S ; B
1 0 i o=
E" * '“—f i
- . e A g ! ; g‘-\"
g g i
: . % 2
4 H- - Lam B % =
T e | e
3 0w ! - il o
. .. e e
r . -2 * i
"I o \S :h B = - - g, ""5
e = Ty o
T § i }‘ e e s
i = ¥ etk 3 = !
A i T '* ."‘_ \.‘: B
Tk, i e P B =
P s e O
i . i N T . nilm e
:J [ ‘;'J'H\' : .); Ene 1
¢ PR 5 ~ L o e “l i e
i = i, e
o : e
-H.T.\ l‘r. |ll‘l " _JB T u
/4 M 5 g '
¢ / ; e ! . i
% o AT
\ B % .J- '{};. i - pi o i L
L e Lo L § A
b b B ]
3""-—,:’?‘ o ‘JI ‘I‘Tff = 2
- o \ L
v i o S 'P"- i e, f i Wy
s i 4 s /‘ J :
Fig. 3.3.7. Powdar-Metallurgy Niohium Anncaled

10 min at 1450°C. Erchont: 50 cm® H,C, 10 cm>
H,50,, 25 cm® HF, § cm® HNO,. 100X,

lies between 1100 and 1350°C, depending on the
omount of cold work te which the materiol is
subjected,

Severol techniques are being considered for
conducting the numerous onnealing freatments
required,  These include vacuum encapsulation
with ond without the addition of getters, treotment
in dynamic vacuums of varicus orders of magnitude,
and trectment in high-purity argon. Each of these
techniques hes on advantage in certain ceonditions:
the orgon otmosphere is recommended for short-
timeg, high-temperature the dynamic
vacuum for intermediate temperature and times, and
the vacuum capsule for extended times ot lower
temperatures. A quontitative survey of the effects

of time and temperoture on the micrchardness and

treatment,

the lattice parometer of powder specimens is being
conducted for each environment,

ldentical microstructures for powder-metallurgy
and arc-cast materiols have been achieved. The
niobium specimens were severely worked (reduced
in thickness by 90%) aond annealed for 1 hr at
1250°C in a dynamic vocuum of approximately 10-9
mm Hg. The resuftant material is very fine-groined,
but the greins shew elongation in the direction of

192

rolling.  Also, the severe reduction limits the
thickness of the material available for the recrystal-
fization study, Further work will be done with
lower reduction ratios and longer annealing times
in an attempt ta develop a more randomly oriented

microstructure in a thicker material.,

METAL HYDRIDES FOR USE AS MODERATORS
IN HIGH- TEMPERATURE REACTORS

R. A. Potter R, E. McDonald'?
T. Hikido'?

The realization of the great potential improve-
ment in performance of circuloting-fuel reactors
which is offered by the NaF-KF-LiF-base fuels
and the nickel-melybdenum olloys will depend to a
large extent on the development of an adequote
Of the

Be(, beryllium canned in

moderator for use at high temperatures,
possible moteriois,
mojybdenum, and the hydrided metals, the laost
appear to offer the mast promise, Research at the
General Electric Company aond other organizations
has shown that certain metals, notably zirconium
and yttrium, are able to abserb and retain large
quantities of hydrogen at high temperatures ond yet
retain their metallic characteristics, However, at
the temperature of interest for advanced reactors,
over 2000°F, yttrium is the only material known at

this time which offers promise of being satisfactery.

Capsule for In-Pile Tests

An octive research and development program on
hydrided metals was initiated, with one of the
first projects being fabrication of the capsule
shown in Fig. 3.3.8 for in-pile tests. The speci-
men consists of hydrided zirconium (NH = 3.5)
conteined ina 0.025-in.-thick arc-cast 0.5% ritonium-
molybdenum alloy capsule. The ends of the capsule
are ]fain. thick ond the over-all dimensions are,
nominally, }2 in, dia x 3 in. long. No protective
cladding was put over the molybdenum, becouse the
test will be made in o helium ctmoschere,

The molybdenum capsule was machined from :",;:1-
in~dia rod. A rod of reactor-grade zirconium was
machined to 0.460 in. in diameter end 2.75 in, in
length ond a .]-’B-in.-diu hale drilled through it
lengthwise.  This zirconium cylinder was iron-

ploted to o thickness of 1.75 mils to aid in ochieving

 

} lUn assigament from Prott & Whitney Aircrofs.
IQOn ossignment from USAF.
ONE INCH

FPERIOD ENDING MARCH 31, 1957

UMCLASSIFIED
PHCTO 40448

 

 

Fig- 3.3.8. A Molybdenum-—ciad Hydrided-Zirconium Capsule Prepared for oo In-Plle Test, (Secret with caption)

a bond to the molybdenum. The ploted zirconium
was inserted into the molybdenum copsule and
hydrided in situ to an Ny of 3.7, calculoted on the
basis of weight gain, The !-é-in. hole was reamed
clean, and a zirconium rod wos weighed ond press-
fitted into the hole. The colculoted Ny

geneization of the rod is 2.5,

after homo-
A l-mil iron-foil
disk was placed over the hydrided zirconium, and
the molybdenum ead plug was inserted and copper-
brozed in place, After hoving been brazed, the
capsule was machined fo size (0.025-in.-wall thick-
ness) and tested at 1850°F for 1 hr under a vacuum
of 3 x 1072 mm Hg, The weight loss of 4.2 mg
under these conditions was not considered to be
serious.  After the capsule was inspecied by
radiography, six Chromel-Alumel thermocouples
were oftached by spot-welding. The completed
assembly waos heated in an argon atmosphere tc
1650°F, cooled to opproximately 800°F, reheated
to 1650°F, and cooled to room temperature, The
thermocouples
this treatment.

A second specimen was processed simultaneously
with the in-pile test capsule in order to provide o
control somple for examination. However, this
specimen wos over-heated during the brozing
operation and is therefore not truly representative
of the in-pile test capsule, Metallographic exomi-
nation of the control semple showed evidence of
the owver-heating in the form of melting ond de-
composition of the hydride ot the brozed end of
the capsule. Although a tight fit was observed

remained securely attached aofter

between the hydrided zirconium and the molybdenum
in the other areas of the capsule, the metailurgical
bond was not uniform,

The tn-pile test capsule was scheduled to go into
the MTR on February 18, 1957, The calculated
maximum operating temperoture will be 1550 to

1600°F, and 100 thermal cycles will be imposed by

moving the test container in and out of the reactor.

Hydriding Equipment

The hydriding was accomplished with ¢ system
bosed on the design developed by the Generol
Electric Company and Sylvania Electric Products,
Inc. On the whole, this system worked very well
on the in-pile test capsules ond other small speci-
mens. However, during these initial operations
the need for several improvements became opparent,
The modifications
include a better temperature control and recording
system, provisions for removing the rsetort from
the furnace to permit rapid cooling from the hy-
driding and equipment to provide
hydrogen at o constant pressure or in meosured
volumes. In addition, a furnace capable of attaining

and these cre now being made.

temperature,

the high temperatures required for hydriding yttrium
hos been built and is now being checked.

Y tirium Metal Production

Research on hydrided yttrium is being seriously
hampered by the unavailability of the material.

193
ANP PROJECT PROGRESY REPORT

Althowgh o considerable amount of work is being
done ot the Ames Laboratory, lowa State Callege,
te cllevictie this shortage, it wos deemed advisable
to produce some yttrium at ORNL. The availoble
experience influoride technology was a major factor
influencing this decision, becayse the conversion
of ¥, 05 to YF, is an important step in the process.
The purity of the fluoride controls the purity of the
final product to a large extent. Accordingly, 15 |b
ot Y,0, was obtoined and 1000 g was converted
to the fluoride. In the reduction process, which
was developed at the Ames Loboratory, the yftrium
fluoride will be reduced with a mixture of calcium,
magnesium, ond calcium chloride. The yttrium-
magnes ium-colcivm alloy thus formed will he refined
by vacuum distillation to remove the magnesium
and the calcium. The resulting yttrium sponge will
then be vacuum melfed.

Fabrication of the equipment for producing the
metal hos been completed, and it is now being
assembled and checked.
mixing of the starting material to the vocuum dis-

The entire process from
tillation of the alloy will be performed under a

vacuum or an inert-gas atmosphere, |t is expected
that the first metal will be produced in March 1957,

194

Cladding Studies

The advanced reactor designs presently being
studied picture the hydride moderater as cylindrical
rods approximately 3’2 in. in diameter and 30 in, in
fength, The hydride will be protected with a duplex
cladding of molybdenum and ¢ nickel-molybdenum
alloy., The molybdenum is required for o hydrogen
diffusion barrier and for compatibility with yitrium,
Since the remainder of the fuel circuit will be
constructed of a nickelsmolybdenum alloy, a similar
alloy must be put over the molybdenum to prevent
dissimilar metal corrosion attack,

Studies
temperatures and amount of work required to achieve
a bond between melybdenum ond anickel-molybdenum
alloy. The test specimens consist of molybdenum
cylinders press-fifted into nickel-molybdenum alloy
tubes, which, in turmn, cre press-fitted into lnconel
tubes, evacuated, and sealed. la the first test,
the sealed capsule was heated te 1600°F ond
reduced 10% by swaging. Metallographic exami-
nation showed thatonly partial bondingwos achieved
under these conditions. In the second test, a
pallodium foil will be inserted between the molyb-
denum and the nickel-molybdenum alloy. Higher
temperatures, different amounts of work, and other
intermediote layers will be tested in the future.

have been initiated to determine the
PERIOD ENDING MARCH 131, 1957

3.4, WELDING AND BRAZING INVESTIGATIONS

P. Patriarca

HEAT EXCHANGER FABRICATION
ART Sodiumsto-NaK Heat Exchangers
G, M, Slaughter

Brazing procedures are being developed for use
in the fobrication of ART sodium-to-NeK heat
exchangers by the Griscom-Ryssell Co. Ex-
amination of a variety of tube-to-header samples
from several brazing experiments performed at the
Griscom-Russell Co. has revealed inconsisient
fiow and fitleting. A typical example of this
condition on o tube-to-heoder joint is shown in
Fig., 3.4.1.

Attempts have been mode ot ORNL to duplicate
the difficulties experienced by the vendor and then
to alleviate them by suitable variotions in brazing
procedure.  For the initial tests, the vendor's
heating rate of approximately 100°F/hr was used.
Samples taken from a typicol test assembly before
and ofter brazing ot this rote of temperature rise
are shown in Figs. 3.4.2 ond 3.4.3, The influence
of the following variobles was studied in these
tests: (1) type of Coast Metals brazing alloy
Ne., 52 ring (ORNL cast ring vs York Corp. sintered
ring), {2) differences in Inconel base material
(ORNL headers vs Griscom-Russell headers),
(3) angle of countersink in headers, and (4) direct
impingement of hydrogen on samples vs slight
baffling. Mo significant differences in braozing
were noted os o result of these vorionces.

An analysis of all the experimental results
indicated thai saotisfactory brozing can be ac-

UHCL ALEIMIED
T-21547

 

Fig. 3.4.1, Tube-to-Header Joint Showing Incomplete
Fllleting.

complished ot 1920°F if oll variables such as
hydrogen purificotion, hydrogen flow rate, material
cleanliness, etc, are corefully controlled. How-
ever, only minor deviations from the optimum
conditions may result in inconsistent brazing at o
100°F /hr rate of temperature rise,

Additional experiments were then conducted at
200°F/hr and ot J00°F/hr rates of tempercture
rise, As con be seen in Figs. 3.4.4 ond 3.4.5,
samples brazed at 300°F/hr had satisfactory
fillets on all joints. With the faster heating rates,
it appeared that brozing could be conducted under
slightly reloxed conditions with a reasonable
assuronce of success. The 300°F/he heating rate,
which should be readily atteinable with most
industrial equipment, provides a high degree of
assurance of successful brazing,

Investigation of Rapid Heating Cycles

E. A. Franco-Ferreira

Tests were run to investigate the feasibility of
vsing very short heating cycles (of the order of
5 min) for the back brozing of heat-exchanger
tube-to-tube sheet joints, The experimental
procedure consisted essentially in heating the
welded side of a small Inconel tube sheet whiie
providing a vocuum or an inert atmosphere for
protection on the opposite side, which was to be
brazed. The heat source used in the experiments
was a salt-bath furnace.

It is felt that certoin advantages might occrue
from this method of brazing, such as minimization
of grain growth in the tubes and reduction of
fube-wall dilution, while avoiding distortion and
the jigging problems associated with conventionol
furnoce brazing of entire ossemblies. One of the
experimental configurations is shown in Figs.
3.4.6 and 3.4.7. A section through a typical jeint
brazed by this method is shown in Fig. 3.4.8.

This method of brazing is not limited fo the use
of g salt-bath furnace as a heat source. It is felt
thot high-energy radiant heat sources based on
surface combustion units could also be used to
advantoge. Although the experiments described
above were prelimincry, the resulis do indicate the
feasibility of using rapid heating cycles for some
high-temperature heat exchanger brazing operations.

195
ANP PROJECT PROGRESS REPORT

AUX. HX.TEST 4
BEFORE BRAZING

 

Flg. 3,4,2, Tube-to-Heoder Tast Assembly Before Brazing,

UHCH ASSIFIED

AUX. HX.TEST 4

e S

AFTER BRAZING

 

Fig. 3-4.3. Tube-to-Heoder Test Assembly After Brazing at 100°F /hr Rote of Temperature Rise.

196
PERIOD ENDING MARCH 131, 1957

.- UMCL ASSIFIED
" ¥-21443

 

PO A &1
P N VA

 

Fig. 3.4.4. Tube4o-Header Test Assembly Brozed at o 306°F /hr Rate of Temperature Rise.

R I 5 S -q
1 UNCL ASSIFIED
i Y-21453
| J R

 

Flg. 3.4,5. Tubeto-Header Joints Showing the Excellent Fillets Obtalned by Using a Heating Rote of 300°F/hr,

167
ANP PROJECT PROGRESS REPORT

1]

-

UNMCL ASSIFIED
Y-20717

 

Fig. 3.4.6. Experimental Setup for Rapid-Heat-Cycle Brazing Experlments.

UNCL a55IF [ED
Youe

 

Flg- 3-‘41?-

Braozing Experiments,

Completed Setup for Ropid-Heot-Cycle

198

FABRICATION QOF YALVYE COMPONENTS
G. M. Sloughter

The developmental work on methods for fabri-
cating valve components of various types con-
tinved. The specialized technigues that were
described in the previous rlt=':|:Ju;~r1‘]"2 for joining the
different compositions of disk ond seat—ring ma-
teriols were used to fobricate the volve com-
ponents described in Table 3.4.1. Two brazing
steps were utilized in the construction of most
parts, Step | consists in the attachment of the

 

IE. A, Fronco-Ferreira, G. M. Sloughter, ond E. 1.
Wilsan, ANP Quar. Prog. Rep. Dec. 31, 1956, ORNL-
2221, p 241,

2F‘. Patriarca and G. M. Sloughter, ANP Quar Prog.
Rep. Sept. 10, 1956, ORKWL-2157, p 198, esp. Figs,
3.4.12 and 3.4.13.
PERIOD ENDING MARCH 31, 1957

 

 

 

) i L LaTE s e :
' UNCLASSIFIED
Q.04
0.02
Hii|o.03
0.04

 

 

 

 

Fig. 3.4.8. A Typical Joint Brazed by Using the Rapid-Heat-Cycle Method. 75X, Etchant: 10% oxolic acid.

Table 3.4.1. Valve Component Fabrication During the Quarter

 

Mumber of Brozes

 

Fype of Yalve Component

 

Step | Step 1l

K-94 disk 2 6
{tungsten corbide plus cobolt binder)

K-94 ssat ring 1 2
(tungsten carbide plus cobelt binder)

K-26 disk 0 2
(tungsten carbide plus cobalt binder)

KM disk 3 4
(‘I"'F'Fi(:2 plus cobalt binder)

Mallory 1000 seat ring Mo step | broze required 1
{sungsten plus copper and nickel as binder moterials)

Thermenal seat ring Mo step | broze required 1

K-138A seat ring 3 2
(TiC-NbTuTica plus cobelr binder)

K-162B disk 0 1
{titonium carbide plus nickel and malybdenum os binder maotericls)

K-1428 seat ring 1 2
(titonium carbide plus nickel and molybdenum as binder moterials)

K-151A disk 0 1
(*itanium corbide pius nickel binder)

Copper seot Mo step | braze requirad 3

Melybdenum seat rings 3 3

Malybdenum disk 1 1

Tungsten disk ] 3

 

 

199
ANP PRQJECT PROCGRESS REPORT

material to a suiteble tronsition loyer, ond Step If
consists in the attechment of the subassembly to
Inconei. The occosicnal rebroze cycles required
to c¢htain components acceptable for use in the
test runs are included in the number of brozes
listed.

ART FILL-AND-DRAIN TANK FABRICATION
G. M, Slayghter

A study of the problems associoted with the
fabrication of the ART fill-and-drain tank is under
way. Specifically, tests are being run to establish
welding and brazing procedures for the tube-to-tube
sheet joints in the tank heads. As a result of the
high degree of restraint exerted by the 1 lé-in.-fhick
tube sheet, tube-to-tube sheet welds made with the
standard fusion procedure used for thin fube sheets
consistently exhibited root cracking of the type
illustrated in Fig, 2.4.2. The direct oim of the
welding investigations has been to arrive at o

E. A. France-Ferreira

welding procedure which "will produce joints free
of such root cracks. A promising joint design
currently being investigated is o trepanned type,
an example of which is shown in Fig. 3.4.1C.

Brozing experiments ore also under way to
determine the proper amount of preplaced brazing
alloy, the type of preplacement procedures, and a
suitable brazing cycle. In the initiel experiments,
the Coost Metals brozing alley Ne. 52 was preplaced
in the form of rings, two per tube, as shown in
Fig. 3.4.11. A photogroph of a typical as-brozed
test sample is shown in Fig. 3.4.12.

The optimum procedures for welding ond brazing
have not yet been determined. Varicus modifi-
cations in technique are to be studied in subse-
quent experiments.

UNGCL ASSIFIED
Y-11545

   

0.0%

0.05

008

 

 

 

 

Fig. 3.4.9, A Root Crack in o Tube<to-Header Test
Weld. 50X. Etchant: 10% oxelic ocid, Reduced 35%.

200

 

Flgl 3'4'104
Design.

Tubesto-Heoder Joint with Trepanned

AC No.l
I L[+, 15

 

Fig, 34.11. Tube-to-Heoder Somple Broze Assembly
Showing Preplecement of Coast Metol Brazing Alley
HNa, 52 Rings.
 

-

UHCL ASSIFIED
Y2129
|

AC No. 1

T

 

Fig. 3.4.12. Tube-to-Header Semple Shown in Fig.
3.4.11 After Brozing.

WELDING AND BRAZING
OF NICKEL-MOLYBDENUM ALLQYS

R. E. Cleusing
Studies of Hostelloy B and W Welds

Preliminary results of an investigation of the
mechanical properties of Hostelloy B and W welds
on '{,!-in. plate were reported previously.d Further
tests were conducted, and the results of all the
0.252-in.-dia tensile-specimen test data obtcined
are presented in Tables 3.4.2 through 3.4.7.

An onalysis of the tensile test data indicotes
that the following generol conclusions moy be
drawn:

1. Hastelloy B welds have good ductility in the
as-welded condition when tested at room temper-
ature  but poor ductility ot 1500°F. In the
transverse weld tests, the specimens tested at
room temperature failed in the weld, while those
tested at 1500°F failed in the base metal. In
aged specimens lested at 1450°F, the failure wos
again in the weld material.

 

3p. Patriarce and R. E. Clousing, ANP Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 22).

PERIOD ENDING MARCH 31, 1957

2. Hastelloy B welds oged 200 hr at 1300°F ol!
failed, with poor ductility, in both the all-weld-
metol and tronsverse-weld specimens. |In the
latter, the failure occurred in the weld metal ot
room temperature and in the fusion line ot 1300°F.
The transverse specimens showed lower ductility
than the all-weld-metal specimens tested ynder the
some conditions.

3. Hastelloy B welds aged 200 hr ot 1500°F
showed a slight increase in strength and decrease
in ductility compared with the as-welded condition
specimens tested at room temperature. However,
at 1500°F the ductility of the aged specimens was
better than that of the as-welded specimens ot
1500°F,

4. Hastelloy W oll-weld-metal specimens show
strengths about the same as those of similar
Bostelloy B specimens and acceptable ductilities
(greater then 10% elongation} in all but two speci-
mens. Aging for 200 hr ot 1300 and 1500°F re-
sulted in considerable loss of ductility,

5. Hastelloy W is notch-sensitive at 1300°F.

6. Failures in Hastelloy W weldments exhibit a
tendency to occur ot the interface between the
weld and the base metel,

7. Hostelloy W weld metal appears to be stronger
and less ductile than the base metal.

8. Weldments of Hestelloy B plate with Hastelioy
W welding wire showed a morked tendency to fail
ot gage morks in the bose meial or ot weld defects,
No opparent advantage is gained by using this
filler wire instead of Hastelloy B filler metal.

?. Vocuum-melted Hastelloy B welding wire has
essenfiolly the some characteristics as those of
commercially availoble Hastelloy 8 wire.

Welding and Braxing Studies of Experimental
Nickel-Molybdenum Allays

The lé-in. plate required for welding studies of
some of the nickel-molybdenum experimental
alloys was received and tests were initiated to
determing the welding ond brazing choracteristics
of these moteriols. The test specimens were
prepored by the methods described previously.3
The observations made in the tests thus faor are
presented in Table 3.4.8. Since alloys INOR-7
and -8 are currently of maximum interest, it is
encoyraging to note that ne cracking tendencies
were found in these tests. Tensile data for other
promising clloys will be reported as they become
ovailable,

201
ANP PROJECT PROGRESS REFORT

Table 3.4.2. All-Weld-Metal Tensile=Specimen Tests of Hustelloy B Filler Metal on Hostelloy B Flate

 

 

Test Tensile | Reduction .
Heat Treatmaent Temparature Stre ngth E Gn-gflflun in Araa ‘
[OF:I {p$i} (% I 1""4 gflge) (%) &
As-welded Room 118,000 25 20
Room 124,000 27.5 23
200 hr ot 1500°F Room 133,600 20 16
Room 134,000 12 12
200 hr at 1300°F 1300 114,300 8 5.5
1300 115,200 & 3.3
200 hr ot 1500°F 1500 67,200 19.5 14
1500 63,800 10 13
As-welded 1500 66,700 0.5 g
1500 62,300 12,5 13
200 hr ot 1650°F 1650 42,100 28 25 !
1650 46,500 27.5 24 "

 

Toble 3.4.3. Transverse-Weld Tensile-Specimen Tests of Hastelloy B Filler Metal on Hastelloy B Plare

 

Test Tenszile Reaeduction
E longaticn ) Lacation of
Heot Treatment Temperature Strength ] i in Area e
(°F) (psi) (% in 1-in, gage) ) racture
As-welded Room 118,304 32,5 21 Weld
Room 124,500 24.5 23 We ld
200 he ot 1300°F Room 144,100 ? 5 We ld
Room 140,500 5 2.5 Weld
200 hr 2y 1300°F 1300 108,300 5 3.4 Fusion line
300 109,100 2 1 Base metal
200 hr at 1500°F 1500 72,100 20 20 We ld
1500 73,700 13.5 23 Bose metal
As-welded 1500 65,900 945 9 Base metal
1500 68,000 2.5 9 Bose metal
T
200 hr ot 1650°F 1650 47,000 26 32 We 14 .
1450 45,600 25 26 Weld

202
PERIOD ENDING MARCH 31, 1957

Tabie 3.4.4, All-Weld-Metal Tensile-Specimen Tests of Haostelloy W Filler Matal on Hastelloy W Plate

 

 

Test Tensile | ' Reduction
Heat Treatment Tempercturs Strangth | olng:uhon ) in Area
a I 1*In. Q0
(°F) (psi) yge (%)
As-welded Room 126,100 40 30
Raom 128,000 36 27
200 hr at 1500°F Room 138,400 7.5 5
Room 138,400 13 14
200 hr of 1300°F 1300 105,300 15 14
1300 117,600 13.5 15
290 hr ot 1500°F 1500 58,500 13.5 23
1500 64,000 7.5 22
As-weldad 1500 71,100 31 16
1500 69,800 20 18
200 hr at 1650°F 1650 42,200 28 45
1650 43,600 21 47

 

Toble 3.4.5. Transverse-Weld Tensile-Specimen Tests of Hastelloy W Flller Metal on Hastelloy W Plate

 

 

Test Tensile £ ) Reduction L . ¢
Heot Treatmant Temparoturs Strength % ; or?uhon in Areo T:u :}h e
“in. Q &
{OF] — {% in 1-in. goge) %) ractur
Az -Welded Room 123,500 38 26 Wald
Room 123,300 35 2% Weld
200 hr at 1300°F Room 136,700 18 19 Base metal at
gage mork
Raom 136,200 20.5 a5 Interfoce
200 hr oy 1300°F 1300 29,000 17.5 23 Base matol at
gage mark
1340 92,100 5 22 Base metal at
gage mark
200 hr at 1500°F 1500 69,500 2645 A0 Interface
1500 70,100 KY:! 43 Interfoce
Asswelded 1500 65,100 22.5 20 Weld
1500 646,200 28 34 Weld
200 he at 1650°F 1650 45,500 47 77 laterface
1659 44,900 35 71 laterface

 

203
ANP PROJECTY PROGRESS REPORT

Table 3.4.8. Tronsverse-Weld Tensile-3pecimen Tests of Hastelloy W Filler Metal on Hastelloy B Plate

 

 

 

 

 

Test Tensile El . Reduction L _ ;
. -
Heot Treatment Temperature Strength ] o:g-'a ren In Area .:::u om o .
(OF) {psi] {% in 1-in. gage) (%) racture .
As=welded Room 119,500 36 34 Weld defect
Rogm 127,700 40 26 We |d
200 hr ot 1300°F 1300
1300 48,900 20 12.5 Base metal ot
goge mork
200 hr at 1500°F 1500 70,100 17.5 37 Base metal ot
goge mark
1500 464,000 17 46 Base metat at
goge mork
As-welded 1500 63,900 10 15 Base metol at
goge mark
15030 64,600 9 3.3 Rose metal at
gage mark
200 hr ot 1650°F 1650 44,700 25 20 Weld defect
1650 45,000 1 36 Weld defect -
Toble 3.4.7. All-Weld-Metal Tensile-Specimen Tests of Vacuum=-Melted
Hostelloy B Weld Metcl on Hastelloy B Plate
Test Tansile | . Reduction
Heat Treotment Temperatyre Strength E ong'ahc-n n Arec
(DF] (o5} {% in 1=in. goge} )
As-welded Room 110,400 18.5 20
Room 125,300 3z 16
200 hr ot 1300°F 1300 117,400 2.5 25
1200 114,900 4.5 3
2Q0 hr ot 1500°F 1500 77,200 29 37
1500 74,400 35 43 "

 

204
PERIOD ENDING MARCH 31, 1957

Toble 3.4.8. Preliminary Results of Welding and Brozing Tests of Experimental Nickei-Moiybdenum Atloys

 

 

 

Alley Wettability Observations of
Designotion* During Brazing™* Weld
INOR-1 Good Badly fissured
-2 Good Ne visible fissures
-3 Yery poor Badly fissyred
-4 Very poor No visible fissures
“5 Good Badly fissured
-7 - Poor Mo visible fissures
-8 Good No visible fissures

 

*See Chap. 3.3, ""Fobrication Research,'' this report, for nominal compositions of these olloys.

**Wettability during brozing determined by using T-joints brozed in dry hydrogen at 1040°C with Coast Metals

brflziflg CI”Oy NO. 52-

Hot-Ductility Tests of Experimental
Nickel-Molybdenum Alloys

Hot-ductility tests were conducted at Rensselaer

in accotdance with pro-
3

Polytechnic Institute

cedures described in the literature?+ on experi-
mental ond commerciol nickel-base alloys. The
results ore presented in Figs. 3.4.13, 3.4.14, and
SIS
areo.

The results of the tests indicate that it is
vsually sufficient to evaluate the hot ductility os

in terms of the percentage reduction in

measured at 2300°F during cooling from some
elevated temperoture ond to compore these dato
with the hot ductility measured at 2300°F during
heating. The experience thus far indicates that
testing at 2300°F during cooling from a peak
temperature corresponding to the point where the
ductility during heating closely approximates
zero provides the best method of testing, The
level of impairment of ductility during cooling
which can be tolerated is probably dependent on
the conditions of restraint ossociated with the
actual welding operotion. However, materiels of
known propensity for cracking have all exhibited
ductility during cooling of less than 40% of the
ductility during heating under the test conditions.
Comgporisons of the test results indicate that
none of the INOR alloys behave in os ductile o

 

e B Nippes et al., Welding §. 34, 1835 (1955).

5E. F. Nippes 2t al,, Further Studies of the Hoi-
Ductility of HighTemperature Alloys, Rensselaer
Polytechnic Institute Report {June 1956).

monner as {nconel and thot they behave more like
Inconel X than like Hastelloy B, Alloys INOR-1,
-2, ond -3 behave in a maaner which appears to be
superior to thot of Hostelloy B (by virtue of high
os-hested ductility), whereas alloys INOR-4, -5,
and -6 ore definitely inferior to Hastelloy B, A
metallographic study is being conducted in on
these observations te micro-

effort to relate

structure and possibly to chemical composition.

HoK-TO=AIR RADIATOR FABRICATION
G. M. Slaughter

Effects of Brazing Yariables on Brozre Adherence

The study of influence of the rote of rise to
brazing temperature on flowability during radictor
brazing wos continued.® Metallographic exami-
notions of tube-to-fin joints brazed at wvarious
heating rates ond with different types of Coast
Metals brazing alloy No, 52 rings were completed,
and the results are presented in Table 3.4.9.
A brozing temperature of 1905°F was used in this
study, and the time at temperature was 30 min,

Tube-to-fin test samples were also brozed with
Coast Metals brozing alloy Neo. 52 at maximum
brozing temperatures of 1830, 1870, and 1905°F
at a rate-of-temperature rise of S00°F per hr.
The results of brazed-joint adherence studies of
these specimens are presented in Table 3.4.10,

 

%P, Patriorca ond G. M. Slaughter, ANP Quar. Prog.
Rep. Dec. 31, 1956, ORNL-2221, p 224,

205
ANP PROJECT PROGRESS REPORT

TESTED ON HEATING

UNCLASEIFIED

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ORNL - LR— DWG 17394
100 | i
_ L ol L] |
¢ L ' |
=~ ap =R T 11
4 “"~+
2 | - T
<l &0 s + i 3 ElL . N IR ) o ki 1
z . ' . ] | L U »
= | [
o 1 - L1
e 40 & INOR 3 |
%) | I R
g :
o 20 = e
& | || : _‘_
Q | |
1400 1800 2200 2600 1400 1800 2200 2600 1400 (800D 2200 2800
TESTING TEMPERATURE (9F)
TESTED ON COCLING FROM 2400 “F
100 e :
- | INEl NN e | [I1]
Q
& po ! ; —r+— —|—-'— — —:
o | [ INOR 3
1al i | i_'_'_'_‘_'_
o
< 60 | ’ __!T |
= L || | i
z | | |
& ag | 11 t 1
o e e
= 1 |
C2 oy . i = i S
* . by | b _
0 | 1 | I' | b
1400 12800 2200 2600 1400 1800 2200 26500 1400 1800 2200 2600
TES]ING TEMPERATURE [°F)
Fig. 3.4.13 Hot Ductllity of Nickel-Molybdanum Alloya INOR-1, -2, and -3.
UNCLASSIF IED
TESTED ON HEATIMNG DRHL -LR- DWG 17493
100 ] T T T T T
_ Y e ) ) =
g | ]
= 80 r r > = =
o o] L :
& | = T | |
80 L
- ot N
3 40 ' L% AL L L =
= _\ INOR 5 ',/ INOR 6
5 ¢
& 20 ————T—— T -
:: L] || |
I i | |
0 e L . . . i 1
3400 800 2200 2600 1400 800 2200 2600 1400 1800 2200 2600
TESTIMG TEMPERATURE (°F)
TESTED ON COOLING FROW 2400°F
oo .
2
< 80
<1
wh
o
< 50
=
Z 20
=
[
=
a 20
L
[ Lo
o

A0

 

 

1800 2200 2600 4400 4300 2200

TESTING TEMPERATURE

2600 1400 1800 2200 2600

(°F)

Fig. 3.4.14. Hot Ductility of Rickel-Molybdenum Alleys INOR -4, =5, and -6.
TESTED ON HEATING

PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED
OANL —~ LR —DWG 17452

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UI Ll | gl g
1400 B0 2200

  

2600 1400 1800 2200

TESTING TEMPERATURE (°F)

 

 

 

i PR g i [ 1
Eg x e - s e Rl
:1 o O
& |
4 =]
= bob—l e ==
=
o L s
£ |
& | 5 e
=
o - o
&
i | I_"‘il [ b 1 |
1400 1800 2200 2600 400 18Q0 2200 2600 1400 1800 2200 2600
TESTING TEMPERATURE (PF )
TESTED ON COOLING FROM 2400°F
| i | i1 M T { I T 1 1 | ! |
; | Y A
a |
& = s o ! = L e
= ] | , | (i 0] |
= ) [INCONEL X | | | | HASTELLOYE |
= | |
o . !
=
L — ~
g |
b =1 1 | T !
m ! —— -4

 

 

 

2600 1400

1B0D0 2200 2600

Fig. 3,4.15. Hot Ductility of the Nickel-Base Alloys Inconel, Inconel %, ond Hastelloy B.

Toble 3.4.9. Results of Tests of the Effect of Heating Rate and Type of Brazing Alloy Ring on Broze Adherence

Heating Rote and Type

Brazing temperoture: 1905°F

Time at temperature: 30 min

Percentage of Joint Areas with

 

of Brozing Alloy Ring 75-100% S0-74% 25-49% 0-24%
Adherence Adherence Adherance Adlterence
Heoting rate: 100°F /hr
ORNL cost rings 77 15 1 7
York sintered rings 15 8 3 74
Heating rate: 200° F /hr
ORML cost rings 27 3
York sintered rings 51 35 4
Heating rote: 200°F /hr
ORMEL cast rings 96 4
York sinterad rings 100
Heating rates 550° F /he
York sintered rings 00

 

207
ANP PROJECT PROGRESS REPORT

Table 3.4.10. Results of Tests of the Effect of Brozing Temperature on Braze Adherence

Heoting rote: S00°F /he

 

Brazing Temperoture and

Percentage of Joint Areas with

 

 

Type of Brazing Alloy Rings 75-100% 50-74% 25-49%
Adherence Adherence Adherance
Brozing temperature: 1905°F
York sintered rings 105
Brazing temperature: 1870°F
York sintered rings g8 Z
ORNL cast rings 100
Brazing temperoture: ]BBOOF
York sintered rings 70 21 g
CRNL cost rings 100

 

Assistance to Yendors

Assistance was given to the York Corp. in the
solution of difficulties in the fabrication of the
high-conductivity-fin NaK-to-air radiators, It wos
recommended thot the retort be boffled to obtain
more positive circulation of hydrogen, porticularly
through the fin bank, during the brozing operations.
The size of the retoct should be held to a minimum
1o lessen channeling of the gos flow and to reduce
the total mass. Very siringent cleaning ard
storage procedures should be followed throughout
the fabrication process. It was suggested thoi the
ossembled fin banks be ploced in airtight con-
tainers before brozing. The fins should be de-
greased immediately prior to assembly. A rate of
temperature rise of greater than 300°F/hr (prefer-
ably 400 to 800°F/hr) should be used through
the temperature range of 1700 to 1900°F. Further
experimentation should be corried cut with sintered
brazing alloy rings in order to study the influence
of processing voriables on the brazing character-
istics, ond experiments with rings produced by
other methods, including ORNL cast rings, should
be continved, The use of mico as a nonwetting
support material wos discouraged. It wos sug-
gested that alumina be used becaouse of its known
inertness.,

208

STUDIES QF GRAIN GROWTH IN
INCONEL TUBES

E. J, Wilsen G. M. Slaughter

The interest in maintaining @ fine-groined
strycture in the |nconel tubing used for reactor heat
exchanger ond radiotor components has led to a
preliminory study of the groin growth that results
from various brazing and heat-treating operations.
The initicl results of this study indicated thot
stress-relief treatments at 1500°F did not result in
grein growth, while significant grain growth re-
suvited from I,i-'nr brazing cycles at approximately
1922°F. It was also noted that growth proceeded
more ropidly in the graing locoted olong the outer
wall of the tube than in those located along the
inner wall,

Since high-temperature back-brazing is required
in the fabricotion of meny components, a more
detailed study was undertaken to defermine the
influence of the different variables on the grain
growth choracteristics of the tubing. The infor-
mation obtoined from an onolysis and evaluation of
growth behavior will be useful in determining the
optimum procedures for fobricating the warious
components,
Influence of Brazing Temperature

Two different lots of CX-900 Inconel tubing were
used for an investigation of the effect of the
brazing temperoture on groin growth, These lots
of tubing were identified as heat 5759 {0.187 in. OD,
0,025 in. woll) and heot 5789 (0.229 in, OD, 0.025
in. wall). One-inch segments from ecch lot were
cleaned with ocetone and subjecied to various
brazing temperatures for 30 min in dry hydrogen,
After the heat treatment the samples were sectioned
and examined in both the transverse and the
longitudinal direction.
structures of the heat-treoted specimens of 0.187-
in.-0D tubing are presented in Fig. 3.4.16 in the
order of increasing brozing temperatures., Although
wide warigtions in grain size occurred along the

Photomicrographs of the

outer peripheries of the tubes, the photomicrographs
shown are typical of the oreos thot experienced
extensive radial grain growth. After the 30-min
hect treatment ot temperctures os low as 1814°F
oend up to and including 1904°F, the grain growth
wos found predominantly in the grains along the
outer surface of the tube. The fine-grained
structure ¢long the inner surfoce appeared to be
relatively stable. After heating at 1922°F and
above, general grain growth wos evident, that is,
grains along both the inner and outer surfaces
had grown to significont proportions and many of
those which originated on the outer surface had
grown completely across the tube wall. As would
be expected, the frequency of large grains in-
creased with increasing tempercture,

The trends observed in these tests are illustrated
in Fig. 3.4.17, in which the grain size is plotied
as a function of temperature for both the outer and
the imner tube surfaces. The grain dimensions
selected for plotting were representative of the
grain sizes in areas that exhibited definite grain-
coarsening tendencies., At the higher temperatures
the curves reflect the coexistence of lorge and
small grains in the tube wall, For exomple, the
tube heated ot 1922°F for 30 min had grains that
penetrated radially from the outside surfoce com-
pletaly through the tube wall, but there were
adjacent arecs thot contcined smaller grains.

Infiuence of Time at Temperature

It was felt that the time at temperature might
also be an important facter in grain growth, and

PERIOD ENDING MARCH 31, 1957

therefore a study was initiated in which both time
and temperature were varied. For these experi-
ments, 0.187-in.-0D, 0.025-in.-wall CX-900 Inconel
tubing from heat 5759 was cleoned with acetone
and sectioned into 1-in. segments, which were then
heat treated at various temperotures for vorious

times. The heat-reated segments were then
sectioned in the transverse and longitudinal
directions for grain-size determinations. Photo-

micrographs of the structures of the heat-treated
specimens are shown in Fig. 3.4.18. As may be
seen the lowest temperature employed (1848°F)
and the 5- and 10-min time intervals produced no
appreciable grain growth along the outer periphery,
The specimens held at 1868°F for 30 min, however,
showed sporadic grain coarsening ot the outer
periphery. More extensive grain coorsening oc-
in the specimen held at temperature for
1 hr, as was expected. After 5 and 10 min ot
1904°F, no grain growth was observed, but ofter
30 min at this temperoture, some growth was

curred

evident. After 1 hr, groins occasionally penetrated
completely through the tube wall, and the inner
surface grains also had grown.

In the tests at 1922°F outer periphery groin
growth bhecame quite noticeable after 5 min at
temperature. Since it hod been estoblished previ-
ously thot excessive grain growth would occur in
30 min at 1922°F, no further experiments were
conducted.

Influence of Tube-Straightening and
Tube-Polishing Operations

A group of experiments was conducted to study
tube-straightening and tube-
polishing operations on grain growth in CX-900
lnconel tubing. Since these operations impose a
smoll degree of cold deformation on the tubing, it
was desirable to determine whether this cold work
was o major confributor tfo the ocuter-surface grain
growth noted upon hecting ot high temperatures,
Samples of 0.187-in.-0D and 0.229-in.-0OD tubing
were obtained in the following conditions: {1) an-
nealed, (2) anneoled and straightened, and (3} an-
nealed, straightened, and polished. The somples
were then examined os-received, of subjected fo @
brazing cycle of 30 min at 1868°F and examined,
or subjected to a brazing cycle of 2 hr at 1922°F
and examined. The specimens were cut transversely
and mounted for microscopic examination,

the influence of

209
oLe

 

 

 

 

LA L SECTIOMS- 3 RSy T L

T DAL ATAD L TTRCLYT W7

1Y0dIY SSIYO0YS LD3T0Hd dNV

Fig. 3.4.16. Tronsverse Sections of 0.187-in.-0D, 0.025-in.-Wol1l CX-900 Inconel Tubing Held 30 min ot Temperatures Shown. Original magnification:
100X. Etchant: oxalic acid (electrolytic), Reduced 67%.
PERIOD ENDING MARCH 31, 1957

UHCLASSFIED
CRNL-LR-DWG 20234

 

 

 

 

 

TT T T T T T T T T T T 1
i | ] , .
26 | —
MAX MU RADIAL GRAIN S1ZE LIMITED 8Y TUBE WaLL THICKNESS
24 |
22 —
20 b—

       
 

-QUTER TUBE SURFACE

RADIAL GRAIN SIZE tmiist
=

3
|

 
  

J

INMER YUBE SURFaCE

m
AT -HECEIVED

 

 

 

 

]

1800 1820 1840 1860 188G 200 1920 1940 1980
TEMPERATURE 19F)

 

Figs, 3.4.17. Effect of Temperature on Grain Growth in 0,187.|n.-0D, 0.025-In,.Wall CX-90 Incone! Tublng.

As may be seen in Figs., 3.4.19 and 3.4.20 the The samples that were cold rolled were ploced in
two sizes of tubing reacted somewhaot differently a retort and subjected to a typicol brezing cycle
when subjected to the brazing cycles. Hawever, (heated at 300°F/hr to 1922°F, held for 30 min at
it is evident that the straightening end polishing 1922°F, and coocled ar 600°F /hr) before they were

operations had only minor effects on the grain prepared for metallegraphic examination of grain
growth.  The as-onnealed tubing olso showed growth,
definite grain growth at the outer surface.
Influence of Cold Work A medium-sized grain structure wos observed
n ;
after heat treatment of the samples which hod been
The recrystallized grain size of a metal de- cold reduced 5, 8, 12, and 15%. Some grain
creases with increasing amounts of celd werk, ond coarsening wos observed, however, in the sample
therefore it was thought thet cold werking could reduced 2.5%. The smallest grain size was ob-
be used lo produce o medium-sized grain structure  served in the sample heat-treated after 15% cold
ofter brazing., Preliminary studies of the influence  ,.diction.
of cold work were conducted on small segments
cut from onnealed lnconel sheet, The segments A length of I,aé—hord {approximately 15% cold
were cleaned with ocetone to remove dirt and reduced) Inconel tubing was therefore acguired
grease ond then examined in the os-received from the Superior Tube Co. for further experiments.
condition and aofter they had been reduced in Samples of this 0.250-in.-0D, 0.035-in.-wall tubing
thickness 2.5, 5, 8, 12, ond 15%, respectively, were prepared ond subjected to various simulated
by cold rolling, brozing cycles.

211
21

 

Fig. 3.4.18. Effect of Time and Temperature on Grain Growth of 0.1874in.-0D, 0.025-in.-Wall
Reduced &§7.:5%.

ALL AT GRS MAMDIELE

 

EPCHANT — OaliC a2 1D FLECTEQLYTICH

CX-900 lnconel Tubing, Original magnitication: 100X,

 

1A0d3Y SS3Y¥00UYd LIIT0Ud NV
el

L L]
R

 

B - § Y-arand

  
 

s ;M-NE.M.ED. STRAIGHTEMED,
ARD HEATED FOR 30 miln AT 1B6E°F

= ]
i

 

i ) . "B ]
ANMEALED, STRAIGHTEMED, POLISHED,

T A0 HEATED FOR 2 b AT 1222°F i,

P

Fig. 3.4.19. Effect of Annealing, $traightening, and Polishing on Graln Growth in 0.229-in.-0D CX-%00 Inconel Tubing. 100X, Etchont: oxalic ocid

and copper regia. Reduced 54 5%.

. AMPEALED AMD HEATED FOR 30 min AT VE&8*F .

e

——

F - - LWL TR
¥ raxm

 

' ANMEALED, STRAIGHTENED, :
-mn HEATED FOR 7 he AT 10324F -

-
ll

i ARMEALED, STRAFGHTEMEL, AND FOLIHED '

 

Tmiy atiErE
- I

 

 
 

ANNESLED, STRAIGHTERED, Foas
AND HEATED FOR 30 min AT 1888%F

    

£561 'IE HOUYW ONIONT QOIY¥3d
riZ

 

Fig. 3.4,20, Effect of Annealing, Straightening, ond Polishing on Grain Growth in 0.187-in,-0D CX-200 Inconel
and copper regia. Reduced 56.5%.

Tublng.

100X. Etchaont: oxalic ocid

L¥0dTY §538008d LD3M0dd dNV
The structure obtained after rapid heoting to
1922°F and holding for 2 hr is shown in Fig. 3.4.21.
The medium-sized grains obtained with the typical
brazing cycle (300°F/hr heoting rate) are shown
in Fig. 3.4.22, The sample wos held at the
brozing temperature of 1922°F for 30 min. A
larger quantity of ],zé-hurd tubing is being acquired
from the Superior Tube Co. ond o more extensive
exemination of its behovior will be mode.

SUBCONTRACT WORK

The University of Florida subcontroct for on
investigation of the diffusion of boren and silicon
from high-temperature alloys was completed ond @
fina! report is being prepared. The Glenn L.
Martin Co. subcontract for « study of electric-
resistance welding of tube-to-heoder joints was
also completed and o final report is being prepared.
The progress made by Rensseloer Polytechnic
Instityte in their hot-ductility studies is reported
in the section of this chapter on ""Welding and
Brazing of Nickel-Molybdenum Alloys."’

The Battelle Memoricl Institute subcontract
for molybdenum welding studies has continyed.

UNCL ASSIFIED
Y-21404

 

Fig., 3.4.21, Mlcrostructure Obtoinad upon Heoting

Y%-Hard Inconsl Tubing at 1922°F for 2 hr. 100X.

Etchant: oxalic acid gnd copper regio.

PERIOD ENDING MARCH 37, 1957

The objective of this program is the development
of techniques for producing ductile weldments.
Carbon-deoxidized arc-cast material was used.
Welds were made with a shielded tungsten electrode
both in o dry box end in the open otmoesphere ond
were tested by o single-point-loaded bend test ot
room temperature and at 75°C. In oddition to
studying the effect of ctmosphere purity, vorious
etching and abrasion techniques were employed
in preporing the surfaces, and several types of
shields were used in welding the samples in the
open otmosphere.

The series of samples prepored by wvorious
cleaning techniques were welded in o dry box,
They showed that acid etching resulted in welds
with somewhat better ductility than those prepared
by other metheds. They exhibited bend angles at
fracture of 40 deg at room temperoture and 76 deg
ot 75°C compared with average bend angles of
28 and 62 deg, respectively, for other samples.
For weldments prepared in air, the samples welded
with a trailing shield showed significantly better
ductility than all other samples,

UNCL ASSIFIED
Y-21395

Fig- 3.4.22.
Ifé-HGIrd Inconel Tubing at o Rate of 300°F /hr to 1922°F
100 X.

Microstructure Obtained ypon Heating

ond Holding ot that Temperoture for 30 min.

Etchant: oxolic acid and capper regia.

215
 

ANP PROJECT PROGRESS REPORT

3.5. MECHANICAL PROPERTIES STUDIES
D. A. Douglas

DYMAMIC STRAIN PROPERTIES OF INCONEL
C. R. Kennedy'

The study of the effect of thermally induced
strain cycles on the rupture life of Inconel was
continued, Two types of tests are being made,
including relaxotion tests and mechanically in-
duced strain-cycling tests. Stress relaxation data
are necessary for determining the octual amounts
of plastic strain that occur under operating con-
ditions; whereas, strain-cycling data can be utilized
in determining the life expectancy of structural
members subjected to strain reversals, Although
the majority of the strain-cycling problems are
cavsed by thermal fluctuations, tests in which the
specimen is mechonically strained under isothermal
conditions give a definite insight to the effects of
thermally induced strain cycles.

 

Yo assignment from Prott & Whitney Aircroft.

24,000 | - ———————r

Ei,OOO! — - -_ | et

 

 

 

 

 

 

 

 

(8,000 |=——18,000 psi AND 0.2% STRAIN-
=—16,000 psi AND 0% STRAIN
16,000 }——' ! L
e, (2,400 psi AND 0.05% STRAIN
& 11,000 psi AND 005 % STRAIN |
W ] = 1
2 12,000 r —_—
2d I
- —r—
Lip}
9000 ¢ =
10,600 psi AND 0.05 % STRAIN
.' | |
|
6000
3000 | | ! —
| |
O i i

 

 

Relaxation tests of Inconel onnealed ot 2050°F
for 2 hr were completed at 1300, 1500, and 1650°F,
and the results are presented in Figs. 3.5.1, 3.5.2,
and 3.5.3. The tests were run at total strains of
0.05, 0.1, and 0.2% at 1300°F and of 0.05 and 0.1%
ot 1500 and 1650°F, One specimen was used af
each temperoture to determine the relaxation rates
for all strains, The procedure was to run four
successive 100-hr tests with 0,05% total strain and
then run the tests with higher total strains. This
was done to determine the effect on the relaxation
rates of prior strain at the test temperature. It was
tound that the prior strain only slightly offected the
relexation properties of annealed Inconel tested ot
1300°F, as shown in Fig. 3.5.1. It may also be
seen that, ot ¢ll temperotures, when the total strain
is greater than 0.05%, which is the strain of the
proportional limit for coarse-grained Inconel at
these test temperatures, the relaxation rate is the
same as that of the 0.05% total strain test. Since

UNCLASSIFIED
ORNL=LR—~DWG 192414

A—NO PRIOR STRAIN
B— 0.04% PRIOR STRAIN! L
| C~0.08% OR GREATER PRIOR STRAIN .

 

 

O 0.02 0.Ca 0.0 Q.08 o 0.2

0.3 i 2 5 10 20 50 oo

TIME (hr)

Flg. 3.5.1. Relaxation Characteristics of Inconel Annealed ar 2050°F for 2 hr and Tested ot 1300°F. Stresssd

to produce o constant strain.

216
PERIOD ENDING MARCH 31, 1957

UNCLASSIFILD
ORNL—LR—-DWG 19942

 

16,000
14,000 — . ]

' I
12,000 12,000 psi AND 0.4 STRAIN-—— | —

Q000 ! i —;
[ | :
9000 ps1 4ND ICI.DS‘.’{. STRAIN

2000 -

STARESS {psi1]

2000 |

 

 

 

 

Gd— - = .
G o0z 004 00 008 o Q2

a5 { 2 5 10 20 50 G0

TIME {hre}

Fig. 3.5.2. Relaxation Characteristics of Inconel Annealed ot 2050°F for 2 hr and Tasted at 1500°F. Stressed

to produce a constont strain. [Confidential with caption)

at these test temperatures strein hardening is
small and recovery occurs rapidly, it follows that
the plastic strain given the specimen on loading to
totel strains greater than the elastic limit has neo
effect on the relaxction rate of the material.
However, at temperatures less than 1300°F, this
may not be the case, ond the motericl may reflect
the effects of strain hardening in its relaxation
rate.

The strain-cycling equipment described pre-
viou5|y2 was in operation during the quarter ond
preliminary results were obtained ot a test tempera-
ture of 1500°F. The greatest number of tests were
run with red specimens in two conditions, os-
received ond annealed at 2050°F for 2 hr, The
test results are shown in Fig. 3.5.4, in which the
log of the plastic strein (€.) per cycle is plotted
vs the log of the number ofpcycles to failure. The
procedure of operation hos been to apply o load
which is of a magnitude that will cause the specimen

 

2], R. Weir, Ir., ANP Quar. Prog. Rep. Dec. 31, 1956,
ORNL-2221, p 246.

to experience ¢ specified plastic strain in 1 min
and then to reverse the load to produce the same
plastic strain in the opposite direction. The
plastic strain vs time wave has the shope shown
on Fig. 3.5.4. It moy be seen thot the coorse-
grained specimens (annealed at 2050°F for 2 hr)
exhibit fewer cycles to failure for all values of
€ . Metallographic examination of the specimens
showed that the failures were intergranular and
that they initiated at the surface of both the
coarse- and fhe fine-grained materiol,

Tests were also run at 1500°F on 0.75in.-0OD,
0.C6-in.-wall Inconel pipe both in the as-received
and the annealed (1950°F for 2 hr) conditions. As
may be seen in Fig. 3.5.5 there is some scatter of
the deto; however, all but three points fall within
the band brocketed by the coarse- and fine-grained
rad data,

Tests of Inconel pipe in the fused solt NaF-
ZrF -UF, (50-46-4 mole %, fuel 30} ot 150C°F are
now in progress in order to determine the effects
of this corrosive medivm on the stroin-cycling
choracteristics of Inconel. Strain-cycle tests will

217
ANP PROJELCT PROGRESS REPORT

20,000 ——————————
18,000 — —
16,000 -

14,000 —_

|
12,000 |-—— S — —
—— 11,000 psi AN? 019, STRAIN
10,000 f—— - : -

STRESS [psi b

8000 |=— 8000 psi AND 0.05 % STRAIN

| | L
6000
4000 |

2000 |-

 

 

 

O e |
o 002 2.049 .06 0.08 34 02

Flg. 3.5.3. Relaxotion Characteristics of Inconel Annealed at 2050°F for 2 hr and Tasted at 1650°F.

ta praduce a constant strain. (Conlidential with caption)

al so be run ot 1300 and T650°F on both as-received
and annecled specimens in argen ond in fuel 30.
Since the stroin-cycle type of test provides an
importont measure of an alloy’s ductility, such
tests will be included in the program for evaluating
potential reactor materials,

ln conjunction with this work at ORNL, addi-
tional data are being obtained uader subcontrocts
at the University of Alabama and Bottelle Memerial
Institute, At the University of Alaboma experi-
ments are being conducted to sfudy the effect of
plastic strain induced by thermaol fluctuations.,
faitially the tests were progromed to provide o
I-hr hold time at gach end of the temperature ex-
cursion; however, all the specimens tested in this
monner exhibited preferentiol growth and ultimotely
tailed by buckling. Recently, tests were conducted
so that a complete temperature reversol was ef-

fected in 1 min. This rate corresponds to the

218

   

UNCLASSIFIED
QRNL=LR—DWG 19943

— T

 

|
L .|
||
e ‘ "i'i‘|
| |
|
i

L

 

. A
L)t ERE
05 i 2 5 1o 20 50 100

TIME (hr)

Stressod

rate used for some of the mecheonicol stroin-cycle
experiments conducted at ORNL, and the type of
Therefore some corcelation
between the two types of tests should be possible,
at least for the rapid cycle rates. It might be
assumed thot o material would react in the some
monner 1o repected plostic strain regardless of the
manner in which the strain wos induced. However,

foilure was similar.

this is true only if no metallurgical changes are
initiated by temperature changes and if the stability
of the structure at the temperature extremes is
independent of time. The main issue to resolve is
whether any such changes occur in the operating
temperature range of the reactor,

The natural desire to obtain information of this
iype as quickly as possible led to an attempt 1o
obtain data before all the testing techniques were
developed. The mony testing difficulties en-
countered have, however, necessitoted a slower,
1000

300

200

 

HH IS FINE - GRAINED ROD

 

% -_ 1 .
oo ] S S
B = i Y _| i ool bl
- | [ M | ] ®
tal . = | = P Ml
o ' COARSE -GRAINED ROD
t'} 20 L | } 1 i ..._il 2
& i
¢ d
2 40l LT s
& - S '
e =13 J
m il [SE—] S— } + - - —_— i
o |
) el - | i
% | i I || |
ey [
_ | |
o | |
W | |
. -
—l
1 L } i it f—— |
== i . |
0.5 | | l
ozl.' Loa g ] 4 g Pyl |
| |
01 | -2 1 340 4 p bbb
0t oz o8 | e 5 10 20

PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED
GRNL—LR-DWG 19944

 

 

 

T e ASRECEWED ||
O ANMEALED AT 2050°F FOR 2 he
| |

 

 

S0 100 200 300 ICO00 2009

LOG OF NUMBER OF CYCLES TO FAILURE (log &)

Fig- 3.5.4. Strain Cycle Charactaristics of Incenel Red Tested in Argon ot 1500°F. (Contidential with coaption)

more analytical approach. As a result time is now
being spent determining the temperature profile
over the gage length, the effective gage length,
and the effect of commencing the test at various
temperature levels. Thus an odditional delay will
be incurred before dota of sufficient accuracy
become ovailable to attempt o correlation with the
ORNL results.

A program wos initioted ot Battelle Memorial
Institute on July 1, 1956 1o fully explore the
slevated-temperature properties of Inconel under
dynomic-stress conditions. The progrom entails
complete reverse stress fatigue tests and fatigue
tests with various constant load stresses applied
in the tensile direction. Such information, together

with stress-rupture dato, con be used to form

“Goodman’’® diagrams.  From these plots the
stress conditions that will result in failure by
¢creep can be estoblished, ond the stress conditions
that will result in failure from fatigue con be
determined. Since the primary interest is in low-
frequency stress reversals, the tests be
conducted at 1 and 10 cps, Preliminary results
show that, ot o temperature of 1400°F, opplication
of constant load stresses up to 4000 psi will not
significantly alter the stress to produce a fatigue
failure in 1000 hr. Data arg to be obtained for

diagroms for 1200 ond 1600°F.

will

 

35, Timoshenko and G. H. MacCullough, Elements o/
Strength of Materials, 3d ed., p 371, ¥Yon Nostrond,
Mew York, 1949,

219
ANP PROJECT PROGRESS REPORT

1200

URCLASSIFIEDR
ORML—LR—DWE 19945

 

 

 

 

   

1L }H—
- -l"r"

 

 

500

 

 

 

 

 

 

 

 

 

 

 

 

 

 

200 ———

 

100 |

 

o
[=

 

 

 

 

 

 

 

i
O

o

 

 

n

- 1 1. L 1l e 4 1111

i 11 |
‘ ® AS RECEIVED

LOG OF PLASTIC STRAIN PER CYCLE {log €,

| O ANNEALED AT 1950°F FOR 2 hr

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  
   

T T

T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I | ‘ -.:
{ — | FE | i :}: ' i | — \‘* | 1
— 1 T — "_.‘_! TN e
= =14 | 1 [ [ T ] —
as T HH __:fjfr; T T L
e 1) Ll (L riiey o 4 L] e A
|—__;_ . 3 — | | I — | | e l —|» ' ! ! 4 _i- —_——— l ‘ i } i -1
| | |
i { 11 1 1yl Ll 1 Ll 17 TR L
a.2 : } | : [ i i ‘ | : | ch- . | T“ H
1 T —
o4 : | ' i | | [} || | | ‘ ! | L i1l
o1 02 05 t 2 5 10 20 500 100 200 500 1000 2000 5000 10,000

LOG QF NUMBER QF CYCLES TO FAILURE (lag M)

Fig. 3.5.5. Stroin Cycle Characteristics of 0.75-in,-0D, 0.08-in.-Wall Inconel Pipe Tested in Argon ot 1500°F,

{Confidantial with caption)

ALLOY DEVELOPMENT
J. W. Woods

Evaluation studies were mode of the creep and
creep-rupture characteristics under a constant load
of two experimental nickel-molybdenum alloys
being developed for reactor use. The environment
used in these tests was the fused salt NaF-KF-
LiF-UF, {11.2-41-45.3-2.5 mole %, fuel 107). The
two olloys tested were vacuum melted ot ORNL.
Heat ¥T-5) had the nominal composition {(in wt %)
17 Mo—10 Cr-0.5 Al-0.5 Mn-0.06 C-7 Fe~bal Ni,
and heat YT-52 contoined (in wt %) 17 Mo-10
Cr-0.08 C-7 Fe—bal Ni. The results of the tests
are presented in Table 3.5.1. The corrosion of the
unstressed surfaces of the specimens from both

220

heats thot were solution annealed at 2000°F for |
hr and aged at 1300°F for 50 hr was 0.5 to 1 mil.

The compositions of the two alloys were virtuclly
identical, but the methods of preporation were
different. Heot VT-51 was prepared from elemental
melting stock, and heat YT-52 was prepared from
Inconel diluted with Ni, Mo, and Fe. Alloy VT-5I
demonstrated an aging response, while alloy VT-52
did net, Becouse the alloys were prepared in
small heats, veriations in mechanical properties of
of the two alloys were to be expected. Both alloys
exhibited good ductility.

Photomi crographs of the specimens are shown in
Figs. 3.5.6 and 3.5.7. The formation of o con-
tinyous network of carbides around each grain is
showrn in Fig. 3.5.6, whereos in Fig. 3.5.7 the
PERIOD ENDING MARCH 31, 1957

Table 3.5.1. Stress-Rupture Properties of Two Exparimentol Mickel-Molybdenum Alloys

Test temperoture: 1500°F
Stress: 8000 psi

 

 

 

 

 

 

a
H N g ; T Rupture Life Elengation
1 ¥ t 1
eat Mo pacimen l[roatmen (hr) ()
¥T-51 Solution anneoled ot 2000°F for 1 hr Bal 76
Soluticn annecled at 200G°F for T Ar 1243 45
and aged at 1300°F for 50 hr
YT-52 Solution annealed at 2000°F for 1 he g20 72
Selution annealed at 2000°F for 1 hr Test discontinuad ot 780 72
and aged at 1300°F for 50 hr
1 1 3 foowm
yogl " e . Ty
. .y UNCLASSIFIED
" : =21710
b o l.,'
; 'i‘ *, X
iy ) ah
. “i
i 'y WS
o A
" _' r“ N 5
‘ ) . ¥
... ‘ .: . L l-_l_.. 5
- ¥ % ‘-';;"‘_‘-.“ I‘. 3 .
’ i'_:l g O I;f- y
VA o 21 ' /
¢ | £, -
i f2 1 1 a
|‘ . .;u;.. } iy
o }
| ..IH‘ 'L
.t !
\1\\’. t . y ‘.‘
| -
1“-‘. d ' ""
1.‘“..-'I

 

 

 

 

 

Fig. 3.5.8. Alloy ¥T-51 (17% Mo—-10% Cr-0.5% Al-0.5% Mn-0.06% C—-7% Fe-bal Mi) Creop Tasted in Fuel 107

ot 8000 psi and 1500°F in the Aged Condition. Ruptured in 1243 hr with 45% alongation. 100X. (Secrst with
caption)

221
ANP PROJECT PROGRESS REPORT

 

   

. - 4 . 4
kv o+
.ri‘. wovhay . "‘r- L !.‘-
i 4 Faoad BT =G
B e i et N
S A ah i
Pl in T J!-':—-
=F . iy TN
. 3 ;-'i . "!“.
: R
= ool g R e
S S
R iRy t-']'_:" %
] -ty £y ik
H gty e
ey R Y
' “hh A ‘; 2
- o 1o e et
SN
Sl ety R
e Tmp e R T R
o3 Fouag ek F K
3 |"'r'{'€ o o R r]i ;
L e M e R &
W S A Sigdh g
Yeus el il
: s I L e i
" ER ; AET .:-‘u-i' 5
5 % ' WY e
B <0 i d g e oty
wk . g . ! ‘ i %
o b PN e 1 FA a1y e
wEan g thed Tt RN TSN oyt
= SR i . & LenE
K o=y l oy } 5
LR, e 1 UNSTRESSED  Ha 7
v BT A 4 e o rmery ® W0 Yo
i Tm W . - e SR o e ol e
v & T L
Sk . oo

  
              
        
    
   

S taa T ML
UNCLASSIFIED i
Y.21708

a

  

  

     

 

    
  

o ™
Ay BB o x -

i : !:”. e ‘Ca"‘r "
ke g i
2% SYRESSED ¥ mit &

ey g RO
w _' ; T '

   

N o Vg
S

A e
erk, S ok
- L
3

st

v
ar
o T

z e ay !
L

A
T

L
v

“la A AT L

= e Ras b

 

 

 

 

 

 

 

 

 

Fig. 3.5.7. Alloy ¥T-52 (17% Mo-10% Cr-0.06% C—7% Fe—bal Ni) Creep Tested in Fuel 107 ot 8000 psi ond
1500°F in the Aged Conditlon. Ruptured in 788 hr with 70% elongation. 100X. (Secret with caption)

carbide precipitote in the grain boundories has a
more random arrangement which could account for
the lower strength exhibited by clloy VT-52, In
general the precipitates were well dispersed in
both specimens.

Hostelloy B hos received considerable attention
in the olloy development program because it is
commerciclly ovailable and it has excellent high-
temperature strength and corrosion resistance.
However, the absence of chromium in Hastelloy B
results in characteristically poor oxidation re
sistonce. One aim of the development program,
therefore, is to introduce sufficient chromium to
provide good oxidation resistonce but not encugh
to result in severe corrosion. Hastelloy C, which
is also o commercially availeble nickel-molybdenum
alloy, centains 15 wt % chromium. |t was recognized
that this amount of chromium would lead to con-
siderable corrosive attock, but, since the composi-
tion of this alioy represents an upper limit of

222

chromium, it was decided that o certain amouynt of
data on its strength properties in fused salts would
be pertinent to the problem., Therefore, stress-
rupture data were obtoined ot 1300, 1500, and
1650°F in fuel 107 for comparison with dota for
other alloys being investigated.

The results of the tests are shown in Fig. 3.5.8.
The dato show quite good stress-rupture properties,
and it oappears that for 1G,000 psi ot 1500°F
Hostelloy C witl have a rupture life of over 1000
hr. The two specimens tested at 1500°F in
Hostelloy B creep test chambers had poorer stress-
rupture properties thon those tested in a Hastelley
C test chamber. This may be explained by the
fact that the creep test chamber made of Hastelloy
B contained no chromium and thus provided,
through alloying and diffusion, a strong ""sink”’ for
chromium and accelerated the removal of chromium
from the specimen.
- The results of the alloy development program
thus far indicote that considerable progress has
been mode toward ochieving the objective of de-
veloping a material that is as fabricable as required
reactor  constryuction, possesses adequate
strength, oand is corrosion resistant in high-
temperature fuel mixtures, liquid metals, and air.
Commercial alloys such as Hastelloys B ond W,
which have excellent corrosion resistonce and

for

PERICD ENDING MARCH 31, 1957

strength, will have to be avcided, since, ot the
temperotures of interest, they precipitate a brittle
phase that causes poor fabricability and low
ductility. Experimentol nickel-maolybdenum alloys
of the single-phase or solid-solution type have the
desired fabricability and corrosion resistance, but
they lack adequate creep strength. In spite of the
poorer creep strength, however, these alloys are

stronger than Inconel.

UNCLASSIFIED
ORNL—LR—-DWG 19946

 

 

 

 

 

 

 

 

   
 

  

 

 

 

 

 

 

 

100,000 L I 1 l T T _‘ | T l T
— =t
— | NE | T
e e L 3 4 — e — L
] | = || | _
| | 1
50,000 F——— e S
I | 27% ELONGATION AT RUPTURE | |
_ | ‘
i | i | i — L { —| -
HASTELLOY C N HASTELLOY C‘CH,TME!:ER AT 1300°F | ‘
< I} |
£ HASTELLOY C IN HASTELLOY C CHAMBER AT {300°F —
20,000 | L '3"'1:? Ll M. -, EE— ; | , 259 _
32% o |
2
» 10,000 F
iy
@ L
m -
'—
s : . 1 4 ) L] i ¢ - I
| i |
| HASTELLOY C IN HASTELLOY € CHAMBER AT {650°F -’r\_l"‘* — ||
; I [ | ) |
5000 — - - ——— S
I | 1'—.4_- {
L = || 4o b Al s =
e \
— ——— 4 i i_ 1 _‘__
|
|
2000 + - | | e |
| .
| | |
| a
| | | |
-. iOOO B | | ' ] | ] J ] ]
. 10 20 50 100 200 500 1000
TIME TO RUPTURE (hr)
3 Fig. %.5.8. Stress<Rupture Properties of Hostelloy C Annecled at 2250°F for 1 hr and Tested in Fuel 107 at

1360, 1500, ond 1650°F. {(Sacret with caption}

223
ANP PROJECT PROGRESS REPORT

Attempts have been made to increase the strength
of the nickel-molybdenum alloys by the addition of
elements that form complex precipitotes and by the
addition of elements that are simple corbide
formers. Additions of very small amounts (< 1%)
of aluminum and titenium do not seem to increase
the strength, but they do increase the final elonga-
tion at rupture, probably because of their affinity
However, fabrication
difficulties are encountered when the additions of
It hos been

for oxygen ond nitrogen.

aluminum and titonium exceed |.5%.
possible, however, to improve the creep charac-
teristics by odditions of carbide-forming elements,
such a5 niobium and tungsten, os wes done in the
case of Hastelloy C, without seriously compromising
the other good characteristics. Several clloys of
this type have exhibited a stress-rupture life of
1000 hr when tested in fuel 107 ot 8000 psi and
1500°F. A considercble amount of odditionol
work is needed to achieve the optimum bolance of
strength, fabricobility, and corrosion resistance,
but the success te date encourages optimism.

TESTING OF BERYLLIUM
D. A. Douglas

The Brush Beryllium Co., under a subcontract,
has for several years been conducting an ex-
tensive program of mechanicol testing of beryllium

224

to explore the effects of variables such as tempera-
ture, surface finish, strain rote, and oxide content
on the mechanical properties. Most of the work
was summarized in o report issued in August 1956.4
In the current contractuol year the work has heen
concerned with measuring the elevated-temperoture
tensile and stress-rupture propeities of samples
trepanned from the blocks being hot-pressed for
the ART. It wos felt desirable to determine
whether the lineor relationship between oxide
content and strength revealed by the past work
would be evidenced in the larger pressings. Also,
some evaluation of the blocks is required to ensure
that the strength does not fall significantly below
the strength on which design calculations were
based. The results of the tests, as reported in
The Brush Beryllium Co. quarterly progress re-
ports,>*® indicate that the strength properties do
not vary appreciably from those predicted,

 

4R. G. O’Rourke et al., Mechanical Properties of

ReactorGrade Beryllium at Elepated Temperatures,
C00-312 {August 1958).

R, G, O’Rovrke ond J. N« Hurd, Mechanical Testing
of Beryllium at Elevated Temperatures, Progress Report
for First Quarter Fiscal, 1357 {undocumented).

®R. G. O'Rourke and J. N, Hurd, Mechanical Testing
of Beryllium at Elevated Temperature, Progress Report
for Second Quarter Fiscal, 1957 {undocumented).

Pt |
 

PERIOD ENDING MARCH 31, 1957

3.6, NONDESTRUCTIYE TESTING AND INSPECTION
OF MATERIALS AND COMPONENTS

R. B. Oliver

EDDY-CURRENT INSPECTION TECHNIQUES
R. 8. Oliver

Tubing Inspection
J. W, Allen

The application of the recently developed Im-
p-«=.u:h}gr-:1ph,\i an instrument which measures both
the a-c resistance and the reactance of its testing
coil over a wide range of frequencies, was in-
vestigated further. A prototype of this instrument,
described by the block diagram in Fig. 3.6.1, was
constructed. It is composed primerily of ean
oscillator in which the bifilar-wound center-topped
testing coil is the frequency-determining element.
Both the a-c resistance of the coil and its in-
ductance cre influenced by the tubing under in-
spection. The amplitude of the clternating voltage
generated by the oscillator varies inversely with
the resistance of the testing coil, and the frequency
of the voltage vories inversely with the square
root of the inductance of the testing coil. The
amplitude detector (Fig. 3.6.1) and its associated

 

'}, W. Allen, ANP Quar. Prog. Rep. Dec. 31, 1936,
ORNL-2221, p 240.

TUBE UNDER TEST

—

 

 

 

 

A. Taboada

amplifier provide o direct coupled signal which
varies inversely with the resistive component,
R, of the testing coil’s impedance. The frequency
discriminator, with its ossociated amplitude
limiter ond amplifier, provides a direct-coupled
signal which varies inversely with the reactive
component, X, of the testing coil's impedance.
The change in impedance of the coil is described
by the curves of Fig. 3.6.2. On this graph, the
short-dashed as o parameter, the
frequency ratio ///_, in which [ is the operating
frequency of the coil and /. is determined by

curve has,

/ 2
B it
wrD,
where
g o= dr x 10-7 henries/m, permeability of
nonferromagnetic tubing,
y = conductivity of paort being inspected,
in mho/m,
D = outside diometer of part, in meters.

£
The four sclid curves drawn from the short-

dashed curve to the point (0, 1) represent the effect

UNCLASSIFIED
ORML-LR-DWG 1TITS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    
     

 

 

 

Ry SELECTIVE
e Y
y TESTING COIL (OPERATING AT NATURAL . .. DIEFERENTIATION
o __J|[_ - _HESONflNT FREQUENCY} CHAMMEL 4 f,';‘_ il
L 3
e == ,—;‘;';.IT—_B ! ' | 7 | -
OSCILLATOR AMPLITUBE e | R RESOLVING |4 R-A)x | poc. | lcHamneL | &
¥ A DETECTOR 0.C. I COMPONENT BRIDGE AMPLIFIER [ | &
& 3 AMPLIFIER N =
- : | | E
FEEDBACK,”Y -~ SENSITIVITY S GaIN A =
g ' K | o
o
] "o
z
CHANNEL 2 o
o
I.-__-:"_"."-_-‘_"‘;-_n- | | t_"‘
AMPLITUDE .| FREQUENCY o : X RESOLVING | FXL(1-0R 0C. | CHANNEL 2| &
LIMITER DISCRIMINATOR k MPONENT | AMPLIFIER o
/ o g AMPLIFIER | | S | ; ‘ | F
! ‘ r .r
P — e ’ £ ;
.'l' . ALY T A — I_ ,rf_ —_— : e J
; £ s
7} FREQUENCY ~A GAIN K. ¢ SELECTIVE
" ) Ty 7y DIFFERENTIATION

Fig. 3.6.1. Block Diagram of Impedograph.

225
ANP PROJECT PROGRESS REFORT

 

UNCLASSIFIED
ORHL~LA~0WSE 139954
o
I | | l [

fo B — ey
¢o €T v s

o = PERMEABILITY DF

HORFERRDMAGNETIC

TUBING thenrizsfm)

¥ = CONDUCTITY OF PaRT
BEING ISPECTED direna/m)

   

  

 

  

07 \ \ O, = DIAMETER OF PART [m)
X % b, DAMETER OF COIL [m)
08 — \ : £ = FREQUENCY -
1 g,
\ b= e

REGCTIVE COMPONENT
o &
F [L]
| |

 

 

 

o3 {
ne
g
aa — Jf“ ‘.(
l.é\?z 3% 9% WLl THICKNESS (N 9 OF OD
@,
o . | 1 — = - 3 4
G Q.4 0.2 Q.3 o4 2.5 06 0.7 Q8 3 10
RESISTIVE COMPONENT
Fig. 3.6.2. |Impedonce Plone for Coils Encircling
Metal Tubes.

of wvarying the tubing wall thickness (as a per-
centage of the outside diometes) for (/] ratios
of 9, 16, 25, ond 100, The long-dashed curves
indicate the effect of varying the outside diameter.

It is important to note that if the proper f/fc
ratio is selected for o particular tube, angular
separation of approximately 90 deg can be obtained
between the effects of wvorying diameter and
varying wall thickness. An example of this sit-
vation occurs at an f/f_ ratio of 25 for a tube
hoving o wall thickness which is 10% of its
outside diameter.

Both the R and X components of any excursion
from the operating point are measured by the
Impedograph. These 1wo components moy be
resclved into any other set of two orthagonal
components by the use of resolving bridges, as
indicated in Fig. 3.6.). Thus the controls may
be odjusted until one channel of the instrument
responds primarily to voriations in diometer and
the other to voriations in wall thickness. Small
defects on the inner wall of the tubing cause
impedance excursions thot are nearly in phase
with those from wall thickness variations, and
small defects on the outer surface cause excursions
that are neorly in phase with those from diameter
variations. Therefere, small inside defects will
appear on the some channel with the woll thickness
variations, and small outside defects will appear

226

on the same channel| with the diameter variations.
Gross defects cause very large impedance ex-
cursions which ore ot some angle between the
effects of diameter and woll thickness; thus the
large resultant signal is present on both channels.

Iflustrotions of the instrument's obility fo dis-
tinguish between diameter and wall thickness
variations are given in Figs. 3.6.3 and 3.6.4.

UMCLASEIFIED
GRAKL=LR~DWG {7798

 

 

 

 

 

 

  

AYERAGE waLlL
THICKNESS tin)

 

 

 

 

 

 

 

RELACTIVE

COMPONENT COMPONENT

REIISTIVE

 

QUTSIDE
AMETER Linl

 

LEMNGTH (o]}

Fig. 3.6,3. Comparison of Impedegraph Signal Traces
with Actvol Dimensionsl VYariations of 0.229-in.-0D,
0.025-in.-Wall Inconel Tubing Specimen 68-1B,

UNCLASSIFIED
QRML—-LA—0W3 1T 746

 

 

 

  
  

 

 

 

 

 

 

 

s
o
§: e VIDIGAGE MEASUREMENT
L 00250 A
§§ 0.0280 :
u¥ 00270
&= ;
- 4
wa l‘ !
— Z TR
CH LT el
2% } e
m <. = —
mg T F o Fe
= s
uwh =
2y
Bg
nE
s
~E0 L
=
Zozun : — ; . :
g”‘ozsoo MICROMETER MEASUREMENT
au ' r e B T e e - —— T
'5% o.ezgoi—‘\cif\.\,/ficjflbfi\:/.}flu \
= o
= 0.2260
o 0 5 0 15 20 25 30 35
LEMGTH Lin?

Fig. 3.6.4. Comparison of Impedograph Signal Troces
with Actual Dimensional VYariotions of 0.229-in.-0D,
0.025-in.-Wall lncone! Tubing Specimen 68-2A.
[ 3]

The signal traces are recordings of the reactive
ond resistive component signols from the Im-
pedograph for 35-in. lengths of 0.229-in.-0D,
0.025-in.-wall Incone! tubing which have been
stretched to occentuate their dimensional voriation.
Above the traces are plots of the overage wall
thickness os measured with the Branson Yidigage
megsurement).
Below the traces are plots of the average outside
diameter as measured with micrometers. |t is
readily seen that close correlation exists between
the instrument signals and the other measurements.

These traces were made at an operating frequency
of 189 ke, for which the f/fc ratio is opproximately
15 for the 0.229-in.-0OD, 0.025in.-wall lnconel
tubing. This frequency and well thickness (11%
of OD) have an operating point on the graph of
Fig. 3.6.2 such thot the resistive component is

{ultrasonic-rescnonce  thickness

ot a maximum as o function of wall thickness.
At this point small changes in wall thickness
cause almost no change in the resistive com-
ponent, but do cause maximum chonge in the
Theotetically, changes in
diameter should affect both channels. However,
the traces of Figs. 3.6.3 and 3.6.4 ¢leorly indicate
that the diameter signals are predominate in the
resistive component and have a negligible effect
Unfortunately, the
variations in both diameter and wall thickness
characteristically follow each other at about the
same period, ond therefore close analysis is
very difficult. Further investigation of this effect
is planned with the use of mochined tubes in
which changes in diameter and wall thickness
moy be introaduced and controlled seporately.

reactive companent.

on the reactive component.

Cladding Thickness Measurements
J. W. Allen R. A. Nonce

The feasibility of determining the thickness of
the cladding on Mark X MTR fuel plates with ¢n
eddy-current probe coil was investigoted. Although
this study hos no immediate application to ANF
work, many of the problems involved are very
similor to problems associated with inspections
of contrel rods ond other clad or shrouded os-
semblies.

The Mark X MTR fuel plates ore composed of
g 0.022-in. nominal thickness core of 52 wt % Al-
48 wt % U alloy cled with o 0.020-in. nominal
thickness of type 8061 aluminum alloy. The defor-
mation choracteristics of the core and the cladding

PERICD ENDING MARCH 21, 1957

alloy are not the same; hence, during the rolling of
the plate to a wniform thickness, the deformation of
the core is frequently nonuniform. The result
is a thick section at eoch end of the core, the
longitudinal cross section of which resembles a
Ydog bone.”” Since the plote is of uniform thick-
ness, the cladding over the thick sections at the
extremities of the core is of course much thinner
than the 0.020-in. nomingl thickness. The pos-
sibility of a 0.003- to 0.005-in. depth of corrosion
attack necessitates the location and rejection of
fuel plates which have excessive ‘‘dog boning.”

To facilitate the investigation of this problem
and to determine the best operating parameters
for the eddy-current test, the prototype instrument
described in the block diagram of Fig. 3.6.5 was
constructed. Since the original purpose of this
instrument was to determine the best coil design
ond the optimum operating frequency, it is more
complex and contains more controls than would
be required on a production inspection instrument.

The instrument consists of o voriable-frequency
oscillator which supplies alternating current to
the probe coil through an isolating network and
to a bolancing network. The isclating network
prevents the presence of the probe coil from
affecting the voltage supplied to the balancing
circuitry, and supplies the probe coil with current
of constant amplitude. Since the probe coil current
is constant in amplitude, the voltage at the
terminals of the probe is directly related to iis
impedoace, which in turn is influenced by the
metal specimen in proximity with it. The baloncing
network is adjusted monually until the voltage
at the netwerk output terminal is equel to the
voltage across the coil, ot the initial test con-
dition, in both amplitude and phase. Any chonge
of the test condition chonges the voltage ot the
coil in phuse or amplitude or both. The difference
in the voltage of the ccil and of the belancing
network is amplified, and hence any change in
coil voltage causes o meter deflection.

After some experimenting, the probe design
shown in Fig. 3.6.5 was established. The coil
consists of approximately 540 turns of No. 40
enameled copper wire in a small, cup-shaped
Ferrocube core (Ferrocube Corp. Type 3C Ferro-
cube).  The coil, ofter conncction with the in-
strument coble, is potted with en epoxy resin into
the end of a short length of micarta tubing. This
coil ossembly is contcined in o heavy brass

227
ANP PRCJECT PROGRESS REPORT

  
   
 

 

 

UNCLASSIFIED
ORNL-LR-DWG {8095

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

AMPLITUDE PHASE SENSITIVITY
FREQUENCY fi —/fi 7{
> VARIABLE~ / . £
~ | FREQUENCY | e Bl SRR R
OSCILLATOR | | % & ; |
{400~y QUTPUT) ] ; / / |
; !
| BALANCE A _(A-B) )
| I.
b B |
T4 | |
COIL SPRING QL ‘ '
coL™~ T ‘
/ LOADING “J~| 1soLATING
1'\
Ea PROBE COIL - ’ | NETWORK ]
_ 540 TURNS NO 40
Bt 5 ENAMELED COPPER. | OUTPUT METER |
| £=3.25mh | Y - ‘
., fags =179 ke |
i OIAT 179ke1 =256 | i
\Z b s ool mmge v e s be ve e e e ]
¥
X PROBE COIL CONTAINED IN PROAE COIL
FERROCUBE CUP el
7 e e 7 R -
?,,: 7 )
L S
3 i SPECIMEN
13%in % 7

 

 

 

 

 

AV
/ £ ,/ff 2
i 4 Probe Coil and Hoider
L- 21"16‘“' _—l
Fig. 3.6.5.

terminotions.

helder which spring-loads fhe probe against the
The use of this mechanism
eliminates spurious signols coused by the wvarying
coupling associoted with a hand-held probe.

metal specimen.

Three considerations were of primary importance
in selecting an optimum test frequency. First,
the frequency was to be sufficiently high to
sensitively detect the difference between the
type 6061 aluminum alloy cladding {conductivity =
45% IACS?) and the uranivm-aluminum alloy core
(conductivity = 22.8% |ACS). Second, it was to
be sufficiently low to penetrate the nominal 0.020-

 

2 -
IACS refers to internoticnal annealed-copper stondard,

228

Probe Instrumen?

Variable-Frequency Eddy-Current Instrument and Probe-Coil Design for Cladding Thickness De-

in.-thick cladding cnd to detect smoll changes
in its thickness. Also, it was to be sufficiently
high that the 0,022-in.-thick core underneath the
0.020-in.-thick cladding would appear infinitely
thick (that is, no eddy-current penetration through
the core). This is necessory to prevent the clad
thickness on the ‘'back side' of the plate from
influencing the thickness determinaticns made
on the side under surveillonce. The optimum
frequency determined experimentally to satisfy
these conditions was 20 ke.

Correlation of the meter deflection with the
cladding thickness was experimentally determined
by first adjusting the instrument to read nearly

L
i

full scale with the probe coil placed on a specimen
of bare uwronium-aluminum core, and to reod zero
with the probe coil placed on an apparent *infinite"’
thickness of the 6061 oluminum alloy. Shims of
the clodding alloy, which varied from 0.003 to
0.020 in. in thickness, were then used to obtain
the information for the colibrotion curve shown
i Fig. 3.6.6.

Three Mark X fuel plates were examined, and
only one showed sufficient variation in cladding
thickness to warrant a more detailed inspection.
This plote was scanned on both sides, with gach
pass of the probe coil overlapping the previcus
pass 1o ensure complate coverage of the surfoce.
The greater portion of the cladding was indicated
to be from 0.017 to 0.020 in. thick. AIl deviations
from this range were marked for loter destructive
examination.

Four of the mesi significant areas noted during
this
to a metallographic examination for correlation ond
evaluation of the eddy-current measurements. The

investigation were sectioned and subjected

locations of these sections are shown in Fig.
3.6.7. Sections 3 and 4 were taken from one end
of the plate where the cladding was very thin os

-28% in_

PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED
ORNL =-_R-DwWG 18096

  
  
 
  
 

BD —"i—‘ e -I i

1
| CLADDING MATERIAL: BOGI Al ALLOY
CORE MATERIAL® 48, U BY WEIGHT,
| ALLOYED WITH &l
CONDUCTIVITY OF CLAD: 45%, 1ACS
| CONDUCTIVITY OF CORE: 22 8% $ACS
[ FREQUENCY=20 ke — |

0
o

  

 

EDOY -~ CURRENT SIGNAL (% FULL SCALE)
B
>

 

 

20
o
o 4 8 2 g 20 24
THICKMNESS CF CLADDING {mits]
Fig. 3.6.6, Eddy-Current Signol-Coltbration Curve for

Cladding Thickness Measurements of Mark X MTR Fuel

Plates,

UNCLASSIFIED
ORNL—LR~DWG 15097

 

 

 

21-"?16 "

 

 

PLATE THICKNESS = 0062 in—"

Locahon of Seclions Token for Metallographic Exommation {Mark X MTR Fuel Plate)

MO it oy~ =
MAXIMUM CLADDING THICKMESS BY EDDY —
CURREMT IMSPECTION = Q.5211n

MAKIMUM CLADDING THICKMESS BY MCTALLO-
GRAPHIC EXAMMATION = O 020 in.

M2 | - e

pAIMIMUM CLADDING THICKNESS BY EDDY -
CURREMT INSPECTION = G.0O10 in.

MIMIMLR CLADDING THICKNESS BY METALLG-
GRAPHIC EXAMINATION = G009 in.

Fig. 3.56.7.
Meosurements Obtained by Metallegrophic Examinotion.

NO 3 fi- Sy, =

NO 4 iR e

FROBE COiL WiIDTH=04381n

_ ~Tu -
MINIMUM CLADDING THICKNESS 8Y EDDY —
CURRENT INSPECTION = Q.003 1n

MINIMURM CLADDING THICKMESS BY METALLD-
GRAPHIC EXAMINATION = 0003 in.

¢ —_—

b

I5.()!32 i,

MIMIMUM CLADDING THICKNESS B8Y EDDY —
CURRENT INSRECTION = 0003 in

MINIMUN CLADDING THICKNESS BY METALLO-
GRAPHIC EXAMINATION = 0002

Comparison of Eddy-Currant Thickness Measuwements of Mark X MTR Fuel Plates with Thickness

229
ANP PROJECT PROGRESS REPORT

a result of the ‘'dog bone'' effect previously
mentioned.  Section | was taken through the
the thickest portion of cladding, while section 2
was tcken through the thinnest portion of ¢ladding,
excluding the very thin clodding over the ends of
the core.

The results of the metcllographic measurements
are summarized in Table 3.6.1 and shown in Fig.
3.6.7. All megsurements shown were mode on
one side of the plate only. The meosurements
for section | were made on the side of the plate
with the thickest cladding, while the remoining
sections were meosured on the side with the
thinnest cladding. It should be noted that in aoll
cases where the metallograghic measurement differs
from the eddy-current measuremens, the difference
is 0,001 in. below the value obtained by the eddy-
current technique. It is therefore concluded that
upon choosing the optimum operating porameters,
the cladding thickness on Mark X MTR fuel plates
can be determined to within £0.001 in.with the
use of an eddy-current probe coil.

Table Lé.1. Comparlson of Mark X MTR Fuel Plate
Cladding Thicknessas Measured by Eddy-
Current Inspectlon Technlques and by
Metallographie Exomination

 

Clodding Thickness {in.)

 

 

Spacimen FPortion of
No Cladding  Eddy-Current Metallogrophic
| Measured Result Result
1 Thickest 0.021 ¢.020
2 Thinnest 0.010 0.009
3 Thinnest 0,003 0.003
4 Thinnest 0.003 0.002

There is no comprehensive theory for the ap-
plication of eddy-current probe coils on metal
surfaces such as the theory which exists for
the use of coils encircling rods or tubes. This
lack of theoretical background requires that eoch
problem be considered separately ond that general
statements regarding the capabilities of the probe-
coil technique be corefully qualified. The ex-
perience gained in the measurement of the cladding

230

thickness of the Mark X fuel plates indicates
that if the two metals, forming a system, have a
ratioc of conductivities in the order of 2:1 or
the thickness of the accessible layer
can be accurately measured. As the conductivities
of the two materials approach the same value,
this system of measurement becomes useless.
Continved experience with the eddy-current
probe-coil technique should resuit in the satis-
factory measurement of cladding thickness of
most alloys on o cermet core, of any of the oxi-
dation-resistant alloys over malybdenum or nichium
cores, of an oustenitic stainless steel over o
nickel-molybdenum alloy core, or of a nickel-
molykdenum alloy over an austenitic stainless
sieel core. The obility of this method to measure

greater,

Incenel on austenetic stainless steel or austenetic
stainless steel on Inconel is currently considered
to be marginal,

The eddy-current probe-coil method of meosuring
independent of the existence or
condition of the bond between the two layers,
and it is best suvited to the measurement of thin
layers. This method cppears to be well adopted
to measuring clodding thicknesses on solid-fuel-
element ossemblies or on moderater assemblies.
It is wiso well adapted to measuring the thickness

thickness is

of the outer layer of duplex tubing, such as might
be used in future heat exchangers and radigtors,
provided the conductivities of the two layers are
sufficiently different. Both the total wall thick-
ness of the duplex tubing and condition of the
bond bketween the two loyers can be evaluated
by vlirasonic techniques.

RADIOGRAPHIC INSPECTION TECHNIQUES
R. B. Oliver W. J. Mason

The voltage ond current settings required for
rodiographic inspection of various thicknesses of
Inconel, stainless steel, nickel-molybdenum, and
similar alloys are not available, and therefore o
program of compilotion of such doto hos been
initiated. Test weldments in ]'é-in. Inconel sheet
are being radiographed through 8-in. of beryllium
to explore technigues for the inspection of the
closure welds of the ART fuel-arnuius core shells.
Beryllium, an element of low atomic number,
causes excessive scatter of the incident x-ray.
The scotter reduces image contrast and shorpness
so excessively thet the rodiograph is not capoble

of revealing any but gross defects. Higher x-ray
energies are being explored, but the reduction of
image contrast with increasing energy appears
to nearly offset any reduction of scattered photons
that might be expected.

EXAMINATION OF LEAK IK SUMP TANK
OF PRATT & WHITHNEY HIGH-TEMPERATURE
CRITICAL ASSEMBLY

R. Heestand

Shortly after the Pratt & Whitney high-temperature
critical assembly was looded with barren fuel-
carrier salt, o leak wos detected in the sump tank.
The insulation was removed and the source of
the jeckoge was found to be severe cracks in the
&-in.-dia cap which had been welded to the botfom
of the tank as o catch basin. Since such cracking
is wunusval for lnconel forged fittings, it was
surmised that the cap hod been machined from
bar stock. A ceall to Prott & Whitney confirmed

 

Figl 3.6-3-
Prott & Whitney High-Toemperature Critical Assembly.

g

PERIOD ENDING MARCH 31, 1957

that forged pipe caps were not immediately avail-
able and that o cap machined from bar stock had
been used. The cop was removed ond ¢ forged
cop was welded to the tank.

Examination of Caolrods attoched to the tank
revealed thot they were glazed with salt-vaper
deposits, and therefore all bottom and side Calrods
the tank was cleaned externally,
and new Calrods were installed.

Visual examination of the faulty cap showed
nine separate cracks which penetrated through
the cop. Examples of the cracks are shown in
Figs. 3.6.8 and 3.6.9, which cre outer and inner
wall views,

were removed,

respectively, of the same crocks.
As moy be seen the cracks run to the weld and
stop, and thus it is apparent thot they were in the
cap before it was installed on the tank. X-ray
exomination of the cap-to-tank weld before service
did not show cracks in the weld, and the x-ray

 

UNCLASSIFIED
¥-21520

o T o e
- A-su
-.:l-ll

o — -‘h-.
;

Cracks on the Outer Sutfoce of the Faoulty Bor-Stock Cap Removed from the Sump Tonk of the

231
ANP PROJECT PROGRESS REPORT

 

FIg- 3-6-9.
Prott & Whitney High-Temperature Critical Assembly.

examination wos not of sufficient sensitivity to
show cracks in the cap.

MATERIAL INSPECTION

The progrom of inspection of materials of con-
struction was continued, with intended use as
the criterion for acceptability. A summary of the
materials inspected is presented in Table 3.6.2,
which gives the vorious types and quantities of
material inspected, the methods of inspection,
and the rejection rotes. All rejected matericl
has been either downgroded for noncritical ap-
plications or used in experimental development
projects in order to minimize losses.

232

"UNCL ASSIFIED
Y-21521

 

Cracks on the Inner Surfoce of the Faulty Bor-Stock Cap Removed from the Sump Tank of the

Tvbing
R. W. McClung

The results of inspections of smoll-diameter
tubing illustrate the high rejection rate that results
when long defect-free pieces are requested. For
example, it wos necessory to get B8-in. pieces
from 120-in. lengths of 0.625-in.-CD, 0.065-in.-wall
Inconel tubing, listed in Table 3.6.2, and hence
a single rejectable defect located more than 32-in.
from either end caused rejection of the entire
length. Many of the discontinuities located on
outer surfaces by eddy-current or immersed-ultra-
sound inspection were grinding marks that resulted
Table 3.6.2. Moteriais inspectlon Symmary

PERICD ENDING MARCH 31, 1957

 

 

Re|ectian
ltem Dascription Quantity Method of Inspeclion Rare Remarks
- (%)

Tubing inconel: randam sizes 419 b Visuol, penetrant 63 Mumerous scrotches, nicks, and gouges
were found thot were coused by
horrdling or shipping; some cuter surs
fece imperfections exceeded 5% of the
woll thickness

CA-Q00A Inconel; 0,187-in, 2432 B Visuol, penetrant, eddy 12 This matericl was primarily for rodiator
0D, 0.0251n. well, and current, ultrasanic manufaciuvre, ond defects greater rhan
miscellaneaus sizes 00005 1n, were cause for rejection

CX-9004 Inconel; 0.229-in. 2600 11 Visual, peneironi, eddy Remaining pertion of a lot of abov
0D, 0.025 1n, wall current, uitrasonic 10,000 11, over-oll rajectron rate 15.6%

Iheanel; 0,25%in, QD, 1000 i1 Ultrasenic, eddy current Fifty-six short defacts found; rejection
004510, woll rote will depend on minimum vsable

length ond defect distribytion; the
shart defects were, in general, less
than 1 an. in length

lacanal; 0.625-in. QD 38390 i Ulirosonic, eddy currem Rejection of 1910 b bosed on minimum
0.085in. wall usable length of 88 in,; if shorter

lengths con be used, the rejections
will be reduced substantially, since
the defects wera, in general, small ond
widely seottered; repolishing will olsa
prabably reduce the rejectians

Hastalloy B; 0.25in. OO, 250 % Ulirosonic, eddy current Seventy-five short defecis found
8.017+in. wall

Hastelloy B; 0,.5in. QD, 250 #r Ultresonie, eddy current 40 Thurty-six short defocts found; 100 f
0.035in. wall rejected

N Pipa Inconel; random sizes 468 It Visual, penetrant, eddy 1 Where possible, defective conditions
current, ond radiagraphic were removed by grinding; 8 It rejected
where necessary becouse of trocklike conditions or de-
fects in excess of permissible
tolerance

Incgnel; 3in. sched 40 258 i Ulteasonie 14 Thirty-six teet rajected because of

cracks

Hastelloy E; %-in. IPS W06 Yisucl, peaeirant 100 Material hod been ceaterless ground, but
sched 40 defects ramained; marerial witl be

usad, as 1s, for rests

Plate and Incanel; random thick- 463 117 Visual, penetrant, ond ultra. &

shee nessos sonic where necessary

CX-900A Lnconal; random 419 1! Visuvol, penetrant, and ulire- i) Sections with defects were ground, hut
thicknesses sonic where necessary if defects exceeded tolerance, 1the en-

tire plate was rejected

Raod Inconel, random diameiers 30 h Visual, penetiant, and vhra- 0 Filty-eight fect rejected becouse of

sonic where necassary seoms and draw marks thot appear te be
inherant in the cold-drown malerial

CX-900A Incanel (billets); 14 f+  Wisual, penetrant, wltra- 0 Defects indicated by ulirasonic inspec-
4 pieces, 17 in. in sonic tion were removoble by mechining
diameier, each 42 in, long

CX.900A Inconel; randam i Yisual, penstrant, ultrg- 0

v diometers sonic
Firtings inconel ells, roducers, etc 57 pieces Visuol, penatrani 0
’ Shafts far  Inconel § pleces Yisual, penetront, ultre- o
» pumps sonie

 

233
ANP PROJECT PROGRESS REPORT

from the explorotion of visual or penetrant in-
dications. Either the defects had not been com-
pletely removed, the grinding sufficiently
disturbed the surfece to produce o further defectlike
indication, This botch of tubing was inspected
to a stondard such that any discontinuity greoter
than 5% of the wall thickness in depth would be
rejectable.  The tubing will be repolished ond
and reinspected, and the rejection rate moy be
reduced considerably. However, net oll the dis-
continuities were on the outer surfoce, as may
be seen in Fig. 3.6.10 which is a photomicrograph
of an inner surfoce gouge opproximately 0.017 in,
in depth, 0.033 in. in width, and 0.375 in. in
length.

The special lots of Hastelloy B tubing listed
in Table 3.6.2 were fobricated by using ¢ balonced
reduction between diameter

or

and wall thickness
ond were drawn with o floofing plug. Despite the
lorge number of discrete defects and the high
rejection, the tubing was better than many lots
of Hastelloy fubing inspected previously. The
most severe cracks found are shewn in Fig. 3.6.11,

   

which is a metallographic section tcken from the
0.25:in.-0D, 0.017-in.-wall subing. The crocks
appear to be up to cbout 0.008 in. deep. The
0.006-in.-deep single crack shown in Fig. 3.6.12
is representative of the mest common defect found.
The very rough inner surfoce, intergranular attack,
ond fine fissures throughout the entire batch of
the tubing, as shown in Fig. 3.6.13, are of course
unsatisfactory,

Plote
R. W, McClung

The requirements for the ultrosonic inspection
of Inconel plate in thicknesses up to 1 in. have
increased, and therefore special steps have heen
taken to increase the inspection rate without
sacrificing inspection quality. The -.».n::.‘rping3
which was prevolent with the thinner sections
of plate wos partially offset by clomping the plate

to o structural frame. With the warpoge reduced,

 

3R W. McClung, ANP Quar. Prog. Rep. Dec. 31,
1956, ORNL-2221, p 269.

 

 

URCLASSIFIED
Y-21279
n F_’ o
A 1 ‘ ¢ o
= — Lt —
Egd ko
f-.-_L = é - » " E
' r
' ' 0.02
Yo, a
: 4. = i s -
k i i " » 4 & & . & = » -
. X » . = 0.03
% - 5 ' . —
. E ; L. E '
l..' e et il = L 5 v = i
‘_' * ) : T A ' P F
3 I\ : : & % i - ik A i i
& ; ‘ O 5 i " - : 5
. g - . = %
= S * v i £ = — O —
. 5 ‘ E |2
! ; K . .
Fig. 3.6.10. Gouge Found on the Inner 3urfoce of 0.625-in.-0D, 0.0465-in.-Wall Inconel Tubing. 100X, Re-
duced 4%,

234
 

 

PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED
¥-21289

I I
mopes B |

LT B

g
o

11

 

 

o
I
o

 

Bl

T
IS?!

1
i

[ Bl

 

 

Flg. 3.6.11. Muitiple Crocks Found in 0.25-in.-0D, 0.017-In.-Wall Hastelloy B Tubing. 150X, Reduced 4%,

UMCLASSIFIED
Y-21281

B

T
NCPES

ELITTE[TTIRITITRD

T
I.":?X

I
1

[ Bl

 

 

Fig. 3.6.12. Typicel Crack on Outer Surfoce of o Hestellaoy B Tube. 150X. Reduced 4%,

235
ANP PROJECT PRCGRESS REPORT

 

¥ ' ® . - " 4 *
o Q', " -
. . ° ¢
* 1
" fl-"
- - - s
F "
- t‘- o ¥ - & T s ' -
* L
a - a
8 -~ r -
Fig. 3.6.13.

Tubing. 530X. Reduced 3%,

less time was required to realign the incident
beam. Another aid in increesing the inspection
rate is the use of the B-scan, which displays the
signal received as on instantaneous cross section
through the plate being inspected. This permits
an increase in linear speed of the transducer from
approximately 10 to 25 fpm.

In general, it appears that for plate of overage
quality, inspected from either face, the inspection
rate will be about 10 2 per shift., This of course
allows for all the downtime necessary for sefup
and may increase or decrease with plate quality;
also, an increase in the inspection rote might be
expected with larger areas.

The rate of rejection increased slightly during
the quorter, ond three prime criteric for rejection
were established: (1) the detection of a dis-
continuity which produces @ signal comparable
to or greater than that from a stondard 4{‘_4-{:1.
flot-bottem hole; (2) the presence of smaller dis-
continuities with their centers nearer than 1 ia.
to each other; and (3) excessive ulfrasonic atten-
uation, which may be due fo porosiiy or precipitates
within the plate.

236

   
    

UMCLASSIFIED
¥-21283
tn
ui
: = I =~
T NG B =
T ‘f.:t’- % . #: -
et e
i '?fl’;fi po 002
'. ™ w B
..fl u ® A : - "
. Srcm g W .
0G3
- & s »
.- - - "
g a
. : - :
! «. B 004
¥ m M T
-
Py 5
3
- A8
' e
-
& -
s / L Hos
k "
. »
O
_0_
a* )
-
- - !
) - «+ " e 008
- . '

 

 

Typical Intergranclar Attack, Fissures, and Rough lnner Surface of As-Received Hastelloy B

Beryllium Hot Pressings

R. W. McClung R. B. Oliver J. W. Allen

One of the six, large, beryllium hot-pressings
fobricated by The Brush Beryllium Co., Clevelund,
Ohie, was vultrasonically inspected at ORNL,
Previcusly, three of these items had been in-
spected at the producer’s plant. However, it was
felt that ¢ more methodical, complete ultrasonic
examination under laboratory coaditions by several
different techniques (some of which could not be
used at the Brush facility} was needed as o check
on techniques which were previously used tn the
field.

The instrument used for the inspection at ORNL
was a Curtiss-Wright 424A Immerscope with o
%—in.-diu Li 50, high-resolution trensducer. An
attenuation curve was plotted which demonstrates
the comparotive signal strength os received from
a constant-gsize discontinuity when seen through
varying thicknesses of beryllium, Fig. 3.6.14,
The instrument was calibrated for all the in-
spections by using a ]zé—in. flot-bottom hole and
applying its presented signal to the appropriate
UNCLASSIFIED
ORNL-LR~D™G 18654

 

 

 

 

 

130 | |
5Me; Y in. Li; SOy TRANSDUCER:
| MODEL 4244 (MMERSCOPE

¥ ?s |
I}
oy
3
=
i
=
= 50
wd
=1
=
@
i

25 |-

ol i i
0 5 10 5 20

BERYLLIUM {in._}

Fig. 3.6.14, Attenuation Curve for Ultrasonic Inspec-
tion of Beryllivm.

thickness on the attenuation curve. Thus the
relative size of any defect could be approximoted
by comporing its signal with that of the standard
ond by using the atienvotion curve. The general
shape of a hot-pressed beryllium piece is shown
in Fig. 3.6.15; however, the contours aond di-
mensions varied from piece to piece.

The section denoted A was inspected completely
by three different pulse-echo ultrasonic techniques.
The first was a simple immersed method in which
the hot pressing wos completely immersed in o
tank of water and the surface was scanned in o
linear fashion with o E*é-in. index between suc-
cessive scans. Three small discontinuities were
located which were estimated to be less than the
rejectable size. For the second inspection the
column-coupled, gimbal-mounted mechanism? illus-
trated in Fig. 3.6.16 wos used. Section A wos
sconned in o series of overlapping circulor scans;
however, the part wos still completely immersed
and this technique merely gave comparative in-
formation. The inspection results were identical
with those obtained by the conventional immersed
technique. For the third inspection the gimbel
mounting, the circuler sean, and also the column-

coupled ‘‘squirter’’ principle were used. More

 

‘R, & Oliver, J, W. Allen, and P. J. Long, Ultrasonic
Inspection of Hot Pressed Beryllium Moderator Blanks,
QRNL CF-56-10-34 (Oct. 3, 1958).

PERIOD ENDING MARCHM 31,1957

f
s & 4 0O

24 5;'9 in.

WNCLASSIFIED
ORRL-LA—DWG n22G

  
   

F-in-DIA PILET

 

[18 in.

NS \\\\ B

L 44 ' in,

SECTION M-

Fig. 3.6.15. Cross Section of a Typical Hot-Pressed
Berylllum Piece.

spurious signals were noted in the third in-
spection, but the results were equivolent to those
obtained by the other two techniques.

The sloping section, B, wos inspected with the
squirter technique, the transducer being aligned
so that the refrocted uliresound beom was per-
pendiculor to the boitom surface. in this portion
ot the inspection, one discontinuity that was larger
than the standord hole was located, but it was in
o region that was to be removed by subsequent
mochining and thus was not cause for rejection.

The hot pressing was olso inspected completely
by measuring the voriation in ultrasonic at-
tenuation throughout its volume.
sistency of back reflections obtained, the pressing
seemed to be quite homogenecus. Since no
rejectable  defects were locoted which would
not be removed by subsequent machining, the
piece was considered to be acceptable.

From the con-

237
ANP PROJECT PROGRESS REPORT

UMCLASSIFIED
FHOTCO 18490

 

Fig. 3.6.16. Conical 5ection of a Beryllium Hot Pressing Being Ultrasonically Inspected with the Use of the
Column-Couplad, Gimbaoi-Mounted Mechanism.

238
In comparisen with the results obtaoined with
the other techniques, it
appears that if careful contrel is exercised the
obtained with
technique are sufficiently

vltrasonic inspection

inspection results the column-
coupled  “‘squirter”
accurate to warrant such inspection of the re-

maining beryllium hot pressings.
Welds
A. Taboada R. L. Heestand

The results of inert-gos shielded-orc weld in-
during  the months are
summarized in Table 3.6.3.

spections past six
The criteria for the
inspection of each weldment are established at
the time of design review. Each weldment is
classified os C, CN, or 5. A C classification is
used for all critical welds that must have optimum
soundness, for example, welds sealing between
corrosive environments, welds that will be sub-
jected to thermal grodients, ond welds that must
have high strength. These welds are made

according to applicable procedure specifications,

PERIOD ENDING MARCH 31, 1957

and they are inspected by visuval, dye-penetrant,
A CN weld requires high
integrity but cannot be x-rayed. Visual inspection
and dye-penetront methods are used in the in-
spection of the CN-type weld. Welds which cre
out-of-specification because of design limitations,

and x-ray techniques.

but which are critical, also fall inte this category.
The S designation is for structural welds, but it
is also used for out-of-specification welds that
are not critical. Welds in this category are in-
spected by dye-penetront and visual metheds, Dye
penetront reveals crocks, but it does not indicote
pin holes or porosities,

In addition to the welds listed in Table 3.6.3,
115 C-type welds on piping manifolds for the
ETL were inspected, with 21% rejected for det-
rimental defects. There were also 60 C-type welds
of Hastelloy B inspected. Thirteen of these
welds were found to be rejectoble for various
types of defects. Another 185 welds fabricated
on thermocouple leads for critical opplication
were 864% rejected.

Toble 3.6.3. Weld lnspection Summory

 

 

 

C Type Weld CHN- and 5 Type Welds
Period Total Number Percentage Tatal Number FPercentage
Inspected  Rejected Rejected Inspacted  Rejected Rejected
Sept. 10 to Dec. 10, 1956 1052 141 13 2048 59 2
Deec. 10, 1956, to March 10, 1957 922 202 22 3103 45 1.5
Totol for the two pericds 1974 343 17 5151 104 2

 

239
Part 4

RADIATION DAMAGE

G. W. Keilholtz

FUEL RECOVERY AND REPROCESSING

R. B. Lindaver
4.1. RADIATION DAMAGE
G. W. Keilhaltz

EXAMINATIONS OF COMPONENTS AND
MATERIALS TESTED IN THE MTR

A, E. Richt
C. Ellis E. N, Romsey
E. J. Manthes E. D, Sims
W. B, Parsley R. M. Wallace
in-Pile Loops

Metellogrophic specimens were cut from in-pile
foop Mo. 4 and mounted. All the specimens were
examined in the as-polished condition, ond some
None of
the etched specimens exomined up to this time
have shown attack greater than 1.5 mils in depth,
and no signs of mass tronsfer were found on the
specimens, Wall thickness measurements witl be
taken on all the specimens from the nose coil end

were examined in the etched condition,

on some samples from straight sections of fuel
lines. The operating conditions for this loop were
described previously.'

All the sections of the fuel-circulating pump,
except
photographed with the remote stereomicroscope.
The condition of the rear holf of the pump sump is
shown in Fig. 4.1.1. As may be seen, no fuel
remained in the sump. A colored amorphous de-
posit similor to thet found in in-pile loop Ne, 3

the impeller and impeller housing, were

 

]A, E. Richt et al., ANP Quar. Prog, Rep. Dec. 31,
7956, ORNL-2221, p 290.

  
   

UNCLASSIFIED
AMG 1591

 

Fig. 4.1.%.

Laap Mo. 4. I{?X.

Rear Section of Pump Sump of In-Pile

wos found on the slinger ossembly and on the
forward bellows, as shown in Figs. 4.1.2 ond
4.1.3. The rotating copper block and its housing
were coated with a similar depesit, os shown in
Figs. 4.1.4 and 4.1.5,

The impeller housing from the pump of loop No. 4,
along with the impeller housings from the pumps
of loops Nas, 5 and 6, will be cut open on the re-
mote lothe at a later date, Disossembly of loop
No, 6 will begin os soon os the disassembly equip-
ment is made ready,

Irradiated Moderator Materials

The Inconel capsuie containing beryliium oxide
slugs, which was recently irrodiated in the MTR,?
was disossembled ond examined. This copsule
was the First of o series of copsules centaining
moderotor materials to be irradioted to investigate
the dimensional stability eand mechonical properties
of the material under simulated reoctor operating

2J. A. Coniin, D. M, Haines, ond C, C. Bolta, ANP
Quar, Prog, Rep. Dec. 31, 1956, ORNL-2221, p %1.

 

— -y
UNCLASSIFIED !
RMG 1594

 

Fig. 4.1.2. Slinger Assembly from Pump of in-Pile
Loop Mo. 4. 2X,

243
ANP PROJECT PROGRESS REPORT

L

* J
-

- &
UNCL ASSIFIED
RMG 15%5

-

 

Fig. 4.1.3. Split Forword Bellows from Pump of In-
Pile Loop No. 4. 2X,

UMCLASSIFIED
RMG 1594

 

Fig. 4.1.4, Rotating Copper Block frem Pump of In-
Pile Loop No. 4. Actual size.

244

UNCULASSIFIED
RMG 1597

 

Fig. 4.1.5. 5plit Copper-Block Hausing from Pump ef
In-Pile Loop Mo, 4. Actuval size.

conditions. The condition of the irradiated test
ossembly after removal of the thermocouples is
shown in Fig. 4.1.6. Unfortunately the high-flux end
of the capsule {the end closest to the reactor during
irradiafion}) was bent slightly while it was being
removed from the bracket surrounding it, but the
BeO slugs were not domaged. Dicmeter measure-
ments were mode on the con before the BeO slugs
were removed, and the results of these measure-
ments ore listed in Toble 4.1.1, The meosurements
were taken ot ];i-in. intervals along the length of
the can, ond two series of readings were token
90-deg aport. The slugs were arbitrarily numbered
1, 2, 3, and 4, with slug 1 being closest to the
reactor and slug 4 being farthest from the reoctor
during irradiatien,

An attempt was made to remove the BeD slugs
with the use of a remotely controlled lathe, Siugs
J and 4 were removed in this manner, but they
were damoged slightly by the lathe taol while the
con was being machined owoy. Figures 4.1.7,
4,1.8, ond 4.1.9 show the condition of the slugs
removed on the lothe as compared with a similar
slug of unirrodiated BeQ. |t was therefore de-
cided to remove the remaining two slugs by using
the remote milling machine. A portion of the cap
on the remaining half of the con was cut off with
an abrasive cut-off wheel to remove the brace
which secured the cop to the can, and the open end
of the coan was trimmed on the cut-off machine,
Successive cuts were then made with o milling
machine, dlong the longitudinei axis of the can,
until the BeO slug surface was exposed ond the
The slugs were then easily
removed from the open con. Slugs | and 2, which

can sprang open.

   

UL ATHR
FELTO . TT R

    
  
   
 
 

| BRACKET FOR HOLDING
ARND POSITIONING
 CAPSULE

INCONEL CAPSLUILE
CONTAINING
Be0 5LUGS

Fig. 4.1.6.
After Removal from the MTER.

Inconel Copsule Contoining BeD Slugs

PERIOD ENDING MARCH 31, 1957

were removed by this method, were in much better
condition than slugs 3 and 4, as shown in Figs.
4.1.10 and 41,11, An unirradicted slug photo-
graphed under the same conditions is shown in
Fig, 4.1,12 for comparison with slugs | and 2 as
shown in Figs., 4.1.10 and 4.1.11, A very small

UNCLASSIFIED
RMG 1598

 

Fig. 4.1.7. BeD Slug 4 Removad from Inconel Capsule
After Irradiation Showing Domage by Luothe Used to

Remove lnconel Con. 2,5X.

Table 4.1.1. Diameter Measurements of Irradioted Inconel Copsule Contolning BeD Slugs Taken at Ié-in.

Intervals Storting at End Foarthest From Reoctor During Irradiation

Diametar before irradiotion: 1.041 to 1.042 in.

 

Dioameter at l;5-i|'1. Intervals

Pesiticn en Capsule

Digameter 90 deg from
Calumn 1 Meosurement

 

lir.) During lrradiation (i)
1.041 End forthest from reactor 1.041
1,042 1.042
1,043 1.042
1.044 1.043
1.044 1.044
1.043 1.044
1.043 1.044
1.042 End closest to reactor 1.044

 

245
= = - i

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
RMG 1599

 

Fig. 4.1.8. BeOQ Slug 3 Removed from Inconel Capsvle
After leodietion Showing Damage by Lothe Used to

Remove Inconel Con. 2.5X.

UNCL ASSIFIED
RMG 1600

 

Fig. 4.1.%. Unirrodioted BeQ Slug for Comporison
with Those Shown in Figs. 4,1.7 ond 4,1,8, 2.,5X,

246

UNCLASSITLED
RMG 1621

 

Fig. 4.1.10. BeO Slug 1 After lrradiation. 2,.5X%,

UNCL ASSIE1ED
RMG 1622

 

Figs 4.1.11, BeQ 5lug 2 After lrradiotion. 2.5%,

 

 
crack was found on the low-flux end of slug 2, but
since this was the only crack found on the slug,
it moy be assumed that the crack was either present
before irradiation or was coused by mechonical

UNCL ASSIFIED
RMG 1623

 

Fig. 4.1.12, Unirradiated Be0Q Slug for Comparison
with 3lugs Shown in Figs. 4.1.10 ond 4.1.11. 2.5X,

Table 4.1.2. Dimensioncl Mecsurements of the Four
BeO Slugs Before ond After Irradiation in the MTR

 

 

 

Slug Dimensions {in.)
Number Before Irrodiation After Irradiation
Length
1 1.000 1.000
2 1.005 1.001
3 1.005 1.000
4 1,005 1.000
Diameter
1 0.9965 0.997
2 0.997 0.997
0.9965
3 0.997 0.997
4 0.997 0.597
0.998

 

PERIOD ENDING MARCH 31, 1957

damage ofter iradiation, The ends of slugs 1 and
2 are shown in Figs, 4.1.13 and 4.1.14, The end
of slug | that was ciosest to the reactor shows a
small crack that was probably coused by mechanicoal
damage raother than radiation damage. No evidence
of BeO powder wos found in the capsule, The
results of the dimensional measurements on all
four slugs cre listed in Toble 4.1.2. No signifi-
cant changes were observed in either the lengths
or diometers of the slugs. Sections of the capsule
will be mounted und examined metcllographically.

UNCL ASSIFIED
RMG 1624

 

Fig. 4.1.13. Ends of BeO Slugs 1 and 2 That Were
Closest to Reoctor During Irradiation. 2,5X,

247
ANP PROJECT PROGRESS REFPQRT !

UNCL ASSIFIED
RMG 1625

 

Fig. 4.1.14. Ends of BeO Slugs 1 and 2 That Were
Farthest frem Reoctor During leradiotion. 2.5X.

CREEP AND 3TRESS-RUPTURE TESTS
OF INCONEL

J. C, Wilson
C, D, Boumann W, W. Dovis
W, E. Brundoge N, E, Hinkle
T. C, Price’

Apparotus for Stress-Rupture Tests of Tubing
in the MTR

Effort has been concentrated recently on com-
pleting apparatus for MTR in-pile tests of the
stress-rupture properties of Inconei tubing, Diffi-

culty was encountered in the ossembly of the

 

3Dn assignment from Pratt & Whitney Aircraft,

248

apporatus in that the pure helium atmosphere be-
came c¢ontaminated by insulating and constructional
materials, and the assembly work during the final
stages was necessarily slow because the size
and complexity of the apporatus limited the number
of people who could work at one time,

The general design of the apporatus was described
previously,? and the nearly assembled specimen
chamber, a sample specimen, and a specimen fur-
nace are shown in Fig. 4.1.15. An end plate and a
water jacket will be welded on in order to complete
the specimen chamber. Water and boral will be
used as shielding in the plug. Each specimen is
a section of 3.8-in, sched-40 Inconel pipe machined
down to o wall thickness of 0.030 to 0.030 in,
over the gage length, and each specimen is pro-
vided with three thermocouples for temperature
A Hfi—in.-OD capillary
will supply gas pressure (to supply the stress) to
each specimen,.

Each of the eight specimen furnaces has three
windings. A rheostat in series with each of

measurement or control.

the 24 seperate furnace windings permits separate
control, Three recorder-controllers {Speedomax
“G" air controilers with pneumatically driven
Variacs) control the temperature in three zones
corresponding to the front end (the end closest to
the reoctor), the middle, and the back end of the
specimen, This system is necessary becouse of
the gamma-ray heoting gradient ia the MTR. The
only alternative would have been the use of 24
controilers or 8 controllers and a system of furnace
shunts thot would be contrelled by recctor power.
The leod $ime on procurement of the exira con-
trollers that would hove been needed was too long,
and it appears now that the present system moy be
the most suitable,

Extensive tests to determine why the Inconel wos
heavily cxidized in out-of-pile stress-rupture tests
in similar apparatus, despite the use of a high-
purity helium atmosphere, showed that some of the
materials of comrstruction were contributing oxygen
or other contominants to the atmosphere. Lavite,
a machinoble ceramic, was eliminated from the
ossembly as the worst offender. The fused quartz
sleeving was found to require special thermal
treatment to rid it of disturbing impurities, Therme-
couple wires with glass insulation were also
suspect, and accardingly all electrical leads and

 

45, C. Wilson et al., ANP Quar. Prog. Rep. Dec. 31,
1956, ORNL-222], p 299,
TEST SPECIMENS (8) Ta—. T
II

 

FERIOD ENDING MARCH 31, 1957

= -
UNCL ASSIEIED
BHOTO 19779
Sy

W5

CABLE TUBING |

Fig. 4.1,15. Partiolly Assembled MTR Stress-Rupture Apperotus,

thermocouples were provided with ceramic insula-
tion. Approximately 50 thermocouples and 35 power
leads (copper} were insulated with a total length
of 12 # of ceramic insulation,

Four racks of power and thermocouple instrumen-
tation and the connecting cables were completed,
and an ovailoble safety circuit was adapted for
use with this apparatus, A rack of helium pressure
control circuits, including 8 pressure recorders to
monitor specimen pressures (ond thus stresses),
was designed, built, and leak tested.

Duplicate specimens are to be tested out-of-pile
at stresses of 2000, 3000, ond 4000 psi ot 1500°F,
The in-pile apparotus will probably be installed
in the MTR during March 1957,

Tube-Burst Tests in the LITR

Two tube-burst stress-rupture tests of Inconel
tubing in a helium atmosphere were completed in

the L_ITR. The thermal and fast > 1 Mev) neutron
Huxes in hole HB-3 where these tests were con-
ducted are both on the order of 6 x 10'% q/cm2.5ec,
Rupture times for the 0.010-in.-wall lnconel tubing
(ART stock) ot 2000 psi and 1500°F were 1050 and
1650 hr ond thus were in good agreement with out-
of-pile data obtained in on argon atmosphers,”
The specimens in corresponding out-of-pile tests
in @ helium atmosphere have passed 1770 hr with-
out rupture, A third in-pile specimen, stressed to
1000 psi ot 1500°F, did not rupture ia 1670 hr,
and the total creep deformation will be measured
upon disassembly of the apparotus,

An unstressed Inconel tube was exposed in the
LITR (HB-3} with nitrogen inside to determine
whether radiction-induced nitriding would occur,
This specimen was also in a temperature gradient

 

33, H. DeVon, ANP Quar, Prog. Rep, Sept, 10, 1955,
ORNL-1947, p 122.

249
ANP PROJECT PROGRESS REPORT

of 1500 to 750°F to oscertoin the effect of tem-
perature, Metalliographic examination of the speci-
men is under way,

Metallogrophic  exomination of an irradioted
apparctus, in which a tubular
Inconel specimen was exposed to NaF-ZrF4-UF4
(63-2512 mole %, fuel 41) on the inside ond sedium
on the outside, was completed. The tube wos
stressed in bending at o maximum stress of 1000 psi
ot 1500°F for 1120 hr so0 that a range of tension
and compressive stresses acted on the metai,
Corrosion on the fuel side was negligible and did
not appear to depend on the sign or magnitude of
the stress; only scottered subsurtace voids to a
depth of 1 to 2 mils were observed. On the sodium
side, attack to a depth of about 3 mils was found.
There were voids at and immediately below the
surface, but deeper intfo the wall there were holes
in the metal that appeared to be filled with a new
constituent, as yet unidentified,

A tube-burst rig is being built for LITR tests of
lnconel exposed under stress to NuF-Zer-UFd
(62.5-12.525 mole %). This type of apparatus has
been successfully tested in the LITR without
stress. Heaters are used with this apparotus which

stress-corrosion

will permit control of the temperature even if a
30% error is made in calculating the fission heat
generation,

EFFECT OF RADIATIOCN ON CORROSION OF
STRUCTURAL MATERIALS BY STATIC
FUSED SALTS

W, E. Browning

R. E. Adams
H. L. Hemphill

J. Lee

Inconel Capsules

Irradiation tests in the MTR of Inconel capsules
filled with NaF-Z¢F -UF, (53.5-40-6.5 mole %,
fuel 44) were continued. Further difficulties of
the type described previously® were encountered
becovse of damage te facilities by the turbulent
cooling woter of the MTR. Some improvement of
the conditien was obtained through the use of
better installation techniques. No design changes
of the capsules or the focilities were made, but it
ts expected that a change to all-stainless-steel
construction of the irradiation facility may be
necessary,

 

GW. E. Browning, R, E, Adams, H. L. Hemphill, ANF
{Quar. Prog. Rep. Dec, 31, 1956, ORNL-2221, p 300,

250

Two irradioted and two contral capsules were
opened and analyzed metcllograghically, and fuel
samples were submitted for chemical analysis.
The operational histaries and the results of metallo-
graphic analyses’ of the capsules are given in
Table 4.1.3. In this series of tests only two
copsules, Nos, 247 and 248, constituted a matched
pair, that is, an irrodiated capsuvle end an identical
heat treated but not irradioted capsule. As may be
seen in Figs. 4.1.16 ond 4.1.17, the attack was
of about the same intensiry in both capsules, The
depth of penetration was so minute, however, in
both coses, that it may be confused with surface
foults and perhops can be considered to be negli-
gible.

Capsule 276 showed (Fig. 4.1.18) corrosive
attack and subsurface voids to a depth of 5 mils.
The area of maximum penetration was near the
mid-point of the capsule, and the distribution of
the atteck tapered slightly toward the ends. This
distribution of the attack wos anticipated on the
basis of the temperature profile of the capsule.
The appearance of the affected areas in this cap-
sule indicates that the corrosion attack is real
ond considerable; however, the contribution by
irradiation control

cannet be ascertained until

data and chemical analyses of the fuel are avail-
able., Copsule 264 (Fig. 4.1.19) will be compared
with its irradiated counterpart in a latter report,

The excellent temperature conirol achieved with
the capsule control system at the MTR is indicated
in Table 4.1.3. Capsule 247 experienced only one
temperature excursion, which was 32°F above the
set point for less than 1 min in an operating pericd
of 676 hr through 16 reactor shutdowns, Capsule
276 was operated for 235 hr with ne excursions
greater than 10°F above the set point,

Hastelloy B Capsules

Hastelloy B copsules 320 and 341 were tested
out-of-pile, Figs. 4.1.20 and 4.1.21, ond two
simifor copsules conteining NaF-KF-LiF-UF , were
inserted in the MTR on Januory 28, 1957, The
design of these capsules is the some as that of
the Inconel copsules.

Hastelloy B corrodes in air at high temperatures
and must be protected during irradiation, Electro-
ptating with chromium over nickel provides the

 

7A. E. Richt and R. M. Wallace, Metallograpbic Exami-
nalion of Static Corrosion Capsules Nos. 248, 264, 320,
341, 247, and 276, ORNL CF.57-2-30 (Fak. 5, 1957),
PERIOD ENDING MARCH 31, 1957

Table 4.1.3. Operating Conditions and Results of In-Pile and Control Corrosion Tests of Inconel and
Hostelloy B Capsules Containing Static Fuel Mixtures

 

 

 

: Time at Number of Metollographic Results
Uranium . :
Caopsule Cartiiit-ol Capsule Full Flux or Excursions Excursions Depth of L stribution
No. Fuel” Material Temperature  Abeve I500°F  Below Melting  Atrack of
thr) Paint of Fuel (mil s} Attack
leradiated Copsules
247 2 mole % UF., Inconei 474 1 min ot 16 0.5 Uniform
(1.5 kw/em®) 1532°F
276 4 mele % UF Inconel 235 None 8 5 Attack concentroted
(3.7 kw/em?) at mid-point of
capsule
Heat-Treated Unirrediated Cantrol Capsules
248 2 mole % UF, Inconel 676 1 min at 16 0,25 Uniferm
1532°F
264 2 mole % UF, Inconsl 218 tNone 28 0.5 Unifarm
320 5.7 male % UF, Hosteiloy 8 621 None 3 0,25 Unifarm
341 Mone** Hasteiloy B 190 Mone To 300°F 0.25 Uni form

13 times

* The fuel mixtures used in all capsules, sxcept Mo. 320, were token from the NaF-ZrF ,-UF ; system. Capsule 320 contoined
11.8 mole % NoF, 44.3 mole % LiF, 38.2 mcle % KF, and 5.7 mole % UF .

“*No solt; helium otmosphere.

 

Fig. 4.1.16. Inner Surface of Irrediated lnconel Capsule 247 After Exposure te ¢ Z!F4-Base Fuel Mixture

Containiag 2 Mole % UF3 for 474 hr at 1500°F in the MTR. {a) Unetched. (b) Etchad with 10% chromic acid
(slectrobytic), 250X, (Secret with caption)

251
ANP PROJECT PROGRESS REFORT

 

. z ) & . 5 5 S .' o ¥ : ; o sk, n.
- ; ; 5, . B e R e R
(a} : - : (-b} : : i o o o ..".r- % i 3 ; _-l.:'. 3 :_ A a b -

Fig. 4.1.17. Inner Surfoce of Unirradioted Inconel Capsule 248 After Exposure to « ZrF4-Buse Fuel Mixture
Containing 2 Mole % UF3 for 676 hr ot 1500°F, {a) Unatched. (&) Etched with 10% chromic acid {elecirelytic).
250X, (Secret with caption)

 

(g) (5) - : e .

Fig. 4.1.18. Inner Surface of Irradicted Inconel Capsule 276 After Exposure to a ZrF,-Base Fuel Mixture
Containing 4 Mole % UF, for 235 br ot 1500°F in the MTR. (2) Unetched. (&) Etched with 10% chromic ocid
(elactrolytic). 250X, (Secrat with caption)

252
PERIOD ENDING MARCH 131, 1957

RMG 1557

 

TN r
- . - & ’
b e it -
' " -
- w— "
v -
L e - ]
. ,‘.1-
- i o LN .d"'"
o L3 a0
- -
= % e s - 3
! -
- = - o
O
- o -
L R -
‘ &

e W= L €

Fig. 4.1.19. Inner Sucface of Unirradiated Inconsl Capsule 264 After Exposure to a ZrF -Base Fuel Mixture
Containing 2 Mele % UFd for 918 hr at 1500°F. (2) Unetched. (5) Etched with 10% chromic acid {elactrolytic).
250X. (Secret with caption)

UNCL ASSIFIED
RMG 1558

  

Tl et bl g o e, T BN e SO
t Za z "f(‘:- L B e g - - s W T
: i < i S 1 '“-"'#".“- ._'-.L;..-. gt
" y 1 .I'l-"-I -‘ 2 .- * e . - - ' o .-" - - - F-.
M P S el e S
e = i 5 5 ¢ L1 ." / " o
I - = T e nd LA - 5‘ 'r.
i il : ~ AN
Ty 2 + .‘ i 9 ' ut - % -',i' z
| = 3 - '\. PR . : Sl
: S
~ % 5 3 dy ¥ '
. £~ p ke
s : 1
= 5 I:-'-. |
o s Il ¢ __ ¥ &
. & o et ol I T i .
N IR ek o T oy 4 B s - "
{{?) ; (b] -, e o ~  f s ot
LY s " 5 3 . -a 5 i

Fig. 4.1.20. Inner Surface of Unirradiated Hastelloy B Capsule After Exposure to NuF'KF-LiF-—UF4 {11.8-38.2-
44.3-5.7 Mole %) for 821 he ot 1500°F. (a) Unetched. (b) Eiched with 10% chromic acid (alactrelytic), 250X,
{Secret with caption)

253
ANP FPROJECT PROGRESS REPORT

 

 

250X, Reduced 13.5%. (Secret with

(g}
Fig. 4.1.21. lIoner Surface of Unirredioted Hestelloy B Copsule Tested in o Helium Atmosphare for 190 hr ot
15009F, (a) Unetched., (5) Etched with 10% chromic acid {slectrolytic).
caption)

basis for formation of on oxide coating which is
protective at all femperatures up to 1500°F.8
Details of the plating of Hastelloy B corrosion
capsules with nickel and chromium by electro-
deposition were reported previously.g

Other methods of apelying nickel and chromium
coatings are being studied in an attempt to improve
the yield of successful coatings. Depositian of
nickel coatings by the 'electroless’’ method
oppedars to be promising, A nickel salt is reduced
by sodium hypophosphite and deposited on the
surface of the metcl to be cocted. The metal
surfece does not enter into the reaction. Nickel
deposited by this method is uniform in thickness
on all surfaces, and the plate is reported to be
relatively nonporaus.'® This method is superior
to electrodeposition in that articles of irregular
shape receive uniform deposits,

The deposition of chromium by vacuum metafizing
i5 also being studied. In this method chromium
metal is vaporized from a tungsten filament, under
vacuum, and deposited on the metal surface to be
plated,'! It is difficult to plate artizles of irregulor

 

5. Inouye, M. D'Amore, and T. K. Roche, Nickel
Base Allovs for High Temperature Service, ORNL CF-
36-4-121 {April 146, 1956).

QW. E. Browning, R. E. Adams, and H, L. Hemphilj,
ANP Quar. Prog. Rep. Dec, 31, 1956, ORNL-2221, p 305,

]GR. M. Burns and W. W, Bradley, Profective Coatings
for Metals, 2d ed., p 205, Reinhold, New Yark, 1955,

_] IS. Dushman, Scientific Foundations of Vacuum Tech-
riques, p 757764, Wiley, New York, 1949,

254

shape with chromium by electrodeposition, and it
is anticipated that vacuum metelizing will over-
come this difficulty. Vacuum metclized coatings
are also expected to have no entrapped liquids,
as might be the cose with an electroplate, and
thus a large percentage of the coated specimens
should survive high-temperature exposure to air,

Tests specimens of Hastelloy B have been nickel
olated by electrodeposition and then chromium
plated by wacvum metalizing, Other specimens
have been nickel plated by the "‘electrofess’
method ond then
metalizing. These specimens were then heated to
815°C in a muffle furnace and coocled to room
temperature daily to study the characteristics of
the oxide film, The preliminory results are promis-
ing, and work is continuing.

chromium plated by vocuum

IN-PILE LOGP EXPERIMENTS
W. E. Browning

R. E. Adams M. F. Osborne
H, L. Hemphill H. E. Robertson
R. P. Shields

Out-of-Pile Test of Pump for LITR Yertical Loop

Operation of the full-size pump designed for cir-
culating fused-salt fuel in the vertical in-pile loop,
which was started in an out-of-pile test loap during
the previous t:]urc:r'fieer_.12 was terminated after

3525.5 hr at 1570°F with NoF-ZrF ,-UF, (50-46-4
mole %, fuel 30}, The pump was stopped for re-
pairs to the cooling coils which had been over-
heated when o water main broke and interrupted
the water supply. The pump could not be restorted,
presumably because the overheating had demaged
the bearings,

The pump wos therefore cut open for examination
and is shown in Figs. 4.1.22, 4.1.23, and 4.1.24.
The desian of this canned-rotor pump may be seen
in the cross-sectional cut shown in Fig, 4,1,22,
The ZrF ,-vapor control features of this design are

 

]2"1'!'. E. Browning, M. F. Osborne, and H, E. Robertson,
ANP Quar, Prog. Rep. Dec, 31, 1956, ORNL-2221, p 299,

UNELASSIEIED
FEaTE

  
     
   
 
 

 
  
 

~—{ EVEL OF LOWER
i s | T geaRiG
9
i i ] i '.?.* |
8 s Sy | COCLING CQIL
- L .
v ! o
4 ‘ "’: :

  
 

. VAPOR-DERGSITION
ISOTHERM

i. ——VAPDR BAFFLE

— __ LIQUID LEVEL

 

 

 

 

OAK RIDGE NATIONAL LABORATORY

 

Fig. 4.1.22, Cross-Sectional Cut Through Pump for
Verticel In-Plle Loop Which Circuloted NoF-ZrF‘-UFd
(50-44-4 Mole %) for 3525 hr at 1570°F Out-of-Pile. The
materiol in the liquid passages is solidified fuel,

[Secret with coption)

PERIOD ENDING MARCH 31, 1957

of porticular interest. The vapor beffle, indicated
in Fig. 4.1.22, provides only a narrow path of
considerable length through which ~vapor can
diffuse, The baffle is made of Graphitar to avoid
stoppage if a piece of solid vapor deposit from

L A b FrE
Ee 0 T0 1RRAT

    

CAR RIDGE NATIONAL LABORATORY

 

Fig. 4.1.23, Pump Shown in Fig. 4.1.22 with Shaft
Removed to Show Yopor Deposits Above Baffle,

 

Fig. 4.1.24. Another Yiew of Pump in Fig. 4.1.22
Showing VYapor Deposits and a Cross-Sectional Cut
Through the Shoft.

255
ANP PROJECT PROGRESS REPORT

above should foll between the baffle ond the shaft,
Every surfoce on the fuel side of this baffle should
be above the vapor-deposition temperature. This
temperature is about 1370°F for fuel 30 at 1570°F
{ref 13}, The vapor-deposition isotherm, shown in
Fig, 4.1.22, was deduced from the location of
vapor deposits in the pump, and it is approximately
at the design point, The small omount of vopor
which diffuses through the baffle is spread over
the large surface provided for that purpose above
the isotherm cnd below the lower bearing. Heavy
local depesits which could stop the shaft after
short running times are thus avoided,

The distribution of vapor deposits is shown in
Figs, 4.1.23 and 4.1.24, The spongy masses
above the boffle are fine needle crystals, loosely
clustered, thot resemble hoarfrost, These de-
posits did not touch the shaft or the baffle. The
deposit just below the bearing wos obout as hard
as cholk and about Ix"]é in. thick. The deposit ot
the top of the fuel chamber was quite dense and
hard. The growing surface haod reached an equslib-
rium position where its temperature was just equal
to the vapor deposition temperature. No wvapor

 

V3, H. Cooper, ANP Quar. Prof. Rep., June 10, 1956,
ORNL-2106, p 65, esp. Fig, 1.4.12.

deposit wos found on the shoft near the baffle, and
ne scrotch marks could be seen on the two focing
surfaces of the shaft and boffle. Hydrodynamic
performance tests, thermal mockup tests, and
bearing tests of this pump were described pre-
viously.”

Examination of the furnaces used in this bench
test, which were identical with the furnoces to be
vsed in-pile, showed several points where foreign
materiols deposited and caused sections of the
furnace to burn out, Analyses of these deposits
showed them to be from soft solder and flux used
in the attempled emergency repair of the cocling
coils. Mo traces of other construction moterials
were found in these deposits,

Assembly of LITR VYertical Loop

The in-pile loop (No. A-8) being assembled for
operation in the LITR was filled in a dry box and
run out-of-pile with temporary heaters and thermo-
couples for 3 hr at temperatures up to 1630°F,
Permanent hecters aond thermocouples were then
The loop is shown in Fig, 4.1.25 ot
at two stages of assembly,

installed,

 

Ty €, Browning e al,, ANP Quar. Prog. Rep. Sept.
10, 1956, ORNL-2157, p 245.

LNCLASSIFIED
PHOTO F9587

 

 

 

MiAPToE T T = ‘fl’

b UNCLASSIFIED
\\ 19470
\ A
| DA
[ \-‘ _'q.\

e
. P

g

PHOTO
{ ',‘1
| ]

N e gk
b LT e 95 .

Fig. 4.1.25. ¥Yertical In-Pile Loop A-8 ot Two Stages of Assembly,

256
The loop control system at the LITR is being
inspected and readied for operation, and two modi-
fications are being made. Provision is being made
for odjusting ond monitering the distribution of
cocling air between the two legs of the loop. Also,
a reservoir 1S being connected into the cooling
water system, and a means is being provided for
cutomatically actuating e circulating pump if the
main water supply is interrupted,

Irradiation Facilities in Oak Ridge Research
Reactor

In  cooperation with the ORNL Operations and
the Engineering and Mechanical Divisions, facilis
ties are being plonned for the future operation of
a minicture corrosion-testing loop in the Qak Ridge
Research Reactor (ORR)., The routing of instru-
ment and power leads, locations of off-ges lines
and carbon traps, locotion of the motor-generator
set, location of a fail-safe nitrogen supply, and
design of the reactor vessel flange seal for the
leads are being worked out, A new control panel,
to be located in a third floor laboratery in the ORR
building, is being designed by the Instrument
Depariment.

INVESTIGATION OF SULFUR CONTAMINATION
OF A DRY BOX BY NECPRENE RUBBER

W. E. Browning
F‘Ei E. AdGmS Ho Lt Hemphi”

During preparations for filling the vertical in-pile
loop, it wos discovered that, while under vocuum,
the interior of the dry box was contaminated by an
unknown gaseous sulfur compound., The presence
of the contaminant was monifested by the appear-
ance of a Cu,5 film on the copper-cooling coils of
the loop ossembly. The Cu,5 film wos identified
by chemtcal analysis ond x-ray diffraction.

The wvaricus lubricants and oils vsed in the
vacyum system and loop assembly were therefore
analyzed for sulfur, and the concentrations of the
sulfur found renged from less than 10 to 92 ppm.
These quantities are to be expected in pump oils
and lubricants, Tests of the foop assembly elimi-
nated it as ¢ source of the sulfur contaminant,
The helium supply was also investigated for sulfur
contamination by bubbling the helium through o
silver nitrate solution, Mo evidence of a silver
sulfide precipitote was noted,

Access to the interior of the dry box is provided
by Meaprene rubber gloves. A glove and a copper

PERIOD ENDING MARCH 31, 1957

specimen were enclosed in a vacuum system inde-
pendent of the dry box, and after several hours
under vacuum Cu,5 wos found on the copper
surfaces, This indicated thot the sulfur contami-
nant was present in the Neoprene rubber and was
Sulfur is not used in
Neoprene as the wvulconizing ogent, but orgonic
sulfides are used as a plasticizing agent.

The Inconel surfaces of the loop assembly were
investigated by electron diffraction, Samples of
the surface film were removed from the loop motor
agssembly which had been exposed to the contami-

liberated under wvacuum.

nated dry-box atmosphere, Simifar semples were
removed from another moter assembly which had an
identical history except for exposure in the dry
box, These surface samples were removed by the
lacquer-stripping technique used in preparing rep-
licas for electron microscopy. The samples were
found to be identical, and crystalline sulfides of
nickel, iron, and chromium that were sought were
nat found.

A small sample of fluoride fuel was exposed over-
night to the contaminated dry-box atmosphere and
then analyzed for sutfur ciong with o control sample
of unexposed fuel,
tent was found to be below the limit of medsure-
ment {20 ppm S5) of the onalytical method being
used., Since the loop surfoces and the fluoride
fuel revealed no sulfur contamination, the filling
operation was completed and the loop was removed
from the dry box,

In the case described the quantity of sulfur com-
pounds present was not considered to be harmful,
but elimination of contominents would be desirable,
and various methods cre being studied. The ovail-
able sulfur content of the gloves might be removed
by chemical treatment without changing the physical
characteristics of the gloves, and, although difi-
cult, the Neoprene gloves might be replaced with
gloves of another suitable material.

In both cases the sulfur con-

EFFECT OF RADIATION ON THERMAL-NEUTRON
SHIELD MATERIALS

J. G. Morgan

R. M. Carroll M. T. Morgan
P. E. Reagon

[rradiation exposures of nine Cu-B,C samples
wera completed in the MTR, and examingtions of

 

135, s, Rogers (ad.), The Vanderbilt Rubber Handbook,
p 47-48, R, T, Yanderbilt Co., New York, 1948,

257
ANP PROJECT PROGRESS REPORT

six of the samples were nearly completed. The
results of the examinations'?® showed essentially
noe changes in dimensions, weight, ond density
ofter irradiation,
data, however, as a result of slight oxidotion of
the somples. Hardness data are given in Table
4.1.4, The dimensions of the shielding elements
tested in these irradiations were 0,1875 x 0.500 x
0,102 in. The core of each specimen was 0.082 in,

There woas some scotter of the

thick and consisted of a copper matrix with ¢
6.6 wt % B,C dispersion. The core was doubly
clad on two faces with 3 mils of copper and 7 mils
of type 430 stainless steel, The core was exposed
on the four remaining edges.

Samples 4, 5, and 6 were irradiated simultaneously
at 200°C for 322 hr in position A-39 of the MTR.
The maximum temperature reached waos 220°C, The
samples experienced three temperature cycles at

 

164ardness and metaliographic dato obtoined by A. E.
Richt ond reported in Metallographic Examination of
Irradiated Cermet Shielding Materials — Report No. 2,
ORNL CF+56-12-105 {Dec. 18, 1956) and Meialiographkic
Examination oé Irradiaied Cermet Shrelding Maferials —
Report No, 3, ORNL CF-57-2-33 {Fab. 7, 1957).

a rate of less thon 100°C/min. The avercge burnup
of the 819 otoms was estimated to be 7%,

No apparent structural changes occurred as the
result of irradiation, as shown in Fig. 4.1.26, The
cores of both irradioted ond control specimens
were examined at both high and fow megnifications,
and no evidences of cracking, blistering, or
porosity were opparent in either the copper matrix
matericl or in the boren corbide particles. Alsg,
there was no reaction between the boron carbide
particles and the copper matrix. The typical
appearance of the stainless-steel-to-copper-te-core
bond is shown in Fig., 4.1.27, There were no
evidences of blistering, ¢cracking, or separation of
the cladding-to-core bonds.

Samples 1, 2, and 3 were irradiated at 420°C for
929 hr in position A-28 of the MTR. The maximum
temperatyre reached was 380°C.  An isetfopic
analysis showed an 18% average burnup of the B'C
atoms,

Hardness measurements made on the stainless
steel cladding (Table 4.1.4, sample 2) indicated o
large increase in hardness os the result of the

Toble 4.1.4. Results of Hardress Measurements of Cu-B‘C Samples 4 ond 2

 

Knoop Hordness (l-kg locd}

 

Locotion of Imprassion

Averoge Before

Percentage incracse
Averaga After e

 

Irradiation lsradiation Change or Decrease After
Irradiation
Sample 4
Type 430 stoinless steel clodding 13% 166 + 27 +19
Copper clodding 64 76 +12 +19
Core
Canter 76 102 + 26 +34
Edge 76 108 + 32 +42
Somple 2
Type 430 stainless steel cladding 139 186 -+ 47 +34
Copper clodding b4 65 =1 -2
Core
Center 76 70 -4 -8
Edge 76 78 +2 +3

 

258
PERIOD ENDING MARCH 31, 1957

- -1

- p' . UNCLASSIFIED
‘ RMG 1529

! UNCLASSIFIED

 

Flg. 4.1.26. Typical Core Structures of (a} Unirradlated and (5) Ireadlated CusB,C Shielding Elements. S00X.
Reduced 13%.

  

~afi— COPPER CLADDING —2m=

1 - o .‘
"' , < [ B e -
b - ¥ _’"
TYPE 430
- — STAINLESS-STEEL 3w -~ - :
CLADDING » Sy
L) (&)

Fig. 4.1.27. Core-to-Clodding Interfoce of a Cu-B,C Shielding Element {a) Befare and (5) After lrradlation.
30X, Reduced 27.5%.

259
ANP PROJECT PROGRESS REFPORT

irradiotion thot was significantly greater than the
increase noted in the examinotion of sampies 4, 5,
end 6, The copper cladding and the core of the
specimen showed no significant changes in hord-
ness. Although the center of the core decreased
in hardness, it is felt that this was the result of
the severe cracking, which could cause erroneocus
measurements.

The ends ond sides of specimens 1, 2, and 3 are
shown in Fig, 4.1.28, The core cracking and the
core-cladding separotion apparent in all specimens
varied in degree from specimen to specimen, Speci-
men 2 was selected for metailographic examination,
and typical sections of the core are shown in Fig.
4.1,29. The particular area shown in Fig. 4,1.29
gives no evidence of cracking or porosity of either
the boron carbide particles or the copper matrix

However, as the etched photomicrogreph
the etching revealed a heavily attocked

material,
shows,
area immedictely adjacent to the B,C particies.
In addition the siructure of the copper motrix
appears to be somewhat disturbed when compared
with the structures obtained in irrodiated specimens
4, 5 and 6, Both these conditions were noted
throughout repeated polishing and etching,

Longitudinal cracks were found in the core ot
both ends of the specimens, which usually started
at the center of the core and progressed approxi-
mately one-fourth the length of the specimen. A
typical c¢rock is shown in Fig. 4.1.30, and the
intergranular ottack in the vicinity of the crack is

shown in Fig. 4.1.31,

Significant core-cladding seporation was found,
as shown in Fig. 4,1,32, [t is interesting to note
that no evidence of seporation was found ot the
interface.,  Occasional
intergranuler separotion

stainless-stee|-to-copper
large wveoids ond partial
were noted in the grain houndaries of the copper
cladding, os shown in Fig. 4.1.33., Examination
of the entire length of a specimen indicated that
the core-cladding separation was dapproximately
70 to 80% complete.

FUSED SALT POLAROGRAFPHY

R. J, Carter

The applicability of polarographic technigues to
the study of the fission-product chemistry of fused
salts is being investigated, The polarogroph which
is being vused is similar to that used in aquecus

260

poforography; however, the hich electrical con-
ductivity of the fused salts as compared with
aquecus sclutions mokes it impractical fo wvary
the potential applied to the polarogrophic cell
linearly with time, For this reason, current-veltage
curves are recorded with on X-Y recorder rather
then with a conventional strip-chart recorder,

Both silver and platinum microelectrodes of
various sizes and shapes ore being tested for use
as cathodes. The platinum crucible which contains
the fused salf is used as the anode. The platinum
crucible is mounted in an Inconel contoiner which
can be filled with helium in order to maintain a dry,
inert atmosphere around the fused salt, The neces-
sary electrical connections are made by using
Kovar seols which are soldered into the lnconel
can, VYertical movement of the microelectrode is
accomplished by mounting the electrode support
on o fexible bellows,

In order to test the equipment, current-voltage
curves were obtained at 350°C for severol solu-
tions containing from 6.4 to 1,0 parts per thousand
of cadmium chloride dissolved in potessivm nitrate.
Five or six current-voltage curves were obtained
for each scolution, The reproducibility of the
curves was good in the initial portions but poor in
the diffusion-current region, where large fluctua-
tions in current were cbserved., The fluctuction of
the diffusion current has since been found to be
caused by vibration of the laboratory floor, and a
suitable meons for shock mounting of the furnace
used to heat the fused salt is presently being
sought,

NEUTRON FLUX STUDIES
D. Binder 1. F. Krause 1’

The rapidity of effects on germanium dicdes and
transistors in the donut of hole 51N of the ORNL
Graphite Reactor led to an investigation of the
variation of neutron flux clong the hole. A pre-
fiminary survey wos mode with sulfur ond cadmium-
covered gold detectors, ond the results are shown
in Fig., 4.1.34, The relative octivities of the de-
tectors are plotted against distance from the con-
crete shield, ond the last point is 11 in, ocut from
the edge of the donut, the position described pre-
viously as outside the donut,'® From this position

 

l:"{.'!n ossignmant from Pratt & Whitney Aircreft.

18p, Binder ond 4. F, Kravse, ANP Quar. Prog. Rep.
Dee. 31, 1956, ORNL-222%, p 308.
PERICD ENDING MARCH 31, 1957

 

- 1
UNCLASSFIED -y UNCLASSIFIED UNCL ASSIFIED
RMG 1578 - BMG 1573 BEMG 1577

 

      

—  UNCLASSIFIED
L RMG 1578A

‘

i UNCL ASSIFIED
- RMG 1574
4 ’Iv"' ' v
U™

 

Flg. 4.1.28. MTR-lrrodicted Cu-B,C Specimens. (a, b, ¢} Ends of specimens 1, 2, and 3, 13X. {d, e, [} Sides
of specimans 1, 2, ond 3. 7X,

261
 

ANP PROJECT PROGRESS REPORT

.. ; -—‘Ii
UNCL ASSIFIED .‘. J \] f} \' 3 UNCLASSIFIED
= RMG 1583 | v RMG 1584

 
   

 

Fig. 4.1.29. Core of MTR-lrrodlcted Cu-fldc Specimen Mo, 2 Showing Heavily Atocked Areas Adacent to
B,C Porticles. {a) As-polished. (b) Exched. 500X. Reduced 19%.

UNGCL ASSIFIED
RMG 1586

 

Fig. 4.1.30. Corea of MTR-lrradioted Cu-B‘{C Specimen Mo, Z Showlng a Typical Fracture. (g) As-polished.
(b) Etched. 250X. Reduced 18%.

262
 

PERIOD ENDING MARCH 31, 1957

7
UNCL ASSIFIED
RMG 1588

UNCL. ASSIFIED
RNG 1587

  
 

  
   
 
 

: ¥ - na{
(2) SR .. rm‘&"fi > .

Fig. 4.1.31. Core of MTR-lrradiated Cu-B,C Specimen No. 2 Showing Intergranular Attack In Yicinity of @

Fracture. {(a) As-polished. (b) Etched. 500X. Raduced 20%.
RME 1590

 
 
  

COPPER CLADDING
COPPER CLADDING

  

TYPE 430 STAINLESS STEEL CLADDING -

  

TYPE 430 STAINLESS STEEL CLADDING

 

Fig. 4.1.32. MTR-lradigted Cu-B,C Specimen Na. 2 Showing Core-Cladding Separation. [(a) As-polished.
(6) Etched. 250X. Raduced 18%.

263
ANP PROJECT PROGRESS REPORT

N —— A W P

L. Y e, a‘ UNCL ASSIFIED =
Yha o RMG 159)
NPT S
‘V AT T
gc;.\"' CORE
- .

== 5

gh &

TS *{}

\,1.4

j “‘Y ‘.;-?; 1.1
' : 5 _-‘ A _i - ¥ R
e — & — ; = -

COPPER CLACDING

TYPE 430 STAINLESS STEEL CLACDING

{g]

Flg. 4.1.33.

T —

UNCLASSIFIED
RMG 1592

   
     

TYPE 430 STAINLESS STEEL CLADDING

(&)

MTR-Irradiated Cu-B,C Specimen No. 2 Showing Velds and Pertial Intergranular Separation In

Grain Boundaries of Copper Cladding. (a) As-polished, {b) Etched. 500X. Reduced 19%.

to the center of the donut, both Np?37 neutron de-
tectors and carrier concentrations in n-type ger-
manivm indicate an increase in activity by o
factor of 3 (ref 18},

The 53:(n,p)P32 reaction has an effective thresh-
old of 3 Mev, and Au'%® hos a strong resonance at
5 ev; therefore the two detectors are seasitive to
the two extremes of the reactor spectrum. The gold
detectors show a smooth increcse with distance
from the shield (the curve is drawn through these
points}. The sulfur points, in general, fall below
the geld peints ocutside the active lattice (24 in,
from the shield). The two detectors are normalized
ot the innermost pesition, [t would be of interest
to compare these wvariations with a variation in
radiation effects. The lotter would presumably
fall between the detector variations,

264

UNCLASSIFIED
ORNL-LR-DWG 16683

 

   

 

5
?
-{
G L{#] |
'-l
> [
£ 5|
1
L
=
k2 2
=1
& |
102 ;
. | Ay (Cd COVERED) — i
* SULPHUR .
2 |
||0_3 | e J
0 20 a0 60 80 100

DISTANCE FROM SHIELD {in)

Fig. 4.1.34, Relotive Activities of Detectors in Hole
51 of the ORNL Graphite Reoctor vs Distance from
Shield Facs.
EFFECTS OF RADIATION ON ELECTRONIC
COMPONENTS

J. L. Pigg

C. C, Robinson
0. E. Schow

0. Curtis

Minority Carrier Lifetime Measurements in
n-Type Germanium

Measurements of the effect of irradiation on the
minority carrier lifetime in n-type germanium were
inved.'? The hole lifetime § | |
centinued, e hole lifetime jor several samples
was meosured before and after a series of irradia-

tions in the ORNL Graphite Reactor and in a Co®®
gamma-ray facility,

Sample Sb-28A-3 was irradicted with fast neutrons
in the animal tunnel of the ORNL Graphite Reactor,

 

Py e Pigg er al., ANP Quar. Prog. Rep. Dec, 31,
1956, ORNL-2221, o 309,

PERIOD ENDING MARCH 31, 1957

This sample, which had on original corrier con-
centration of 3 x 104 carriers/cm?, was irradiated
until it became intrinsic,

Initially, very slight irrodiation produced large
changes in the hole lifetime. Since the initial
lifetime was 1010 psec, it can be seen from Fig,
4.1,35 thot the hole lifatime changed by about a
factor of 5 for a dose of 101! fast neutrons/cm?,
which produced a change in carrier concenfration
of only 0.1%, The plot of hole lifetime vs fast-
neutron dose in Fig., 4.1.35 shows a simple power
dependency, but the stroight line produced does
not have o slope of unity, but rather 0.7,

The departure of the dota from the straight line
at the higher doses was to be expected as a result
The
minority carrier lifetime depends inversely upen the
carrier concentrafion when the matericl is near the
range,  lhe of carriers thus
becomes important for this sample after a dose of
1077 neutrons/cm?,

Variations in the surface recombination velocity
would tend to give scatter in the dota for fong

of the change in the carrier concentration,
removal

infrinsic

UNCLABSSIFIED
ORNL -LR-0OWG 20382

 

10} ——

 

 
 

 

 

 

 

 

 

 

  

 

 

 

 

 

—=—— —+H e —— —— —
- I EF T ] 1 R - B I i "“J:
R o L il I S W | | ‘ I L Pl
59 ¥ 1 S S 1 T
_ (1] | 1]
| I I . I | ]

2 —— R <L T T 1 ~t GERMANIUM SAMPLE: SB-28A-3— 11171
| \ | ‘ | INITIAL LIFETIME: 1010 psec [ |
| | | | | |

2 5 I | | | |

s SO0 =y ! 1 - | ] I —
a L = | = i | 113 =1 T
3 — 1 . I e
= = L [ - L
E 9 = | I | |
= L = L L1l
: |
t..._LJ.l' | |
g L
10 _ e ! SN et B Ll
e A == L
= t ___ 1 _._._i_ 3 114
5 i y .y Y | . |
e I — T __'_‘I 'I‘
o e b — P
1 | | L
o'¢ 2 5 o' 2 5 i0'? 2 5 10'? 2 5 i0' 2 5 {09

FAST-NEUTRON DOSE (neulrons/cm®)

Fig. 4.1,35. Effect of Reactor frradiotion on Hole Lifetime in n-Type Germanium,

265
ANP PROJECT PROGRESS REPORT

minority carrier lifetimes., Therefore, the initicl
points in Fig, 4,1,35 show scatter for points where
the diffusion length of the minority carrier is of the
same order of mognitude os the cross sectional
dimensions of the sample,

The results of a series of irradiations in a Co®?
gamma-ray source of two samples of a second ingot
obtained from The Eagle Picher Company are shown
in Fig. 4.1,35, The original carrier concentration
of the material used was about 1,5 x 1074 corriers/
em?. Somple EP2-3 was irradioted until it became
The behavior is quite similar to that
for neutron irradiction, except for one striking
ln order to obtain the same change in

intrinsic.

ditierence,
hole lifetime with the use of gomma roys, o much
larger percentage of the carriers must be remaoved,

The reason for this should be indicated by the

03

2
_ 10t
|n}
&
vi
A
w 5
=
-
Ll
L
|
w2 =
|
o
T i
10

o't > 5 o 2 5

 

type of damage produced, The slope of the curve
is somewhat less than the slope of the curve for
nevtron irradiotion, being in this case about 0.6,

The role which the initial carrier concentration
plays is being investigoted with samples of widely
varying resistivity, Measurements are also being

made with a Cs'37 gamma-ray source,

Diode Irradiations

It was shown pr\evin:-u&‘.l-;.»'2“:l that the reverse
voltage-current characteristic curve of a germanium
diode is sensitive to Co®® gamma-rays. In the
previous experiment, the
currents were meosured ot a fixed voltage bios, In

forward and reverse

 

2':'J. C. Pigg and €. C. Robinson, ANP Quar. Prog.
Rep. fune 10, 1936, ORNL-2106, p 244,

UNCLASSIFIED

CRNL-LR-DW6 20383

EAGLE PITCHER COMPANY INGOT NQ. 2
|

A SAMPLE EP2-3
® SAMPLE EP2-4
| |

10'® 2 5 10'7 2 5 108

Co®9 GAMMA-RAY DOSE (photons /em?)

Fig. 4.1.36. Effect of o Gumma=-Ray lrradiation on Hole Litetime in n-Type Germanium.

266
order to differentiote between barrier current ond
surface-leckage current, o IN 238 dicde was ex-
posed ot a dose rate of 500 r/hr in a Co®? focility,
Data for a complete characteristic curve were
obtained up to a maximum bias of 1.5 v,
these data, the soturation current intercept, |

From
0 ond
the surface-feakage curreny, f;, can be cbtained,

The soturation current intsrcept is cbtained by
extrapelating the saturation part of the reverse
characteristic back to the origin. This value, | go
enters into the diode equation for the reverse biased
condition as follows:

} om !U“ _e—qukT] ,

1 = current through barrier at voltage V,
g = electronic charge,

Boltzmann constant,

T = absolute temperature,

A
it

The parameter [, appears in tronsistor considera-
tions as the cutoff current {_ , where I__ is the
collector current with zero bias on the emitter,

The change in 7, as a fuaction of the Es5*
gamma-ray dose for a IN 38 dicde is shown in
Fig. 4.1.37. The change in i, is rapid during the
initial part of the irradiation but tends to saturate
as the exposure continues., Previous experiments
with 1N 38-A diodes showed a maximum in the
reverse current ofter about 1.7 x 107 ¢+, The 1N 38
showed no maximum in Iy after 4 x 107 r, Since

UNCLASSIFED
ORML-LA-DWG 20384
Pl

 

  

REVERSE $ATURATION CURRENT, /, amp)

 

0 ———l = e
O 107 2n0’
GAaMMA-R LY DOSE {r)

Baiy’ q 107

Fig. 4.1.37. Effect of Co®0 Gomma-Roy Irradiation
on the Reverse Saturction Current of a 1N 38 (Mo. 10)

Germaniom Disde.

PERIOD ENDING MARCH 31, 1957

these diocdes are commercial vnits, there is no
indication of the initial carrier concentration or the
initial mobility. Thus there is no reason to expect
that the moximum observed previously?® would
occur of the same integroted exposure for a different
type of unit,

Upon removal from the gamma-ray source, the
change in I, indicated that an annealing process
was taking place in the barrier. As con be seen
from Fig. 4.1.38, recovery of the unit was not
complete, ond the onnealing was not o simple
first-order process, The minimum in Iy after 125 hr
of annecling may be real,

The surface-leckage current, I, is obtained by
subtracting /) from the reverse current measured at
a bias of 1 v, The value 7, may thus be considered
to be the leakage conductance. The chonge in i,
as a function of exposure is shown in Fig, 4,1.39,
It can be seen that !, saturates after a ropid initial
increase of about one order of magnitude. This
saturation moy indicate that the depth of surface
attack is limited.

The annealing of I, upon removal from the gamma-
ray source, shown in Fig, 4.1.40, is complicated,
as it was in the case of /. A definite minimum
can be seen at cbout 40 hr, The fact that annealing
occurs indicates that the surfoce teckage is not
caused by simple chemical attack alone, Since o
commercial unit is being studied, if is not practical

UNCLASSIFIED
ORNE -LR -0 20345

ROCK TEMPERATURE ANNEALING TIME {hr)
5 0 15 z0

s ©
3% = — T T

TR TP

1 ®

 

REWERSE SATURATION CURRENT, fplampl

 

0 - -— o — i e
Q o0 100 {50 200
ROCM TCMPERATURE AMNEALING TIME [hrl
Fig. 4.1.38. Room-Temperature Annealing of the

Rovorse Saturation Current of a IN 38 (Ne. 10) Germa-

nium Diode Subsequent to Irradiation to o C:-:'r“JI Gamma-
Ray Dose of 3.24 x 107 -

267
ANP PROJECT PROGRESS REFPORT

to isclate the different possible sources of surfoce
change or to study the surface directly,

Transistors

A G-E 2N 44 PNP fused-junction germanium

transistor wos exposed in the ORNL Grephite
UNCLASSIFIED
- QREL-LR-GWG 20306
10 L T T

 

 

 

 

a4 |-

SURFACE-LEAKAGE CURRENT, Z lamp)
8
ch

 

 

 

 

0 e 1
) 167 Zx10’ 3x107
GAMMA-RAY DOSE (A
Fig. 4.1.39. Effect of gt Gomma-Rey lrrediation

on the Surfoce-Leakcge Cutrent of o IN 38 {Neo. 10)

Germanivm Dode.

UNGLASSFIED
DAL L R=CWG 20T
FOCM TEMPERATURE AMNEALIMG TIME [he)
50 00 1540 200

SURFACE -LEAKAGE CURRENT, J, (omp)

 

 

  

 

15 2_0
ROCM TEMPERATURE ANMCALING TIME (he}
Fig, 4.1.40. Room Temperature Annecling of Suifoce-
Laukuge Current of ¢ TN 38 {No. 10} Germanium Dicde

Subsequent to Irradiction to a Coéo Gamma-Rey Dose
of 3.24 x 107 ¢,

268

Reactor, The leo values for different voltoges
were recorded os o function of expasure, Both
grounded-base and grounded-emitter connections
were used, The feg values for the grounded-bose
connection are the I, values for the collector
It can be seen from Fig. 4.1.41 that the
barrier 1, drops until ofter an exposure of about
8 x 10'" neutrons/cm?.  The leckage current, Iy
mentioned above would be reflected in the spocing
between the curves, The change in [, is smell
compared with the change in /.

The f_, measurements for the grounded-emitter
connection are shown in Fig., 4.1.42. The general
shape of the curve is the same os that for the
grounded-base connection, but the change in leck-
age current is much more prenounced in the grounded-
emitter connection. The velve of I, in terms of
leakage conductance, as defined above, varies from
3.4 x 1078 amp/v initially to o minimum of 1.4 x
107% amp/v ot the minimum of the curve and in-
crecses fo approximately 4,2 x 107% amp/v after
an exposure to a thermal-neutron dose of 3.45x 1012

barrier,

L

neutrons/cm?,

From the observetions to dete, if appears that the
manner in which the choracteristics of a junction
device change depends both upon the nature of
irradiation to which the device is exposed ond the
method by which the junction is made. Qualitative

UNCLASSIFIED
6 ORNL—LR—DWG 20398

CUTOFF CURRENT, [,, (omp)

 

|
| |
o i 401% 2x102 52102
THERMAL-NELUTROM DOSE (adutront feme )

4 240l E

Fig. 4.1.41. Effect of Recctor Irradiation on Grounded-
Bose Cutoff Current of o G-E 2N 44 PNP Fusad-Junction

Germonium Tronsistor.
 

 

   

 

UNCLASSIFIED
P CANL-LR-DWG 20389
310
A5y
&
£
= |
D a ,
3 2410 |
e
= 225w |
¥
e
2w
3 I
L
S 0t A et AT . 3o |
3
|
o o o _—l
0 10'2 20t 32102 4107

THERMAL- NEUTRON DOSE (neutrons/em?]

Fig. 4.1.42, Effect of Reactor Irradiation on Grounded-
Emitter Cutoff Current of o G-E 2N 44 PNP Fused-

Junction Germonlum Transistor.

trends in behavior can be established by using
commercial units, but the quontitative date neces-
sary for reliable interpretation of the effects will
require the use of ingots or samples specifically
prepared for such on investigation. In porticular,
it must be possible to control the history of the
sample surface, The somples would be constructed
so that the conductivity of the base maoterial
could be determined,

A Safe for Removal of Samples from the Reactor

The removal of samples from the ORNL Graphite
Reactor subjects the personnel engaged in the work
to beta ond gomma irradiation from the octivated
material being removed and to contamination by
the radicactive dust associated with the decay of
insulation.  This problem is particularly acute
when samples must be removed from the reoctor
loading face, As a first step in solution of this
problem, o sample removal sofe was designed,
Fig. 4.1.43, The sofe can be handled by two men,
and vyet it is thick enough to provide odequote
shielding for the levels of activity normally en-
countered. All access facilities can be operated
by one men., The device can be used either at the
reactor loading face, or, by transterring to another
dolly, at holes 50, 51, and 52. Provision hos been
made Sor vacuum removal of rudicociive dust ond
particles through numerous access ports.

PERIOD ENDING MARCH 31, 1937

Modifications to Hole 30 of the ORNL
Graphite Reactor

In order for measurements fo be made con semi-
conductor components under irradiation, the ex-
posures must be conducted at flux levels one to
two orders of magnitude lower than those found in
the central region of the ORNL Graphite Reactor,
Unless such low flux levels are used, the measure-
ments cannot be made in o time interval that is
short compored with the fime interval which pro-
duces @ significant chonge in the parameters being
measured.

If exposures are conducted in hole 51, where q
fission spectrum, o removal focility, and provisions
for instrumentation are available, the reactor must
be operated at power levels between 50 and 400 kw,
The time available for such operation is limited
and the reproducibility of flux levels s difficult.
Exposure near the periphery of the reactor couses
a positioning problem and provides an undesiroble
ratia of thermal to fast flux,

In order to eliminate these disadvantages and to
supply several advantageous features, an exposure
device has been designed for use in hole 30N of
the Graphite Reactor, as shown in Fig, 4.1.44, The
device consists of a cart which may be moved from
a loading shield at the reactor shield foce to a
fission chamber locoted inside hole 30N, The sam-
ple can be moved from the external shield to the
irradiation position in about 10 sec.

The fission chamber position may be adjusted to
give a range of flux levels. The external shield is
designed to facilitate removal of the fission
chamber should such action become necessary.
The sample moy be instelled ond removed without
breaking the electricel connections. This will
make possible continuous readings from the pre-
exposure to the postexposure periods.

DIELECTRIC CONSTANT AND LOSS TANGENT
OF IRRADIATED PLASTICS AT MICROWAVE
FREQUENCIES

R. A, Weeks J. C, Pigg

Previous work indicated thet reactor irradiation
causes changes in the dielectric properties of
polyethylene and Teflon,2! The data showed
lorge changes in loss tangent after fast-neutron
doses of up to 10'® nevtrons/em?.  The samples

 

2p. A Weeks, Sohid State Semiann.
Aug, 30, 1954, ORNL-1762, p 113.

Prog. Hep.

269
0LL

UNCLASSIFIED
- & LEAD-FILLED DQORS - ORML-LR-OWG 20390

     
    
 
  
   
 

Lol EH LEAD-FILLED DOOCRS
i,
- -LEAD-FILLED BOGDY
—

 

 

m g

 

 

 

MOTE -
SHIELD WEIGHT — APFROX. 24001

 

8

[ DOLLY

     
  
 
 

|
|
|
1
oo A

i

 

 

 

 

SwIVEL CASTER

‘RIGID CASTER
POISITION LOCK

 

1012345678910
SCALE IN INCHES

ey

Fig. 4.1.43. Sample Removal Sofe.

1¥0d3Y S53820¥d 1D370¥d dNY
Lig

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UNGLASSIFIED
ORHL-LR-DWG 203591

 

 

 

 

 

e A

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

   

[ i i

_ T 1; L]
== ‘ . i 5
1 | Fes
bl —} :‘ ——r———— |
qI—l- _I_ & s o
..r L s T =n ]
= ] | | gl e
| T 1 B 1———' ||
| - :‘F SECTION 88 L S fi “ SAMPLE REMOYAL PORT
c
4 b
_‘:jgfi
1 - = | L} REMOVABLE SECTION
- 2 ' || s
VIEW G-C !L |,‘f|--*-:$ Q
FISSION CH.&MBER\\ /m_ummum LINER Ty )i \ ! i : _? e, o baty?
| N “ |
SAMPLE GART " e R
1 ___I_L e e — _3 CHAMBER P-osmoi NG
// : /7////7/ ///‘ i / e - e /fi,
i / L & CONCRETE SHIELD OF s fif“‘
< i /' A 4

 

 

o

GRAPHITE REACTOR

SECTION 4-A

 

REEL SHIELD FOR
INSTRUMENE;I'ATION

     
 

.
. |

 

 

i /
f
@‘DRIVE MOTOR

Fig. 4.1.44, Exposure Facility for Use in Hole 30N of the ORMNL Grophite Reactor.

 

1561 'IE HOYVYW ONIGNT dold3d
ANP PROJECT PROGRESS REPORT

were irzadiated in o water~cooled facility in the
QRNL Grophite Reactor. Examinction after re-
moval from the reactor revealed evidence of wofer
in some of the containers. The presence of
moisture roised a question as to the amount of
water absorbed by the specimen during the irradia-
tion. The ovailable evidence indicated, however,
that the observed changes were not due to water
cbsorption.

These studies were resumed during the quarter
with exposures of Teflon, polyethylene, polysiyrene,
unfilled phenol formaldebyde, and aylon in a Co%°
gammo-ray focility, The irradiation temperature
The only moisture
The same

was opproximately 25°C,
present wos thot in the otmasphere.
sample was vsed for an entire series of irradia-
tions. The low temperature, absence of moisture,
and use of the same somple throughoui the experi-

served to increase the reliobitity of data
21

ment
compared with that reported previously.

The chonge in loss tongent of polyethylene ot
variovs frequencies as o fuaction of the Co®?
gamma-ray dose is shown in Fig. 4.1.45, Similar
data for Tefon, polystyrene, nylon, and phenol
formaldehyde are shown in Figs, 4.1,46 through
4.1,49, It is evident thot the irradiation had a
large effect on the loss tangent of polyethylene
and Teflon but only o small effect on polystyrene.
The nylon and phencl formaldehyde samples have
not yet received sufficient irradiation to produce
a marked change in less tangent. On the bosis of
the results in polyethylene, the nylon ond phenocl

 

 

UNCLASSIFIED
tx ™M
|
25
I
— 20
=
Ladd
L]
=
3
8
S
A0 |
all
o ey - — e —
QL.. e - el bl !
G102 05 0 20 50 00 200 500 (k100

GAMMA-REY DOSE (r)

Fig. 4.1.45. Effect of Co®0 Gamma lrradiation on the
Loss Tangent in Poiyethylane.

272

 

 

LINCLASSIFIED
- ORNL-LA-DWE 203593
A T e e T Y
| ! |
20 | A BODme &
000 me
& | » 3000 me
&8 * 8500 me
Z |
}...
L |
& 10 : | PREIRRADIATION ;
& > VALUE
!
S A
TI—!—'I'I". =t
| 0| |l"lj l
O i ] - L= = i =3 Sahiy i el 25
ot 02 05 10 20 50 100 200 500 (xig)

GaMMo-RAY DOSE ()

Fig. 4.1.46. Effect of Co®® Gommo Irradiation on the

Loss Tangent in Teflon,

UHCLASSIFIED
DANL=LA-DWG 200354

 

 

 

HO.—Q] T _“| . "'_"'_‘_| "l" —"_
| § | 1
24 | fact B 4 I | _, e b o { —tfet |
| | & 800 me ‘ |
| 4000 me ||
7 | 3ol g Lt gy s 4
o 8 | [ = 3000 me | |
z | * BROO me |I
o 16 = | | e | Lokt =
= |
g
o N et e e e
& |
g |
g
4 |e— T i 1
| ||
O - * il e bl 1 ——— ks — . — ——————— ?
o1l 0.2 s 0 20 50 100 200 500 (w109

GAMMA=-RAY DOSE (r)

Fig., 4.1.47. Effect of Co%? Gamma leradiction on the
Loss Tengent in Polystyrene.

UNCLASSIFIEDD
ORNL =LR=DW; 20335

 

tx10™) — - e =
|
£00 -
oc | _.& BO0 mc
i ? ' 1000 me
5 « 3000 me |
Haon Fe bl e B800O mMe |
z .
i |
w 300 : -
e |
o - PREIRRACIATION waLUE |
200 A : et
. | | s ® gla LLL gl |
B
106 = EREr 1 : =it |
| | | |
C, L Ll R 2N b et N I i
of D2 D5 10 20 50 100 200 500 (07

GAMMA-RAY DOSE (r)

Fig. 4.1.48. Effect of Co®® Gamma lrradiation on the
Loss Tangeat in Mylon,
formoldehyde will need approximately twice the
present integrated dosage to produce a significont
change,

The changes in loss tangents of polyethylene
and Teflon appear to be frequency sensitive; the

WHCLASSIFIED
OFML—LR—0ws BOR0E

 

 

 

 

 

ST mms CTTTIT T 11T | 1T
S 1 4 I. |_ SIS S I. 1= .!..I.I
| ) |
1000 | i s i
= | '
& I ! :
% 300 i kL SN = T AR | I. . 1 4 et ek —_— i =&
2 | |
soo |4t UL Lodlet] | LY
g ‘ ikpéilhmpmflqw VALLE | ‘ ! I | |
qoo il il | - - i _: |\ : -
|
( & BEIJOan
00— T —- 0 3000 me i |
| ¥ B?(}On'bc
5 | L | L LIHE L L]
o f i 100 {x107)

GamMME-RAY DOSE ()

Fig- 4.1,49. Ef#fect of Coén Gamma lreadiation on the
Loss Tangent In Phencl Formaldehyde.

PERIOD ENDING MARCH 31, 1957

greatest change occurred at 800 mc, There was
evidence of this frequency dependency in the
teactor-irradiated specimens discussed above. On
the basis of the change in attenuation observed??
in polyethylene cable measured ar 4 mc and the
results shown in Fig. 4.1.45, it seems ciear that
the loss tangent of irradiated polyethylene has o
maximum in the frequency range between 4 and
800 mc. Such a maximum probably exists also for
Teflon. Measurements in this frequency range are
needed to substantiate these conclusions.

Any chonges in the dielectric constants of the
five materials studied are less than the error of
measurement, Since the limit of accuracy of the
present mecsurements of dielectric constants is
5%, changes smaller than this volue would not be
observed, In the cose of Teflon there is a con-
sistent trend in the data which suggests that o
change in dielectric constant may occur,

 

22p. A. Weeks and D. Binder, Effect of Radiation on
the Dieleciric Constant and Altenuation of Two Coaxial
Cables, ORNL-1700, p 11 (Morch 2, 1954).

273
ANF PROJECT PROGRESS REPORT

4.2, FUEL RECOVERY AND REPROCESSING
. B. Lindaver

D. E. Ferguson

PILCT PLANT

J. E. Bigelow W. H. Lewis
F. N. Browder J. T. Leng
W. H. Carr F. W. Miles
R. B. Keely 5. H. Stainker

C. L. Whitmarsh

Construction of the fused salt—fivoride volatility
pilot plant is essentially complete, with the
exception of facilities for handling ARE fuel,
which will not be installed until runs with aon-
irradiated materiol ore completed.
of the fluorine disposal system it was found that
fluorine can be satisfactorily scrubbed from the

In studies

off-gas stream at normal process flow rates with
the use of only three of the seven spray nozzles
in the scrubber,

The molten-sait handling equipment was tested,
and the stainless steel vessel for melting the
chorge suffered heovy corrosion. After three test
melts were prepared, the salt outlet of the vessel
plugged. The stainless steel vessel will be
replaced with a nickel vesset. The remainder
of the molten-salt system {fluorinator, waste can
and carrier, avtoresistance heated lines, and
freeze valves) handled the molten salt in a manner
which will be acceptable for operations.

Heating- ond cocling-cycle studies were started
on the MNaF cbsorbers and on the cold traps.
Resuits of the first tests indicate that the heating
time for the absorbers as designed may be ex-
cessive and that neither the 5-in. K-25 cold trap
nor the &-in. ORNL.designed cold trap will reach
design temperature. Further studies are in progress
in order to determine the modificotions and op-
erating techniques necessary for satisfactory
operation.

In one of the test runs, the molten salt NaF-
ZrF -UF, (56-37.5-6.5 mole %, fuel 108) was
sparged in the fluorinator ot 600°C for 7.5 hr
with N, at a rote of 2 scfm. (The pleat flueri-
nation ¢ycle is expected to be 2 hr of sparging
with F, at a rote of 1.7 sefm.) An anclysis of
the deposit in the vopor trap indicated that salt
entrainment ond ZrF . vaporization from the salt
were of the some order of magnitude. Seme moterial

274

¥. K. Eister

H. E. Goeller

passed through the nickel mesh of the vapor trop;
this material could not be blown off the walls,
but it was easily wiped off.

SULFUR EMBRITTLEMENT OF NICKEL

M. R. Bennett E. E. Hoffman
G. I. Cathers R. L. Jolley
L. R. Trotter

The embrittlement of nickel and of high-nickel-
content alloys by sulfur was investigated chem-
ically and metallurgically to determine the
importance of the effect in volotility processing.
Experiments with fused salts to which known
amounts of sulfur were added showed that nickel
foil is sensitive to sulfur concentrations as low
at 5 ppm.  Under ordinary conditions, however,
in equipment with heavy nickel walls, a sulfur
concentration of 200 1o 500 ppm is necessary
to produce severe embrittlement. The sulfur effect
is cumulotive, as indicated in o -consultation
with an Internationsl Nickel Company repre-
sentotive, | ond there is danger of sulfur buildup

in fused-salt treatment vessels os a result of
processing many batches with small sulfur contents.
Evidence has been obtained, however, that the
sulfur content of the fused salt can be reduced
to a harmless level by one exposure to nickel
equipment. The mechanical effect of sulfur em-
brittlement wos found to be much less on lnconel
than on nickel.

Metallurgical examinations were mode of spec-
from 0.035-in.-wall nickel test capsules
thot had been exposed to sulfur-contemincted
fused salt mixtures for 3 hr at 700°C under the
conditions given in Table 4.2.1. The sulfur was
added in the elemental form and as Ne,SO, to
salt mixtures that initiclly contained less than

imens

50 ppm. The inner waolis of the capsules used
in all the tests showed a characteristic type of
cracking when bent, as shown in Fig. 4.2.1. No
apprecioble difference was evident in the degree

of damage to the walls of the capsules (specimens

 

]Con.*.ulmtiOn with E. M. Skinner of internotional
Nickel Company.
PERIOD ENDING MARCH 31, 1957

Table 4.2.1. Conditions of Tests of Nickel for C ond E) exposed to Na,50, or to free sulfur
Determining the Effect of Sulfur on the added to NaF-ZrF -UF, (50-46-4 mole %, fuel 30),
Mechanical Properties An odditional heat treatment for 50 hr ay 700°C in

Tast ismperatore: 7009 vacuum resulted in typical homogeneization (Fig.
Test poriod: 3 hr 4.2.2). Further evidence of the degree of sulfur
Sulfur added: 500 gpm embrittlement is shown in photomicrographs of
as-received moateriai and of specimens from two

 

 

Test Fused Solt of the tested copsules (Figs. 4.2.3, 4.2.4, ond
Copsule Compesition Sulfur Added As d’2'5)' ] ) ) )
SRR CoHaTE) In additicnal tests at G00PC, 50-mil nickel wire

specimens were exposed in nickel copsules to
A NGF-ZFF4 (55_50) E]Emflflffll SUHUf NGF'Zer (50'50 mOIB %, SUIT 31) WIT|:1 knOWH
additions of free sulfur or Na,50,. The nickel
B NaF-Z¢F , {50-50) Na, 50, wire was moderately embrittled in o 1-hr fest

in a fused solt mixture that contained 250 ppm

W NaReAr iy Ha,30y of sulfur os anhydorus Ne,30,, whereas 100 ppm
ey of sulfur caused slight embritttement in the some
E NeF -ZrF -UF, Elemental sulfur period. Repeated additions of 100 ppm of sulfur
(50-46-4) as N02504 ot l-ht intervals to the some salt

 

mixture produced severe embrittlement af o total

UNCLASSIFIED
Y-20136

 

Fig. 4.2.1. Results of Band Tests of Specimens Cut from Walls of Nickel Test Copsules Thot Were Exposed

to Sulfur-Contalnlng Fused Solt Mixtures, Specimen F is as-received moterial. {(Confidantial with coption)

275
ANP PROJECT FROGRESS REPORT

| 0D

' '1!!II:1T||']1I1!!H]I' .I"']L

UNCLASSIFIED
Y-20200

 

Fig. 4.2.2, Bend Test Specimens Shown in Fig. 4.2,1 After on Additional Heot Treatment of 50 hr ot 700°C

in Yacvum. Mote typicel homogeneization that resulted from the heat treatment. {Confidential with caption)

 

- - A
/ [
e o f
! o
-
. -
| S
-
» . = i
\..‘
o > >
- \' ;:
| -
- b =
-'"_-‘_-_\_
"
r
— = &
- - 1

Fig. 4.2.3. Specimen from Inner Wall of an As-Received Nickel Test Capsvle. Etchant:
500X,

276

003

a9

 

006

 

 

o
<
S

KCN-(NH,)58,0,.
 

PERIOD ENDING MARCH 31, 1957

 
  
 

  
 
  
  

UNCL ASSIFIED
¥- 20122

  
 

  

UNCL ASSIFIED
¥-20624

 
 

UNCLASSIFIED § 5o
Y- 20583 |

 

 

 

(@) ' -84 () 1

Fig. 4.2.4. Specimen from lnner Wall of Mickel Capsule A {See Table 4.2.1) That Was Exposed for 3 hr at
700°C to NuF-Zer (50-50 Mole %, 5alr 31) Contoining 500 ppm of Elemental Sulfur, (g) As-rested wall — note
nickel sulfide compound formation in grain boundaries. 500X. {6) As-tested ond bent, 250X. {(¢) As-testad, heat
treated at 700°C for 50 hr in vacuum, and then bent. 250X, Etchant: KCN-{NH,),5,0,. Reduced 11.5%. (Confi-

dantial with caption)

UMCL ASSIFIED 2
Y-20124

      
  
 

UNCL ASSIFIED
Y- 204625

UNCLASSIFIED
Y- 20582

  

   

s

t

tNEHE

?

3

 

g

= EEEER

o

&
2

T
250x

 

 

250X 1

(@), (&) i 4 - 4¢)
. I -

Fig. 4.2.5. Specimen from lnner Wall of Nickel Copsule B (See Table 4,2.1) That Was Exposed for 3 hr ot
700°C 10 Nch-sz‘l (50-50 Mole %, Salt 31) Centoining 500 ppm of Sulfur as Nazsod. fa) As-tested woll. 500X.
(6) As-tested and bent. 250X, (c) As-tested, heat treqted ot 700°C for 50 hr in vocuum, and then bont, 250X,

Etchant: KCN'(NHAJLFS?OB' Reduced 13%. (Confidentiol with capticn)

277
ANP PROJECT PROGRESS REPORT

sulfur concentration of 300 to 500 ppm. These
of 50-mil wire, which indicated that the
was slight at concentrations of

tests
embrittlement
less thon 100 ppm sulfur, were not considered
to be sufficiently seasitive because the em-
brittiement was limited to the outer surfoce of
the wire and there wos o large loss of sylfur inte
the wolls of the nickel container. Therefore a
more senstive type of test was performed ia which
a strip of 5-mil nickel foil (weight of 1- by 2-in.
sheet, ~1.5 g) wos immersed in 150 g of salt
contoined in o graphite reactor. A nifrogen at-
mosphere wos maintained above the salt. In
this test the sulfur concentrates in the nickel
foil, since the salt-to-nickel rotio is maintained
at about 100:1. A major portion of the Semil foil
is available for sulfur penetration in o few hours
in this test if a diffusion rate ot 0.4 mil/hr is
assumed. '

A 2-hr test of foil ot 600°C in a solt mixture
containing 100 ppm of sulfur os Na,50, resulted
in essentially complete loss of mechanical strength,
and the foil broke under very slight stress. The
embrittlement became progressively less as the
svlfur concentration was decreased to 1 ppm.
When ne sulfur was added to the salt, o foil im-
18 hr retained its mechanical strength,
but there was evidence of some embrittlement
A comporative test
Inconel foil in a salt mixture con-

mersed

at the salt-vopor interface.
with  S5-mil
100 ppm of sulfur resulted in almeost ne
embrittlement, in confirmation of previous ob-
servations that some types of aickel olloys are
more resistant than pure nickel to the sulfur
eftect.

The effect of sulfur on nickel is believed to
be dependent on the initial chemicol state of the
sulfur and its subsequent reactions, os well as
on the rate of diffusion of sulfur or nickel sulfide
through Comparafive time studies
indicate thot in the case of N02504, 1 to 4 hr
is required for moderately complete reaction.
This was confirmed by studying the behavior of
tracer S°% in an experiment in which o totol of

taining

the metal.

15 ppm of sulfur was added to the salt mixture.
The disapoearance of the 33 activity wos 90%
complete in 2 hr. Less thon 5% of the sulfur that
disappeared was found in the walls of the nickel
capsule, and thus it oppears that only a small
part of the sulfur present in the fused salt actuolly
causes the embrittiing effect. A sulfur material

278

balonce experiment with an 'nconel capsule also
showed that only a fraction of the sulfur actually
transferred from the salt to the metal. With 100
ppm of sulfur in NoF-ZrF, at 600°C and o salt-to-
metal ratio of 160:1, the sulfur actually feund in
the Inconel after a 4-hr exposure accounted for
only 3% of the sulfur in the original salt mixture.
The sulfur content of the Inconel increased from
less than 130 ppm to more than 630 ppm, as
determined by activation analyses.

OXIDATION RESISTANCE OF ALLOYS IN
FLUCRIDE SYSTEMS

M. R. Bennett G. I. Cathers R. L. Joliey

The gas-phase oxidation resistance of types
316 and 347 stainless steel was found to be quite
unsatisfactory in tests of their use as passible
container materials for molten NaF-ZrF,-UF,
at  1200°F. Inconel was also attacked to an
unusval extent in these tests, Nickel showed
the least corrosion effect. This study was mode
to evaluate the corrosion effects observed in one
of the stainless steel vessels for melting salt
The low
oxidation resistonce of stainless steels in contact
with fluoride-bearing vapor is in merked centrast
to the wusual resistance of stoinless steels at

mixtures in the volatility pilot plaat.

high temperatures. The fluoride presumably acts
as o destructive agent for the protective film of
chromium oxide believed to exist on stainless
stee| surfaces.

The tests ceonsisted in helding comrosion speci-
mens partly immersed in molten NoF-ZrF -UF,
{nominal composition, 48-48-4 mole %) at 650°C
while bubbling gos through the solt mixture. The
first test wos made with air that hoed been passed
through Drierite to determine whether the stoinless
steels retoined any of their typical corrosion
resistance under these coaditions (Fig. 4.2.6).
After 2 hr all the iron-centaining alloys (stainless
steels and Inconel) had scole above the salt line.
The scale, which was strongly magnetic, was
thick on the stainless steels, in contrast to the
small amount on the Inconel ond none on the
A second test with nitrogen dried with
Drierite showed similar results after 16 hr (Fig.

4.2.7).

out with helium, there was, unfortunctely, some

nickel.
ln this test, as in several runs carried
back-diffusion of air into the reactor as o result

of the low flow rete and o leak at the flange.
Another test with nitrogen at a higher flow rate
 

I
3 3

NICKEL INCONEL TYPE 316

 

f 2

PERIOD ENDING MARCH 31,1957

UNCL ASSIFIED
1 FHOTO 19471

490 °C

520°

580°

Bl0°

—~ SALT LINE ,650°C

 

TYPE 347

STAINLESS STEEL STAINLESS STEEL
3

3

OaK RIDGE NATIONAL LABCRATORY

 

Fig. 4.2.6. Specimens That Were Partly Immersed for 2 hr in NuF-ZrF‘-UF4 Through Which Air Wos Belng

Bubbled.

to minimize air bock-diffusien resulted in prec-
tically no buildup of scale in 26 hr. For this
test the nitrogen was dried with Anhydrone and
then passed through copper foil ot 750°C to remove
any traces of oxygen.

In all tests in which there was much scale
formation, there was alse an accumulotion of
solids around the bottom of the specimen below
the salt surfoce (Figs. 4.2.6 ond 4.2.7). This
material contained ZrO,, as determined by petrog-
raphic examination, possibly as a resuit of iron
or chromium oxides dropping into the sclt from
the gos phase,

POROSITY AND DENSITY OF NoF

M. R. Bennett G. |. Cathers
R. L. Jolley

- Porosity values were colculated from density
measurements on three lots of NaF cbtcined from

The Harshaw Chemical Co. No significant dif-

ferences were found among the three lots, one
of which wos 12-20 mesh materiol to be used
in the pilot plant and the others were }‘é-in.
pellets.  Particle, bed, ond over-all porosities
of 48, 33, and 65%, respectively, were found for
the 12-20 mesh granular material. A crystalline
type of NoF obtained from Allied Chemical & Dye
Corp. was observed to be much denser, heving
an over-all porosity of 35 instead of 65%. The
over-ctl porosity is defined os the percentage of
totel void space (bed voids plus particle voids}
in o bed of the material, The porosities were
colculated from deferminations of the bulk densities
and the porticle densities as given in Table 4.2.2,
along with the use of the crystolline or x-ray
density of 2.79 g/Cma. Weight-volume and dis-
placement methods of meosuring density were
used in determining both the butk and the particle
densities of the .'ré-in. pellet material. Mercury
displacement wos used in finding the porticle
density of the 12-20 mesh NeF from Harshaw.

279
ANP PROJECT PROGRESY REPORT

! 3 : 490°C
530°
580°

‘ : -] 810°

! o * g

n'i ‘ I-. ‘fi?‘

3 J v
NICKEL INCONEL TYPE 316 TYPE 347

STAINLESS STEEL STAINLESS STEEL

i T
! Z 3 .
0ax RIDGE NATIONAL LABORATORY |

 

UHCL ASSIFIED
PHOTS 19470

SALT LINE ,8650°C

Fig. 4.2.7. Speclmens That Were Partly Immersed for 16 hr in NoF-ZrF -UF , Through Which Nitrogen Was

Being Bubkhled,

Toble 4,2.2. Results of Density Detarminotions on Several Lots of NaF

 

Farticle Density (gfl:m3}

 

 

Bulk Density
Materiol Weight-¥olume Displocement (gj,.-cm3)
Method Method

The Hershaw Chemicol Co.

%-in. pellets, batch 2 1.47 .48 0.94

Y-in. pellers, batch 4 1.49 1.47

12-20 mesh pilot plont moteriol 1.45 0.97
Allied Chemical & Dye Cerp.

20-40 moesh material, bateh 2955 2,62 1.81

 

280
The benzene displacement methed had to be used
with the finer Allied material. The weight-volume
and displacement methods gave resylts that were
in close agreement. It is onticipated that the
porosity data will be useful in interpreting the
heat «nd material transfer problems associated
with the absorption-desorption process.

RUTHENIUM DECONTAMINATION

M. R. Bennett . |. Cathers
R. L. Joiley

An experiment was conducted to fest whether
the plating out of ruthenium on metal surfaces
could be used to achieve odditional decontam-
ination in the wolatility process. The column
used for the experiment consisted of a 12-in.
length of 1-in.-0OD nickel tubing fitled with 192 g
of %z-in.«OD 30-gage nickel Fenske packing
with o total area of 0.36 m?. This column was
placed between the second NaF bed and the UF
celd trap and was held at 120°C. It was vused
for about 130 min in o desorption run, and it was

PERIOD ENDING MARCH 21, 1957

not particularly effective in removing activity
from the UF, product coliected. The decontam-
inaticn factor, calculated on the basis of the
activities found in the nicke! column ond in the
final product, was only 1.6. The activity in the
nickel packing was removed with an alkaline-
tartrate-péroxide solution, and the effectiveness
of this method has not been fully evaluated. About
0.025% of the processed uronivm was found in
the nickel trap.

The low ruthenium decontamination in the nickel
trap is perhaps not too surprising in view of the
small amount of activity involved and the small
surface area used. The product contoined only
28 ruthenium gamma counts per minute per mil-
ligram of vrantum. The activity removed by the
nickel was calculated to be 200 counts per minute
per square centimeter of nickel surface. It is
anticipated that the removal of ruthenium with
a nickel ¢olumn moy be made more effective by
using a greafer surface area or by operating at
a higher temperature.

281
Part 5

CRITICAL EXPERIMENTS AND REACTOR SHIELDING

E. P. Blizard
5.1. CRITICAL EXPERIMENTS
A. D. Caillihan

REFLECTOR-MODERATED REACTOR
EXPERIMENTS AT HIGH TEMFERATURES

The significant results of the experiment are as

follows:

Critical concentration

D. Scott E. Demski'
R. E. Malenfant W quer'l At 1258°F with rod out 10.97 wt % U (5.1 mole %
J. J. Lynn D. A. Harvey' . UF,)
W. C. Tunnell E. V. Sandin At 1259 F with rod in- 12.20 wt % U
W G Aiwungl sarted te midplane
Rod value with red in- $2.47

The measurements of some of the nuclear charac-
teristics of the circuloting-fuel reflector-moderated
reactor under study by Pratt & Whitney Aircraft
were completed. The ossembly consisted of o
cylindrical beryllium central section or ‘‘island"”
surrounded by an anaular fuel region, which was
enclosed in o reflector of beryllium. Shells of
Hastelloy X separated the fuel from the beryllium,
Control and safety of the experiment were effected
by a rod which could be positioned along the axis
of the island. The neuvtron-acbsorbing section of
the rod c¢ontoined a mixture of 30% rare-earth
oxides (somarium ond gadolinium) and 70% nickel.
The reoctor section of the assembly, which was
described in detail in an earlier report,? was
mounted, with its oxis vertical, above o sank in
which the molten mixture of sodium, zirconium,
and uranium flyorides was stored. The liquid
could be roised into the reactor by pressurized
helium. The equipment was operated ot tempera-

tures neor 1250°F.

ln  the experimental procedure followed, the
system was first filled with NaoF-ZrF, (50-50
mele %) to which increments of NGF‘UF“ {66.7-
33.3 mole %, contoining uranium enriched to
93.2% U23%) were added until the system became
criticol. The concentration was further increased
stepwise as the control rod was inserted to the
midplane. The value of the rod over the upper hal
of its travel (Fig. 5.1.1} was determined from
period measurements and the inhour equation. The
temperature coefficient of reactivity over the range
1200 te 1350°F was linear, within the precision of
the medasurements.

 

]On assignment from Pratt & Whitney Aircroft.

2D, Seott ei al, ANP Duan Prog. Rep. Sept 10,
1956, ORNL-2157, p 264.

serted to midplane
Temperature coefficient
in ronge 1200 1o 1350°F
(based on a delayed
neutron fraction B =

0.0073)

—0.47 ¢/°F, =3.6 x 10~
(A R/E)/CF

A similar experiment was performed in 1955 with

the ART core.?

The critical

concentration ob-

served then with the rod removed was 6.30 wt % U

(2.87 mole % UF4} and the tempercture ccefficient
at about 1250°F was ~2.3 x 10—° (Ak/k}/°F.

 

A, D. Collihan et al., ANP Quar. Prog. Rep. Sepr. 16,
1955, ORNL-1947, p 58; ANP Quar. Prog., Rep. Dec. 10,

1955, ORNL-2012, p 71.

 

 

 

 

 

 

e
ORNL-LR-OWG 20263
250 : |
/
7
f
=
g 150 | - { .
S | L
?
= =[
: ‘ g
o 2
g =
=g
I&J'IOOl - —— | i 1
| ;/’ ‘
/|
i
0 |
o 8 10 t5 20 25
ROD POSITION, MEASURED FROM TOR QF ISLAND BERYLLIUM fin)
Fig. 5.1.1. Effectiveness of the Contral Rod in the

Pratt & Whitney High-Temperature Reflector-Moderated-
Reacter Critical Assembly.

285
ANP PRDJIECT PROGRESY REPORT

5.2. SHIELDING THEORY

SUMMARY OF MONTE CARLO CALCULATIONS
OF GAMMA-RAY PENETRATION IN
MULTIREGION SHIELDS WITH
SLAB GEOMETRY

S, Auslender! A. T, Futterer'

Some of the calculations of gamma-ray shield
penetration thot were performed with the use of an
Oracle cede based on Monte Corle techniques have
been published at intervals? during the past year.
A topicel report is now being prepored which will
describe the Oracle code used for these calcula-
tions.2 In addition, the results of all the colcula-
tions are being collected in o separate repert con-
sisting of several volumes.? Since the results are
especially opplicoble to the ANP program, o
summary of the various types of problems is pre-
sented here, along with corresponding references
to the data.

The summary {(Table 5.2.1} designates the ma-
terials, thicknesses, gamma-ray energies, and
ongles of incidence investigated. For oll the
problems the information obrained directly fram the
machine calculetions includes dose rate, energy
flux, and energy deposition; for some problems it
olso includes energy spectra. Dose- oand energy-
buildUp “factors, as well as heating in the shield,
may be computed directly from the infermation, and
in some cases they have been,
of dum!“ comparisons of the shield atfenuctions
are also made.

Two studies, which have been designated the
CE137 and COB&0O

carried out to determine the penetration of Cs

In the compilation

respectively, were
137

series,

 

on assignment from Pratt & Whitney Aircroft,

25. Auslender, ANP Quar. Prog. Rep. March 10,
1956, ORNL-2061, p 223; Juwe [0, 1956, ORNL-2106,
p 266; Sept. 10, 1956, ORNL-2157, p 271; see olso
5. Auslender, Appl. Nuc. Phys. Ann. Rep. Sept 10,
1956, ORNL-2081, ¢ 180; see olse 5. Ausiender, A
Monte Carlo Study of the Gamma-Ray Ernergy Flux,
Dose Rate, and Buildup Factors in a Lead-Water Slab
Shield of Finite Thickness, ORNL-2194 (lan. 10, 1957).

3. Auslender, Instructions for the Operation of an
Oracle Code for a Monte Carle Solution of the Transport
Problem for Gamma Rays [ncident Upon a Slab, ORNL
CF-57-6-1 (1o be published).

S, Auslender, Compilation of Monie Carlo Calcula-

tions of Gamma-Ray Penetration in Multiregion Shields
with Slab Geometry, ORNL-2310 (tc be published).

286

and Co®® gamma rays through slabs of polyethylene
and lead. The primary purpose of these calcyla-
tions wos to establish the volidity of the calcula-
tional method by comparing calculated results with
The porticular
experiment chosen was performed by the National
The results were in good

results obtained in an experiment.

Bureay of Standords.®
agreement.

Ancther calculation, designated the FAB series,
was performed to determine the effect of the extent
of lamination in a shield consisting of a total of
16 ¢m of water and 4 cm of lead. The materials
were subdivided and alternated {2 to 16 regions) to
mock up the progression toward a homogenecus
shield.

In another group of problems, the DO2C5S series,
the effect of chonging the relotive amounts of lead
ond water in a two-region shield that had o con-
stant mean free path thickness was investigated.
In this case no transmitted or reflected energy
spectra are avoilable, as yet.

Two sets of problems identified as the AD and
the AB series, respectively, represent an exten-
sive study of shields consisting of 7, 14, ond
21 cm of water with and without a 0.1-in.-thick
lead hacking.
seven gamma-ray energies and seven incident
angles. Inifial histories range from 1000 to 2500
for incident angles less than 60 deg. For greater
angles the number of histories ranged from 400 to
800 10 avoid excessive machine time. In o]l cases
a doubling technique was used which tends to
maintain an even populotion throughout the shield.
In this way statistics were obtained near the finol
boundory for which excessively lorge
histories would have been required if this tech-
nique had not been used. The trensmitted and
reflacted energy spectra from this study have been
stored in o systematic order on magnetic tope and
are available for calculations. The energy spectra
give the fraction of the incident energy which may
be expected within each of 120 energy—solid angle
intervals (that is 10 epergy by 12 angular inter-
vals) for both the reflected and the transmitted
energy of each of the 294 cases.

Caoleulations were performed for

initiel

 

SE. S, Kirn, R. J. Kennedy, and H. 0. Wycksff,
Radiology 63, 94-104 (1954),
187

Table 5.2.1, Summary ond Index of Monte Carle Caolculatiens of Gammae-Ray Penetration in Multireglon Shields with Slab Geometry

 

 

Gammae-Ray Energy

0, Angle of Gammo-Ray

 

 

Series Shield (Mev) Incidence (deg) References
CE137 (for Cs'37)4 0,2,5, 9, or 14 in. polyethylens 0.661 0 ORNL-2310
+0,0.25, or 0.5 in. Pb
COB6D (for Co®?) 0, 2, 5. 9, or 14 in. polyethylene 1.25 6 ORNL-2310
+0, 0.25 or 0.5 in. Pb
FABb 16 em HZD + 4 ecmPh
(8 em H,0 + 2 cm Pb) x 2
(4 cm H,0 + 1 cm Pb) x 4
(2 em Hy0 =0.5 em Pb) x8 1.0, 3.0, 8.0 0 ORNL-2310
4 cm Pb 4 16 em HQU
(2 em Pb +8 cm H,0) x 2
(1 em Pb+ 4 cm H20] ® 4
{0.5 cin Pb+ 2¢m H,0) %8
DO2CS5 1 mfp Phb ORNL-2061, p 223;
0.75 mip Pb + 0.25 mip H,O 0, 60, 70.5, 75.5 ORNL-2081, p 180;
0.5 mfp Pb + 0.5 mfp H,0 1.0, 3.0 (sec O=1, 2, 3, 4) ORNL-2310
0.25 mtp Pb +0.75 mip H20
1 mfp HQD
0.75 mfp H,0 +0.25 mfp Pb 0, 60, 70.5, 75.5 ORNL-2061, p 223;
0.5 mfp H,0 + 0.5 mip Pb 1.0, 3.0 (sec 0=1,2,3, 4 ORNL-2330
0.25 mfp H,0 +0.75 mfp Pb
2 mfp Ph
1.5 mfp Pb + 0.5 mfp H,0 0, 60, 70.5, 75.5
1 mfp Pb + 1 mp H,0 1.0 (sec =1, 2, 3, 4) ORNL-2310
0.5 mip Pb + 1.5 mfp H,C
2mfp H,0 ORNL-2061, p 223;
1.5 I'I"I'FP HZD +0.5 I'I'IFP Pk 0 D, 60, ?0-5, ?5-5 0RNL'20813 P 130,
1.0 mip H)0 + 1.0 mép Pb I (sec B=1, 2, 3, 4) ORNL-2310
0.5 mip HEO + 1.5 mfp Pb
8 mip Pb
& mfp Pb + 2 mip B,0 0, 60, 70.5 orNL
4 mfp Pb + 4 mfp Hy0 3.0 {sec 6=1, 2, 3) NL-2310
2 mfp Pb + &6 mip H,O

 

L6561 “1f HOYVYW ONIONI AQoid3d

ZNg "*s' available on machine.

®Constant total shield thickness with layers censecutivel y subdivided and alternotad to mock up progression from a two-region shield to o homogeneous
shield.
882

Table 5.2.1 {continued)

 

 

' _ Gaomme-Ray Energy 0, Angle of Gamma-Ray
Series Shield (Mewv) [ncidence (dag) References
AQ 7,14, or 21 cm HIO 0.5, 1.0, 2.0, 3.9, 0, 20, 40, 60, 70, QORNL-2310
4.0, 6.0, B.0 80, 85
AB 7, 14, or 21 cm H,0 + 0.1 in. Pb 0.5, 1.0, 2.0, 3.0, 0, 20, 40, 0, 70, ORNL-2310
4.0, 6.0, 8.0 80, 85
B¢ 10,0r 20 in, HQO 1.0, 2.23, 2.70, 4.0, 0, 60, 70 ORML-2310
6.814, 9.0
40 in. H,0 1.0, 2.70, 6.814, 9.0 0, 60, 70 ORNL-2310
2 in. Pb +0, 10, or 20 in. H,0 1.0, 2.70, 4.0, 6.814, 0, 60, 70 ORNL-2310
9.0
2.23 0, &0 ORMNL-231%0
2in. Pb + 40 in, H'ZO 1.0, 2.70, 6.814, 2.0 Q, 60, 70 ORNL-2310
4 in. Pb+ 49, 10, or 20 in. HED 1.0 o ORNL-2310
2.23, 4.0, 6.814, 9.0 0, &0 ORML-2310
2.70 0, 80, 70 ORNL-2310
4 in. Pb +40 in. H,0 1.0 O ORNL-2310
2.7 0, 60, 70 ORNL-231%0
6.814, 2.0 a, &0 ORNL-2310
4 in. Pb +0, 10, or 2D in. Hzo 2.70 0, 60, 70 ORNL-2310
4.0 0 ORNL-2310
6.814, 2.0 0, a0 ORNL-2310
6 in. Pb 4+ 40 in. H,0 2.70 0, &0, 70 ORNL-2310
6.814, 9.0 0, 60 ORNL-2310
Undesignot ed 1158 em Pb + 35.81 cm H20 1.0, 3.0, 6.0 0, &0 ORNL-2157, p 277;
ORNL-2081, p 184;
ORNL-2194; ORNL-2310
Undesignated 3.79 cm fuel? + 0.976 cm Incone 1.0 0, 45, 40, 70.5, ORNL-2106, p 266;
+21.19 em Na + 0.488 cm 75.5 ORNL-2310

Inconel (0.7 mip fuel® + 1.0 mtp
Inconel + 1.0 mip Ne + 0.5 mfp
Inconel}
“This series hod a maximum of 800 histories ond a minimum of 400; a doubling technique was used to improve statistics. Even so, the statistics cre
poor for the thickest slabs ot slont penetrations os far as energy spectral dotu are concerned,
dNoF-Zer-UF‘s {52.5-42.5-5 mole %).

| 1
N + »

l¥0d3Y SS5FHD0Ad LI3f0Hd dNV
Another set of problems, the B series, is a study
of thick lead and woter shields. The energy
spectra for this series were also stored systema-
tically on magnetic tape. It is to be noted thot
the low number of initial histories combined with
the large thickness of the shields tends toword

PERIOD ENDING MARCH 31, 1957

poor statistics for the energy spectra informetion
despite the doubling technique.

Two additional problems, one for a fuel-Irconel-
sodium-Inconel shield and another for a lead-water
shield, were also computed. The fuel mixture

studied was NaF-ZrF ,-UF, (52.5-42.5-5 mole %).

289
ANP PROJECT PROGRESS REPORT

5.3. LID TANK SHIELDING FACILITY

NEW INSTRUMENT CARRIAGE AT THE LTSF
W, Zobel

The necessity of toking progressively more pre-
cise data ot the Lid Tank Shielding Facility (LTSF)
made it obvious thot one of the uncertainties which
had to be eliminated was that associated with the
location of the detectors, While the position of a
detector in the tank could be reproduced to within
about £2 mm, the absolute position was open to
doubt within about +5 mm, It was therefore de-
cided to replace the detector holder and positioner.

The new equipment was required to be more
rigid than the old so that the absolute position and
the indicated position would always coincide, and
it had to hove an inherent accuracy of about £0.5 mm.
The system (Fig. 5.3.1} which wos adopted is very
similar te that in use at the Tower Shielding
Focility (TSF). This system provides for remote
contrel of the carriage from the control room and
thereby greotly increases the ease of operation,
The new instrument carriage has been installed
and is now in use. |t appears to give very satis-
factory service, and it is expected that the con-
fidence level of future data taken at the LTSF will
be increased.

The detector carriage is driven in the y-direction
by two ball-bearing lead screws which, however,
do not support the carriage. Thus greater accuracy
and longer life can be obtained. The use of the
ball-bearing nut eliminstes backlash. The load
is carried on the channet base, which con be made
as strong as required, A guide rail ¢n one side
assures that the carriage will run true,

In the z-direction the load is supported by two
large shafts, which also serve as guides. The
drive is ogain provided by a ball-bearing lead
screw.

The drive in the x-direction is of the rack-and-
pinion type, with the rack cut info the vertical
drive shoft, This is o large tube ond was chosen
to provide sutficient rigidity for the holder.

Contrel of the drive motors in the three directions
is manual at present. The position of the detector
suppert, located at the end of the vertical shaft,
is indicated at three panels in the control room,
Two indicators are provided for each direction.
One is a dial for coarse control, driven by a torque
transmitter system geared to the drive motor; the
other is a fine control, indicated by an Amerac

290

dial, which uses a selsyn system to drive the dial
by the usuel null technique., The indication of
the fine control is to the necrest millimeter, with
interpolation on the dial to 0.5 mm eosily possible.
Provision has been made to permit later conversion
of the conirols to a closed-loop system by which
the desired position on the detector con be selected
and the corrioge will then “‘hunt’’ until it reaches
that location. Provision is also made for future
installation of ¢ readout system for use with an
isodose plotter,

Preliminory tests showed that the instrument
carrioge, as finally constructed does not deviate
from a given plane by more than 0.006 in. over the
full range of the water tank at the LTSF. This
indicates that the aobsolute position of an instru-
ment can be determined to at least +5.5 mm and is,
at present, limited by the accuracy with which the
indicators can be read. It is {elt, however, that
+0.5 mm is sufficiently accurate for the purpose,
and therefore no changes are contemploted,

INVESTIGATION OF THE FHYSICAL
PROPERTIES OF LITHIUM BOROHYDR IDE -
AMMONIA UHDER SUBCONTRACT NO. 390

E. A, Sullivan S. Johnson
Metal Hydrides, Inc.

One of the materiats which has been suggested
as a neutron shield for a mobile reactor is lithium
borohydride,  This relatively new compound is
particularly attractive because it contains 18.5wt %
hydrogen, which is the most efficient lightweight
neutron abseorber known, as well as lithivm and
boron, both of which suppress secondary gommo
radiation arising from neutron capture by the shield
material, in oddition to being good neutron ab-
sorbers, Lithium borohydride is a white crystalline
selid which is stable to its melting point of 275°C
and has essenticlly no vapor pressure even of this
elevated temperature, In on attempt to achieve a
liquid shield which could be cooled externally and
would avoid difficult fabrication problems, empha-
sis has been cltered to consider shield components
utilizing this compound in combination with an
ingxpensive diluent that would not change fhe
Ammonia, which
contains 17.8 wt % hydrogen, is ideclly suited for
this purpose.

shielding properties appreciably,

Since the physical properties of
(62

UNCLASSIFIED
ol QRHL-LA-DWEG 2379

"XTAXIS DRWE SHAFT —-

Ty

  
 
 
    
 
    

"r™AXIS DRIVE 80X -

- ZANIS GUIDE AND LOAD
| CARRYING SHAFTS

r-a%IS COORDINATE BO)CT

  

W' COORDINATE
and DRIVE BOX

 

P - THIS SIOE FACES REACTOR WaLL

 
 

"WAXIS DRIVE PINION MATING WITH

RACK IN "¥'LAXIS DRIVE SHAFT - "FAXIS LEAD SCREW

     
 
   
 

  

"Z'-a%iS LEAD SCREW —
"F-p0s ORIVE BOX ——

 
   
 
 
    

e ADDITIONAL "ZAXIS LOAD
CARRYING MEMBERS

“Z"-AXIS COORDINATE BOX
e

"YUAxIS LEAD SCREW MUT -

   
 

11\\

“Plaxi§ LEAD SCREW -~ — - CHARNEL BASE

  
   

HORIZONMTALLY ~MOUNTED
Y-85 LOAD CARRYING RAIL: PCIIRRER

"r'2xi% GUIDE RAIL - ——k

Fig. 5.3.1. New LT5F Detector Carrioge.

£S61 “1E HOXYYW ONIGNI 00CI¥3d

 
ANP PROJECY PROGRESS REPORT

shield materials must also be known, an investi-
gation hos been carried out to determine the equi-
librium vopor pressure, the thermal stability, and
the density of varicus lithium borohydride ammoni-
ates. The results of the investigetion have been
reported in detail! ond ore summarized below.

Pressure-Temperature-Compos ition Relationships

The ammonia vapor pressure in eqguilibrium with
mixtures of lithium berchydride and ammonia was
determined os ¢ function of concentration ond of
temperature, The lithium borohydride mole frac-
tions used ranged from 1.0 to slightly less than
0.2 in the temperature range O to B0°C. Lithium
borohydride forms four stable adducts with gmmenia:
lithium borohydride monoammonicte, LiBHd-NHB,
which has a melting point of 56.5°C; lithivm boro-
hydride diammoniate, LiBH .2NH,, which melts
ot  17.0°C; lithium borchydride trismmoniate,
LiBH ,-3NH,, which displays an incongruent
melting point ot 32.5°C; and lithium borohydride
tetroammoniate, LiBHd-dNH}. which melts ot
approximately 68°C,

The phase diagram in this temperature ronge has
been constructed from the experimental dato, The
behavior of lithium borohydride —ammonia adducts
is such that two eutectic compositions, LiBH .
1.?5NH3 ond LiBH ,-2.35NH,, ore completely
liquid abave 13°C, A solid solution of the ftri-
ammoniate in the tetraammoniate was found to
exist which is stoble vp to the melting point of
the tetracmmoniate.

The ammonio pressure can readily be calculated
for gny temperoture and composition from eguations

of the type
A
log P{mm Hg) = —-—?:1- B .

In studying the lithium borohydride —ammonia
system, it was found that an interaction occurred
to a smaoll extent between the components, which
This
ammonolysis reoction was studied briefiy in the
temperature range of 30 to 100°C, and it was
found that the rate of ammonoclysis of fithium boro-
hydride: (1} decreases with time, {2) increases

gave, as a byproduct, gaseous hydrogen.

 

]Eu A. Sullivon and 5. Jdohnason, The Lithium Doro-
hydride—Ammonia System, Part |, Pressure-Temperature-
Composition Relationships. Part I, Density Relation-
ships, Metal Hydrides, Inc., report identified as
CF-57-2-134 ot ORNL (Feb. 12, 1957).

972

with temperoture, (3) decreases with increosing
ammonia concentration {and is essentially consiant
beyond an ammonia-to-lithium borohydride mole
ratio of 2:1), (4) is approximately on exponential
function of composition at a given temperatue,
and (5) is approximately on exponential function
of reciprocal temperature ot o given composition,
Where necessary {0t temperatures >40°C), small
corrections were made in the equilibrium data to
account for the effect of cmmoneolysis on the com-
position of the somples. When precautions were
taken, this reaction did not interfere with equilib-
rivm measurements,

Density Relationships

The densities of liquid lithivm borohydride
ammoniates were determined as a function of both
composition and femperature. Five ceonditions
between LiBH ,+1.06NH,; and LiBH .2.74NH , were
studied. The temperature range in which each
compaosition wos studied depended on the liguidus
temperature at the given composition and on the
vapor pressure encounterad, Two of the composi-
tiens studied, Li|:.‘>H4~'f.7"3NH3 and LiBH4-2.36NH3,
correspond to the ewtectics on either side of the
diemmonigte of lithium borohydride. These com-
positions hove the lowest liquidus temperatures
in the entire phase diagram, both ot 15°C, and
therefore have the longest temperature ronges for
the liquid state. la generol, the study encompassed
temperature ranges between 15 and 85°C.

The density for o given composition was found to
be a linear function of temperature, os expected,
and the experimentally determined equeotions are,
where ¢ is in °C:

(1) LiBH +1.06NH,:
(2) LiBH - 1.73NH,:
(3) LiBH,+2.04NH:

d,=0.651 - (2.537 x 10~
d,=0.663 — (3.683x 10~*):
d,=0.675 - (4.664 x 10~*);
{4y LiBH,«2.36NH,: 4, =0.676 - (4.201 x 10~4);
(5) LiBH+2.74NH: o, = 0.684 — (4.427 x 10~ *)

With the exception of L]BH4-2.04NH3, a simple
relationship was found to exist between the com-
poesition and both the slope and intercept of the
density equations. The following empirical equa-
tions were found to relate these foctors:

(6) Inteccept = 0.629 + 0.020(c, . )
where ¢, o is the composition expressed as the
mole ratio of ammonia to lithium borchydride and
(7) Slope = —1.97 x 10~* + 8.80 x 10-4(NHH3}
where 1"‘«',.”43 is the composition expressed as mole
fraction of ammonia. From Eqs. 6 ond 7, the
constants for density equations were calculated
which cover the entire liquid range.

SURVEY OF SHIELDING MATERIALS UNDER
SUBCOMNTRACT MO, 931

R. Arcnson K. Held
Technical Research Group

A recent survey of shielding moterials confirms
that lithium hydride and polyethylene are the mast
effective neutron shields known to dote. Lithium
hydride is by for the best solid shield material in
curved geometries. The use of Li®H is expected
to lead to a further weight savings of approximately
12%. Lithium hydride is the only lightweight neu-
tron shield material which can be used in the range
of 700 to 1200°F without developing high pressure
or decomposing. This is the range of temperafure
which may be expected in some parts of the direct-
cycle recctor shield. Neutron shield moterials for
use at temperatures above 2000°F may be required

for the direct-cycle reacter, Reoctor outlet air at

PERIOD ENDING MARCH 31, 1957

o temperature of 1850°F is used to cool the neutron
shield at the rear foce., Here the requirement is so
severe that the problem is one of obtaining ma-
terials that will withstand the high temperature,
rather than a minimum weight problem. M a liquid
is required because of the need for shield augmen-
tation or some other reason, cne of the substituted
organic compounds 2 must be used.

Polyethylene of high density (>0.95 g/cm?) and
of high strength is now available. Some analyses
of wvarious crew shield designs indicate the very
fow requirement for additionol structural support
for a lead-polyethylene shield, Other materials
including wverious lithium hydride—polyethylene
mixtures are under consideration,

The high-density materials required for the re-
actor gamma-ray shield still center around lead,
bismuth, tungsten, uranium, etc. Possible appli-
cation of hydrides, borides, etc., of these elements
await clarificotion of the secondory gammo-ray
picture,

Contact has been maintained with Metal Hydrides,
Inc., on their studies on the lithium borohydride—
ammonia system,

 

2Ff.:)r example, boran or lithium substituted into the
organic molecule,

293
ANP PROJECT PROGRESS REPORT

5.4. TOWER SHIELDING REACTOR Il

C. E. Clifford

As was stated previously,! a reoctor which will
have a higher and mere uniform fast-neutron and
gamma-ray leakage flux than that ovailoble with
the present Tower Shielding Reccter is needed for
future aircraft shielding experimentation. To fill
these requirements a 5-Mw solid-fuel Tower
Shielding Reactor |l {TSR-11} has been propesed
which will be spherical in shape. The TSR-Il will
be controlled by umbrella-shaped beoron-looded
plates located in an interna!l water reflector, in
which position they will give a minimum perturba-
tion to the reactor leckage flux,

In order to establish the dimensions of a core
configuration that would fit the requirements of
the reactor, the following calculational procedures
were followed. A nuclear parometer study was
carried out first to determine the critical moss
and control available for various core sizes,
internal woter reflector thicknesses, and aluminum-
to-woter volume ratios. From this study o configu-
ration was selected which appeared to be the
optimum configuration from the viewpoint of critical
mass ond small reactor diameter. This configuration
was then examined to ensure thot the surface
density of the U239 in the fuel plates would be
betow 0.0283 g/cm?, a density which was estab-
lished during operating experience with the MTR.?
Finally the configuration was exomined to deter-
mine whether the waoter flow available fo remove
the heot from the core was sufficient to prevent
local boiling. The results of these tests are
presented below,

NUCLEAR PARAMETER STUDY

Colculations,? for which the three-group, three-
region recctor code (the 3G3R code) wos used as
a two-group, three-region (2G3R) code, were per-
formed on the Oracle to determine the critical
mass and control ovailable in reactor configurations

 

lsee, far exomple, C. E, Clifford ond L. B. Helland,
ANP Quar, Prog. Rep. Dec. 31, 1956, ORNL-2221, p 352,

2R.‘ 4. Beover, ORNL Metaliurgy Division, private
communication,

¥ These colcutations were performed by M. E. La¥erne;
pravious calculations by E. G. Silver were helpful in
setting up the present problems, see Appl Nuc. Phys.
Ann. Rep. Sept. 10, 1956, ORNL-2081, p 47.

294

L. B, Holland

having internal water reflector diameters of 32,
40, and 50 cm and core thicknesses of 18, 20, ond
25 cm. The oluminum-to-water volume ratios in
the cores ranged from 0,250 1o 2.000. All possible
reflector-core combinations were investigated for
eoch oluminum-to-water ratio. For the calculations
the external water reflector was considered to be
20 cm of woter rather than the lead-water combi-
nation that is proposed for the actual reactor. A
sketch of the geometry used in the calculation is
shown in Fig. 5.4.1, ond socme of the constants
are presented in Toble 5.4.1,

Calculations were made first to determine the
effective multiplication factor for severcl concen-
trations of U233 in the fuel plates. Typical results
for one geometry ond all aluminum-to-woter volume
ratios are shown in Fig. 5.4.2, Earlier calculetions
for this reacter showed that 4% excess reactivity
would be necessary to compensate for temperature
rise, xenon poisoning, and burnup ond to provide
1% for control. The concentration that gave an
excess reactivity of 4% in a perticulor configuration
was used to compute the mass of U233 that would
be required for thot canfiguration.

In Figs. 5.4.3 through 5.4.5 the critical mass is
shown as o function of internal reflector radius for
voricus aluminum-to-woter volume rotios and oll
As expected, the
smaliest internol reflector and lowest aluminum-
fo-water voelume rotio will give the minimum critical
mass.

core geometries considered.

The omount of excess reactivity that it is pos-
sible to control for the various gecmetries was
determined next. This calculction was performed
in o stroightforward manner (os provided in the
3GIR Orocle code) by the insertion of a boron-
loaded sheil between the internal reflector and the
core region. The results shown in Fig. 5.4.6 are
for a very light loading of boron in the shell
(0.0022 gz”cmE). Unfortunately, time did not permit
o more involved calculation for o heavier boron
shell; nowever, volyes for heavier boron concen-
trotions will be obtained in later calculations ond
experiments, Although it is possible to obtain a
larger amount of control by using higher aluminum-
to-water velume rotios ond larger internal reflector
regions, the greatest ingrease in control is reclized
DIME NSIONS
A 16,20 OR 25cm

8 16,20 OR 25cm
L 20cm

 

 

 

 

 

 

 

WATER FLOW

 

 

 

 

 

 

 

PERIOD ENDING MARCH 11, 1957

UNCLASSIFIED
owe 2-04-080-F

LOCATION OF CYLINDRICAL
TUBE IN 4CTuaL REACTOR

IMNTERNAL WATER REFLECTOR

MOVABLE BORAL SHELL

CORE REGION (WalER PLUS
URANIUM - ALUMINUM FUEL
PLATES)

EXTERNAL WATER REFLECTOR

Fig. 5.4.1, Vaerticol Section of Reactor Core Mode! Used in Nuclear Parameter Study for TSR-1L.

Taoble 5.4.1. Muclear Poromeaters Used with the 3G3R Orocle Code far Caleulations of the Critical Mo=sz
of the T5R-1l with an Aluminum-to-Water ¥Yolume Ratio of 0.707

 

Crass Section*

 

 

 

Region
za
f
Internal reflector 0
Core 0
External reflactor 0

0.01535 + 0.4712702

/

5

Transfar
Cross Sic:iflfi** Temperufure
(em™ ) (%C)
3 3
x!, Wi
0.0361 0 50
0,0210 0 60
0.0361 0 20

 

*Volues of Ea obtoined from BNL.325, Neutron Cross Sections, by D, J. Hughes ond J, A, Harvey (July 1,

1955), with correction applied for {1} non-1/v behovior, (2) Maxwellion distribution of neutron speeds, ond {3)

temperature effect on density.

**Values of D, D‘sJ Ex ., and E.x taken from 100-18133, Reactivity Effect of Reducing the AI/H20 Ratio in the

/

5
MTR Core, by J, W, Webster {Oct. 26, 1953).

295
ANP PROJECT PROGRESS REPORT

by decreasing the thickness of the core region.
This is fortunate since the thinner core region
also reduces the heat transfer problem.

A calculotion performed to determine the amount
of separotion required between the fuel and the
control plates fo reduce the effect of the control
plates sufficiently to permit operation of the re-
acter involved o more subtle use of the 3G3R
code. In this cose it wos necessary to artificially
maintain the neutron flux and current at the outer
water reflector ond core boundary so that the
internal reflector could be divided into two regions
with a boron shell between them, the core making
the third region. |t was then possible 1o vary the
thickness of the water regions in the internal
reflector ond effectively simylate movement of the
The calculation was mode for only
ratios, 1.0 ond
The resvlts plotted in

control plates,
two aluminum-te-water volume
0.707, and one geometry,

UNCLASSIFIED

 

2--06a0-~-8
1.o| - _— —
| L:EGMETHY rNTERN.flL REFLECTOR RADIUS = 20cm
_ CORE THICKNESS = 16 ¢m . '
EXTERNAL REFLECTOR = |
"‘HICKNESS 20crn
15 s —— —_

| sz e {VU’T’:‘NOD . ! { d
| Moo= ATOMS OF U?* PER cc
|~ v=NEUTRONS PER FISSION|

oy = MICROSCOPIC FISSION | 7
ceosg gECTIDN ;J.’ o

2
n

v%, (PROPORTIONAL TO UZ3S DENSITY)
(o]

 

0.02 |- ' |

0.0 | = A |
D6 .7 0.8 09 1.0 1.1 12 13

A, EFFECTIVE MULTIPLICATION FACTOR
Fig. 5.4.2, Effective Multiplicotion Foctor in the

2345

T5R-I! o0s o Function of the U Concentration for

Yartous Aluminem-to-Water Volyme Ratios. Reoctor
geometry: 40-cm-dio internal reflector, 16-cm-thick core,

20-em-thick externol water reflector.

296

Fig. 5.4.7 show that no more than 2.5 in. (~6.5 cm)
of travel will be required to reduce the control
plase effectiveness to 0.1 of its full value. The
results of an earlier experiment of the Bulk Shielding
Facility (BSF) recctor, which hos on aluminum-
to-water ratio of 0.7, normalized to the colculoted
value for the TSR-Il are also shown in Fig. 5.4.7.
The change in reactivity of the BSF reactor was
observed as o cadmium rod was moved laterally
away from the BSF reactor. The fact that the
experimental curve is steeper thaa the calculoted
curve shows that the use of the calculoted volues
in the design is conservotive,

Although the information was not required to
establish the dimensions of the TSR-II, the effect
ot the control plates on the fast- and thermal-

UNCLASSIFIED
2=0i=—080=9

 

7 . . i —
——— CALCULATED VALUES
—_— |NTEHP0LMED vm_uasf
shomaan £
&y ,f

16 - —

CRITICAL MASS (kg}

 

 

 

 

 

i2 ig 20 24 28
INTERNAL REFLECTOR RADIUS {cm)

Fig. 5.4.3. Critical Moss in the T3R-1l as 0 Function
of the Internal Water Reflector Rodius for Various
Aluminum-to-Water Volume Rotios. Reoctor geometry:

14-cm-thick core, 20-cm-thick external water reflector.
UMCLASSIFIED
2—0i—060=10

(5 gk
—— CALCULATED
——— INTERPOLATED /s
{4 l— -~—t i .

i
|

   

 

 

CRITICAL MASS [kg)

 

 

 

 

12 28
INTERNAL REFLECTOR RADIUS (em)

Fig. 5.4.4. Critical Moss in the TSR-H! o5 0o Functlon
of the
Alyminum-to-¥arer Volume Ratios.

Internal Woter Reflector Rodivs for Yorigus
Recctor gaometry:

20-cm-thick core, 20-em-thick external water reflector.

neutron flux throughout the reactor was also colcu-
lated as shown in Fig. 5.4.8. Cases & and ¢ show
the effect on the flux of having the boral shell
adjocent fo the core or 5 cm from the core, respec-
tively. Cose a represents the same geometry
shown in cose & end is included oaly to show the
validity of the method used, that is, leaving off
the external water reflector region and artificially
maintaining the neutren flux ot the outer core
boundary. The effect of the boral shell position
on the fast- and thermol-neutron leckoge from the
TSR-Il for two aluminum-to-woter volume ratios,
0,707 and 1,000, is shown in Table 5.4.2. ln this
calculation the internol water retlector was 40 em
in diameter, ond the core region and exterpal water
retlecior were each 20 ¢m thick, Chonging the

PERIOD ENDING MARCH 31, 1957

UNCLASSIFIED

 

 

 

2—0!—060—”
16 . | : —
2
|
15 T — | e T A 4| - %I_Vi
—— CALCULATED /
——— INTERPGLATED /7 /
B A R g

 

 

 

 

 

 

 

CRITICAL MASS (kg)

 

 

 

 

 

 

 

 

 

 

iz 16 20 24 28
INTERNAL REFLECTOR RADIUS {cm)

Fig., 5.4.5. Critical Mass In the TSR.1f as a Functlon
of the Internal Water Reflector Radius for Various
Aluminum-ta-Water Yolume Ratios. Reoctor geometry:

25-cm-thick core, 20-cm-thick externol water raflector.

aluminum-to-water volume rotio produces less than
1.5% chenge in the fast-neutron leckage. Over the
porobable operating ronge of the contrel plates
(3 to 5 cm), the variation in fast-neutron leckage
will be less than 1,5%. '

The spoce required to house the proposed controf
mechanism and olso allow it to be withdrawn 2.5 in.
from the core will necessitote that the internal
water reflector be no smaller than 17.5 in. in
diameter. With this parameter fixed it was then
necessary fo carry out o study to determine the
corresponding core parameters which would give
a minimum fuel loading. This was done by plotting
the critical mass as a function of core thickness
for various cluminum-to-water volume rotios (Fig.
5.4.9). As would be expected, the moss decreased
with a decrease in the aluminum-to-water volume
ratio; however, it does not seem practicoble to go

297
ANP PROJECT PROGRESS REPORT

WNCLASSIFIED
2-0-060—-12

 

 

AL/ H,0
| VOLUME RATIO=
et 2,000
l

50

CORE

.-"""'.-'.—.
THICKNESS - 1.000
lemy T psit s roiea i B
| e e
18 4= --""_’_ -——-— 0500

1000
~""""* e 707
_fl--' 2.000

REACTIVITY (% A kfk)

 

 

20 25
INTERMNAL REFLECTOR RADIUS fem)

Fig. 5.4.6. Chonge in Reoctivity Produced in the
T5R-H by the Presenca of o Beral Sheli Between the
Internal Woter Reflector and the Core as o Function of
the Internol Reflector Radius for Varicus Core Thick-

nesses and Aluminum-to-Water Yolume Ratios, Externcl

UNC LASSIFIED
E—H=060-13

 

 

10 — =
GECQMETRY: INTERNAL REFLECTOR RADIUS = 20cm
CORE THICKNESS = 20cm
EXTERMNAL REFLECTOR THICK-
MESS =20¢m

      
 

ey |
S
T |
= f = =
= 2
gt = 0707
)..
P- }
= : -
= | 7 .
Q| BSR EXPERIMENT, + = |
ul = - GADMIUM EXTERNAL 3 -
05 -TOLATTICE, m,’dzn =072 -7 N i
P prn iy T s > \_\_\\_
| TSRil ,-:a.l,fHZG-1.t::rc:+=;k
0.2 = —

0.1 — : ——
0 { 2 3 4 5 6 7

DISTARCE BE TWEEN SHELL AND FUEL (em Ha0l

 

Fig. 5.4.7. Change in Reoctivity Produced in the
T5R-H by the Presence of o Boral Shell in the Interncl
Water Reflector os o Function of the Distance Betweesn
the Shell and the Core for Aluminum-~to-Water Volume
Ratios of 0.707 ond 1.0.

intarnol water reflector, 20-cm-thick core, 20-cm-thick

Reacter geometry: 40-em-dig

woter reflector thickness:

20 em.

externu| water ref'rec!or (ef’f'ective}l.

Table 5.4.2, Predlicted Fast- ond Thermol-Neutron Leckage from the TSR-{l for 0,707 ond 1,000
Aluminum=-to-Water Yolume Ratios ond Yerious Core~Boral Shell Seporation Distonces

 

Shall Separation

Core—Boral

Qisrance (em)

Fast-Meutren L aokoge

Thermal-Neutron Leakage

 

Ab:H

2

O Ratio = 0.707 M:HQD Ratia = 1.000

 

0

5

1.0
13
2.0
25
3.0
3:5
4.0
4.5
5.0

No shell

A!--.Hzl) Ratio = 0.707  Al:H,O Ratio = 1.000

0.107¢0 0.1083 0.01294 0.01322
0.1064 0.1080 0.012%1 0.01325
0.1057 0.1074 0.01284 0.01318
0.1051 0.1067 0.01276 0.01306
0.1044 0.1060 0.0126% 0.01300
0.1038 0.1053 0.01258 0.01291
0.1033 0.1048 0.01232 0.51285
0.1028 0.1043 0.01247 0.01276
0.1024 0.1038 0.01243 0.0121
0.1020 0.1035 0.01235 0.01267
0.1018 0.1032 0.01235 0.01264
0.1083 0.1120 0.0142 0.0169

 

298
 
   
 
   

Lo — =
N

oORAL

| SMELL

FAST (RO |
FLLx

  
     

 

Fig. 5.4.8. Neutron Flux os a Function of the Radial
Position in the TSR-I for Yarious Positions of o Boral
Shall in the Internal Water Reflector end o 0.707 Alumi-
num-to-Water Yolume Ratio. Reactor geometry: 40-cm-dia
interncl water reflector, 20-cm-thick core, 20-cm-thick

external woter reflector {effactive).

below the ratio of 0.7 used in the BSF reactor.?

Also, if this ratic is used the standard arrangement
of 60-mil-thick fuel plotes separcted by 120 mils
of water can be used. Since the amount of control
availoble increases with a decrease in core thick-
ness, an approximately 1d4-cm core thickness
{5.5 in.) appecrs to be the optimum. This thin
core, which requires less thon 6.5 kg of U3 for
criticality, will alsc be an advantage with respect
to the heat removal preblem (see below).

The configuration having a 17.5-in.-dia internal
woter reflector and 5.5-in.-thick core region wos
then examined to ensure that the U233 fyel plate
loading would not exceed the 0.0283 g/cm? surface
density previously designoted. It wos found thot
the maximum loading pessible for this surface
density wos 9.5 kg. This allows for an error in
the critical mass obtained from the two-group
Oracle calculotions to be as high as 45%. The

 

4In the BS5F reuctor the fuel piute:s und wolar coalant
hove on aluminumr-to-water ratic of 0.5; the side plates
required to hold the fuel plates raise the owver-all
aluminum-ta-woter volume rotico to approximately 0.7.

PERIOD ENDING MARCH 31, 1957

UNELASSIFIED
2« OH--060=—1%

T2.0

1o -

CRITICAL MASS (kg)

8.0

6.0

 

40l - - = =
10 15 20 25 >0
CORE THICKMNESS (cm)

Fig. 5.4.9. Critical Mass in the TSR-ll us a Function
of the Core Thickness for a 44.5-cm-dio {17.5-In.-dia)
Internal Water Reflector ond Yarious Aluminum-to-Woter
¥olume Rotios. External water reflsctor thickness:
20 cm.
critical mass will be calculated agoin with a

multigroup, multiregion code in the near future.

HEAT REMOVAL STUDY

The heat flux from the fuel plates will be approxi-
mately 5 x 104 Btu/ft2.hr, To remove this heot o
two-pass cooling system hos been proposed. |t
wos previously determined® that a 1000-gom flow
rate with a 40°F temperature rise of the water over
the inlet water temperature would provide sufficient
cooling to prevent beiling or excessive fuel plote
temperatures in d reactor core with higher alumi-
num-to-water volume ratics. It was clso stated
that o larger internal reflector ond thinner core
region would make conditions even better. The

17.5-in.-dia reflector and 5.5-in.-thick

internal

 

SF“. Staver, USAF, private communication,

R Dol -
 

core meet these conditions, but the larger water
volume dictated by the lower oluminum-to-water
ratio i5 somewhat detrimental. It is felt that if
turbulent water How could be maintained in each
channel, odequate cooling could still be achieved
with the 1000-gpm flow rate. Using o log meon
flow aoreo ond ossuming equel flow per channel,
the Reynolds number was calculated ond found to
vary from 11,000 in the inner chennels to 8500 in
the outer channel. It appears, therefore, thot
sufficient lotitude will exist over which to adjust
flow rates to obtain uniform outlet temperatures
and still maintain turbuient flow.

FUTURE STUDIES

Preparations are being made for several develep-
ment experiments to further estoblish the TSR-II
design. Several control system mockups consisting
of sclid stainless steel plotes, perforated stoinless
steel plates, or boron-loaded cluminum plates will
be tested in a é-in.-square water-filled hole in the
center of the BSF reactor. A prototype of the
actual control design is olso being constructed
for other dynamic tests. In addition, the fuel
element drawings are now complete, and a model
will be fobricated to establish a heoder arrange-
ment tor the cooling system,

o "
 

THE

 

AIRCRAFT NUCLEAR PROPULSION PROJEGT

 

 

 

 

 

 

 

 

AT
THE OAK RIDGE NATIONAL LABORATORY
aeRIL 0, 19T

ANP PROJECT BHEECTOR W, W KIRDAN a0

CODIRECTOR £ A CHOMER AD

ASSRTANT IEECTOR & L MILLER &D

O WL v ER, SEC, RE¥
PRATT & WHITHEY AIRCHAFT REPORTS

E. R 0¥ TXO, ASST, PRO, ENE, ena . ARE
e ET«G::?;.D' ADw, AS5 T i:E rY ??;?II.E::S ARF

 

 

 

 

 

 

 

 

AMECELFT REAMCTOR ERDINEERIMG DOVISH0H

§ 1. CROMED, (R ECTOR RE
A E LIVIMGSTON, SEC. ~FE

 

 

 

 

 

SLE PR TG RESEARCH

¥ K, JORDAN
Ao L WIELER

WP ATEERING OO TTER
o JDSDAN, CHAIRMAR

E P, BLiZARD
BOUDSELLE
ELVD

  

 

 

 

 

 

 

MSFIAHT TO GIRECT R

R % CARLSM TH

?&GOEEE GEREARCH MASDTIATE

 

.

 

FRAAL W K. ERGIM ARE
CRIES

 

 

 

 

 

 

 

EXPRRMENTAL ENGINEEFIHE
H_ % SavAGE ARE

 

 

 

POwER PLAMT EnGMEEE MG
h P OFRALS ARE

 

 

 

 

AT ERTA TP AMD COMT ALY

 

 

 

 

 

 

 

 

 

 

 

 

R, MANM i
ERCIMEE RIMG DERCH
E.J. BREECMG ARE
EELCTOR FHYSHI
AL W, FERRY EHR
rein
w_F BOUDREAL ART

 

 

 

 

 

HEAT TRAHIFER AMD PHYL{ICAL PROFEATIE:
H &, HOFF AN ARLC

 

 

 

KEILHQETD

 

HILLER
FEREY
SANVACE METALLUREY

ERIFLEY WD WAHLY, STAFF ASRATANT L

1o b, SWARTOUT

 

 

« F
G E
5k
K
AP
AR
G
& [ waNLY
Al
4 W
H oW
E
A
A M

WELKBERL-

 

 

 

 

CHEWSTHY
W GAES, STAFF ALEILTANT <

 

 

 

 

APPLIED HUCLESN PHYEHL
HITE, THIS CHART Spdwh GRLY THE LINES &F TECHRICAL DOORGIMATICH OF THE aNP € b BLI , ATAFF ASHSTNT "
PROJECT. THE YARICGH (HIVIPUALE AN GROUFS OF FEQRLE LISTED 1k THIS AND THE

 

 

FOLLEWING CHARTS ARE EHGAGTD FITHER wHILLY DR FART TIME O RESEARCH AND
PECIGH WHICH S COXICIMATED FOR THE BENEFIT OF THE 4NP PROJECT IN THE WAMNER
INDICAT & Ot THE CHART. EaCH GROLF, HOREVER, ¥ ALY RELPONSIBLE TO (TS MVISICH
CRECTOR QR THE PETAILED PACGRESS OF TrE RESEARCH ApD FOR ACWlniSTRATIVE

 

RATTERS. RACHA THIH DAMACE
TrE KE¥ ¥ THE ABEREVIATICNS LSED 15 GivEN BELCY Gow, KEWLROL T SIAFE ARSOATANT 35

 

 

 

AC Arbdl F IS CHENTSTRY DIVISION - ORI

AR AN ESEATOH MadUF AT FUBIHG CmP AFY

AR APELIED HUCLERR PIOYSICE O 0m - ORML
AIRCRAFT REAL TOR EnSIMEERIWG DiYi% 0N -~ ORMHL

 

 

BENMA PRODDCTS DN OF BEHDLL %Le TION CLRFORATION FYEL REP ROCFAHG

CHEWETEY DIVISMON - ORKHL R B LIMPALUER, STAFF ASSISTANT Y

 

 

 

 

 

CANVALR, & DEEIDN OF GERERAL Drmasds
FEHG O RING a0 WECHAN ICAL DIVTGH - oRpL
EMR  ELECTRINLGLEAR RELYEATCH OIVWHOH = QR
Gl GLEMHN L WAT DM COMPANY
IC AMETRUMENT ATIDN AMD OOk TROL 3 Thv 15100 - QRML
L ME TaLLURGY CIVISION — ORML
Pwa  pRATY A mGQTHEY AMCRAFT, CIVEIICH OF LAC
RCA  READKD CORPORATION OF AERICA
RD  RELEa®IH (HRECTOR'S DEAARTMENT - ORML
Wt STABLE NSQTAFES DIVISION « OFNL
5 SOL 10 STATE DHYISGN = GRNL
LSRE LRITED STATES AIR FORLE
AT WRIGHT A4F DEVELOPMENMT CENTER

ARE

Ep

<

&1 CEbiCaL TECHRALDGY DIVISICH - QN
o

Em

 

 

 

 

 

 

301
 

 

THE AIRCRAFT NUCLEAR PROPULSION PROJECT

AT

THE CAK RIDGE NATIONAL LABORATORY

APRIL T, MK

 

AIRERAFT REMTOR EHC HEERIKG DEYISIOH

5 ). CROMER, FIRECTOR 22
A B LiYINGITON, JIC ARE

 

 

 

 

ALUESTANMT T0 DIRECTOR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R & CARLSMITH ARE
R, E. THOMFEOM ARE
5 AMHSOH, SEC LRE
RELCTOR PHTHCE PWEAT mmm:#:::nmm TEID DEPARTWENT FONER FLANT ENGINEERING EAF ERGHEHT AL EMOIHEERING ENGINE ERING DELIGH LS T R EMTATION AME COUTEGL
L e ¥ F BOUDREAL ARE AP FRAMS ARE M ¥, SAVAGE 4RE £, J, BREFONG ARE £, &, bANN &
W, SEC. ARE MW HOF Falan ARE ) PARKER, SC ARE & R BEMNETT, 3EC. ARE D ALEXARDER, SEC ARE L. £ FERGUSOM, S5, ARE C % WALKER I
L) KEYES ARE & BENDER ARE F ol OTAVLGH S0C ARE W, Tl SEC. "
H. ¥, BERTIN| aRE M. O EDER SEC ARE MONVERTON, SEC ARE
' ADUIWISTRATION
. M COPERHAVER ARE
E £, HALGERT ARE HEAT TRANAFER ETU AND ART ASSIHBLY w L s0att ARE REACTOR CESIGH SYSTEM DENGH
. TaAGLRG ARE 51 QOHEN ARE € b WHITMAN /e ¥ R OGN, BRATTR e Mo S R e e b3
A E Y ESTABROOH LRE WG BECKER &
b gLt ::E B, K HARRINGTON, SEC 4RE HTDRODTHACS ARD THERSIGTHANLS T HELa AR e F oL owAY e
e H i &BELE ARE ¥ T, FUR J. V. MEADGR ARE G C CUERRANT [
o7 oo ::; : % oot } puessaen e L o " E St &
Y E wott s E.x.mt{@m’mm ARE L& ePeL ARE W A SYLVESSER PHA
: _ % PEM ARE C - . R, HHOKS ARE THSTRUMEMT A TICH
.G ST roa TH ARE :- E h:g:; ::g TEST OPERATION PUMP BEYELOPMENT L R LARRABEE P,
L L WANTLARD ARE ¥ E THMAS ARE 2 s RSOH R c R G, AFFEL 14
4, LONES ARE SMIELS ARE W, B, MCDOHALD ARE R D TR FHh ®_E, AMOERSR EAF
Y. M JDWES ARE 8. H FITZGERALD, COMPUTER  ARE D P WARRIS 3EC. ARE & & GRINDELL ARE PLUAP ARD PIFING DEOGH . i BURGER v
R. L, MILLER ARE \RT DF E“*"mm‘“fi‘“m“" 12 TUDGR, DRAFTEMAN ARE A, MOM TOAMERY, SEC ARE ; ;ug:?-'(a x: %, G CORB ARE €. M BURTON e
HYGRODYHANKS £ F WL KM ARE APPLIED MECHAMICS AND TEST FACILITIES ASSEMELY O L GRAY ARE :E ff;?:ffir ::: ;‘: ,;"Y',ffiqm .'g
P 2. LncH ARE A A ABBATRELLG PrY TTRESS AHALYEIS COOADIMA ® & JEWHESS ARE L, # XOFFMAR ARE LW KREWSOM I
6. L AULLER twi ¥, F. FERGLDON ARE B V. LECHRESL Lan 4RE € STORIQ ARE T i L v, siLson ARE A & LEPPERT 1
R, D, PEAK [ L. LOvE v M AUBUCHEH, SEC art E 4. ALMGREN ARE B o s AR M. I METZ 1o
¥, ). STEY IMadl “RE G € ROBINGDH ARE e ¥ COTTON ARE ) L CRAWLEY ARE PG aan puif ELECTRICAL DESIGH ¥ R wILLER I
LB, WK Ko WMLLEY EW. 8 L GREEWITREET ARE ., . R, E, PiOGEDM, JR. RiCa
B e ARE T. £ CRAGTREE ARE M LuRiE ARE P SANDERSH P MG rouns ~ TR bt Rson ARE c L :;‘”‘I:E' o Ree
FHYSICAL PROPERTIES :— : fif; ::': CORRCAIDH LOCPS $YSTEMS DEVELOPUENT B. & GARRETT ARE ‘UJ' by AD%ES%‘H ‘rf?
* 0. POWERS ARE FGAPAEHT CONTRO L0 PALRER, /R, ARE ¥ . KELLEY ARE 0. B TRADIER ARE fi SF;T;JLM ::: R 5. HEMSLEY <
G £ BLALDOK ARE B B BACHULIS ARE O H PLATUS SAF A DLS0H ARE 5. K, PURMEY, SEC. ARE "5 F. HOWELL ARE J R FCMES i<
L) CLAIBORRE MAE H W, HOVER ARE DL PLATYL A F M P RUSZKOWE, SEC ARE 1o O HICHE S0 ARE A ¢ #os i<
L) PLATZ LRE F I STAMEY ARE FLECTRICAL JEMVICES H,: ;:”:xt I
3R TALLACKEON [ & a LEES ARE HEAT ENCHANGERS GEMERAL DE S b 572.,.9‘:0” :E
A _E.
HELT YRAHSFER L e ) e AR K. €. DAMIELS ARE ¢ v, wRIGHT B
% 5 HARRM ARE .
R, 0. SCHA THEISH ARE R L SENN ARE X F UARCUARDT ARE
F . ROBERTSCH, SEC ARE TECHHICIAN GROUS 1 CABLE RAE MUCLEAR AND $F ECIAL PEYICES
. B CLARKE ARE R HELTEH ARE VALVE AHD COMPOMERT DEVELOPMENT R H. XotdES ARE
W 4, FONTAHA ARE T ARNWING ARE © G, MICHELSOH ARE 5 opareR ©
), FOSTER ARE 6% EHILT R 4k, COMLIM ARE - . L C SHUFSRD Pwa,
k.1 giay PuA € 5. BwER prid N . € F e AR r B ¥ERTS "
R 4 MOLCOMBE ARE Y o CoRUn ARE BoM LOOPLR P & B IECLS 4AE
E. A JAGCERS LRE J 8 CUNNING A ARE & RRESS PWa ' LUILATOR
Th :EwL:I.'[mm ::E 5 ; g;‘:l:‘ :‘q“f_ IN-PILE LOOP OPERATION RECURDY AND PRINTING F P OGREEM i
T. K waLTERS ARE E £. CHAMBERS ARE
) R CHACK WORTH ARE 4 A, SONLIN 1RE
WM YARDSH ARE ¥ 1 OLCRAERTH ARk & & BoLta ek L ) FOSTER ARE
CRARTY-TIME LOAM PDOENMCH ARE
. FOUST 4RE REACTOR GPERATION AND DI FASSEMBLY
R. K FRAMELIP ARE
WD GHIRMLEY ART w 3 COTIRELL ARE
C ) GHEEM ARE ¥, €, BPORHING® 55
CONSUL TaNTS T L SREGONY ARE C N CUNNRGHAR ARE
F A AHDTRIOH, UHIVERSITY OF WISSSSPPL R. A& RAMEICE ARE 5. 4 DECAMP ARE
k. F_BAILEY, LRIVERSITY OF TENWESSEE €. G. HEMLEY ARE R, &, DREISIACH Fua
A H, FOY, UMION (OLL EGE LW WINGSLEY MRE R P SHIEL DS” u
1. FRENCH, UNWVERSITT OF TEANESSEE 1 R LOVE &RE
L LISTGR, PURAJE UMFYERMTY 6.C wmiLis ARE ENGIHE ERING SERVICES
W, LOWEN, UMION COLLEGE 8. H, AHIGOME RY ARE
PL . MASWELL UNIVERSITY OF TENNESSEE 0 G FEMH ARE e msv SEC :EE
PM mnc'.:.:fm LMIVERETY OF $0UTH :. :Erocm.mn ::;’. F. W LIl SEC. RE
CARDL TLaGy &
2. W, FRUNMEL., STATE LHIVERST Y OF 10%A F. . CHnFER ARE j ;' EVANSE -'«;E
W K. STAMR UMVERSITY OF TENHESSEE . A, suGarT ARE T. v MAYES ARE
J.o B WEST, RLAHCMA AR Y COLLEGE o E :':;:JELL fizi 5 B ASHTOM ARE
G l::bn'lm:EEms. P ERHETLYANIA STATE E: TALLiCE :RE € WAETEMS ARE
LLEG
8. ¢ wILLIAMS ARE
oW, il Som ARE

 

 

 

 

 

 

 
 

THE AIRGRAFT NUCLEAR PROPULSION PROJEGT
THE OAK RIDGE NATIONAL LABORATORY

APRIL 1, 1957

 

SUPPORTING RESEARCH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

w. H. JORDAN
A, J. MILLER
RESEARCH ASSOCIATE
W. K. ERGEN ARE
STAFF ASSISTANT STAFF ASSISTAMT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT
E. P. BLIZARD AP w. R. GRIMES C W. D, MANLY M G, W. KEILHOLTZ 55 R. B. LINDAUER cT
APPLIED HUCLEAR PHYSICS
. P BLIZARD® AP
R. RICKMAN,* SEC. AP CHEMISTRY ANALYTICAL CHEMISTRY HMETALLURGY CERAMICS RESEARCH RADIATION DAMAGE FUEL REPROCESSING
L. 8- ABBOTT AP W, R, GRIMES c 4. C. WHITE AC W. D, MANLY*® W L. M. DONEY M G. W. KEILHOLTZ 3 R. B. LINDAUER" cT
D. E. CALDWELL, SEC. ¢ E. M. ZARZECK], SEC. AC T. HKIDO USAF C. A. PRDAPS, SEC 55
3. R. CUNEQ c M. R.HILL** M
A. L. ATCHLEY, SEC. c J. MORRISON, SEC. " . HARRIS
RESEARCH AND DEYELOPMENT ‘fi' A POTTER ’,a MTR LIAJSON CHEMICAL DEYELOPMENT
SHIELDING THEORY TOWER SHIELDING FACILITY SPECIAL PROBLEMS A.S. MEYER, JR. AC GENERAL CORROSION J. A GRIFFIN M H. V. KLAUS 55 D. E. FERGUSON cT
J. P. YOUNG AC G.l. CATHERS cT
A. SIMON* AP C. E. CLIFFORD*~ AP R. F. NEWTON RD R. E. FEATHERS, JR. AC E. E. HOFFMAN M HOT LAB FACILITY M. R. BENNETT cT
V. GLIDEWELL,* SEC. AP E. MCBEE, SEC. AP R. E. MINTURN USAF J. R, FRENCH AC W. H. CODK M _ R. L. JOLLEY cT
H. B. MORAN PWA V. R. CAIN AP T. W, GILBERT, JR. aC D. H. JANSEN M CONSULTANT i- : g:él:\'L;KE :2 C. J. SHIPMAN cT
C.D. ZERBY AP C. R. FINK . GLM PHASE STUDIES G. GOLDBERG AC J. L. GRIFFITH M T. N. MCYAY T
F. N. GREEN GLM B. L. MCDOWELL AC J. W. HENDRICKS M C. E. ELLIS S5 UNIT QPERATIONS
SHIELD DESICN 2L HULL aP C. 1. BARTON C W, ). ROSS AC 1 E. POPE o E. J. MANTHOS 55 | C. BRESEE* cT
W. ). MCCODL. PWA R. J. SHEIL c L. R. TROTTER M W. B, PARSLEY s 1T LoNG cT
C. E. CLIFFQRD** AP F. J MUCKENTHALER AP R. E. THOMA c SERVICE R.N. RAMSEY 58 5. H. STAINKER cT
F.L. KELLER AP W, E. PRICE PYA H. A. FRIEOMAN C DYRAMIC CORROSION E. 5. SCHWARTZ 55 G. JONES, JR. cT
8. OLINGER,* SEC. AP M. G. WELCH AP 8. S. LANDAU c W. F. YAUGHAN AC E. D. $IMS 58
w. G BLESSING cy 5. M. ARGO AP B. C. THOMAS ¢ R. F. APPLE AC J. H. DEVAN ) J. 0. STIEGLER S5 GESIGN
L. BOWMAN waDC C. C. BARRINGER AP R. E. MOORE c D. E. CARPENTER AC G. D. BRADY ) METALLOGRAPHY R. M. WALLACE §s
A_T. FUTTERER PWA ¥. A, BIRD Ic "R E. MEADOWS ¢ H. L. HOLSOPPLE, JR. AC E.J. LAWRENCE M R . GRAY M H. E. GOELLER" cT
C. A. GOETZ PWA E. D. CARROLL Ic "R K BAGWELL ¢ R, C. BRYANT AC M. A. REDDON M : RADIATIGN METALLURGY R. P. MILFORD T
J. HILGEMAN WADC b D. COMMNER g o t : IH;OLJ :g F ABRICATION 1 €. WILSON ss J. P. RUCH cT
S. K. PENNY AP J. N. MONEY AP LE. N
N abery - ¢ b RICHARDSON o FUEL PREPARATION RESEARCH AoH. MATTHEWS ac R. S. CROUSE M C. D. BAUMANN ss PILOT PLANT
o 6. G, UNDERWOOD ic G. M. WATSON c A. D, WILSON AC J. H. CooBS M E.L. LONG W ¥. E. BRUNDAGE 5
-2 - W, BLOOD ¢ C. M. WILSON AC M. R. D'AMORE PWA J. R, RIDDLE M %. w, DAVIS 55 R. B. LINDAUER* T
LID TANK SHIELDING FACILITY ;' A GILBERT c H. INOUYE M D. F. STONEBURNER M N E. HINKLE 58 . H. LEWIS* cr
TOWER SHIELDING REAGTOR (f 1. H. SHAFFER < R. E. MCOONALD PWA J. E.VANCLEVE " J. €. ZUKAS ss W. H. CARR cT
W. ZOBEL AP V. SMITH c 3. P. PAGE M M. D. ALLEN " F. N BROWDER €T
4. LUTTRELL, SEC. AP C. E. CLIFFORD** AP R. A STREHLOW c T. K. ROCHE M L. A, AMBURN M REACTOR EXPERIMENTS J. E. BIGELOW cT
D, CADY USAF L. B. HOLLAND AP W T. WARD . R. W. JOHNSON M N M, ATCHLEY M R. 8 KEELY cT
' M. E. LAVERNE® AP - T. WAR ¢ J. F. NEWSOME M E R W. . BROWNING 55 F. W MILES cT
L. JUNG AP G. A. PALMER c - BOYD M R. E. ADAMS 55
J. M. MILLER AP SPECTROGRAPHIC AMALYSIS W R.L. FITZGERALD M O E. OUSS USAF C. L. WHITMARSH cT
E. A WARMAN PWA CRITICAL EXPERIMENTS CHEMICAL EQUILIBRIA ELEING AND BRAZING E. P. GRIGGS " H. L. HEMPHILL 55
L. BECKHAM AP J. R, MCNALLY* sl J. T. HENSON " e
. A. D. CALLIHAN* AP 1. A. NORRIS® AND OTHERS 5 P. PATRIARCA M E H LEE " M. F. OSBORNE 3%
D. R. HENDRIX AP . L. C. OVERHOLSER c LA R . H. W & RORE .
M. L. RUEFF " SEC. AP e . E. CLAUSING M B C LESLIE u . E. ROBERT3ON s
H. JARVIS 1 W. G ALWANG PRA - E. KETCHEN < E A. FRANCQ-FERREIRA M B J REECE M R. P, SHIELDS 58
J. R TAYLOR AP £. SEMSKI PHA 4. 0. REDMAN < G. M. SLAUGHTER M 4 J. E. KRAUSE PWA
B. J. STURM C L. G. SCHRADER M
J. F.ELLIS* AP E ). WILSON M C b ZRCHARY "
W. ). FADER P, CORROSION STUDIES B. MCDOWELL M CHEMICAL EFFECTS OF NUGLEAR REACTIONS
BULY SHIELDING FACILITY V. G. HARNESS® AP E C. E. SHUBERT M
. D. A. HARVEY PWA F. KERTESZ c R. G. SHOOSTER M M. T. ROBINSON 55
F.C. QAL%JSSSSLNSEC :r’ PR P S, ¢ MASS SPECTROME TRY L. C WILLIAMS " R 1. CARTER 55
' ’ ' R. E. MALENFANT AP F A KNOX c J.R. SITES sl
T. V. BLOSSER* AP E. R. ROHRER* AP 1. M. DIDLAKE c MECHANICAL PROPERTIES ELECTRICAL COMPONENTS
G. T. CHARMAN AP E. V. SANDIN PWA D, A DOUGLAS " 1 c PIGe s
G. DESAUSSURE" AP on SCOTT, JR. ARE A " T . €.
G. ESTABRODK* AP W O TUNNELL® AP MATERIALS PRODUCTION €. W. DOLLINS M — ETALLURGY CONSUL TANTS 0. L. CURTIS 55
K. M. HENRY* AP MK ALBRIGHT® o C. R. KENNEDY PWA N. CABRERA, UNIVERSITY OF VIRGINIA C. C. ROBINSON 5§
E. B. JOHNSON® ap D E MCSALTHYT “P G. 1. NESSLE C 1. W, WOODS o N. J. GRANT, MASSACHUSETTS INSTITUTE OF 0. €. SCHOW ss
J. D. KINGTON* AP ' . M. R. SWAN, SEC. c F.l.BEELER M TECHNOLOGY
J. LEWIN GLM C. R. CROFT c K. W. BOLING M J. L. GREGG, CORNELL UNIVERSITY CERAMICS
* F A, DOSS c E BOLLING M W. D. JORDAN, UNIVERSITY OF ALABAMA
R.W. PEELLE AP 1. 6. MORGAN 55
N 4. E. EORGAN c J. T. EAST M E. F. NI®PFS, RENSSELAER POLYTECHNIC Ce
E. G. SILYER AP R. M. CARROLL S8
0. W. CHRISTIAN* AF CONSULTANT xjn' ERE:LLTD E J. D. HUDSON M " LNSST;\I’%E RENSSELAER POLYTECHNIC W T MORGAN s
D. K. KIRBY* P -G V. G. LANE M . F. , - T
C. E. MCNEW" A H. A BETHE, CORNELL LNIVERSITY C. N. ADKINS c B. MCNABB, JR. M INSTITUTE P. E. REAGAN 55
J. SELLERS® i€ D. H. BOWMAN C E. 8. PATTON, JR. M
R. M. SIMMONS* AP J. P. EUBANKS C €. K. THOMAS M
D. SMIDDIE™ Ic CONTRACTORS W.K.R. FINNELL C C. W. WALKER M
G. G, STOUT* Ic METAL HYDRIDES, INC. B. F. HITCH c
H. YEAVER® AP NUCLEAR DEYELOPMENT CORPORATION E HOLTY c MASS TRANSFER STUDY
OF AMERICA W. JENNINGS C - "
TECHMICAL RESEARCH GROUP W. M. JOHNSON C s
L. MASSENGILL c H. :‘ L—IEAVENWORTH PWA L METALLURGY CONTRACTORS
g' gEm_EERY g PART TIME - W BARNETY M BATTELLE MEMORIAL INSTITUTE
o “*DUAL CAPACITY BRUSH BERYLLILUM COMPANY
NONDESTRUCTIVE TESTING MNEW ENGLAND MATERIALS TESTING
THERMGDYNAMICS R. B. OLIVER " LABORATORY
F. F. BLANKENSHIP c J. W. ALLEN M RENSSEL. AER POLYTECHNIC INSTITUTE
M. BLANDER C R. W. MCCLUNG M SUPERIDR TURBE COMFANY
S. CANTOR c R. A. NANCE M UNIVERSITY OF ALABAMA
S LANGER c O E. CONNER M UNIVERSITY OF TENNESSEE
M. B. PANISH C R. A. CUNNINGHAM M
L. E. TOPOL c W. 1. MASON M
INSPECTION
4. TABDADA M
CONSULTANTS D. K. DAYIS PWA
k. INSLEY A. E. GOLDMAN M
T. N. MCVAY R. L. HEESTAND PWA
G. M. TOLSON "
R. M. EVANS M
J. B. PHILLIPS M

 

 

 

 

 

303