ORNL-1896.txt

£

brary wi

ni,

cume

do

,.
e i

,

doumvany &g

h

G Mao
in name wi

<F

b

5N

YO VLS
wrange @ loan

T
¥
5

MARTIN MARIETTA ENERGY SYSTE

KA

 
 

 

 
 

 

)

)

 

ANP Authorization Required
ORNL-1896

This document consists of 246 pages.

Copy/(aa of 179 copies. Series A.

Contract No, W-7405-eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT

For Period Ending June 10, 1955

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

DATA RECEIVED BY INFORMATION AND REPORTS DIVISION
(JUNE 28, 1955)

DATE ISSUED

JUL 28 958

OAK RIDGE NATIONAL LABORATORY

 

' Operated by
CARBIDE AND CARRON CHEMICALS COMPANY il
A Division of Union Carbide and Carbon Corporation
, . Post Office Box P

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“VOQkk_R_idgle, Tennessee .

    

 

{ 3 4yu5L 0250999

8

-~

 

 
o oo b, e

 

 

 

0O NG AW

  

Z““m“3005?%>>W?§F?PFW??P?OP?POFP?T?PWNPWOfif

o0z

   
  
 
  
 
 
 
 
 
 
 
 
   
  
 

MASTOMIrYQoEwMIPAMSETOAO> MU T E e

. H. Jordan &
. W. Kellholtzfr

. Adamson
. Affel
. Baldock

. Barton

. Callihan
. Cardwell
Cathcart
. Center (K-25)
. Chapman

. Charpie

. Clewett

. Clifford

. Cottrell

. Cowen

omer

Crouse

. Culler

. Emlet (K-25)
. Ferguson

. Fraas

Frye

. Furgerson

. Grimes

. Hoffman

. Hollaender

. S. Householder

T. Howe
W. Johnson

P. Keim

. T Ke[ley

. Ker'resz

I;j:son

M. Klnj‘
A,
. E.
E!L erne

    

 

ORNL-1896

Progress

INTERNAL DISTRIBUTION

   
 
 
 
 
 
    
 
    

46, R. S. Livingston
47. R. N, Lyon

 

48. F. C. Maienschein
49. W. D. Manly
50. L. A. Mann
51, W. B. McDonald
52. F.W.McQuilken g
53. A. J. Miller &
54, K.
55. E.
56, J.F
57. G.
58. R.
59. P.
60, A.
61. H.4&
62, P
63 '-» W. Savage
? A. W. Savolainen
E. D, Shipley

_ . O. Sisman
%67. G. P. Smith
"88. A. H. Snell

6 R. . Strough
70.‘%% D. Susano

Whltmun R
. Wigner {cons
. Williams
. Wilson
83. C. E. Winters
84, C. D, Zerby
85.94. X-10 Document Reference ki
95-114, Laboratory Records Depar'rm-
115. Laboratory Records, ORNL R-:.
116-118. Central Research Library

fln‘uo

     
  

|6
A

 

 

 
 

 

EXTERNAL DISTRIBUTION

119. AFDRD Jones §
120. AFDRQ .

4, 121, AFSWC
W 122, Aircraft Lab WADC (WCLS)#
4123, Argonne National Laboratéfi%y
04 ATIC
]25—1'
12‘“ BAGR - WADC

129\.‘ah‘e||e Memorial In&'afute

 
 
  
 
 
  

  

 

141

142,

143. ofd

144. i boratory — WADC (WCLE)

145-148. |
149.
150.
151, Knollgg'A\‘omlcfif_‘E ower Laboratory
152, Loc heed Bugbank
153. Loc .;

  
  
  

...
r,

| : 172, SAM
}73-174. Technical Information Serv , Oak Ridge
Operations Office

,.[}175-177 WADC - Library

E  178. WAPD — Bettis Plant
179. Wright Aero

  
   
 
   
   
   

 
 

 

 

 
 

 

ORNL-528

ORNL-629
ORNL-768
ORNL-858
ORNL-919
ANP-60

- ANP-65
ORNL-1154
ORNL-1170

ORNL-1227
ORNL-1294
ORNL-1375
ORNL-1439

- ORNL-1515

ORNL-1556
ORNL-1609
ORNL-1649
ORNL-1692
ORNL-1729
ORNL-1771
ORNL-1816

ORNL-1864

 

Reports previously issued in this series are as follows:

Period Ending November 30, 1949
Period Ending Fébrudry. 72'8, 1950
Period Ending May 31, 1950
Period Ending .Augus-fr- 31, 1950

Period Ending December 10, ]950

Period Ending March 10, 1951
Period Ending Junelr'lfl, 1951

Period Ending September 10, 1951

Period Ending December 10, 1951
Period Ending March 10, 1952
Period Ending June 10, 1952 |
Period Ending September 10, 1952
Period Ending December 10, 1952
Period Ending March 10, 1953
Period Ending June 10, 1953
Period Ending September 10, 1953
Period Ending December 10, 1953
Period Ending March 10, 1954
Period Ending June 10, 1954
Period Ending September '170, 1954
Period Ending December 10, 1954
Period Ending March 10, 1955

 

 

 

 

 
 

 

1.

iy

)

 

FOREWORD

Th:s quarterly progress report of the Aircraff Nuclear Propulsron Project at ORNL records the
technical progress of the research on c:rculchng-fuel reactors and all other ANP research at the
Laboratory under its Contract W- -7405-eng-26. The report is divided into three major parts:
[. Reactor Theory, Componen’r Developmenf and Construction, I, Materials Research, and

IHl. Shielding Research

The ANP Project is comprlsed of about 475 technical and screnhflc personnel engaged in
many phases of research directed toward the achievement of nuclear propulsion of aircraft. A
considerable porhon of this reseqrch is performed in support of the work of other organizations
participating in the national ANP effor’r. However, the bulk of the ANP research at ORNL is
directed toward ’rhe development of a circulating-fuel type of reactor.

The design; construction, and 6pércrtion of the Aircraft Reactor Test (ART), with the coopera-
tion of the Pratt & V‘Whitney Airc':rafi-D'ivision, are the specific objectives of the project. The
ART is to be a power plant system that will include a 60-Mw circulating-fuel reflector-moderated
reactor and adequate means for heat disposcl.' Opercrfion of the sysferfi will be for the purpose
of determining the feasibility, and the problems associated with the design, construction, and

operation, of a high-power, circulating-fuel, reflector-moderated aircraft reactor system.

 

T ST T T

P T

 
 

 

SRR T

 
 

 

 

 

 

 

 

 

 

 

 
 

SUMM/\I EY ............................................................................................................................................................

PART I. REACTOR THEORY, COMPONENT DEVELOPMENT, AND CONSTRUCTION

1. REFLECTOR-MODERATED REACTOR oo e eeer e
| Aircraft Reactor Tesf Design e ettt et et e b et s eee et ens
ART Control o, et byt eeenres rerrre s et eareabaeaaenne et naes

Control Principles ..ottt
Operating Procedure ........ OO OO OSSOSO PRO P PRTUOROSIN e

Reactor Physics ..., e ettt e e e -
Probable Effect of Replacing Inconel by Columbium in the ART Core Shells................o
ART Temperature Coefficient ......co.iviieeiiricirieis ettt sssb st
Reactivity Effect of a Heat Exchanger Leak ...,
Burnup and Gamma-Ray Heating of Control Rod ...

2. EXPERIMENTAL REACTOR ENGINEERING ...

In-Pite Loop Component Development ..o
F lUX MEOSUREIMENES o iiiiiiiiiieeiieit et eeiorte s teiseeeeasseeeesanraeeeestsarteas e b e et s sbab e e esermbaas s abatsseansm e ee s s baaaeeaabiaees
Fission-Gas Holdup.....ccooocovenninini et — et e e e bt e nbe it e e nreens e et ete e et e b e e e neaas
Bench Test........... e e e e e ereeeeba et e e s e st et sae e renee ettt
Horizontal-Shaft Sump Pump ...ccoveieiie e, eetteere it i bo et e ersenenaeataesanats
IS rUMENTATION ..ottt ettt bbb e e

" Assembly of MTR In-Pile Loop Nou Tt seesesssssomsenes s srons

Development and Operation of Forced-Circulation Corrosion and Mass Transfer Tests .................
Operation of Fused Salt-Inconel Loops ...
Sodium in MUBIMETal LoopS oot e

i

Pump Development ..o e et
Water Performance Tests ... s
Bearing and Seal Tests ... i s

" 'Mechanical Shakedown Tests......c.ccocvrnnnne. s bttt
fShorf—C:rcunf F’ump Test Stcmd.......' ..... ettt bbbttt a st ettt i

 

 

    

'Heat Exchcnger T eSS o e et eee bt enes
Heat Exchanger Tube-Spqcer Pressure Drop Tests ..................................................... et benes

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

arsasssncen sseuss 4teE N Lt NPT It e s eaassansb et seentactndnnebbbbbatat b

Thermal Cyclmg Te ts of Sodlum-inconel Berylhum Sysfem ettt et
Gas-Flred Hecn‘ Source ettt eraeen e e

    
   

   
 

o Room Temperature Ref
| , “Reactivity Measurements ...
| - Power Distributions................ et e eetteereitteteeniareteee hrateeiae_aeeeeieteteeeitnsaeeannteaeeaaneteseanteesee s tanreseneraes

Critical Experiments.....cccocooniieirinnacn.

 

Assembly for High- Temperature Critical Experament ............................................................................

 

[ OREWORD ........................................................................................................................................................

15
15

17
18
19

20
20
20
21
22

26

26
26
26
27
27
29
29

30
30
31

32
32
33
34
35

35
35
37
37

37
39
42

43

43

43
46

48

Vil

 
 

 

 

 

| - |
PART lI. MATERIALS RESEARCH | b
B
4. CHEMISTRY OF REACTOR MATERIALS .....ooovvvocrurreemiemmnienececensssssemsimsesseeseseresessssssssssbosssss s 51
 Phase EqUiliBrium STUBIE@s ...o.oouieeee oo ee ettt st st e .. 51 b
The Binary System LiF-ZeF f.oooo s 51 T
The Ternory System NaF- LlF Z e 52 '
" The Quaternary System NaF-LiF- ZrF UF s e 55
_BeF ~Bearing SysStems ... e e arre e 56
Visuai Observation of Fluoride Melts ... 59
- Phase Separation by Zone Melting ...t 59
Chemical Reactions in Molten Salts ..., et e 60
Equilibrium Reduction of FeF, by H, in NaZrF oo e 60
Reduction of UF by Structurcl Metals ................................................................................................ 60
S’rablhfy of Chromlum and Iron Fluorides in Molten Fluorides ......ccoooniviiniininicecincee, 63
The Disproportionation Pressure of U g e 64
Reduction of UF, with Uranium in Alkali Fluorides ..o, e ereeans 67
. Effect of Filter Medium on Stability of U e, 69
* Stability of PbF -BeF, Melts in Inconel ..o 71
Soiublhfy of Metals in Molten SIS ot ne e s ee ey 71
__ Produchon Of PUrified FIUOFIAES ..oveoeeeeeeeeeeeeeeeeeeeeeeeeeerse et ssn s v essseesesneneannnes 12
3 - Fuel Purification and Preparation Research ..., et 72
Pilot-Scale Purification Operations .........ccociiviervieiviieniiee et esseeeeesre st e seesaasen e saesseanes 73
i Production-Scale Operations ...ttt e ettt e st e eatean e baaten 73
] Batching and Dispensing Operation ..........cccieeriieriireie ettt e O 75
Loading and Draining Operations ...........c.ccoioiiiiiiiiiiiiiii ittt et 75
P A SOIVICES cuuiiiviiiie ettt ettt ettt ge et at e s aateseeane e eraeeneesaeenseenneetans 76
Experimental Preparation of Simple Fluorides ... 76
Fundamental Chemistry of Fused Salts ..ot 77
EMEF MeaSUrEMENTS ..oiviiiiieiieiiii ettt e s bb s e b e bbbt et es b st e sb e nb b 77
Vapor Pressures of LiF-ZrF , Mixtures ... s 80
Solubility of Xenon in Fused Salts ... 81
X-Ray Diffraction Study of Liquids ..o 8]
High-Temperature X-Ray Spectrometer Studies ....oooviiioiiiiioiiie e eeeer e srasssine e 81
5. CORROSION RESEARGCH ..ottt ettt bbbkttt ese s enenee 83
4 Forced-Circulation Studies ..o ettt 83
; Fluorides in Inconel . ..., e - 83
Sodium in INCONEI .. oo et cereeeieereeens 86
5 Thermal-Convection StUIEs .....ooooiiiiiicee ettt eae e et srs e srs e eans et 88
Alkali-Metal Base Mixtures with UF and UF in lnconel oo et eeeeteeraaaaareeaans 88
3 Zirconium Fluoride Base Mixtures W|th UF, and UF, in Inconel oo, 90 -
1 Effect of Temperature on Mass TranSFEr ..o ovoiieeeeeee ettt enns 91 E
i Effect of Loop Size and Shape on CorroSion .......c.ccccoeerivieieeisresienseseesesscesss et ssssssss e 1
{ Evaluahon of Control Loop Results ..ttt e 93
1 Generdl Corrosion STUGIES ..ot e 94 s b
; Brazing Alloys on Type 310 Stainless Steel and **A’’ Nickel in Sodium and ‘ o F
in Fuel Mixtures ... s . 94 :
Screening Tests of Solid-Phase Bonding ..ot e s 9% v
Sodium in INCONE] .. ..o e e ettt b ettt eenes e 101

 

 

O vildi

   
 

 

Lithium in Type 347 Stainless Steel. ... 101

- , Versene Cleaning of Beryllium-Inconel Sysfems ................................................................................ 102

Fundamental Corrosion Research et e bRk SRR bbb e et s 105

= Film Formation on Metals ... cvvreernsnneenes 105

| Mass Transfer and Corrosion in Fused Hydroxide ... 110

Chemical Studies of Corrosion .......ccccoevvvvevienvvinencnneenn, e et et e e ettt e e ara e st e e et 114

Corrosion of Inconel by LiF-BeF, and by LiF-BeF -UF , s 114

Effect of UF ;-UF , Mixtures on Corrosion of Inconel by Various Solvents ... 114

Studies of the Sodium Hydroxide—Nickel Reaction ........cocooiiinininiiiii, 115

6. METALLURGY AND CERAMICS . oot sre s er e e eb e st 119

Development of Nickel-Molybdenum Aoy s .......cociviiiiviiieieieiee st 119

Fabrication SHUAIES .ovv ottt ereb ettt e er e st bbb b anes s 119

Oxidation Studies . ....o.ocevoiiiiieiice e, e 123

Stress-Rupture Studies of Nickel-Molybdenum Alloys ... 124

HOSTEIOY B oottt 124

Modified Nickel-Molybdenum Alloys ... 125

Tensile Properties of Hastelloy B ..o 125

Development of Brazing AHoys ..., ettt 128

High-Temperature Oxidation Tests ..., 128

Phy sical Property Tests oo e 129

Fabrication of Test Components.....c.ccoociiiniiiiicicciiccrric e e 129

Twenty-Tube Heat Exchanger ... 129

Intermediate Heat Exchanger No. 2. ...t 131

- NAK 10-ATr RAAIGEOIS ...oiiiiiiiiiiiiiicieet et e ettt s e sttt et et s s e st ebe s sbe e b b eas s enan 134

" Cornell Radiator No. T oot ettt es bbbt s i ea e s s ensn st s e 139

- - Special Materials Studies......cccoovirrennn. ettt ettt ettt anenes 140

3 Special Aoy EXIUSIONS ottt 140

' Clad-Columbium Fabrication .ottt et bbb e st s ene s 140

B6C~Cu Shielding Material .. ..c.ooiviiiii e 142

Magnesium-Lithium Aoy .o 143

Welding of Columbium Thermal-Convection Loops ..o, 144

Dimensional Stability Test on an Inconel Spun Core ..o 144

Brazing of Cermets to Inconel ... ... 145

Ceramic Research .....c.ioovevennenne. ettt teeststateasateansiteteseshaEeEetat e aeate et e Rt et et e see e e R e sarat e e bearerearennerene 145

’ RAre=Earth CeramiCs oo eeeieeeeeiesree e et st e e s ta ettt etaseese s as e e ebeabe e smecamebeeaaesnesssressasarn s bn st e arabasabaeabes 145

UO, Particles Coated with ZrO ..o 147

Graphite-Hydrogen Reactions and Erosion lnvestigation ..., 147

7. HEAT TRANSFER AND PHYSICAL PROPERTIES ... 149

Fused Salt Heat Transfer.................... s et e et 149

ART Fuel-to-NaK Heat Exchcnger ..... ettt e e s 149

ART Core Hydrodynamlcs e ' ' 151

Reactor Core Heat Trcmsfer s 154

- Free Convection in Fluids Having a Vo[ume Hecf Source et eibeeas e b ettt et er et b s st en ettt s et e 154
3

HEG CaPACHY .ooiiiiiii et bk sa s e S0 h bbb 156

ix

 

 

 
 

 

 

 

10.

1.

12.

VESCOSIEY oottt ettt ettt ettt bbb h st st et et b e bbbttt bs bt e aa bt s bbbt a ettt be Rt e et b e s - 157
Thermal COndUCHIVITY «.eomeeereeeeeeseeeeeeeesseeeeseeeeeeseeeseeeeresessoeseeseeeeees et 159
Electrical ConducHiVity ..ot ....... 160
RADIATION DAMAGE .ottt bbb es bbb s ettt 161
MTR Static Corrosion Tests .............. e ettt eteanes e 161
LITR H'ofiiontal-Beam-Hole Fluoride-Fuel Loop ....cocoooiii e e . 163
Deposition of Ru'03 in LITR Fluoride-Fuel Loop .ooooivorieeeeeeeeeoeseoseeoes e eeeeeeeeeseeeeesesssese s eseeer e 167
Miniature In-Pile 00D ettt et ettt ettt e 168
Delay of Fission Gases by Charcoal Traps ..o 169
~ Creep and Stress-Corrosion Tests ..o, e 170
" A Theoretical Treatment of Xe 135 Poisoning in the ARE and the ART ... 171
ANALYTICAL- CHEMISTRY OF REACTOR MATERIALS ..... 174
' Défé'r'min'ation of Uranium Metal in Fluoride Salt Mixtures ..o..oooooevooovevoooeosees e 174
Det.err.niinction of Trivalent Uranium in Fluoride FULS w.....ovvovrorooeoeeoeeoeoeoeoeoee 175

Oxidation of Trivalent Uranium by Methylene Blue ..o, 175

Simultaneous Determination of Trivalent Uranium and Total Uranium........c.ooooii 176
Determination of Lithium in LiF-BeF, and LiF-ZrF4-UF4 ................................................................ 176
Volumetric Determination of Zirconium in Fluoride Fuels with Disodium Dihydrogen

Ethylenediaminetetraacetate .. ..ottt ettt e 177
Determination of Oxygen in Fluoride Fuels ........cocoooiiiiiii e 178
Determination of Oxygen in Metallic Oxides by Bromination .........c.cccoeiiiiiinnnnccececnnes 178
Determination of Oxygen in Beryllium Oxide by Acidimetry ........c.ooiiiiiiiiiieceee 179
Determination of Trace Amounts of Nickel in Fluoride Fuels with Sodium

Diethy Idithiocarbamate ... e e e b e er e s 179
ANP Service Laboratory . ..ottt et eb e ere et ebe et e asEr st et ar et ans 180
RECOVERY AND REPROCESSING OF REACTOR FUEL .ooooooooooooooeoeooeeeeeeeeeeeeeeeeeee oo 181
Pilot Plant Design ....coiiieiiiiiceiee ettt ettt st s et e enere e te e se s eneeneteensens 181
Process DeVElOPMENE o .ottt ettt et n et et e ettt eaee e e 181
Corrosion Studies 182

PART Ill. SHIELDING RESEARCH
SHIELDING ANALY SIS Lottt et eass et s e st e aeas - 191
 Gamma-Ray Distribution in a Circulating-Fuel Reactor and ShIEld oo 191
Enérgy Absorption Resulting from Gamma Radiation Incident on a Multiregion Shield | ._ -

With SIab GeomMEtry ..ottt e nes e, 192
Energy and Angular Distribution of Air-Scattered Neutrons from a Monoenergetic Source......... 192
Analysis of the Constant-Velocity Transport EQUation..........ccocooovmeeeoeeccrecereereorereseesecessisscercrssine 193
LID TANK SHIELDING FACILITY ........... . 194

| Reflector-Moderated Reactor and Shield Mockup T SES oo 194

st

 

 

 
 

 

 

 

|

)

Gamma-Ray Dose Rate Measurements ...t e ve e 194
NeUtron MeaSUFEMENTS ...ttt ettt se st s et et e eae st st reee e e aeene 197
Sodium Activation in Heat Exchanger Region ... 198

13, BULK SHIELDING FACILITY oottt eb ettt s 200
GE-ANP Air-Duct Mockup EXperiment ......c.ocooioiiiiiii et 200
The Spectrum of Fission-Product Gamma Rays ... 203

14, TOWER SHIELDING FACILITY Lot 205
The Differential Shielding Experiments at the TSF: Phase | ... 205
Measurements in the Detector Tank ...t e 205
Measurements in the GE-ANP Crew Compartment ..., 205
Analysis of the Differential Shielding Experiments ... 206
Definition of Dose Scattering Probability ... 208
Evaluation of Probability from TSF Experiments ..o, 213
Evaluation of Direct-Beam Integral ..o e 214
Calculation of Scattered Dose ... 215
Effects of Direct-Beam Collimation ... e 216
Effect of Neutron Energy Spectrum ... e 216
Application of the Differential Shielding EXPEriments ....coooiiooiooeoeeee oo eeeeeeee oo 217

PART IV. APPENDIX
15. LIST OF REPORTS ISSUED FROM FEBRUARY 1955 THROUGH MAY 11, 1955 __..coooooooovceeeeenee. 227
ORGANIZAT'ION CH AR T S et e 231

Wfif xi

 
 

 

it

o

\"16 5%, on:4 k

savmg Il"l urcmlum concen’rra’rlon “of the fuel cnd a

shREY

ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART I. REACTOR THEORY, COMPONENT
DEVELOPMENT, AND CONSTRUCTION

_1.' Reflector-Moderated Reactor

The development of the reactor layout is con-
tinving. New features that have been incorporated
because of stress, fluid flow, or fabricability
considerations include an elliptical fuel expansion
tank, a rounded dome to enclose the top of the
reactor, a newly designed sodium pump impeller,
and other related items. Recently completed heat
exchanger tests yielded consistent data from which
a series of heat exchangers is being designed.
The most promising of these will be chosen for
the ART.

The preliminary layout of the interior of the ART
test cell which shows the major items of equip-
ment and the recommended provisions for support
was completed. A similar layout for the NaK
piping and radiators was also completed so that
drawings for the building and concrete work could
proceed. More detailed drawings of the cell that
will show the disposition of the small items and
the instrumentation are being prepared.

Information block diagrams showing the basic-
control actions desired for the ART have been
prepared and will be used as a basis for selection
of the hardware types and the control techniques
to be used and for determination of areas of con-
trol that will require new component development.

The room-temperoture critical experrmenfs for

‘the ART have shown ’rhot 'rhe critical mass de

" creases’ very rapldly as the neutron obsorphon in

~ _the core shell decreases. Hence, replacemenf of

- the hlghly absorbmg Inconel by a low-cross- sectlon '

“:'material, such os columblum, would brlng about 0

substantial scvmg in uranium ‘investment and,

mudenfclly, make the temperature ‘coefficient more

~ ““negative. For fabrication | reasons, only about 50%

of the Inconel could be replaced by columbium. Itis
esflmofed _that the effect of a 0% replacemenf

ical mass by about

   
    

3 wh C:h 'o : onds Yo a 0.7 mole %

 

U235 ivestment.

Calculations were made of the temperature coef-
ficients of the ART that check well with previous

7-kg saving in total

multigroup calculations. The results support the
postulated negative over-all temperature coef-
ficient. It has also been found that, if as much as
5 wt % lithium were added to the NaK in the
secondary coolant circuit, the reactivity of the
ART would be a fairly sensitive indication of a
NaK ledak into the fuel circuit. Burnup and gamma-
ray heating of rare-earth oxides being considered
as control rod materials are being studied.

2. Experimental Reactor Engineering

A series of design changes have been made in
the nose, or heater, section of the in-pile loop as
more information concerning the neutron flux in
the MTR beam hole and the flux depression of the
materials of loop construction has become avail-
able, The present nose section consists of a
2V sturn coil (an increase of 1 turn) with its axis
parallel to the beam hole center line. The flux
seen by the fuel in the in-pile loop is now expected
to be 30% of the unperturbed value, and the average
power density in the nose section will be 0.7
kw/em3.

A working mockup of the in-pile loop has been
completed and is being operated with the fuel
mixture NaF-Zer-UF4 (53.5-40-6.5 mole %) at a
Reynolds number of about 5000, a temperature
differential of 175°F, and a maximum fuel temper-
ature of 1500°F. Resistance heating is employed.
The experience gained will aid in operation of the

in-pile loop at the MTR.
The fission gas holdup system for the loop is

" being tested with flow rates of 0.15 and 0.03 scth
of helium with about 0.13% krypton, After 15 days

“"of operation, no significant amounts of krypton

were  getting through the liquid-nitrogen-cooled
charcoal-adsorption traps. Assembly of the first

loop for the MTR in-pile experiment is under way,
“ with operation scheduled for the next quarter.

Three additional test stands were installed for

‘the operation of resistance-heated high-velocity
- forced-circulation |oops with iorge tempero’rure

d:fferenfluls for mveshgahng corrosmn "and mass
“transfer of Inconel by fluoride fuel mixtures. Also,
three more gas furnaces were placed in service as
heat sources for these loops, bringing the total

 
 

o il o e bifig

 

 

 

 

- Golhio . FEiieobio S

ANP PROJECT PROGRESS REPORT

humber of test stands to ten. Operation of 16

loops was terminated during the quarter. The
majority of these loops operated for 1000 hr in
the Reynolds number range of 1,000 to 15,000 and
with temperature differentials of 100 to 300°F.

 The ‘éx‘éés's'iv'e' high wall temperature in the bends

of electric- resnstonce heated loops was corrected
by relocation of the heating elements so that heat
is applied in only s?rmghf sections of tubing.

Four test stands are now in operation for studying

 éotrosion and moss frcmsfer of sodium-Inconel and

sodlum—lnconel-s’rcmless steel systems. Oper-
ation of six such loops was terminated this quarter
at the conclusion of periods of either 500 or 1000
hr with maximum sodium temperatures of up to
1500°F. Appreciable deposits of mass-transferred
material were found in the cold legs of these
loops, and fherefore a controlled series of experi-
ments’ was “started. The loops will provide infor-
mation on ‘fHé‘éffecfs of the oxygen content of the
sodlum, the use of a cold trap, the use of a lower
temperature, and the use of an all-stainless-steel
system,

- The ART fuel pump (model MF-2) was operated
in water performance tests. Some cavitation noises
were present in all tests, but no serious effect of
the apparent cavitation on performance could be
found. It is estimated that the efficiency of the
pump at design point, exclusive of seal and bearing
losses, is 70%. New inlet volute configurations
are currently being designed and tested. The
bearing and seal, cold mechanical shakedown, and
high-temperature test stands are being fabricated
and assembled.

An extensive program of heat exchanger testing
is under way for obtaining reliable heat transfer
data on fuel-to-NaK heat exchangers of the general
type and configuration for ART application. Con-
siderable information will be obtained on the effect
of corrosion and mass transfer on materials of
fabrication in high-temperature, high-heat-flux,
NoK and fluoride systems. An opportunity will
also be provided for ascertaining the structural
integrity and reliability of fabrication of heat
exchangers and radiatou:s supplied by outside

vendors.

_The program involves the operation of three
intermediate heat exchanger (IHE) test stands and
two small heat exchanger (SHE) test stands, such
as the SHE stand now in operation. The IHE

stands will be used to test large tube bundles

(about 100 tubes) of the general size and configu-
ration of ART heat exchangers (arranged for
regenerative operation), while the SHE stands will
be used to test smalier, more easily fabricated,

tube bundles (20 to 50 tubes). The SHE stand now

in operation will be modified for use as a general

test loop following termination of the current test.

Design of all IHE test stands is complete, and

procurement and assembly are well under way.
Design of the SHE test stands and procurement of
equipment are under way. -

Water tests with a full-scale aluminum model of
the 21-in. ART core and the entrance header region
were initiated. Without inlet guide vanes, the
fluid was observed to enter the core at an angle of
about 70 deg from the vertical. Flow reversal at
the island was observed. Data from these tests
are being used in the design of turning vanes and
vortex generators to correct this unacceptable flow
condition, ' I .

A second thermal-cycling test of a sodium-
Inconel-beryllium system was completed. Over
100 thermal cycles were applied to the beryllivm
piece between the range of the high power level,
61 w/cm® to the beryllium, and the low power
level, 2.5 w/ecm®. The sodium temperature from
inlet to outlet at the high power level ranged from
1050 to 1200°F. Inspection of the beryllium piece
after the test {total operating time, 1030 hr) re-
vealed three axial cracks on the outer surface of
the hot end of the beryllium,

The 100-kw gas-fired heater was tested and was
found to perform very satisfactorily. Minor modi-
fications can be made to this heater that will
increase the capacity.

3. Critical Experiments

The critical assembly of the reflector-moderated
reactor consisting of the beryllium island and
reflector enclosing the fuel region and having axial
extrusions simulating the exit and entrance flow
channels was loaded with sufficient uranium to
give several per cent excess reactivity. This
overloading was used to evaluate some of the
materials of interest in the design of the ART prior
to dilution of the fuel to the critical uranium
density. ' o

Measurements were made with samples of a
mixture of the oxides of the rare-earth elements

being considered for the absorber material of the
control rod of the ART. A cylinder of the mixture

FRERET T

 

 

 

 

 

 

 
 

=f

)

e cal properhes ‘l_oeh‘
:..,f:v_{:NaF ZrF4-UF4*system' has been’ found . -
Study of the analogous NaF-LiF-BeF, system

“has shown fhcn‘ adequateiy low’ melflng pomfs are
“Yespect 16 pure UF, and uranium metal at temper-

” cn‘ures below 1400°C

0.79 in. in diameter and 21 in. long decreased the
reactivity nearly 2%; measurements with shorter
lengths of the material indicated that the total
value of the rod could be mcreosed by 60% if the
diameter were increased to 1% in. Tests made on
tubes of several different dimensions gave data

for the design of the control rod guide thimble.

In other experiments, columbium was shown to
be somewhat less poisonous than Inconel in a
neutron spectrum similar to that expected in the
ART, and it was found that a layer of beryllium in
the center of the fuel channel of the critical
assembly reduced the critical uranium concen-
tration by a few per cent. Fission rate distri-
butions were measured across the flow channel at
several locations in the central reactor region and
in the end duct.

A high-temperature critical experiment embodying
the nuclear characteristic of the ART has been
designed which will operate at zero nuclear power
at about 1200°F. The purposes of the experiment
are to measure the critical uranium concentration,
the temperature coefficient of reactivity, and the
effectiveness of control rods.

PART |.| MATERIALS RESEARCH
4, Chemlstry of Reactor Mufenals

Interest in obfqmmg fuel mixtures more suitable
for use in a C|r¢u|ahng -fuel reactor than those
available in the N'cx'F-Zi'F“'-UF“ system has led to
evaluation of the NaF-LiF-ZrF, ternary and the
NaF-LiF-ZrF,-UF, quaternary systems. The NaF-
LiF-ZrF, system has been reasonably well de-
fined, but much work remains to be done on the
quaternary system. Phase-equilibrium data show

that quite low melting pomfs are available in the

ternary system ‘and that adequa'rely Tow meltmg"-'}"’
points are available at Zrl':4 cohcentrations as low
. as 21 mole %, however, no composition ‘with physr- '
fter than those avollable in fhe_'_ ;

 

available over wide areas. Phys:cal properfy dafa
- -have not yet been ‘obtained in sufficient detm[ for
:"compquson wu’rh avo_[lable fuel. mo’rertal Pre-v:
- .llmmary dofo on the éo '{fblllfy of UF3 in BeF
k becrmg composmons e »
data show a trend of increasing UF3 SOIUblIIfy-M-

 
 

obfumed The sc cm‘ered

with increasing temperature and decreasing BeF,
content, but in no case does it appear that the

 
 

"PERIOD ENDING JUNE 10, 1955

solubility of UF, at 600°C is sufficiently high to
provide more than. a fraction of the concentration
needed for an ART fuel.

Previous measurements of the partial pressure of
HF at equilibrium during the reduction of FeF,
by hydrogen in NaZrF, showed higher valuves than
would be predicted from thermochemical data and
ideal solution behavior, and it was postulated that
the higher values were due to a lowered activity
of the metallic iron because of alloying with the
nickel apparatus. The postulated alloying has
been confirmed by low values for final FeF, con-
tent of the melt and by chemical analyses of
portions of the nickel apparatus.

Additional data were obtained on the reduction
of UF, by structural metals. Data on the reduction
of UF, by metallic chromium with NaF-Zer-UF“
(51-45-4 mole %) showed that, in comparison with
the data obtained with NaF-ZrF“-UF4 (48-48-4
mole %) as the solvent, an increase in the final
mole fraction of NaF from 0.48 to 0.51 in the melt
containing uranium causes a significant decrease
in the equilibrium CrF, concentration. Data for
the reaction of UF, with metallic iron in these
mixtures and in NaF-KF-LiF (11.5-42-46.5 mole %)
agreed very closely and were somewhat higher at
600°C than ot 800°C. Some, as yet inconclusive,
data were obtained on the reaction of UF4 plus
UF. with chromium metal and Inconel in NaF-KF-

LiF (11.5-42-46.5 mole %) at 600 and 800°C. Ad-

ditional data were obtained that confirmed previ-
ous findings that FeF, is relatively stable in the
NaF-KF-LiF eutectic at 600 and 800°C and that
CrF, is not stable.

Previous evidence that UF, was more stable at
elevated temperatures than free energy estimates

‘"had indicated was substantiated by vapor pressure

measurements on UF, in the temperature range
1270 to 1390°C. The disproportionation pressure

"curve that was ob'romed shows that the dispro-
~ “portionation of UF_ is far from complete under the
“conditions prevallmg in the vapor pressure cell

and that UF, is thermodynamically stable with

In the’ mvestlgahon of varmbles ‘affecting the

“reduction of UF4 with-uraniom in alkali fluorides,

the effects of the surface area of nickel or copper

“exposed to the melt and of adding excess uranium

metal were studied. |t has become apparent that
alloying of nickel and uranium can occur ot temper-

 
oI

 

 

 

ANP PROJECT PROGRESS REPORT

atures far below the minimum nickel-uranium
liquidus temperature (732°C) and, consequently,
that disproportionation of UF, can be expected
to occur at the temperatures of interest at nickel
surfaces or at any metallic surface with which
uranium can alloy.  Preliminary evidence was
obtained which indicated that UF, dissolved in
an alkali fluoride mixture is more stable in copper
than in nickel.

Investigations of methods for rapid purification
of fuel mixtures included attempts to use electroly-
sis under a hydrogen atmosphere to remove oxides
in order to avoid the container corrosion that
results when HF is used and to use metallic zir-
conium to replace most of the hydrogen in the
stripping operation. The use of zirconium metal
was demonstrated on a 5- and a 50-1b test scale
and was found to be a quite rapid and effective
method for purification if small quantities of UF
in the product are tolerable or desirable. In

- electrolysis experiments, the ZrF ,-bearing mixtures

could be electrolyzed smoothly, but the alkali
fluoride mixtures gave variable results. Attempts
to prepare mixtures containing UF, and no UF,
were unsuccessful, the largest UF, content at-
tained being 85% in an NaF- ZrF, base.

Fifty-six pilot-scale preparonons totaling 630 ib
of material were produced in various compositions
for small-scale corrosion studies, for physical
property determinations, or, in many cases, for
use as purified intermediates in phase-equilibrium
studies. Uranium trifluoride was a component of
nearly 25% of the materials requested. Production
operations were resumed on March 1, 1955, on a
three-shift, five-day-week basis to provide test
material for the greatly accelerated ANP engi-

~ neering program. A total of 4800 b of purified

material was prepared during the quarter. Attempts
to flnd a commercial source of ZrF, are under

~way. |f a commercial source is not found, it will

be necessary to expand the Y-12 production facili-

" ties immediately, Two batches of enriched fuel

were prepared for use in an in-pile loop, and prepa-
rations are being made for the production of the
materials for the proposed high-temperature critical
experiment. '

Potential measurements were made with combi-

* nations of s'-ever‘ol__ half cells consisting of metal
- electrodes bathed in solutions of the corresponding

metal ion in the molten salts. The temperature

) rdh’éé"'stqdi'éd.' was, in general, 550 to 700°C.

Cells consisting of zirconium rods immersed in
various NaF-ZrF, melts and cells consisting of
metallic nickel electrodes immersed in solutions
of NiF, in molten NaF-ZrF, melts were studied.

Vapor pressure measurements of mixtures in the
LiF-ZrF, system were started. The data showed
the vapor pressures of the LiF-ZrF, mixtures to
be considerably higher than those of the corre-
sponding NaF-ZrF, mixtures.

An x-ray diffractometer for studying the structure
of liquids has been constructed and is undergoing
final testing. A high-temperature attachment for
an x-ray spectrometer has been used for studies
of compositions in the systems NoF-ZrF,, LiF-
ZrF,, and NaF-Ber.

5. Corrosion Research

Several Inconel forced-circulation loops that
were operated with fluoride mixtures and with
sodium as the circulated fluids were examined.
The fluoride mixtures included ZrF ,-base mixtures
with UF, and with combinations of UF3 and UF
and an olkch-mefol base mixture containing UF
Favorable results were obtained with the ZrF -
base materials in that the depths of attack were
no deeper than have been found previously in
thermal-convection loops. Attacks as low as 5
mils in 1000 hr appear to be obtainable. The
conversion of some of the UF, to UF, decreases
the attack. The most important variables appear
to be maximum wall temperature and the hot-zone
surface-to-loop volume ratio, whereas large vari-
ations in velocity and Reynolds number have very
little effect on the depth of attack. The alkali-
metal base mixture containing UF, caused exces-
sive mass transfer and a very heavy intergranular
concentration of subsurface wvoids to a depth of
35 mils. The mass-transferred deposit in the cold
zone was up to 85 mils thick.

Mass transfer of large amounts of nickel metal
was found in the Inconel forced-circulation loop

- that circulated molten sodium at 1500°F. Layers
of dendritic metal crystals up to 26 mils thick

were found to have formed in 1000 hr. The use of
type 316 stainless steel in the cold portions was
found to reduce the mass transfer slightly, but
further study of the variables in the process is
needed to confirm this finding. :
Alkali-metal base mixtures containing combi-
nations of UF, and UF, were circulated in Inconel
thermal-convection loops, and, when about 2 wt %

 
 

 

4y

)

kY

yranium was present as U3+, low depths of attacks
and no deposits were found. Higher U** concen-

trations resulted in decreased attack, but hot-leg
layers were found.

Inconel thermal-convection loops in which ZrF ,-
base fuels containing about 2 wt % uranium as
U3* were circulated did not show so large a re-
duction in attack as was found in the loops that
circulated alkali fluoride fuels also containing
2 wt % uranium as U3¥, However, some reduction
in depth of attack and a fair reduction in amount
were found. The effect of the hot-leg temperature
(1200 to 1600°F) on mass transfer was investigated
in several Inconel thermal-convection loops oper-
ated for 1500 hr. A definite increuse in depth of
attack with an increase in hot-leg temperature was
noted that may be attributed to mass transfer,
inasmuch as loops operated previously for 500 hr
did not show the effect of temperature on depth of
attack. The mass transfer effect is masked by
the effect of impurities and nonequilibrium conditions
during the first 500 hr.

Considerable work is being done in an effort to
find the cause of the increases in depth of attack
and the nonuniformity of results now being ob-
tained in Inconel thermal-convection loops operated
as control loops under standard conditions. Con-
tamination during filling or operation does not
appear to be the cause of the difficulty.

Corrosion tests of brazed type 310 stainless
steel T-joints in static sodium and in static NafF-
ZrF4-UF4 (53.5-40-6.5 mole %) showed the brazing
alloy 9% Si-2.5% P-88.5% Ni to be satisfactory

in both mediums. Similar tests of brazed ‘A"

nickel T-joints showed the following brazing alloy's

to be satisfactory in both mediums: 90% Ni~10% P,

80% NI—IO% Cr—]O% P, and N:crobrqz Seesaw
_corrosmn fesfs in both mediums on brazed |ncone| ‘
T-|01nfs showed the Coast Mefcls alloy No 52t0

- have the best resus'rcmce in sodium and in 'rhe"_," \
ol uorlde mixture. |
" Since fhe sfrucfurai mefcl alloys ‘that have beenhw
: *'-*:proposed for use in fl'\e fabrication of reactor fuel

and coolant lines have a tendency to form solid-
phase bonds at elevated tempefatures in liquid

metals and in fused salts, they are unsun‘ob[e for
use in valves bearlngs,' and seals. Therefore
cermets and ceramics that do not bond to each
~ other but that can be bonded to the structural metal

alloys are being investigated. The cermets (metal-
bonded ceramics) appear to be the more promising

~ PERIOD ENDING JUNE 10, 1955

because of their high corrosion resistance and
other satisfactory chemical and physical properties.
Several Kentanium cermets tested in NaF-ZrF -
UF, (53.5-40-6.5 mole %) at 1500°F for 100 hr were
found to have good resistance to solid-phase
bonding if the contact pressure between the speci-
mens did not exceed 50,000 psi. The compositions
of the cermets tested were:

80 wt % TiC-10 wt % NbToTiCs—IO wt % Ni
70 wt % TiC=10 wt % NbTaTiC3-—20 wt % Ni
64 wt % TiC—6 wt % NbTaTiC,~30 wt % Ni
64 wt % TiC~6 wt % NbTaTiC, -
25 wt % Ni=5 wt % Mo

Lithium was circulated in two stainless steel
thermal-convection loops for periods of 1000 and
3000 hr, respectively. The hot- and cold-leg tem-
peratures were 1000 and 550°F, respectively.
Operation was satisfactory throughout the test
periods, and macroscopic examination revealed no
mass-transferred crystals in the loops or in the
{ithium drained from the loops. Metallographic
examination revealed subsurface voids and a
ferritic surface layer 0.3 to 1.0 mil thick in the
loop operated for 1000 hr and 1.0 to 1.5 mils thick
in the loop operated for 3000 hr. Lithium metdl
had penetrated to the depth of the subsurface voids.
The weld zone of the pipe was attacked to a depth
of 3 mils in the 1000-hr test and 4 to 5 mils in the
3000-hr test. A few small (0.2-mil) carbide par-
ticles were found attached to the wall in the cold-
leg sections of both loops.

Determinations of the oxidation rate of sodium

" have been made at —-79, —-20, 25, 35, and 48°C.

The experimental results indicate that, contrary

~to current oxidation theory, the oxide films formed

on sodium are highly protechve in the absence of
water vapor. The rate curves do not conform to

any of the ‘“standard’’ rate equations reported in
 the fiterature, but they are quchfohvely comparable
to low-temperature curves for copper. It is hoped
‘to clarify the oxidation mechanism associated with

a linear rate law through a careful study of the

structure and composition of the oxide films formed

on columbium in the neighborhood of 400°C.
Studies of corrosion and mass transfer by fused

’-Thydromdes indicate that both nickel and Inconel

may possibly be compatible with hydroxides at
temperatures of about 600°C. Hastelloy B is
unsatisfactory because of its poor corrosion re-

 
 

 

 

e A i RO s b o i B i s SR

ANP PROJECT PROGRESS REPORT

sistance. In the temperature range 600 to 700°C,
there is evidence of an accelerated rate of mass
transfer with both nickel and Inconel, as well as
corrosion of Inconel.

‘The Ber-Bearing mixtures LiF- -BeF, (69-31
mole %) and LiF- BeF,-UF, (67.3-30.2-2.5 mole %)
were tested in Inconel capsules in 100-hr tilting
furnace tests. No evidence of attack by either
mixture was found. A study of the effect on cor-

~rosion of the rcfrid of UF3 to UF, in various

solvents revealed that increasing the UF_ content

‘up to 50% was beneficial and that any further

increase had little effect.
v '6. Metallurgy and Ceramics
~Investigations were continued in the study of the

properties of nickel-base alloys containing 15 to
32% molybdenum, ternary alloys with a nickel-

molybdenum base, and Hastelloy B. Attempts are

being made to improve Hastelloy B with regard to
fabricability, oxidation resistance, and mechanical
properties. Additional evidence has been obtained
which indicates that the poor high-temperature
fabricability of the commercial material is related
to the impurity content; however, it is felt that the
superior strength of commercial Hastelloy B may
be derived from the impurities. Therefore mechani-
cal property tests are under way on an alloy with
the nominal Hastelloy B composition, 4% Fe-28%
Mo—~68% Ni, but without the tramp elements va-
nadium, silicon, manganese, cobalt, chromium,
tungsten, and aluminum.

A tube blank extruded from a vacuum-melted cast
billet of commercial Hastelloy B fractured during
the first step of a reduction operation; however,
a blank made from a wrought billet was success-
fully reduced from 1.5-in.-OD, 0.250-in.-wall to
0.187-in.-0D, 0.017-in.-wall seamless tubing. Two
impact extrusions of as-cast vacuum-melted com-
mercial Hastelloy B were made at 2000°F with
good recovery of sound rod. Attempts to roll the
rod at 2000°F were, however, unsuccessful; the
material cracked severely. ‘

'De'sign curves were prepared from the results

. Vof creep-rupture tests of Hastelloy B sheet in the
‘solution-annealed condition in an argon atmosphere

at 1500 and 1650°F. A comparison of these data

wnth preliminary ‘data from tests in fused salts

shows that properties of the alloy in the fused
salts are actually superior to the properties in an

argon atmosphere. Creep tests in air, in argon,

and in hydrogen at 12,000 psi showed the effects
of argon and hydrogen to be similar, but the effects
of air followed closely the pattern observed for
Inconel and ““A” nickel in air, insofar as reduced
creep rate and longer rupture life are concerned.
However, the final elongations of Hastelloy B in
air are equivalent to or lower than those in argon,
whereas the elongations of other nickel-molybdenum
alloys are markedly greater in air than in argon.

A program has been initiated to determine whether
aging treatment results in serious embrittlement
at service temperatures. Typical microstructures
have been obtained from specimens heat-treated
at temperatures from 1100 to 1600°F for times from
100 to 1000 hr, and a correlation is to be made
between physical properties and the microstruc-
tures. The relative merits of various preaging heat
treatments of Hastelloy B are also being studied,
and it is hoped that, as a result of this work, a
procedure can be developed that will stabilize the
microstructure sufficiently to reduce the sensitivity
to high-temperature aging.

The binary nickel-molybdenum alloys that have
been studied have included 15 to 32% molybdenum,
and the ternary alloys contained 20% molybdenum
plus nickel and one of the following: 3 to 10% chro-
mium, 2 to 10% columbium, 2% aluminum, 1% tita-
nium, 2% vanadium, or 1% zirconium. The results
of stress-rupture tests have shown low strengths
and ductilities for most of these alloys and indi-
cate that vacuum melting alone is not sufficient
to obtain optimum properties. The possibility of
increasing the strength and ductility through cerium
additions appears to be promising, since the ad-
dition of cerium has been shown to improve the
physical properties of the alloy. It has become
apparent that improved deoxidation practices are

necessary in melting these alloys, and efforts are

being directed toward the production of sounder
ingots.

Oxidation tests of nickel-molybdenum alloys
containing 3 to 10% chromium have shown that
about 10% chromium is necessary to form a non-
spalling protective oxide on the alloy. However,
the oxidation rate under static conditions can be
reduced 50% by the addition of 3% chromium.

Cyclic tests consisting of 190 air cools from

1500°F in 500 hr have now been completed for

brazing alloys previously evalu‘otec{' in static
oxidation resistance tests at 1500 and 1700°F.
Most of the alloys tested showed good resistance

v

 

 

 
 

 

 

to oxidation under both static and cyclic con-

. ditions. Cyclic tests at 1700°F are now under way
on these brazing alloys, which include commercial
alloys, experimental nickel-base alloys, and ex-
perimental precious-metal base alloys. Physical
property tests showed that the physical properties
of Inconel are not impaired by the brazing process.
The fabrication of several major experimental
heat transfer test components was completed. The
items fabricated included a 20-tube fuel-to-NaK

<

Inconel heat exchanger, a full-scale 200-tube fuel-

to-NaK heat exchanger (intermediate heat ex-
changer No. 2), two 500-kw NaK-to-air radiators,
and a full-scale integral-fin liquid metal-to-air
radiator designed by the Cornell Aeronautical
Laboratory. The combination welding and brazing
procedures used in the fabrication of 1rhese units
“are described. :

Three billets of vanadium were extruded at
2000°F. Tubing prepared from these blanks is to
be clad on the outside with stainless steel, and
the clad tubing will be used in corrosion studies.
Four high-purity molybdenum billets containing
0.7% titanium were extruded into rod to be used in
welding studies. Flow pattern studies of the
extrusion of duplex and three-ply composites were
continued. The three-ply materials are to be used

- in the production of stainless-steel-clad seamless
tubular fuel elements. Attempts were made to clad
columbium with the following heat-resistant alloys:
types 446 and 310 stainless steel, Inconel, and
Hastelloy B. Columbium was found to be ade-
quately protected by each of these alloys, and the
effects on ductility were slight; however, interface

@

reactions occurred in all combincfi'ons, and there

~was, as d result, separahon at the mterfaces

Dn‘fusnon barrlers for assuring bonding at 'rhe
_interface were fherefore studied. Preliminary evi-
dence indicates that copper W|H be a sutfcble 7

barrier for ?he columbium.

 

 

1 __ sensitive; at 200°F the strength was a factor of 4
g™ : below that at room temperature.

S fnveshgaflons of the fabrlccbxllfy and mecham_.
e éu] Proper’flesl"""f_boron con’ralnlng shleld materials

were confmued, ‘and an evoiuahon is bemg mcdef’;f"

. of a mogneswm ||'rh|um clloy for use as a struc-

- 'fural shleld muferlcxl The addmon of 20% ll‘l‘hlumb,::.l:

. _,_‘ro mogneswm reduced 'rhe tensile s?rength and‘j—_-‘:,

- T mcreaSed the ‘ductility of “the mcgnes:um.- Data ™
| " obtained at room temperature and at’ 200°F “ahaw
the strength of the alloy to be very temperature

PERIOD ENDING JUNE 10, 1955

Several type 310 stainless-steel-clad columbium
thermal-convection loops were fabricated. A two-
stage process was used in which the columbium
and stainless steel were welded in separate oper-
ations. A typical specimen of a spun Inconel
configuration fabricated to simulate the ART core
shell design was tested for dimensional stability
after thermal cycling. No recordable diametrical,
axial, or thickness instability was observed. A
method was developed for producing cermet-to-
Inconel joints with adequate ductility.

Ceramics composed of rare-earth oxides, which
combine the property of high absorption cross
section for thermal neutrons and the usual ceramic
properties of high density, strength, corrosion re-
sistance, and high melting temperature, were pre-
pared in the shapes required for testing in critical
assemblies to determine their possibilities as
control rod materials. The possibility of coating
uo, particles with ZrO2 to protect the UO2 from
reaction with molten silicon in an SiC-Si fuel
element is being investigated. Also, graphite-
hydrogen reactions are being studied.

7. Heat Transfer and Physical Properties

The friction factor as a function of Reynolds

‘modulus was determined experimentally for the

case of turbulently flowing NaF-ZrF,-UF, (53.5-
40-6.5 mole %) in Inconel tubes; the results are
in agreement with conventional friction data. A
full-scale ART fuel-to-NaK heat exchanger was
studied as a water-to-water heat exchanger. Pre-
liminary measurements indicate that the heat
transfer coefficients on the fuel side of the ex-
chcnger which has the controlling thermal resist-
ances in the system, are about 1.6 times lower than
would be obfomed by the conventional relation for
turbulent flow in circular-pipe systems; also, the
corresponding pressure drops were two times as

"high as those for flow in smooth pipes. Some
~velocity profile data were obtained for the 18-in.

ART '><‘::_o'|;'e for the case where the fluid enters the

core with d_ rotational velocity component; the
influence of turbulence-promoting screens at the

core inlet on the flow was also studied. A study

‘was made of fhe Temperature ‘and ‘corresponding
Tiensile stress flucfucmons in the Inconel walls of
.k'the ART core for momentcry periods of flow
) stagnahon near the wall. The results of a theoreti-

cal and an experimental study of a free-convection
system containing a volume heat source are given.

R T T

 
terials were determined: namely, NaF-ZrF,-UF

 

 

 

 

ANP PROJECT PROGRESS REPORT

The 'én’rhdllpies and heat capacities of five ma-

4
(50-46-4 mole %), NaF- Z:'F‘1 (50-50 mole %), NaF-

7-)7'_‘ZrF -UF4 (56-39-5 mole %), NaF- LIF-ZrF UF
- (20- 55 21-4 mole %), and lithium hydride. The hea’r
_ccpacmes in the liquid state of the 17 fluoride

‘mixtures that have been studied to date are repre-

' séhfea by the simple equation

where N is the average number of ions and M the
average ‘molecular welgh’r The viscosities of seven

~ fluoride mixtures were determined: namely, NaF-

LiF- ZrF UF -UF (20.9-38.4-35.7-4-1 mole %),

_NGF ZrF -UF (50 46-4 mole %), NGF-UF (66.7-
'33 3 mo[e %) NaF KF-UF, (46.5-26.0-27.5 mole%)

NaF- LuF ZrF UF UF (20 55-21-3.6-0.4 mole %),

-NGF LIF ZrF (22 55. 23 mole %), and NaF-LiF-

BeF2 (56 16- 28 mole %). From the viscosity
measurements that are now available for BeF -
bearing fluoride mixtures, a relationship between
BeF. concentration and viscosity was investigated.
In general, the viscosity decreased as the BeF
concentration decreased. Some preliminary thermal-
conductivity data of a ZrF ;-bearing fluoride mix-
ture in the liquid state were obtained with a new
conductivity cell and are in agreement with previ-
ous data obtained with a different type of cell.

8. Radiation Damage

The program of MTR irradiations of Inconel cap-
sules designed for comparing UF,- and UF ,-base
fluoride fuels has continued. The results of exami-
nations made thus far have shown no evidence of
radiation damage in that there is no corrosion, no
significant segregation of uranium, and no changes
in the impurity content of the fluoride mixtures.

" The fluoride-fuel loop that was operated in the

'LITR horlzom‘cl beam-hole has been disassembled,
; and parts of it have been examined metallographi-

cally. Subsurface void attack of the Inconel tubing
used for the fuel loop was limited to less than 1
mil in depth. The fission products Ru'®? and

'Nb?% were found to have plated out in two sections
. of the loop, and thus partial substantiation of a
- Vsumlldr occurrence in the ARE was obtained.

The small “loop intended for operation in a verti-

cal hole in fhe LITR has been charged with fuel

and is now in the final stages of assembly for the
in-pile test, Data which were obtained for three
different sizes of charcoal traps to delay fission
gases from the small loop in the event of rupture
were positively correlated. Two charcoal traps
have been incorporated in the cooling-air off-gas
line, -

A tube-burst stress-corrosion apparatus has been
assembled and awaits filling prior to irradiation
in the LITR, and an LITR-irradiated stress-cor-
rosion rig is ready to be examined in the hot
cells. The creep apparatus installed in the MTR
has just completed six weeks of irradiation and is
ready to be returned to ORNL for postirradiation
measurements.

An equation describing the behavior of the xenon
poisoning in a fluid-fueled reactor was derived
and was applied to the ART design. The calcu-
lations indicate that the removal of xenon by
sparging with helium will be a satisfactory means
of controlling xenon poisoning in the ART. Also,
no difficulties appear to exist in connection with
shutdown poisoning at the sparging rates selected
for ART operation.

9. Analytical Chemistry of Reactor Materials

Apparatus was calibrated for the determination
of vranium metal in fluoride salts by the method
in which the metal is converted to UH, and subse-
quently ignited in oxygen at 400°C to form water
and UQ,. Samples of UF3 and KF-UF3 were then
analyzed. The coefficient of variation was 7% for
the range 6 to 60 mg of uranium. No interference
from the presence of either fluoride salts or tri-
valent uranium was encountered, and therefore the
procedure should be applicable to all types of
fluoride fuels.

A comparison of the methylene-blue and hydro-
gen-evofution methods for the determination of tri-
valent uranium in LiF-Ber, NaF-LiF, and NaF-
BeF, base fuels was made, Satisfactory agreement
of the methods was observed for these materials;
however, the results for trivalent uranium in a
KF base obtained by the methylene-blue method
showed negative bias when compared with those
from the hydrogen-evolution method. Methylene-
blue solutions which were 1.5 to 6 M with respect
to HCl were shown to be reduced to methylene
white at room temperature by finely divided metallic
chromium, iron, nickel, and uranium-nickel alloy.

Studies were continued on the simultaneous

 

 
 

 

o

1)

 

lron, chromlum, erCO i

determination of frivalent uranium and total uranium
in fluoride salts. The postulation of an interaction
species of pentavalent uranium and methylene
white was investigated.

By using an anion-exchange resin in the hy-
droxide form to retain zirconium, beryilium, uro-
nium, and sulfate ions, the quantitative separation
of alkali metal ions was rapidly effected. Determi-
nation of the alkali metal concentration was made
by titration of the free base which results from the
anion resin exchange. When more than one alkali
metal was present, the 2-ethyl-1-hexanol procedure
was applied to the determination of lithium, and
the tetraphenyl boron method was applied to the
determination of potassium. Sodium was determined
by difference methods.

A rapid volumetric method for the determination
of zirconium in fluoride salts was proposed. In
this method, which is a modification of the method
of Fritz and Johnson, a standard solution of
disodium dihydrogen ethylenediaminetetraacetate
is used to complex zirconium, and the excess re-
agent is back-titrated with iron(lH), with disodium-
1,2-dihydroxybenzene-3,5- d:sulfonm‘e as the indi-
cator.

The bromination method for the defermin'afion, of
oxygen as oxide was applied to samples of CrF,
and Na,ZrF .  Further tests were made on the
electrolysis method for this determination. A
modification of the Winkler method for the determi-
nation of oxygen in water was applied to determine
the oxygen in the off-gases from the electrolysis.
The oxygen is absorbed in a solution of Mn(OH),
and Kl, which, upon acidification, liberates a
quantity of iodine equivalent to that of oxygen.
The lodlne is determined spec’rrophotometrlcally

by ex'rracflng it w:fh orthoxylene,

- An mves’rlgcmon was made of the de’rermmahon '

of trace quantities of nickel in fluoride by the use
-~ of the reogent sodium’ dlethyldlth:ocarbomafe. The

" present experiments ‘were not successful because

‘Vr.lf“’_:O'F the interference of uranium and | iron. Absorbonce,_

' specfro of the carbon tetrachloride extracts were
determined for aqueous solutions containing sodium

diethyldithiocarbamate cmd such cations as nlckel
m, g: d urcxmum

   

   

 

10 | Rficovery and Reprocessmg‘of Reucfor Fuel“ -

     

The feasniblllfy of ‘repeated use ¢ of a \nxlcke| S

action vessel in the fluoride volatility—fused salt
process for recovery of ARE-type fuel was demon-

 
 
 
 

PERIOD ENDING JUNE 10, 1955

strated in 20 laboratory-scale runs. Nickel test
coupons held in the nickel reaction vessel during
the 20 fluorination runs showed corrosion of the
solution type that was even over all surfaces,
including welds, in contact with the molten salt,
Severe local pitting was noted that varied in depth
up to 19 mils on the fluorine gas inlet tube in the
vapor zone above the molten salt. The attack on
this tube in the liquid zone was more uniform, and
varied from 4 to 7.5 mils in depth. The reaction
vessel showed nonuniform attack of the solution
type that varied from 5 to 9 mils in both the liquid
and gas zones.

At either 200 or 650°C, CaF, was much less
efficient than NaF at 650°C in removing volatilized
ruthenium from the 'UF6-F2 gas. However, results
of runs made under various conditions indicated
that the temperature, size, and conditioning of the
NaF bed are very important,

The engineering flowsheet for the ARE fuel
recovery pilot plant is 85% complete. Design of
10 of the anticipated 29 process equipment pieces
is complete, and the pieces have been ordered.

PART IH. ‘SHIELDING RESEARCH
11. Shielding Analysis

A semianalytical Monte Carlo calculation has
been initiated to determine the history of all gamma
radiation born within a circulating-fuel reactor.
All shells of the core, reflector, and shield will
be taken into account, but the calculation will be
simplified by the assumptions that the reactor has
spherically symmetric geometry and that all regions
are homogeneous. The results will include de-
terminations of the energy absorbed and of the
energy spectrum and angular distribution of the

. gamma rays penetrating the shield.

The coding of a Monte Carlo calculation of the
energy absorption resulting from gamma radiation

... incident on a multiregion shield with slab geometry
_is nearing completion. The original code has been
-.revised, and the calculation should now provide

good statistics for gamma rays incident on a
shield with a thickness of approximately seven
mean free paths,

~ The codlng of a Monte Carlo calculation of the
energy and angular distribution of air-scattered

. .neutrons from a monoenergetic source is also
nearing complehon Modifications in the original

problem moke it possible to perform calculations
for surface sources having angular variation in

 
B g SR

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

strength propomoncl to various powers of the
cosine of the angle to the normal and then by a
sun’rable combmcmon to duplicate the distribution

~ from the surface of a sphere such as a circulating-
~ fuel reactor.

The analysis of the constant-velocity transport
‘equation has been extended with the aid of eigen-
functlons of fhe various media.

12 le Tonk Shleldlng Facility

The smtlc source tests in the second series of

experlments “with mockups of a cireulating-fuel
" reflector-moderated reactor and shield are in

progress. The mockups consisted of the following
regions: Inconel core shell, beryllium reflector,
first boron curtain, heat exchanger, second boron
curtain, pressure shell, and a lead-water shield.
Gamma-ray dose rate measurements beyond mock-
ups with variations of these regions have shown
that (1) / -in. thickness of boral for the first

- eurtain decreases the dose as much as a 1% -in.-

thick boral first curtain would; (2) the dose is
decreased with an increase in heat exchanger
thickness; (3) there is appreciable gamma-ray
production in and beyond large uniform thicknesses
of lead in the mockup; (4) only an attenuation
effect is observed when pressure shell material
(that is, nickel) is added; and (5) the dose is
increased 10% by the addition of 0.125-in.-thick
Inconel cladding on the first boron curtain. A
measurement of the distribution of the thermal-
neutron flux within the 12-in.-thick beryllium re-
flector showed that the flux peaked at about
41/4 in. of beryllium. Neutron measurements as a
function of the lead region thickness revealed that

~an increase in lead thickness effectively only

moved the thermal and epithermal flux outward
from the source. A study of the activation within
the heat 'exc}-'ucnger was also carried out, and, on
the basns of this study, a calculation for a 300-Mw
cnrplcne reactor indicated that a 2000-curie acti-
vation is fo be expected in the NaK. For an un-
shielded NaK-to-air radiator this would give 16
rem/hr at 60 fl'.

13. Bulk Shleldlng Facility

A mockup of the swept-bock air-duct system for
’rhe GE ANP reactor was tested at the Bulk Shield-
ing Fccuhfy (BSF). The mockup consisted of a

“pair of annular ducts that-represented a segment
of the duct system around the airplane reactor. A

10

long inlet duct and a shorter outlet duct were
placed against, and on opposite sides of, the BSF
reactor, which was modified with GE-type transition
sections. Both ducts curved to the right, and the
shorter one nested within the longer one so that
they ran roughly paorallel. Measurements of the
radiation ledking out the end of the large duct as
the distance between the ends of the ducts was
varied (that is, as the large duct was moved out)
showed no evidence of streaming down the entire
length of the large duct. Fast-neutron dose rate
measurements beyond the shield and parallel to
the last leg of the large duct were not affected by
the position of the duct. Angular distribution
measurements around a fixed point beyond the
ducts showed no evidence of fast neutrons escaping
from any portion of the large duct.

14. Tower Shielding Facility

Differential-type shielding experiments have
continued at the Tower Shielding Facility (TSF)
with emphasis on measurements of the fast-neutron
dose rate distribution within the detector tank and
the GE-ANP crew compartment held at a separation
distance of 64 ft from the reactor tank. This
distribution was determined as a function of vari-
ations in the reactor shield thickness, crew com-
partment thickness, and angle of radiation emission
from the reactor tank. The measurements, which
were made at a height of 195 ft, are presented in
this report.

A method is presented for interpretation of the
TSF differential shielding experiments in terms of
the probability of fast neutrons scattering into the
sides of a crew compartment of an aircraft divided
shield. The effect of beam collimation on the
experimental results, as indicated by single air-
scattering calculations, is discussed. An indi-
cation is obtained from the experiments of the
effect of the neutron energy spectrum on the dose-
scattering probability.

The results and interpretation of the TSF dif-
ferential shielding experiments were used to
calculate the fast-neutron dose rate for a divided-
shield mockup which had been measured previously
at the TSF. The agreement between mockup
experimental results and the calculations based
on the differential experiments was good, in view
of the various uncertainties existing at this time.
In moking the comparison, procedures were de-
veloped for using the experimental results in dose

fay

3.4

i)

 

 
 

v

 

2y

 

predictions for aircraft divided shields. Some
experimental information is presented on the re-
lation between dose within a cavity and the source

 

 

 

PERIOD ENDING JUNE 10, 1955

strength at the surface of the cavity, where the
source strength is taken as the dose rate at the
position on the surface with no cavity present.

11

 
 

 

 

 

 

 
 

 

 

 

P

Part |

REACTOR THEORY, COMPONENT DEVELOPMENT,

AND CONSTRUCTION

 
  
  
  

 

 

 
 
 

[ M]

 

¢y

 

"'f_;'con stitute

1. REFLECTOR-MODERATED REACTOR

E. S. Bettis
W. K. Ergen

A. P. Fraas
A. M. Perry

Aircraft Reactor Engineering Division

AIRCRAFT REACTOR TEST DESIGN
A, P, Fraas

Aircraft Reactor Engineering Division

The preliminary layout of the interior of the
reac'r'or cell for the Aircraft Reactor Test (ART)
which shows the major items of equipment and the
recommended provisions for support was completed.
A similar layout for the NaK piping and radiators
was also completed. [t has been established that
five bulkheads will be required in the tank and
cell walls for instrument and control wiring, piping,
and miscellaneous service lines. Consideration
has been given to the design of the addition to
Building 7503 required for the ART, the layout of
the blower house, and the electrical power system,
distribution, and auxiliary equipment. Mounting
arrangements have been described for the radiators,
for the fill and drain tanks, for the inner and outer
reactor cells, and for structurally attaching the
reactor assembly to the cell, dump lines, NaK
piping, and other service and instrumentation lines.
The type of equipment by which oxygen will be
removed from the inner cell has been specified.

The basic requirements of the heat dump system
and the instrumentation and control thereof have
been specified. The ART heat dump system is to
provide heat dissipation capacity of 60 Mw of heat

with o mean temperature level of ]300°F in the

NaK system The most convenient, inexpensive,
and compcct heat dump has been found to be a
round-tube, p!ate-fm radiafor core. This basic

type of heat transfer surface has been found fo be’
'_sufficnently rehable in hect exchanger test rigs.
- Five separate systems will be used; four will
»;mcnn heat dump system, w['ule fhe“
Cfifth will be the modera'ror heat dump system.
Layouts have been prepared of the air duct and

   

stack associated with the heat dump radiators,

and air flow ra'res qnd pressure drop hcve been'
o _"-:cc:lculoted o e o
« .~ Design ; cdhs'ldefd}“io'hs hove been tem‘utlvely* )
-‘-;estcblrshed for the prevention ‘of failure of the

“Inconel core and reflector shells durmg operahon":

  
 

of the ART. The core shells are to be /8 in.-thick

Inconel. They are to be cooled on one side with

  

flowing sodium in order to maintain the shells
everywhere below 1500°F, The reflector shell
is to be maintained below 1300°F by surrounding
it with a layer of boron-bearing material, a gas
space to serve as a heat dam, and an Inconel
cladding layer to separate it from the hot fuel in
the heat exchanger. To prevent buckling of the
outer core shell because of external radial pres-
sure differences and to prevent excessive de-
formation of the reflector shell because of ex-
cessive pressure differences, the pressures in the
sodium coolant system and in the fuel are to be
adjusted to be approximately equal at a point
halfway through the primary heat exchanger. In
order to limit cyclic thermal stresses, the tempera-
ture difference across the shells is not to exceed
300°F. Stress and heat transfer analyses are far
from complete, but tentative pressures have been
established for the fluid circuits.

In case of a failure of apart of the reactor
system, for instance a fuel pump, during operation,
it may be necessary to maintain emergency con-
ditions for short periods in which the pressure
differences across the shells will be much greater
than those existing during normal coeration. Calcu-
lations indicate that these conditions can be
maintained without sudden failure of the shelis
by buckling or rupture. However, the total life of
the reactor for normal operation will be dispro-
portionately shortened.

Details of the basic design of the reflector-
moderator cooling system have been set forth.
Final design awaits confirmation of the values

- _used in the design by critical experiments, heat
. exchanger optimization, core hydrodynamic tests,

and beryllium thermal stress and corrosion tests

now in progress (cf. Secs. 2, 3, and 7).

The xenon-removal system has been designed,
as described previously,! and the fuel and sodium
pumps are being tested (cf. Sec. 2, “*Experimental

- ~Reactor Engineering’’). New features that have

been incorporated because of stress, fluid flow,

... ot fabricability considerations include an elliptical

 

1G. Samuels and W, Lowen, ANP Quar, Prog. Rep.
Dec. 10, 1954, ORNL-1816, p 21.

15

 
 

 

 

ANP PROJECT PROGRESS REPORT

fuel expansion tank, a rounded dome to enclose
the top of the reactor, a newly designed sodium
pump impeller, and other related items.

A number of fuel-to-NaK heat exchanger designs

have been calculated by utilizing the hecn‘ transfer
data obtained experimentally (cf. Sec. 2, ““Experi-
‘mental Reactor Engineering,”’ and Sec. 7, *“‘Heat
Transfer and_Physmal Properties’’). The most
 promising of these designs are described in
Table 1.1. The values given in the table are
intended 'fdrfd'éilifdte comparison of typical heat
'e:xlo‘:hahgers. The values listed for heat exchanger
fhlckness and shield weight are only relative and
were based on a mmtmum-fhlckness tube array
svited to the present design layouts. Final
selechon of fhe heat "exchanger must be con-
tingent on a composnte evcluchon of stress limi-
tations, geomefry, mass transfer and corrosion
- effects, and heat transfer requirements.

Conservchve colculc’rlons were made of the

neufron ‘and gcmma ray dose rates at various
locations in the ART facility. |t was determined
that radiation levels in the building during power
operation would everywhere be below laboratory
tolerance. Further, with the exception of a few
locations, the initial dose rate from a reactor
catastrophe would not be more than a few roentgens
per hour. These low dose rates are the result of
the reactor being located 16 ft underground and
shielded (in addition to the lead-water reactor

shield) by 3 ft of water in the cell annulus and
substantial thicknesses of concrete and dirt. The
radiators are to be shielded so that, in the event
that 10% of the fuel entered the radiators, the
radiation dose would not exceed 1 r/hr at any
point outside the main air duct or reactor cell.

Research is under way that will provide much
of the information needed as a basis for final
design. In the experiments that are now being
made with room-temperature critical assemblies
of the reflector-moderated reactor, materials of
interest in the design of the ART, such as beryl-
lium, Inconel, and rare-earth oxides, are being
evaluated. The information being obtained will
be of benefit in making final decisions on reflector-
moderator and island dimensions, configuration,
and cooling requirements. Data on power distri-
butions in the core will be used in determining the
final configuration and size of the core and the
end ducts, as well as the thickness of the Inconel
shells and the cooling required. Measurements
being made on rare-earth oxides are expected to
provide data needed for design of the control
rod for the ART.

Intensive experimentation is in progress in at-
tempts to find fuels superior to the ZrF ,-bearing
fuel used in the ARE. As yet, no positively su-
perior fuel mixtures have been obtained, but the
physical properties of the ZrF,-bearing mixture

have been improved to some extent by varying the

TABLE 1.1. ART HEAT EXCHANGER DESIGNS

Tube diameter, in. 3/16
Number of tubes per bundle 143
Tube spacing, in. 0.030

Tube wall thickness, in. : 0.025
Fuel temperature range, °F

NaK temperature range, °F

Tube length, ft 6.0
Fuel AP, psi 41
NaK AP, psi 39.6
Heat éxchcngér rhickhess, in. o 2.62

' Reoctor shleld Welg['\‘l' water plus lead, Ib 72,800

"V:-;,‘-Tota| number of we!ds 6864
’ P?wer, Mw ' 55
Relat.iv.e Nq-i-( ocfi.vofion ” 1

16

1

1250 to 1600
1070 to 1500

%6 ! '

143 130 130

0.030 0.020 ' 0.020

0.035 0.025 0.035 |
1250 10 1600 1250 to 1600 1250 to 1600 °

1070 to 1500

1070 to 1500

1070 10 1500

6.5 5.45 5.81
45 28.4 30.1
95 5.5 9.6
2.62 3.78 3.78
72,800 74,100 74,100
6864 6240 6240
55 55 55

1 2 2

 

 

 

 

 
 

v

 

 

composition, and the corrosiveness has been
lowered by the addition of even small amounts of
trivalent uranium.

The operation of forced-circulation, high-tempera-
ture-differential, Inconel loops with fuel mixtures
of interest has demonstrated that the velocity of
the circulated fluid has little effect on corrosion
and mass transfer and that the results of the
numerous thermal-convection loop tests are appli-
cable to the dynamic ART system.

One in-pile forced-circulation Inconel loop that
circulated a fuel mixture in the LITR has been
examined, and radiation was found not to have
a detrimental effect on corrosion and mass transfer;
a second loop will be operated in the LITR during
the next quarter. An in-pile loop that will more
closely simulate the ART conditions is to be
operated soon in the MTR.

Experiments with sodium in Inconel loops have
thus far indicated wunsatisfactorily high mass
transfer, but it is thought that better purification
and handling of the sodium will greatly decrease
the mass transfer; alsc, the use of stainless steel
in the cold portions of the system is being in-
vestigated.

Cermets for use as valves, bearings, and seals
are undergoing intensive self-bonding and cor-
rosion resistance tests, and procedures for brazing
cermets to Inconel have been developed.

Studies of expenmentcl nickel-molybdenum base
alloys and commercial Hastelloy B, in particular,
have indicated possibilities of finding a structural
material with properties superior to those of
Inconel for c:'ircula‘ring fuel reactor dpplicafion,

PERIOD ENDING JUNE 10, 1955

Shielding Facility will provide much of the data
needed for final design of the reactor and, in
particular, the shielding. A mockup of the reactor
and shield is being used to determine activations
within the heat exchanger and the effect on dose
rates at various locations of varying materials and
thicknesses of materials. The activation of the
NaK within the heat exchanger is of particular
importance because of the effect it will have on
dose rates at the NaK-to-air radiators, which are
outside the reactor shield,

Design of the engineering flow sheet of a fuel
recovery pilot plant is 85% complete. It is hoped
that the rapid reprocessing and recovery of fuel
that would be reguired under service conditions
can be demonstrated with this pilot plant.

The many calibration and performance tests
scheduled for the ART are summarizedin Table 1.2,
Some of the tests can be run in the course of the
endurance tests, which should consist first of
25 simulated flight cycles (16 hr at full power and
8 hr at from 1 to 10% power) and then of 100 hr
of continuous operation at full power. This will
give a total of 500 hr at full power and 200 hr at
low power during the endurance test period. It is
expected that the preliminary calibration and
performance tests can be carried out during the
first 400 hr of operation, with allowances for
servicing and maintenance. In reviewing the
design requirements and the tests planned for the
ART, it is evident that many modifications could
be made in the test conditions to vary the severity
of the test. The key variables, together with the
desired values for the test, are given in Table 1.3.

but some of the probfems involved are of such ,

a mqgmtude “that it is ot thoughf 'rhaf such ma-

terials can be made avmiable in time for use in
the ART. The weldlng and brozmg matermls and
 techniques needed for fabrication of the ART have
"«'_rffbeen developed qnd thoroughly‘ 'rested |n the con-“";&"w

 
 
  

A study of The. behowo_ : nonp on ng n

a fluid- fueled reactor has |nd|ca’red ’rhat fhe?:_:
removul of _xenon by sporglng “with helium will
7 sfactory means for controlling xenon
po s‘en g in 'rhe_"ART The sfudy alsdhmdncated\it
" that no dlfflcul’ry wou[d exist in ‘connection with
~*“shutdown poisoning af the sparging rates selected

 

for ART operation.
Experiments now under way at the Lid Tank

 

ART CONTROL
J. M. Eastman
Bendix Products Division

El Ro Mdhn

Ins’rrumentatlon and Confrols DIVISIon

rJflm‘orma'uonfi"l:)lock dmg"“ms showmg the basic

tfi-c'ontro! ac, |<'>hs'(de5tred (w:thouf reference to hard-

ware or fechnlques) have’ been prepared for the
ART They will be used as a basis for selection

of ’rhe hardware types ‘and the  control techniques

to be used cnd for determmahon of areas of control
’rhuf w:ll reqmre new COmponent developmenf.”
Information for reactor simulation has been supplied
to Pratt & Whitney, and they have built a simulator,
It will be used to check the adequacy of the basic

17

 
i

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 1.2. NUCLEAR EXPERIMENTS ON THE ART

Zero-Power Experiments

Criticdli?y '
Demonstrcte procedure for going critical

o Determme crmcal mass

Reochvn‘y Experlmem‘s

Determine fuel temperature coefficient

- Determine reflector and island temperature coefficient

Defermme mass reachvnty coefficient

Measure reachvn‘y as a function of flow rate

' Defermme delqyed neutron loss to xenon purge system

Shle]dmg
Survey surface of shield for radiation

.. Determine NcK achvahon in heof exchanger

High-Powe: Experiments

Demonstrate That the Réa_qctor Is a Slave to the Load Demand

- Simulate s'ubdrdén.d;ér'r‘achld for increased power by turbojets

" Simulate failure of one turbojet
Determine Reactivity Effects of Other Transients

" Results of sudden flow stoppage

Results of NaK flow stoppage

Determine Effectiveness of Xenon Removal System

Determine Compensation Required for Fission-Product Poisoning and Burnup

Demonstrate Afterheat Removal upon Shutdown

Obtain Heat Balances at Various Powers to Determine Extracted Powervs Nuclear Power for the Core, Heat Exchanger,

Island, Reflector, Pressure Shell, Gamma Shield, and Neutron Shield

control actions and to determine control component
design data, Additional simulator work will also

be done at ORNL.

Control Principles

The control system is to provide automatic cor-
rective action for emergencies requiring action too
rapid to permit operator deliberation. Automatic
interlocks will prevent inadvertent dangerous oper-
ation, with minimum operator limitation. Operation

in the design power range will be independent of

nuclear mstrumenfc’rlon and will have only limited
dependence (for safety) on other instrumentation
during' power transients. Three classes of emer-
gencies will be provided for:

1. those requiring automatic rod insertion, load

" removal, and fuel dump,

' ,'2. those requmng only automatic rod insertion and

18

load removal (followed by manual dump at oper-
ator’s discretion), '
3. those requiring only operator warning and manual
action.
One-half the wind tunnel blowers and one-half

the pumps (of each type) w:ll be pOWered by com-

mercial supply. The remaining blowers ‘and pumps
will be powered by a deisel system. Duplicate
fuel and sodium pumps have been provided for
safety, but if any one of these four pumps failed,
the output of each of the remaining fhree pumps
would have to be decreased to prevent pump cavi-

tation or pressure overstressing of the Inconel core

shells in the current design, The basic control
actions provide for this depression of output, but
it is questionable whether pump speeds can be de-
creased fast enough to prevent adverse transient
conditions. For example, the Inconel may be over-

 

 

 

 
 

 

)

TABLE 1.3. KEY ART PERFORMANCE VARIABLES

 

 

Variable Desired Performancg

Time at full power | 500 hr
Total time critical 1100 hr
Total time }hgrmany hot 1500 hr
Power level 60 Mw

Peak fuel temperature 1600° F
Peok metal temperature 1540°F
Peak NaK temperature 1500°F
Temperature difference 100°F

between fuel and NaK
Number of power cycles 30

Rate of change 7°F/sec

of mean fuel temperature

Number of dump cycles 5

 

stressed momentarily, with a consequent increase
in creep rate. To prove at least that the reactor
can be safely scrammed after such a pump failure,
it is planned to include cutting off the commercial
power (or diesel) at design point (60 Mw) as a part
of the test procedure. This will constitute a class
] emergency, as defined above,

The control rod cannot be abruptly inserted in
the case of an emergency, because to do so and
yet prevent the fuel from freezing or getting too hot
would require accurate transient matching of ab-
sorbed power and flux power, which would be im-

practical to accomplish. Therefore a scram (class

1 or class 2 emergency) will abruptly insert the rod

“only enough to reduce Ak/k by 1%. This will be
followed by automatic insertion at the rate of 1%
Ak/k in 30 sec. This Speed ws“ be the ‘‘fast’’
'normai rcn‘e of rod movement. The opemtor mcy”
_-a!so choose ‘ slow” rod movemenf of 1% A]e/k in
5 min, Below 10-Mw power operahon, the opercn‘or
will be automchca”y limited to the slower rate.
The 1% scram will automatically cut off one-half
‘the blowers to reduce the load. The load will then

be further reduced by automatic closing of the

_radiator, shu’r’rer in response fo a 1050°F low limit

signal for the NQK -to-air radiator outlet temperature.

Since the de5|gn point occurs at the hlgh temper-
ature limit (1600°F) for the fuel, accurate limiting
will be required, |n addition to thermocouples in
the fuel, the maximum fuel temperature will be

PERIOD ENDING JUNE 10, 1953

indicated by continuous computing from NaK
thermocouple signals by using heat exchanger
calibration data, The dependability of fuel-temper-
ature-sensing thermocouples is questionable be-
cause of the poor conductivity of the fuel and the
high gamma-ray heating. '

Operating Procedure

The reactor is to be initially filled with barren
(without uranium) fuel carrier to check out the
system. This will be done with the system iso-
thermal at 1200°F and all NaK and sodium pumps
operating., Electric heaters will supply the heat
needed beyond that produced by pump work dis-
sipating, After the checkout of the system, part
of the barren carrier will be removed and replaced
with fuel-enriched material to provide 80% of the
calculated critical amount of U235, Criticality
will be checked, and the fuel will be further en-
riched in steps until criticality is obtained. The
control rod will be calibrated in the process, A
flux servo will then be available for holding the
reactor at very low power levels (under 10 kw)
during low-power experimentation. At these levels
the temperature coefficient will be inadequate for
good control because of the high thermal capacity
of the system.

The minimum power level for control by the
temperature coefficient is expected to be about
300 kw ~ the estimated heat removal by the radia-
tors with the heat barrier doors open and no air
flow. To take the reactor to the 300-kw level, it
will manually be put on not less than a 20-sec
positive period, and the barrier doors will be
opened when the flux level reaches a set value
equivalent to something less than 300 kw. |[f the
doors were opened too soon or too late, excessive

power surges might occur before the flux stabilized
‘and there might possibly be damaging thermal

shocks, After stabilization at 300 kw, the control

rod will be slowly adjusted to bring the mean fuel
temperature to 1200°F (it will have gone somewhat
" above this).

To increase the power above 300 kw, the blowers
will be turned on and the shutters opened as de-

sired. The control rod will be used to adjust the
"mean fuel temperature. Automatic limiters will

close the shutters if the minimum NaK temperature
drops below about T050°F (to prevent fuel freezing)
and will open them if the minimum NaK temperature
goes above 1300°F (to dissipate afterheat and pre-

19

 

 
~ and femperature surges.

 

 

 

ANP PROJECT PROGRESS REPORT

vent loss of ZrF,). To take out more than 20 Mw
will require that the mean fuel temperature be set
above 1200°F (it must be set at 1425°F to take out
60 Mw). Below about 6 to 10 Mw, the heat dump
will have to be adjusted slowly to avoid high flux
' Above this power the
temperature coefficient is expected to permit
reasonably fast (for jet engines) load changes
without excessive transient surges.

VREACTO R PHYSICS

Probable Effect of Répldcihg Inconel
by Columblum in fhe ART Core Shells

Ww. K Ergen

Aircraft Reactor Engineering Division

The cold critical experiments for the ART have
shown? that the critical mass decreases very
rapidly as the neutron absorption in the core shell
decreases. Hence, replacement of the highly ab-
sorbing Inconel by a low-cross-section material,
such as c:olumblum,3 would bring about a sub-
stantial saving in vuranium investment and, inci-
dentally, make the temperature coefficient more
negative. For corrosion and fabrication reasons,
only about 50% of the Inconel can be replaced by
columbium. The effect of this 50% replacement has
been calculated on the basis of recent data from
the cold critical experiments and measurements of
resonance capture integrals.?

From the critical experiments it is concluded
that the epicadmium flux is about the same for the
core shell and for the fuel. From the capture
‘resonance integrals of Inconel and columbium,
respectively, and the fission resonance integral of
U235, it is computed, for the epicadmium range,
that the columbium absorptions amount to 1.6% of
the fissions, whereas the absorptions in the re-
placed inconel amounted to 2.4%.

For the ‘‘below-cadmium’’ neutrons the cold
~ critical experiments are consistent with the model
that the shells absorb neutrons according to the
thermal cross sections and that all neutrons pass
through the shells perpendicularly. By using the
'l/u extrapolation to get the 1400°F obsorpfion

 

A D. Callihan et al., ANP Quar. Prog. Rep. Mar. 10,
1955, ORNL-1864, p 43.

3C B. Mills and H. Reece, Jr., Design Study of an
ANP Circulating Fuel Reactor Nov. 30, 1954, WAD-
1930, p 44.

—4R. L. Macklin and H. S. Pomerance, Resonance

Capture Integrals (to be published).

20

cross sections, the columbium absorbs 0.5% of the
neutrons, whereas the replaced |nconel absorbs
3%.

The experimental result that the epicadmium
neutron group and the subcadmium neutron group
each contribute about 50% of the fissions then
indicates that effectively 1,65% of the neutrons
would be saved by the replacement, If

(AE/R)/(AM/M) = 1/10

is assumed, the saving in critical mass would be
about 16.5%, or 4 kg, corresponding to 0.7 mole %

in uranium concentration of the fuel and 7 kg in

total U235 investment.

As a check, the above method can be used to
compute the effect of completely eliminating, in the
cold critical assembly, I/] in. of the core shell.
The result would be a 41% saving in critical mass,
compared with the experimental value of 45%., In
another room-temperature critical experiment (cf,
Sec. 3, ‘‘Critical Experiments’’), 127.4 g of the
Inconel core shell was replaced by 98.5 g of
columbium. The effect calculated by the above
method would have been 2.5 cents, and the experi-
mental result was 3.6 cents,

ART Temperature Coefficient
L. T. Anderson

Aircraft Reactor Engineering Division

A calculation of the ART temperature coefficient
showed a positive contribution resulting from the
decrease of the absorphon cross section of the
Inconel core shells3 with increasing neutron energy.
The smaller Inconel cross section allows more
neutrons to enter the fuel annulus and thus increase
the reactivity, It seems justified toassumeal/v
cross section for the Inconel on the basis of the
experimentally measured value of the resonance
integral, and therefore, if an increase, AT, of the
beryllium temperature causes a neutron energy
increase proportional to AT, the fractional cross-
section decrease of the shells is l/2 AT/T, and
the number of fissions is increased by the factor

1/ (AT/T)Z,, x shell fh:ckness X cadmlum frachon

if only the subcadmium neutrons are cn‘fected and
the neutrons are incident normal to the shell,
Actually, the thickness of the Inconel that the
the neutrons go through is somewhat greater than
that of the shell, on the average, but this is at
least partially offset by the cross section of

 

 
 

v

~cases. 
B '_'\.fculchon

Inconel having a less-than<thermal value for some
of the subcadmium neutrons. The value of Ak/k
calculated in this way is +1.7 x 103 AT (°F) for
]/-ln. Inconel shells. As pointed out by Mills and
Reece and Ergen (above), core shells of materials
of lower absorption cross section would reduce
this posmve coefficient contribution, If ]/16 in,
of the /8 in, Inconel shell were replaced by colum-
bium, the Ak/k value would be reduced to

 

17 « 10-5 {(Inc. + Cb)
’ f(Inc.)
0.035
= 1.7 x 10— = 1.0 x 10=3/°F ,
0.060

where [ is the fraction of thermal neutrons incident
on the shells that are absorbed (cf. preceding
paper by W, K, Ergen, this report). '

The effect of beryllium expansion is considered
by wusing, as a model for the reactor, a sphere
(rd_di'us', R,) of noncapturing beryllium with a high
energy “neutron shell source at R.. The source
neutrons slow down according to age theory, and
the slowing down density at thermal age is used
as a source for the thermal diffusion equation. It
is assumed that at thermal energy every neutron

‘striking the shell at R is absorbed and causes a

flssmn, By using this model Ak/F turns out to be>
—2KR, [T AT

7 R, (R, = Ry) _RZJ;

where K is the coefflaent of Ilneqr ‘thermal ex-
pansmn for fhe bery”:um and 7 is the fhermai age.

 

By using K = 93x ]0 6/°F 7 =90 cm?, R, R, =25
~cm, ‘and R2 56 cm, |'r is found thai Ak/k = —] 2 x
' ']O SAT .
i The net of these ’rwo contrlbuhons is then 0. 5 ><

10-5/°F for. 1/8-m. Inconel core “shelis and Zo2x T

10‘5 for core_shells of /16""' [nconel and ]/]6-|n.
- :;_f‘f'coltmelum. |
f._'f,_:r_.;whlch is based ona conset "ctlv‘eya?ye “of 0.1 for'l o

 

 

'efflment of -1. 7 X ]0" /°F,

 
 
  
 
   

 

 

 

S, VAnderson, Calculation of the Beryllium Contri-
bution to the ART Temperature Coefficient of Reactivity,
ORNL CF-55-5-76 (May 11, 1955).

6W. D. Manly, private communication,

PERIOD ENDING JUNE 10, 1953

less, the results support the postulated negative
over-all temperature coefficient of the ART,

Reactivity Effect of a Heat Exchanger Leak
A. M. Perry

Electronuclear Research Division

If the pressure in the NaK coolant circuit ex-
ceeds that in the fuel circuit, a leak in the inter-
mediate heat exchanger may first become apparent
through its effect on the reactivity of the ART.
The effect will be due to dilution of the fuel and
to increased parasitic absorption. It has been
suggested by the Advisory Committee on Reactor
Safeguards that the reactivity effect can be greatly
enhanced by adding lithium or Lié to the NaK,
Preliminary experiments® indicate that perhaps
5 wt % lithium can be added to the NaK eutectic
without raising the melting point above 25°C,
The loss of reactivity due to parasitic absorption
is calculated on the assumption that

e M= iy T

 

where / equals (neutrons absorbed in fuel)/(neu-
trons absorbed in core), AX, is the change in
poison cross section, and 2. is the fuel cross
section.

The loss of reachvn‘y due to fuel dilution is
calculated on the assumption that

— TET cmt—— — N —— i

where V is the initial fuel volume, and AV the
volume of NaK leaked into the fuel circuit, The
factor 1/]0 is a conservativé estimate based on

~eritical experiment results.

. The over-all reactivity loss, if AV is in cubic
feet of coolant leak into the fuel circuit, is given

by

Ak
T = ~2.4 AV(%) ,

| nc ‘rhe 'crorelic}lnt is Nak; by

Ak
S —(2.4 + 2.2 W) AV(%) .,
the coolan;]s NCIK plus 'W wt % h'rhlum, and by
Ak
”7;“- = (2.4 + 29.6 W) AV(%) ,

2]

 
o chonges

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

lf 'rhe coolam‘ is NaK plus W wt % Lié (Of course,
the opproxumahons made are valid only for small
in reactivity and for small percentage
“additions of lithium to the NaK.)

The maximum volume available before an over.
flow would occur will be 0.37 #3, This volume of
 NaK in the fuel circuit would produce a reactivity
change of 0.9%. The same volume of NaK plus

0 5 wt % lithium Would produce a reactivity change

of 5%. A leak of 0,007 ft* of NaK plus 5 wt %

' Ll would produce a reactivity loss of 1%,
| Burnup and Gomma-Ray Heating of Control Rod

- W. Fader
Proh‘ & Whitney Alrcroff

Burnup of Poison.
~ a slab of neutron absorber with thermal neutrons
incident on one face are, for the case of an ab-
sorber with n isotopes with high thermal capture
~ cross sections o,

Z
E + 2(;5 = 0 ’
d2;
'—gg— + aiziqb =0,
where
] = net neutron current normal to slab face,

linear coordinate normal to the slab,

H

%, + 2, + ... + £ =macroscopic thermal
neutron absorption cross section,

flux, '

exposure time,

I

i

x
2
é
t

By making use of the relationship | = ¢, where
I is the average value of the cosine for the angular
distribution of the neutron flux, the equations be-
come

 

(1 .Uax Z],— 0,
azz.
(2) I +o2] =0 .

ot

~ In the region where most of the burnup takes place,
11 is constant if the macroscopic absorption cross
- section is much greater than the macroscopic scat-
| fermg cross section, lts value may be expected
to range between unity at points deep in the slab
~to 0.5 at the surface, the latter value being the
diffusion theory approximation to the boundary

The equations of burnup in

condition at the plane interface of a scattering
medium and a strong absorber,

Solutions for J(x,) and X(x,t) were found for the
case of a single absorber and the results have
been plotted in Fig. 1.1 for a slab of Sm,0,. For
large values of o] /i, where J, = ](0 t), the
curves of J(x,t) ond 2(x,t) assume rigid shapes
that are propagated in the direction of increasing
x with apparent velocity

0’]0
2 I

 

0

which, for the case of the curves in Fig. 1. ]. was
v = 2,9 x 108 em/sec. This suggests the use of
the linear formula

(3)  X() T 0.012 + (29 x 1022 [ 1) em

for the thickness of a slab of Sm,0, that will
transmit 10% of J neutrons/cm? osec mcudenf on
its front surface after an exposure time of t sec,

Equations 1 and 2 were integrated numerically
for the case of gadolinium with two isotopes with
high thermal absorption cross sections, and the
results are shown in Fig, 1.2 for Gd ;03 The
linear relohon between X and ¢ for large cr] ot/

for Gd O

(4) X() = 0.0025 + (1.4 x 10=22] ) cm .

Formulas 3 and 4 may also be used as an ap-
proximation of the burnup of a cylindrical shell of
absorber, provided that the inner radius of the
shell is much larger than the thermal neutron mean
free path of the absorber; for, in this case, Eq. 2
remains unchanged, while Eq. 1 becomes
d
(3) i 'éi = 3] - IT”‘!‘

r r
for neutrons incident on the outer surface. For a
shell of this kind, the ferm T]/7 in Eq. 5 is always
negligible compared with %] in the region of burn-
up. In the ART, a 20-in. control rod with a ]?’/B-in.
outside diameter must absorb an average of at
least 2 x 1014 neutrons/cm?:sec in order to be 5%
effective in reactivity. Thus, if a cylindrical shell
of Sm,0, absorbs 2 x 1014 neutrons/cm? sec for
1000 hr, the neutrons will penetrate a distance

X(t) = 0,012 + (2.9 x 10=8) (7.2 x 106) = 0.22 cm.

A shell of szo3 must have at least this thickness

 

 
 

e it e e B . it

i s

.

 

PERIOD ENDING JUNE 10, 1955

-CONPIERITAT
ORNL-LR-DWG 7551

 

 

N ™ T~

 

%
JF—
14eBOx 0y Vg _madyifp ]

 

N ~
\ #=0 \,«f_=4,5 % 10° sec \v/f 9 x 10 sec Jo = 10'% nsem? - sec
\ A=t \ a=Yy \ E=l 020 ¢

 

o=58x/

 

0.6

Sy
//
-~
//
Z
7

2(0)=200 cm™f FOR Sm,05;  —

 

 

0.4 \ \\ \\ \\?<fi=_?uossec

 

a1
. zO />\\ \ \

 

 

 

N\ L

/
L N
AT I S N N \

0 S~ T ~_

~ //7/
0.8 // ,// /
/Lf 4.5 x10% sec / f=94“05 sec

/ H= i/2 // /‘—"z
/ Z{O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.6 - / 5
/ / - i e(chof “TOIx}/E ez«»x/fi
s /
R 04 A
“ /
/Q\f 9 x 10° sec
/ / p=lp
0.2 / pd
/ /
0 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 0.055

%, DISTANCE FROM FACE OF SLAB (cm)

Fig. 1.1. Burnup of a Samarium Oxide Slab.

to remain black to neutrons after 1000 hr with

Jo =2 x 10%4, For Gd,0,,
X(t) = 0. 0025 + 1 4>< 10-8(72x 106) 0 10 em .

To obtain an eshmcn‘e of fhe error of the approxn-
mation, note that for a shell with an outside dlam- '
eter of ]38 in, = 1.76 ¢m and a length of 20 in, =
50.8 cm, an absorption rate of 2 x 104 nev-
trons/cm%:sec at the outer surface will burn up
4,05 x 1023 u'roms or 100 g of Sm'49, This corre- .

sponds fo 868 g of Sm, 03. ‘the volume of burnedfi
up material would be 116 em3. A shell with an

outside diameter of 1.76 c¢m, a length of 50.8 cm,'jf‘“
and a voiume of 116 cm3 has a wall thlckness of

been calculated, and the results are presented in
Fig. 1.3.

Gamma-Ray Heating. Two effects are expected
to contribute to the internal heating of control

rods of rare-earth oxides: absorption of gamma
radiation emitted immediately after capture of

~ neufrons in the rod and absorption of gamma rays

from the reactor fuel.

‘Preliminary calculations of the gamma-ray heating
‘in the control rod indicate that the average energy
absorption will be at least 100 w/em3, More de-
tailed calculations made by Pratt & Whitney Air-
¢raft indicate that the gamma energy absorption

may be as h|gh as 170 w/cm3, For the higher

value the maximum interior temperature rise above
V'rhe surface temperature is 460°C for a 1%.in,-OD
hollow rod of rare-earth oxide with a 14 -in, wall,

. - if the thermal conduchvni‘y is assumed to be 0.0048
0.22 em. The burnup ‘of Eu 0 ‘slabs has olso”‘

“cal/sec:°Cocm. If a cermet of rare-earth oxide and

iron is used in place of the oxide, the maximum
temperature rise is about 45°C, if a thermal con-
ductivity of one-third that of iron, or 0.050 cal/
sec:°Cecm, is assumed,

23

bt i ouoniNoh < e

 

,,,,,,
ANP PROJECT PROGRESS REPORT

BONFIRLNTTIL
ORNL-LR-OWG 7552

; _ JO=10'4 nyeme - sec
; N

) . ‘ \ \ \ o =7 10-2° cm?
E . \ N +=6.5 2107 sec \ \ o =16 x 10 20¢m?2
4 =45 107 sec \ B=1 2
3 , p=1 % \ %,(0)=253 cm™*

A 2,(0)= 620 cm™'

 

 

/
/
/
|
J

 

 

 

Z

 

 

 

‘06

 

i
W N
-0

 

 

 

t=4.5x10° sec
E=lr

>

 

1=9x 105 Sec
A=

I/,
o
D
] '—-/r-—-—-—-

& O

\
JEE L AN o
\\\ \\\\\ \\ N
/
/

 

 

\_

 

 

% | 0.8 /

 

 

 

_fi' 0.6 Y
§ ' | —+t=4.5 xiossec r=9x105 se¢
3 m=1 fi=1

%, /2,0

 

 

 

 

/

 

»
AL
/ _

 

 

 

b /

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S y
: Y +=4.5x10° sec e £=8 % 10% sec
o4 ~ a=1 a=1
p ~
3 C 0.4 /
_ 0.2 // 1
= —
; 0 0,002 0.004 0.0086 0.008 0.010 0.012 C.014 0.016 0.018 0.020

x, DISTANCE FROM FACE OF SLAB (cm)

PTT

Fig. 1.2, kB-urnup of a Gadolinium Oxide Slab.

24

 

 
PERIOD ENDING JUNE 10, 1955

SOMFITEN T
ORNL-LR-DWG 7664

 

 

 

 

 

0.90 '
_ ) - L i g =1x 10" n/em? - sec
o 0.80 2,(0) =120¢em ™ (Eu'®")
| _ \ —7=360 hr 2, (0= 5.2 et (Eu'5?)

F=500hr

 

t=0hr
t=250hr

 

0.60

N\

 

0.50

| o4 AN
L NN

Jxh
%
/|

 

 

 

 

j ) 0.0

 

 

 

 

 

 

o 0008 T Cgoio” o 0.015 ' ©0.020
S B "y, "DISTANCE FROM FACE OF SLAB (cm) ' '

  
   

3. B

urnup qf. a Europiumfi(l):‘cji"é‘lé Slub. |

 

P

 

4 >
1
25

 

 
 

 

 

~“ANP PROJECT PROGRESS REPORT

2. EXPERIMENTAL REACTOR ENGINEERING

H. W. Savage

E. S. Bettis

Aircraft Reactor Engineering Division

~ Design work on the in-pile loop for operation
in the MTR has been completed, and the final loop
-is being fabricated and assembled. A bench test
of a working mockup is under way. Twenty-two
high-velocity forced-circulation large-temperature-
- differential loops were operated for investigating
the corrosion and mass transfer of Inconel by
fluoride fuel mixtures under dynamic conditions.
Six similar loops were operated to test mass trans-
fer in Inconel and stainless steel loops in which
sodium is circulated.

A full-scale model of the ART fuel pump was
tested with water, and performance characteristics
were obtained. A test stand for high-temperature
tests has been designed. A test stand for inter-
mediate heat exchangers (100-tube bundle) is being
assembled, and tests are under way with a stand
designed for testing small-scale (20-tube bundle)
heat exchangers,

Flow patterns are being studied in a full-scale
model of the proposed 21-in. reactor core and en-
trance header. Several modifications are to be
tried in an attempt to prevent flow separation in
the core. -

A fhermol-cyclmg test was made on a sodium-
Inconel-beryllium system, and apparatus for «
third test is being assembled. A small-scale gas-
fired heat source was operated successfully at a
power output of 100 kw, and minor modifications
were planned that will increase the capacity.

IN-PILE LOOP COMPONENT DEVELOPMENT
D. B. Trauger

Aircraft Reactor Engineering Division

Flux Measurements

D. M. Haines
Pratt & Whitney Aircraft

- Flux measurements, as previously described,’

were carried out in the HB-3 beam hole of the MTR,

Cobalt foils installed in various assemblies that
“simulated the loop were irradiated for 3 hr at about
5 Mw. Plans to make measurements at full power
- 'wei'e aba'ndbned because an excessively long ir-

 

: ]D M. Humes, ANP Quar., Prog. Rep. Mar. 10, 1955,
ORNL 1864, p 32.

T

radiation was required to override the effects of
the minimum period in which the reactor could be
brought to full power. Gold foils irradiated at the
low power and again at full power in a vertical
hole adjacent to HB-3 provided a means for extra-
polating the data to full power.

This measurement made it possible to evaluate
the effect of materials and geometry of the in-pile
loop on the flux. Values obtained from foils inside
the fuel tube were consistent with other data taken
from foils irradiated in Inconel tubes by using a
rabbit facility for irradiation in hole HB-3. Other
data on the depression of flux in Inconel tubes
with a fuel mockup present were also used to esti-
mate the flux for the MTR in-pile loop.

The flux seen by the fuel in the in-pile loop is
now expected to be 30% of the unperturbed value.
The heater loop, or nose section, originally de-
signed for a depression of 50%, has been modified
to obtain more power, The present nose section
consists of a 2]/2-turn helical coil shaped, in out-
line, somewhat as a truncated cone, It is mounted
with its axis parallel to the beam hole center line
and with the small end forward. This will permit
the nose to be placed in the most forward position
possible by utilizing the concial end of the water
jacket. The developed length of the coil is slightly
over 3 ft. A total power generation of 24 kw is
required for the design conditions: Reynolds
number, 5,000; temperature differential, 200°F,
The average power density in the nose section,
on this basis, will be 0.7 kw/cm?; however, this
value may be conservative. Provision is being
made to change the loop position in the beam hole
during operation and thus utilize the flux gradient
to adjust the power. A higher power density may
be feasible for later loops.

Fission-Gas Holdup

D. W. Magnuson
Aircraft Reactor Engineering Division

The adsorption of krypton from helium by acti-

vated carbon at liquid-nitrogen temperature is being
tested in adsorption traps designed for use with
the MTR in-pile loop. The traps contain 280 g of

Columbia ACA activated carbon. A helium cylinder

 

 
Sh e fadur giovGoob < fo s

 

@

_ '_mcenner as fhey wn“ in fhe m-plle fes'r

: h‘i_n 14 ft_of leng it

   

containing 0.13% krypton is being used to supply
five times the design flow rate through one trap.
The same supply furnishes the design flow of 0.03
scth ’ro a second system confaining two carbon
traps in series. The second trap wilt be isolated
at the conclusion of the experiment, and the ad-

- sorbed gas will be analyzed for total krypton. After

15 days of operation, the krypton concentration in
the effluent stream was approximately 1 ppm in
both systems, or the fractional breakthrough was
less than 0.001. A temperature-sensing element
filled with oxygen, with an automatic fill device,

is being used to keep a constant liquid level in

the metal Dewar which contains the traps.

Bench Test

L. P. Carpenter
Aircraft Reactor Engmeermg Division

A bench test for the |n-p||e loop Yas been oper-
ated for more than 666 hr. The purpose of this
bench test is to determine the feasibility of con-
struction techniques, to tesi the suitability of the
various materials of construction, to aid in estab-
lishing control and operational procedures and in
training operators for the in-pile test, and to deter-
mine the endurance of supporting equipment.

The design conditions as sut forth for the in-pile
loop were cdhered to as closely as possible in the
bench apparatus. Modifications were limited to
incorporating a resistance-heater coil for power
generation. The loop is enclosed in a plug that
differs from the ccfucl in-pile loop plug only in
that ’rhe forwcrd end is open to accommodate con-

nections w:th 'rhe resnstcnce heater. AH servnce .. the alarm system. A leak in the drive pump suction

. line has been found to be the cause of cavitation
~and_has been corrected. Tl in-pile power. unit

lines to the foop run fhrough ‘the plug m the scrne

 

TR

 

 

 

st

ing parts durmg construction, and the drawings
have been revised in some instances.

  
   
 
 
 
 
  

  

On-e"Ba q'i‘equre ent fof'cn:""'_‘:
| of less ’rhan }’6 in, - -
=y \iffl Jlfy was experlenced o
‘becc’”se of WQ"PI'"Q at fherm‘?,wbpie wells and poor

fits. F:xtures dre being fabricated for use in align=""

PERIOD ENDING JUNE 10, 1955

The commercial Kovar-glass seals used to bring
power and thermocouple wires through bulkheads

and to form leaktight closures have proved trouble-
some by not being adequately leaktight. A seal
has been developed at the Gaseous Diffusion
Plant, K-25, which is quite satisfactory, except
that it is vulnerable to breakage during loop ase

- sembly.  Fourteen thermocouple leads or eight

power leads can be brought through a 1/ -in.~dia
glass cup seal.

Bench test experience showed that the use of
a poured barytes concrete shield in the rear sec-
tion of the shield plug required that the lead wires
be protected from moisture condensation and the
concrete. Glass-braid insulation was applied to

individual wires, which were then bunched and

encased in plastic tubing.. After the concrete was
poured, the wires were found to be shorted. Heat-
ing and pumping on the concrete made the con-
nections usable, but other steps are being taken
to prevent recurrence of this trouble for the in-pile
loop. ‘

The operating conditions for the bench test are:
temperature differential, 175°F; Reynolds number,
5100; power input, 21 kw; temperature differential
for air, 250°F at 250 scfm. Control is achieved
by regulation of the air flow through the heat ex-
changer by an automatic controller that maintains
a nearly constant temperature, +10°F, on the fuel
tube at the pump. The electrical power input is
manually controlled. o

Operation of the loop is proceeding satisfactorily.
Cavitation of the hydraulic power unit has caused
momentary fluctuations in the pump speed that trip

elng" operq'red and checked prior to ship-

 

 

" ment 'ro_ NR.TS

Horlzonfol Shuff Sump Pump
o JUA. Conlln o
Aircrcff Reacfor Engmeerlng DIV!SIOn |

iously described,? was dlscssembled and lnspecfed

after complehng ]OOO hr of opercmon. As can be

'seen in th. 2.1, there was no erosion ‘énd only

d sllght mdlcahon of rubblng befween the pump

 

2}, Conlin, ANP Quar. Prog. Rep. Mar. 10, 1955,
ORNL-1864, p 32,

27

 

s ciout il g

o e

 

 

 

 
s

 

 

 

' 'ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
PHOTO-23330

‘ W

 

Fig. 2.1. Ht;fi.ion'ful-SHin Sump Pump for In-Pile Loop After 1000 hr of Operufio.nuc'.ircul-afi.ng N;:F-
Z(F,-UF, (53.5-40-6.5 mole %) ot 1400°F and 1 gpm in an Isothermal Loop.

shaft extension and the pump housing. This rub-

bing is known to have occurred during preheating
of the pump and to have been caused by thermal
stresses induced by an unduly rigid pump mounting.
There was no evidence of zirconium fluoride vupor
in the rear pump housing beyond a point / in,
from the pump sump proper, and no salts were found
in the vent line after the test. A continuous helium
purge of this area of about 0.3 cth was used.
There was no evidence of oil in the fuel portion
of the pump, the measured oil seal leakage being
0.15 em? for the 1000 hr,  However, the pump
sump gas pressuie was higher than the bearing

" housing pressure, which would tend to cause seal
leakage to be toward the bearing housing. The
~ shaft seal faces were in good condition and ap-
parently could have opera’red for an additional

B 1000 hr, or more,

The pump impeller that W|l| be used in the in-

- pile experiment is identical with the one used in
- -the test described above. The bearing housing
- -and seal designs are different, however, in order

to overcome the problems of radiation damage to

, --;fhe oil and to prevent leakage of fission gases,
" Instead of the oil cooling, the shaft is cooled by
30 copper spool on The shm‘t that operates in a he-

8

lium atmosphere with a close clearance to the

water-cooled housing. The pump bearings and seals

are drop-lubricated. To seal fission gases from
the oil in the hydraulic drive motor, two shaft seals
are employed, one between the pump sump and the
bearing housing and one between the bearing hous-
ing and the hydraulic motor. These are facetype
seals with metallic bellows for the flexible member.
The gas volumes in both the sump and the bearing
housing are purged with helium to further reduce
the possibility of oil contamination, The pump
housing also serves as a leaktight bulkhead in the
water jacket to seal the loop end from the bearing
housing and motor section of the water jacket.
This section between the pump bulkhead and the
intermediate bulkhead will be used to accumilate
the waste bearing- and seal-lubricating oil. An
exploded view of this pump is shown in Fig. 2.2.

Difficulty is still being expérienced with failure
of the rotating seals. This trouble is principally
associated with the bellows and with breakage
of the carbon rings. Replacement of the brass
bellows with stainless steel bellows and the ex-
ercise of greater care in assembly seem to have
improved the situation. The sliding surfaces have
given little, or no, trouble.

 

 
 

 

PERIOD ENDING JUNE 10, 1955

  
  
  
      

UNCLASSIFIED
PHOTO~23679

 

 

 

Fig. 2.2. Exploded View of Pump to Be Used for MTR In-Pile Loop No. 1.

 

lnsfrumenfoflon B ' ' ment, it has prov:ded a good test of fhe recording

‘R, G Affel ‘_ B A Gnodt _V:{M»L)C“n‘cxon'frOI SYSfems |
Alrcraff ReGCfOr Englneer|ng D'VISIO“ o A A T

   

 

 

_ lnsfrumento‘hon _design has been completed, and =~ W Cunmngham -
~ the panel for MTR operanon is necrly assemb!ed 4 Asrcraff Reactor Englneerlng Division
Operatmg exper:enc’ wn‘h fhe bench test pqnel has L Mosf parts “for the flrst m-plle loop ‘have been

- fabrlccn‘ed‘ and ussembly IS under way. A wooden

     
    

/ NRTS. Thi‘s,‘lf is hoped wnlruénsure a proper fl’r
ond minimize on-the.site msfalicmon time.

29

  

 

 

— e et AR
Lot e

b i

 

 

 

ANP PROJECT PROGRESS REPORT

DEVELOPMEMT AND OPERATION OF FORCED-
CIRCULATION CORROSION AND
MASS TRANSFER TESTS

Operation of Fused-Salt-inconel Loops

W. B. McDonald
C. P. Coughlen P. G. Smith

Aircraft Reactor Engineering Division

J. J. Milich ~ R. A, Dreisbach
Pratt & Whitney Aircraft

The operation of high-velocity forcedscirculation
large-temperature-differential loops for investigat-

ing the corrosion of Inconel by fluoride fuel mix-
tures under dynamic conditions has become so
routine that operation can be scheduled, and the
schedules can be maintained for long periods with-
out serious interruption. Twelve loops have been
terminated following scheduled operation; five
loops were terminated short of scheduled operating
time because of various failures; and five loops
were started and are continuing in operation. A
summary of the typical operating conditions is
given in Table 2,1,

During the early part of the quarter, loop failures
were encountered because heater lugs were welded

TABLE 2.1. SUMMARY OF OPERATING CONDITIONS FOR 22 FUSED-SALT-INCONEL
FORCED-CIRCULATION CORROSION AND MASS TRANSFER TESTS

Maximum Fused Salt Temperature: 1500°F

 

 

 

: : T Maximum o .
Loop " Method of ype of Reynolds T?mperat?re Recorded Tube Fused Salt p;lzafmg Reason for
No. Heating Heaf.ed Number lefeorenhol Wall Temperature Circulated me Termination
. Section . (°F) (°F) (hr)

4950-1 Direct resistance Straight 5,000 200 1650 Na F-ZrF4-U F4a 1000 Scheduled

4950-2  Direct resistance Straight 5,000 200 1565 Na F-ZrF4-U F4‘z 1000 Scheduled

4950-3 Direct resistance Straight 10,000 200 1690 Na F-ZrF4-U F4b 1C00 Scheduled”

4950-4 Direct resisfunce.e Straight 10,000 100 1600 Na F-ZrF4-U F4‘z 1000 Scheduled

4950-5 Direct resistance Straight 10,000 200 1575 Na F-ZrF4-U F4a 1000 Scheduled

4950-6 Direct resistance  Straight 8,000 300 16206 . Na F-ZrF4-U F4a 1600 Scheduled

4930-A Direct resistance Coiled 1,060 300 1695 Na F-ZrF4-U F4C 1000 Scheduled

4695-1 Direct resistance Coiled 10,000 300 1720 NaF-ZrF4~UF4b 385 Leak.

4695-2  Direct resistance Straight 15,000 200 1670 Na F—ZrF4-U de 887 P ump-bearing failure

4695-3  Direct resistance Straight 10,000 300 1640 NaF-KF-LiF¥ 630 Leak

4695-4A Direct resistance Straight 10,000 200 NoF-ZrF4-U F4a 0 Terminated after
cleaning operation
as first of a series
for determination
of effect of time

4695-5A Direct resistance Straight 10,000 200 Na F-ZrF4-U F4“ 10 Scheduled

4695-4B  Direct resistance Straight 10,000 200 NaF-ZrF4-U I=4‘I 50 Scheduled

4695-4C Direct resistance Straight 10,000 200 NaF-ZrF -UF 100 Scheduled

4695-5B Direct resistance Straight 10,000 200 NaF~ZrF4-U F4a 241 Leak

4695-5C Direct resistance Straight 10,000 200 NaF-ZrF UF 500  Scheduled

4935-1  Gas-fired heater  Coiled 1,000 300 1670 NaF-ZrF -UF,© 1000  Scheduled

4935-2  Gas-fired heater  Coiled 5,000 200 1675 NaF-Zer-U F4“ 1000 Scheduled

4935-3 Gas-fired heater  Coiled 1¢,000 100 1540 Na F-Zer-U F4a 1000 Scheduled

4935-4 Gas-fired heater Coiled 10,000 100 1690 Na F-ZrF4-U F4a 486 Leak 7

4935-5 Gas-fired heater Coiled 10,000 200 1645 NaF-ZcF -UF ,* 1000 Scheduled

4935-6  Gas-fired heater  Coiled 8,000 200 1550 NaF-ZrF -UF,% 1000  Scheduled

EComposition: 50-46-4 mole %.

bComposifion: 50-46-4 mole % with 2 wt % of total uranium converted to U3,

“Composition: 53.5-40-6.5 mole %.
dCOmposifion: 11.7-59.1-29.2 mole %.

30

s

 
col R i | o el SRR e

.":‘; ) R - \,4.:9,_51‘-42'

 

to the tube wall in such a manner as to cause high
thermal stress concentrations and high electrical
current density in the lugs. Three identical
failures resulted before butt welds were sub-
: stituted for axial bead welds to alleviate the con-
dition that caused the failures.

Two bearing failures resulted in an investigation
of the thrust bearings used in the pumps (model
LFB). This study indicated that the fits were too
tight, and new bearings with looser fits between
balls and races were obtained. No further bearing
failures have occurred. -

Metallurgical examination of early direct-re-
sistance-heated loops in which the heated sections
were coiled revealed excessive attack on the com-
pression side of the bends. Temperature measure-
ments showed the compression side of the wall to
be approximately 100°F hotter than the tension
side. Since the wall of the tubing on the tension
side of the bend would have become thinner during
bending and the wall on the compression side
thicker, it was thought that the greater current
density of the thicker wall on the compression
side could have caused the overheating and the
resultant excessive corrosion, An inyesfigation3
of flow revealed that poor fused-salt flow distri-
bution at and past the bends resulted in poor heat
transfer and thus also caused overheating. A new
heated section was therefore designed that elimi.
nated bends from the sections that carried high
current and provided approximately 40 diameters

 

3H, W, Hoffman, L. D, Palmer, and N, D. Greene,
Electrical Heating and Flow in Tube Bends, ORNL
CF-55-2-148 (Fe,-l:vt 22, 1955).

     

PERIOD ENDING JUNE 10, 1955

of straight section following each bend. In addition,
the length of the heated section in the resistance-
heated loops was increased from 12 ft to approxi-
mately 17 ft in order to reduce the maximum tube
wall temperature to a more tolerable level, These
modifications appear to have corrected the over-
heating and resulfant excessive corrosion,

The gas-heated loops are being run to evaluate
the effect of the heating method and the effect of
tube wall temperature on corrosion and mass transe
fer. Loops 4935-5 and 4935-6 are the first two of
a series of three loops fo be run with wall temper-

atures of 1550°F, 1650°F, and 1800°F.

Sodium in Multimetal Loops

C. P. Coughlen D. R. Ward

Aircraft Reactor Engineering Division

R. A. Dreisbach
Pratt & Whitney Aircraft

Four loops with sodium in Inconel and two loops
with sodium in Inconel and type 316 stainless
steel were operated and terminated. The important
operating conditions for these loops are given in
Table 2.2, Since appreciable deposits of mass-
transferred material were found in the cold legs
of these loops, a controlled series of experiments
was initiated, The operating conditions for these
loops, which are now operating, are given in Table
2.3. These loops will provide information on the
effects of the oxygen content of the sodium, the
use of a cold trap, the use of a lower maximum
sodium temperature, and theuse of an all-stainless-
steel system. '

  

 

 
  
  

   

emperature ' Operating

-~ Ditterential Time

 

 

951
Inconel

4951-3 / Inconel

  
      

>,.‘|5’b.06 S T R
> 15,000 180 500

(°F) (hr)

 

7 oo
B [ .
00 4sd
‘300 500

 

31

 

 

 

oy e

IR

 
ANP PROJECT PROGRESS REPORT

TABLE 2.3. SUMMARY OF OPERATING CONDITIONS FOR LOOPS NOW CIRCULATING SODIUM

Reynolds number: 15'00'0
Temperature differential: 300°F
Scheduled operating time: 1000 hr

 

Maximum Sodium

 

 

 

 

 

 

L oop Loop Material Temperature Controlled Variation
N°t o
("F)
4951-4  Inconel 1300 Maximum fluid temperature
4951-5 inconel 1500 0.15% 02 added
4951-6 Incone! 1500 Special high-purity sodium cold trap used
4951-7 ' Type 316 stainless steel 1500 Loop metal
PUMP DEVELOPMENT A series of ten experiments has been performed,
E. R. Dytko to date, and the test conditions and results are
Pratt & Whitney Aircraft given in Table 2.4. The performance data from
. « : . experiment No. 3 are plotted in Fig, 2.3, These
A. G. Grindell G. D. Whitman

Aircraft Reactor Engineering Division

Water Performance Tests

G. D. Whitman
R. L. Brewster M. E. Lackey
Aircraft Reactor Engineering Division

A full-scale model of the ART fuel pump design
designated MF-2 was built, and water performance
data were obtained for several impeller designs
and pump suction conditions for a given discharge-
volute design. The discharge volute and the test
impel lers were fabricated from brass, and a bearing
housing used for the model MF-1 tests? was adapted
for use in the rotary assembly. The unit was
driven by a direct-coupled, 15-hp, variable-speed,
d-¢c mofor. The pump was installed in a test loop
built of é-in. pipe with head and flow measuring
instrumentation and a throttling valve,

The first experiment was performed by using a
pump-suction configuration that simulated the re-
actor design. A box containing a flat plate 1 in,
below and parallel to the pump-suction flange was
used. For the second test, the flat plate was
removed and an 8-in. pipe was connected directly
to the pump suction. Since pump performance was
about the same in both tests, the remainder of the
experiments were conducted with the 8-in. suction
line.

 

- 4A, G. Grindell and W. C. Snapp, ANP Quar. Prog.
Rep, Muar, 10, 1955, ORNL-1864, p 34.

32

data are representative of the best operation ob-
tained during the tests. The pump efficiencies are
not considered to be accurate on an absolute basis,
because the motor was not calibrated and the motor
efficiencies were obtained from the manufacturer’s
computed data. It is estimated that the efficiency
of the pump, exclusive of seal and bearing losses,
is approximately 70% at the design point.

Data obtained by varying the running clearance
between the lower impeller shroud and the volute
indicated that the clearance could be in excess
of 0.040 in. without loss in performance at design
point. At low-flow high-head conditions, there was
approximately a 10% loss in head because of re-
circulation in the pump.

A cavitation problem persisted throughout the
experiments. At flows of over 400 gpm and speeds
in excess of 2000 rpm, a slight noise was detected
in the pump, and the intensity of this disturbance
increased with increased flow or speed above the
threshold values. However, no cavitation damage
has been detected in the pump, and the data do
not indicate a decrease in performance in the cavi-
tation region. The performance of the pump was
not altered by varying the suction pressure over
a range of =5 to +15 psig; however, the intensity
of the noise could be suppressed by increasing
the system pressure. The suction conditions were
altered by changing the radius of the inlet eye
from 1/2 to 1 in., and the impeller nut was rede-
signed to give better fluid guidance. Neither of

-

)
 

L)

*y

PERIOD ENDING JUNE 10, 1955

TABLE 2.4. CONDITIONS AND RESULTS OF WATER PERFORMANCE TESTS OF
ART FUEL PUMP MODEL MF-2

Design point: 620 gpm, 35-ft head

 

Experiment

 

    
 
     
 

Impeller Design Suction Conditions Remarks and Results
Number

1 Five vanes; blade tip Suction box simulating  Design point met at approximately 2800 rpm
angle, 22 deg reactor; ]é-in. radius

on suction eye

2 Same as above 8-in. pipe with four No appreciable change in performance

antiswirl vanes; ]/z-in.
radius on suction eye

3 Five vanes; blade tip S‘qme as above Design peint met at approximately 2600 rpm;
angle, 26.5 deg approximately a 10% increase in efficiency

with respect to experiment 1, with the peak
efficiency shifted toward higher flows

4 Five vanes; blade tip Same as above Decrease in pump performance with respect
angle, 22 deg; leading to experiment
edges cut back on an
80-deg cone angle

5 Five vanes; blade tip Same as above No appreciable change in pump performance

- angle, 26.5 deg; in- at design point
creased lower shroud
clearance from 0.010
to 0,025 in,

6 Same as above with Same as above No appreciable change in pump performance
radial clearance in- at design point; 10% loss in head with re-
creased to 0.040 in, ' spect to experiment at flows below 300 gpm

' ' and speeds above 2000 rpm

7 Six vanes; blade tip Same as above Similar to experiment 1

angle, 22 deg
Same 'as;i‘arb::vé _' " ‘ Similar to expérimehts 1and 7

._'ane Yanes, blade hp _
Ie', 26 5 deg T

Tongue cuf back on volufe, performance

N S|m||ar fo thut |n éxper:ment 3 T

      
    
  
   
 
   

Soma as above Same as above Tongue cut back on velute; mod'lfled :mpeller

   

'performqnce sumllar to Thqt m ex-

per:meni 8

 

 

: Beurmg and Seal Tests
these chdnges']hqd any’ detectable effect on the _
- ,'caw'rahon problem. It is believed, at present that T D.R. WF’F_‘.*
" the cavitation may be a re\syfih of poor fluid guid-" _ W, C Tunnell ‘ J. W. Kingsley
" ance at the leading edges of the impeller vanes.“; 7 Alrcraft Reactor Engineering Division
A new vane design has been completed and will A test has been desibned for studying the func-

be tested soon, ' tion of the interference fit between the journal

   

  

 

33

 

AP v

s

 

 
 

 

 

ANP PROJECT PROGRESS REPORT

EORE

 

 

 

 

" ORNL~LR~—DWG 7553
10
100
90
_ 4——SURGE BOUNDARY
80

 

/ "4\‘ — — — PUMP EFFICIENCY
- / / / \

 

 

70 e 6\0/
.-.-"'--. O/ —_——————
st~ O T
£e0g e 4\\ T =X
$0 S T - 7 =3, -
Ll
I

 

 

 

40 A=

50 / 3 A& O)/ ” : T .
/
' / // // ’ 30/ - \ // X DESIGN POINT
/s
/) <

 

 

 

 

\fi ///g)j]o/
= L———
O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30 iz — \

_ ' -~ o
===l i |7\ °%
. —em T N T\

v o
N0 o,
10 | ~ A%, o
05 1000 o >
800 rpm ND\
o .hi'—.-..___.‘ -
0 100 200 300 400 500 600 700 800 900

CAPACITY (gpm)

Fig. 2.3. Performance Characteristics of ART Fuel Pump Model MF-2 in Water Tésts. For test con-

ditions see Table 2.4,

bearing and the face seal in the model MF-2 pump
at elevated shaft temperatures and under simulated
bearing loadings. The test apparatus consists of
an MF.2 rotary element to which a loading device
is attached at the impeller location. The pump
sheft rotates freely within the side-loading device.
Heat is applied at the lower region of the pump to
simulate gamma heating under actual operating
conditions. One phase of the test will consist in
an endurance run under simulated pump design
conditions, and the other phase, for which a dupli-
cate test assembly is to be used, will consist in
short tests for studying the following variables:
bearing loading, pump speed, lower shaft temper-
ature, rate of coolant flow, and time.

Mechanical Shakedown Test

W. L. Snapp J. J. W. Simon
Pratt & Whitney Aircraft

A few additional mechanical shakedown tests
were performed on the model ME-1 ART fuel pump.

34

The rotary element was operated continuously for
a period of 300 hr at 3800 rpm without achieving a
successful seal at the lower journal-bearing region.
[t was found that the oil leakage rate was nearly
constant at 1.8 in.% per day. Although this rate is
not excessive, the goal of zero leakage was not
achieved, probably because the face of the journal
was not flat, No further testing of this model is
planned.

As part of the renovation of ARE-type sump
pumps for use in heat exchanger and other tests,
100-hr cold mechanical shakedown tests of the
rotary elements were conducted on five units. To
assist in the development of the metal-to-metal
seal for the model MF-2 pump, the floating Graph-
itar ring at the lower seal assembly was removed
and replaced, on two units, by a modified upper
seal bellows assembly. This modification made
a metal-to-metal lower seal assembly. One ele-

ment failed in the cold shakedown test because of

a faulty bellows convolution, but the other element

 
D e e e i

 

i

ik o it s

e, siva b i

 

was found to be very satisfactory in that no leak-

3 age was detected, The latter element was then
' placed in operation at high temperature on a heat
. exchanger test stand. After about 200 hr of hot

- _ operation, a bearing hum developed and the unit

e was replaced; however, up to that time, there had

been no oil leakage detected at this lower metal-
to-metal seal. Upon disassembly of the unit, the
seal surfaces were found to be in excellent con-
dition. As a result of this test, all additional
ARE-type pumps will have metal-to-metal seals
at the lower journal.

Design work has been completed, and fabrication
and assembly have been started on two mechanical
shakedown test stands for testing model MF-2
rotary assemblies. '

‘Short-Circuit Pump-Test Stand
S. M. DeCamp

Aircraft Reactor Engineering Division

The study and design of « short-circuit loop for
testing model MF-2 pumps were completed, and
fabrication was started. Tests are to be made with
this loop at operating temperatures in the range
1100 to 1500°F, The tests have been designed
s ~ for determining the following:

1. proper operation of the lubricating and coolant
system at operating temperatures,

2. leakage rates of the upper and lower seals
during actual pump operation at elevated tem-
perafures, ’

3. temperature gradients along the pump support
cylinder and in the rotary element,

e

4, proper clecrcm(':'es' between the test impe”er
~and pump’ casmg at operating temperatures,
5. 'proper f:’mng of parts at operahng fempercfures

cmd |ocds,

 

N'pressure-breckdown Tabyrinth fpr leakage around
the pump impeller. This labyrinth was designed to
simulate pump-suction and discharge-volute con-

 

 
  
  
 
 
 
 
 

 

 
   
 
 
     

.‘fh‘” thé PU“?P mG)'f; be operated at flow rates of Up
to _600 gpm and a head of cpprox1mcn‘e|y 50 ft. /

PERIOD ENDING JUNE 10, 1955

ditions as presently envisioned for the ART fuel
pumps.

The removal of pumping power, approximately
40 hp at the isothermal steady-state condition, is
to be obtained by constructing a water wall around
the pump tank. A movable furnace will be installed
between the pump tank and the water wall, It
should be possible to dissipate the pumping power
at an isothermal operating condition by lowering
the movable furnace and exposing a portion of the
hot tank wall to the water wall. The fluid flow in
this very short loop will be observed through ports
during operation with water,

HEAT EXCHANGER TESTS
E. R, Dytko
Pratt & Whitney Aircraft

R, E. MacPherson

Aircraft Reactor Engineering Division

" Heat Exchanger Tube-Spacer Pressure-Drop Tests

R. D. Peak
Pratt & Whitney Aircraft

J. W, Kingsley

Aircraft Reactor Engineering Division

In order to select the best spacer configuration
for the ART heat exchanger tube bundle, a series
of tests on various spacer arrangements has been
conducted. The test apparatus consisted of a
representative ART type of tube bundle composed
of 25 aluminum rods 0.191 in. in diameter and 6 ft
long contained in a square aluminum duct 1.12 in.
inside, The tube spacers were formed from flat.

tened copper wire 0,028 by 0.046 in. Tube-to-tube

’ ‘spacing was 0.028 in., and tube-to-wall spacing
'"‘was'_" 0. 025 fo 0028 m.'

50 ered_ pn ‘opposite sides of the bundle into

The spacer rods were

. thick 'i:opper plafe ‘/ in. wide. Water

wos me'rered through the qppamtus by a Rotameter,
"""""_"and total pressure drops were foken across the
f\_tube bundmlrpby means of a monomefer.

Six confugurahons were tested, and the results

of each test are shown in Fig. 2, 4. The results

"':ndlcate ‘that the pressure “drop is least when the
spacers “afe’ mcllned at 45 deg 'to the dlre»chon of
”_'lflow. Vertlcal, m-lme Spocers, of . cour'.e, gave

- :.’”'ifh \

lghest pressure drop Sfuggerlng the spccers

 

inclining these staggered spacers prowded no
improvement, as would be expected.

35

 

saiibainitu: ol il o

T

 
 

jEo dbaiEe G dbaa b

 

 ANP PROJECT PROGRESS REPORT

SEGRET

ORNL—LR—DWG 7554

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 
   
 
 

 

 

 

50
/‘
STAGGERED INCLINED
o O 00 OO
A 80° 0°
m 45° 45°
A Q° 459
® SPECIAL 450
0 74 /// //
17
: A
O /
5 /L\‘——TUBES WITHOUT SPACERS
wl
W 50
-
W
2 A/W /
@
a
4 /
=
2 o B / |
/ 0° STAGGER,
® 0° INCLiNE—y
A
2.0
.l
.
N
I A
O
AQ W
e
T
.0 7/ < TT—-45° STAGGER,
~ 0
. 2 45° INCLINE
>~ 0° STAGGER,
45° |NCLINE
4 /
&
/ ~~——45° INCLINE , SPECIAL
05 l I | | [
. 500 1000 2000 5000 40,000

- F|924 L llé;as:ulfé of Pressure Drop Tests of Vari-d‘hs Tdbe;Spocer Arrungelfienfs for unARTHeot

. Exchflngér Tube Bundle.

  

REYNOLDS NUMBER

 

 

 
e i

 

ote (T LRSS i e ol

 

Intermediate Heat Exchanger Tests

. _ R. D. Peak

. Pratt & Whitney Aircraft

: Construction is about 70% completed on the test
T stand described previously® and shown here in
o Fig. 2.5. The two 1.3-Mw 100-tube heat exchanger
e bundles have been completed,® as well as the two

500-kw high-conductivity-fin radiators. The NaK
pump (DANA), fuel pump (DAC), and the radiator
blower are salvaged equipment from the ARE. The
1-Mw gas-fired heater for this test loop is now
being fabricated by the Struthers Wells Corp.

Small Heat Exchanger Tests
J. C. Amos M. M. Yarosh

Aircraft Reactor Engineering Division

R. I, Gray
Pratt & Whitney Aircraft

A test of a small fuel-to-NaK heat exchanger
was started April 10, 1955. Pressure drop and
heat transfer data have been taken through the
fluoride mixture Reynolds number range 500 to

’ 6300. Preliminary analyses of the data indicate
that the heat exchanger pressure drops are in good
agreement with pressure-drop and friction-factor

o information obtained from recent water-pressure-

) o S .:;", 5 ST ST
' o 1955, ‘ORNL- 1864, p 36,

drop tests carried out on tube bundles with similar
geometry. '

Basic heat transfer data are presented in Fig.
2,6, which compares the recent results with the
theoretical relationship Nu/Pr0+4 = 0,23 Re0-8,
The equivalent diameter used in calculating the
Reynolds numbers was based on the total wetted

perimeter of the tubes plus the s:de-wall crea,rr
" while the equrvolen’r diameter used in compu’rmg
the Nuseelt_number uhhzed only fhe ’rube perlmeter. h

 

  
 
 
 
  
     

 

 

6P, Patriarca et al., ANP Quar., Prog. Rep. Mar. 10,
1955, ORNL-1864, p 131.

 

 
  
  
 

 

\Inconel “tubes 3/] 6 in. in outside dian'letfier\,-r wrth o

The tubes, WhICh are qp',,\,_‘.. flows below this value.

e
fow separa‘flon is encountered at the outer core
shell surface above the equator, turbulators will

PERIOD ENDING JUNE 10, 1955

clearance., The fluoride mixture NaF-ZrF4-UF4
(50-46-4 mole %) is circulated outside the tubes,
and NaK (56% Na-44% K) is circulated in the tubes,

The assembly is now operating on an endurance
run at a fluoride mixture Reynolds number of ap-
proximately 3000, It is planned to take data over
the Reynolds number range 500 to 6000 after 500 hr
of operation and again after 1000 hr.

CORE FLOW TEST
G. D, Whitman R. L. Brewster

Aircraft Reactor Engineering Division

A full-scale model of the proposed 21-in. reactor

core and an entrance header have been fabricated

and checked for fluid flow reversal and/or stag-
nation, The model was installed in a loop con-

taining a 1000-gpm water pump, and two entrance

lines were provided at the header to simulate the
reactor design. Data were taken at flows that
provided a Reynolds number correlation of 1:1
between water flow rates and design fuel flow,

The core model was machined from cluminum
castings and was instrumented at 72 static pressure
and 18 probe points at nine elevations along the
vertical axis of the core. Directionsfinding impact
tubes were used to traverse the core-shell region
at the probe points. There were two prcbe loca-
tions at each of the nine elevations and in the
entrance header. A traverse was made from the
inner to the outer wall of the fluid passage to
obtain the direction of flow and a total pressure
profile. Static pressure measurements were taken
at the walls at each elevation, and the fluid ve-
locity was computed from the total and static
pressure data,

~~The data were taken in the upper half of the core,
dnd a region of flow reversal was encountered
“around the inner core shell.
" tended approximately one quarter of the distance

This reversal ex-
between the inner and outer shell surfecces and
was not sensitive to flows down to one-half the

design Reynoids number, Data were not taken at

. The entrance geometry produced a large rotational

. component of velocity in the fluid entering the core,

and turning vanes are to be inserted to direct the
xially through the core volume. If intolerable

 

be attached to the surface in an attempt to reduce
this trouble.

37

I

 

 
 

 

 

 

 

“ ANP PROJECT PROGRESS REPORT

. . o
TWO 500-kw RADIATORS ORNL-LR-DWG 7555

 

I P
Tl

   
 
   

       
 

  
  
  

NakK PUMP

-

   

4

mmfl"
Qo
o

 

5y ol
i
FUEL. PUMP

 

STl 7 VENTURI FOR FUEL
. PRESSURE MEASURING DEVICES

- L I

S

t/ |
¥ - Is\:\ (\/:\.\\.‘:\ VENTUR| FOR NaK
T “\é“ PRESSURE MEASURING DEVICES
S

FREEZE VALVE

 

   

i—;jt
FUEL SUMP TANK
m__,fi
“W‘n-fi—\—j-f—r( 7777
{-Mw GAS-FIRED HEATER ——— | :

 

 

  
 

FREEZE VALVE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NaK SUMP TANK

   

o Flg 25 Isometric Drawing of Intermediate Heat Exchanger Test annd; o o

38

 

ool
b o ._

 
 

 

Nu /PO

100 200 500

 

PERIOD ENDING JUNE 10, 1955

CTTRET
ORNL—LR—DWG 7556

Nu/Pro% = 0023 Re®®

0 GROUP 1 (RUNS 1-19)
& GROUP 1T (RUNS 21-38)
o GROUP 11 (RUNS 40-53)

10060. 2000 5000 10,000

REYNOLDS NUMBER OF FLUORIDE MIXTURE

Fig. 2.6. Heat Transfer Data for 20-Tube Fuel-to-NaK Heat Exchanger.

THERMAL-CYCLING TEST OF SODIUM: |
INCONEL-BERYLLIUM SYSTEM

R. D. Peak M. H. Cooper
F’ruh‘ & Whutney A:rcrcft '

 

   

-

tabulated in Table 2.5.
71bid., p 134

 

“The ey-

 The beryllium piece was cycled 104 times be-
tween full power and low power, with 20 cycles

" having a cycle period of 20 min and 80 cycles

“having a cycle period of 4 hr; four cycles were

- required for instrument checks, power failures
Opercmon of the second sodlum-berylhum—lnconel r P ’

_ compatibility testing apparotus, prewously de-
e scribed,” has been completed The tes’r stand |s' |
" shown in the isomefric drawin
. lindrical berylllum tesf'plece was mounted between;
s the secondary connections of a 250-kw transformer

- and dlrectly heated by passmg electrical current
o fhrough it. Hot sodium was pumped through the
T berylhum 'tesf plece cmd fhen to a rodlqtor. The

and startup. The time required for a change from
~ full power to low power or back was 2 min. Upon
completion of the required cycles, the apparatus
. was run at full power to achieve a total operating

.. time of 1030 hr. The beryllium test piece is shown

in Fig. 2.8 after the unit was disassembled, The
beryllium was found to have grown from 0.0003 to
0.0041 in. on an average outside diameter of 1,125
in, There were three axial cracks approximately
1/16 in. long on the outer surface at the hot end of
the beryllium. Inspection by the Dychek method
revealed no other cracks on the outer surface.

39

 

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 7557

SURGE TANK

==
=
==

 

 

 

 

FLOW METER

 

 

 

 

 

    
  
  
 
  
  
     
     

RADIATOR

 

ELECTROMAGNETIC PUMP

COLD TRAP

  
   
 

 

T
SN
o

-

 

 

 

   

BUS BAR
TO TRANSFORMER AN

e

J

BERYLLIUM TEST SECTION

‘l'
BUS BAR
TO TRANSFO@ ‘

FREEZE ValVE

 

 

 

-§ I = i

1

N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

| Fig. 2.7. lsometric Drawing of Lodp for Thermal-Cycling Tests of a Sodium-lnconelaBeryliliu'r.h System,

40

 

 
e

PERIOCD ENDING JUNE 10, 1955

TABLE 2.5, SUMMARY OF OPERATING CONDITIONS FOR THERMAL-CYCLING TEST OF
SODIUM-INCONEL-BERYLLIUM SYSTEM

 

 

 

il

b i ke e

WaaEE L a i M

 

 

 

High Power Level Low Power Level

Beryllium power, w/cm3 61 2.5
Current through test piece, amp ' 11,100 1,900
Sodium flow, gpm ' _ ‘ 4.0 4,0
Sfidium' inlet temperature, °F 1050 1050

Beryllium-sodium interface temperature at sodium inlet, °F 1095

Beryllium outside-diameter temperature ot sodium inlet, °F 1225
Sodium outlet temperature, °F 1200 1050

Beryllium-sodium interface temperature at sodium outlet, °F 1245

Beryllium outside-dianeter temperature at sodium outlet, °F 1375

UNCLASS
BERYLLIUM TEST PIECE PHOI%(?-—Z.I’.»ZEI#D

 

-s—— DIRECTION OF SODIUM FLOW

 

 

 

Fig. 2.8. Beryllium Test Piece from Thermal-Cycling Test of Sodium-Inconel-Beryllium System,

41

   

 

 

 

 

 

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

‘The beryllium pieces and sections of the loop are
undergoing extensive metallographic examination,
Apparatus is being assembled for a third test.
- The test stand and test program will be similar to
those used for the test described above, The
same full-power density, 61 w/c_m3, will be em-
ployed, but the average sodium temperature will
be increased from 1125 to 1225°F. The test is to
be completed during the next quarter.
" GAS-FIRED HEAT SOURCE
R E.'MacPherson
Aircraft Reactor Engineering Division
| R, Curry
Pratt & Whitney Aircraft
The small-scale gas-fired source, described
previously,® has been tested at a power output

level of 100 kw, The operating conditions for a
‘power output of 100 kw are given below,

Sodium flow rate, gpm 7.9

Sodium temperature differential through 380
heater, 7

Furnace output, kw 103

Gas flow rate, scfm at 14,7 psia and 70°F 20.5

Chemical heat input rate (based on - 354
980 Btu/scfm), kw

Furnace efficiency (furnace output vs 29

chemical heat input), %

42

Testing was not attempted at higher outputs be-
cause a modification of the gas injection ports to
reduce the pressure drop would have been neces-
sary, Preheating of the heat exchanger section
prior to sodium filling was accomplished by oper-
ating the burner at low power output and adding
cooling air downstream of the combustion chamber.
It was possible to maintain heat exchanger tube
temperatures in the 800 to 1200°F range by this
means. o

No difficulties were encountered except for the

failure of the spark-plug igniter and several plastic
thermocouple connectors because of the high
radiant heat emission from the combustion chamber
barrel. The gas fiow to the burner was momen-
tarily interrupted several times in the course of
the test of reliability of relighting. No problems
arose prior to failure of the igniter mechanism,
mentioned above,

With a few minor repairs, the furnace can be put
into operation as a vutility heat source. By en-
larging the gas injection ports (a minor modifica-
tion), the capacity of the furnace may be increased
considerably for any future application.

 

8R, E. MacPherson and R. Curry, ANP Quar, Prog.
Rep. Mar, 10, 1955, ORNL.-1864, p 37.

 
e e

 

Jo ‘Srn CrUdeIe
J« W, Noaks

ROOM-TEMPERATURE
REFLECTOR-MODERATED-REACTOR
CRITICAL EXPERIMENTS

A series of experiments have been done on
several critical assemblies of the circulating-fuel
reflector-moderated reactor., These experiments
embodied a number of different configurations of
fuel and reflector, which were described in detail
previously.!] The latest of these modifications
consisted of a beryllium region surrounded by a
fuel layer, which, in turn, was enclosed by the
beryllium reflector. This configuration was ex-
tended, with reduced dimensions, on opposite sides
: of the central reactor region to form ‘‘end ducts®’
which simulate the entrance and exit fuel channels,
Some results of a variety of experiments which
were done with this assembly are presented here.

 

e G ks

ki . o s,

Reactivity Meqsuremenis
The fuel loading of the critical assembly was

such that it contained about 3% excess reactivity,
which was utilized for the evaluation of reactivity

S|gn of the ART.’

  
 

on the crn‘:cal posmons of control rods.

  
   
  
  

 

" the beryllium refle [ the ¢
‘the assem_kn:.]y ‘was reduced'from}”
remov ing fhé ou’rer (essenha!ly cyhnc?rlcci annulus,

sund he‘cenfrol region of

'j-'rvcmnulus was coaxial with the fuel region. The
o decre -

  

 

1A, D. Callihan et al., ANP Quar. Prog. Rep. Mar. 10,
1955, ORNL-1864, p 41.

 

 

chcnges incurred by u!ferlng 'rhe sfrucfure ond byi

It was feasible to mcke 'rhese_;_'_" 7
“solid cylinders and on a cylindrical annulus were

“made along the axis of the critical assembly in
* the test volume described above,

rements prlor to the reduction of the fuel\:if‘
, , ration “to thaf requ:red to determme the
“eritical fuel loadmg. “The evaluations were based
“iIn” oné’ e‘“'enmenf ‘the thickness of
to 85 ine by

.V\;hICh was 27 in. wide and’ 18% in. long. This

y wcs ]65 cents. /_ o

PERIOD ENDING JUNE 10, 1955

- 3. CRITICAL EXPERIMENTS

- A, D, Callihan, Physics Division
’ Do Scott Wo Cn TUI’lne“

S Aircraft Reactor Engineering Division
' , R. M. Spencer, United States Air Force

J. J. Lynn, Physics Division

E. V. Sandin
S. Snyder

Pratt & Whitney Aircraft

ART Control Rod Materials. To provide bases
for the design of the control rod for the ART,
measurements were made on samples of several
possible components. A test volume was formed
by removing beryllium from a region 1/16 by 17, e
in. in cross section that extended 21.5 in. along
the axis of the assembly from one end to a point
0.72 in. beyond the equatorial plane. A loss in
reactivity of 19.1 cents resulted,

Consideration was also given to the thimble in
which the control rod would operate. Tubes of
different dimensions and materials were placed
within the test volume, aond the concomitant de-
pressions in reactivity were determined. The
results are given in Table 3.1. |t is indicated that,
in the range of the measurements, the change in
reactivity depends more upon the quantity of
material present than upon its shape. A plot of
the loss in reactivity incurred by a sample as a
function of its linear density, that is, its mass
per unit length, is given in Fig. 3.1.

An evaluation was then made of the neutron ab-
sorption properties of a mixture of some of the
oxides of the rare-earth elements being considered
for use as the poison in the control rod, Tests on

The composition of the mixture of rare-earth

“oxides tested was the following:

. . Amount in Mixture
Oxides in Mixture "M

(wt %)

Sm 203 63.8
G450, 2.3
Dy,0, 4.8
Nd203 0.9
Y1,0, (and others) 4,2

43

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT
TABLE 3.1. REACTIVITY EFFECT OF CONTROL ROD TH!MBLE MATERIALS
Dimensions _ ¢
Material (in.) Linear Density Reactivity L.oss -
Qutside Wall (in.) {cents) )
Diameter Thickness
Inconel _ 0.871 0.035 12.27 25.5
1.255 0.045 20.72 41.7
1.243 0.062 | 32.68 64.7
1.255 0.045 |
} 32.99 63.3
0.871 0.035 o
1.250 0.G85 41.75 80.7
1,250 0.085 -
| 54,02 97.5
0.871 0.035
;  Type 304 stainless steel 1.253 0.028 14,10 o229
' Type 302 stainless steel 1240 0.060 27.91 41.7
EeRET The mixture was pressed into solid cylinders 0.450
100 ORNL-LR-DWG 7377 and 1.375 in. in diameter and into annuli 0.790 in. j
/ in outside diameter and 0,140 in. in width, Each .
o / piece was about 3/4 in, in height, and the pieces
were combined to give various-length samples,
In the first set of experiments with the small-
80 / sized cylinders and the annuli, the Inconel thimble,
/ which was 1,250 in. in outside diameter and 0.085
70 in. in wall thickness and which decreased the
- reactivity 80.7 cents, was in place in the axial
§ 60 void, The absorber samples were contained in an
E  INCONEL Inconel carrier tube which could be inserted into
z the thimble. The outside diameter of the carrier
2 tube was 0,871 in., the wall thickness was 0.035
- }/ in,, and the end was closed by a 3/32-in.-fl1icl< plug.
o 1O The small cylinders were centered in this carrier
w
S — STAINLESS STEEL tube by a third |nconel sleeve, with an outside .
30 / diameter of 0.500 in. and a wall thickness of 0.020
/ in., which, in turn, was centered by aluminum
20 spacers. The effects of the carrier tube and of
/ assemblies of cylinders and annuli, both separately
o / and together, are summarized in Table 3.2, The
reference zero for the reactivity changes is taken
with the thimble alone. |n all cases the test rod
| 0 _ . . . . .
i B 5 o 20 o 40 o o o was inserted 21.5 in, into the thimble; fhu.'r is, the .
] LINEAR DENSITY (g/in) | end extended 0,72 in. beyond the equatorial plane .
, | of the assembly,
_ Fig. 3.1. Reactivity Effects of Thimble Materials. In a second series of experiments a comparison
i
1 44
i
A
3

 

 
G i R S

 

i i &

- TABLE 3.2, EFFECT OF SAMPLE
CONTROL RODS ON REACTIVITY

 

 

e . o Reactivity
| lL.ength Diameter
Sample _ . Loss
(in.) | {in.) (cents)
Carrier tube only | 16.8 £0.5
Cylinders*  23.0  0.450 155.0 £ 15.0%*
Annuli 2.0 0.790 230.0 +20.0**
(outside)
0.510
(inside)
Cylinders and 21,0 0.790 250.0 +20.0**

annuli together

 

*The innermost centering tube alone decreased the re-
activity about 5 cents.

**Obtained by exfropo!chon of a series of measurements
with shorter samples, effect of carrier tube mcluded
was made of the effecf on reactivity of the dlameter
of the samples of oxide mixture. The diameters
of the samples were 0.450, 0.790, and 1.375 in.,
the one of intermediate size being formed by plac-
ing the small cylinders in the annuli described
above. The samples were about 5.5 in. long, and
each was tested at the same position in the as-
sembly, The distance from the end of the sample
to the midplane of the assembly was 2.0 in. |t
was necessary to remove the 1.25-in.-dia thimble

- for these measurements, and the absorbers had to

be wrapped in aluminum foil. The changes in
reactivity measured, referred to the array with the
test section void (without the thimble), are given

in Table 3.3.

TABLE 3.3 EFFECT OF ABSORBER
i _DIAMETER ON REACTIVITY -

  

 

   

 

    

| Diameter 0 Lenath Loss.
(m) - (_'nf) L .. (cents)

PERIOD ENDING JUNE 10, 1955

a 3-inclong cylindrical sample, 0.450 in. in

diameter, was measured at a point where the
cadmium fraction measured by gold-foil activation
was about 0.5. The measurement was then re-
peated with the sample covered by cadmium 0.02
in. thick. The values were 55.9 and 61.7 cents,
respectively. The cadmium alone reduced the re-
duced the reactivity 37.7 cents, From these data
it appears that about 43% of the reactivity decrease
is due to neutrons which penetrate the cadmium.
Inconel. A measurement, similar to that just
described, was made with an Inconel sample in
order to estimate the dependence of the reduction

in reactlvny upon neutfron energy. The sample was

11/2 43/8 X /32 in.; it weighed 84.14 g; and it could
be provided with a 0.02-in.-thick cadmium cover.
The effects, as poisons, of the [nconel and the
cadmium were measured singly and together at the

center of the beryllium island and in the fuel region

adjacent to the fuel reflector interface. (The I/8-in.
Incone! core shell was reduced to % in. in thick-
ness in the region of this measurement.) The gold-
activation cadmium fractions at the two positions
were 0,57 and 0,33, respectively. The reactivity

changes are given in Table 3.4, The fractions of

the total change that are caused by neutrons

penetrating the cadmium are also tabulated,

TABLE 3.4, REACTIVITY CHANGES FROM
BARE AND CADMIUM-COVERED INCONEL

 

 

 
 
  
      

 

  
 
   

 

 

 
  

An estimate of the ‘spectral distribution of the
neutron absorption by the rare-earth elements was
also made. The change in reactivity produced by

 

At
In Fuel-Reflector
Island
Interface

“Reo.‘o‘fivi'ij loss (cents)
- Ificonél ‘ 18.0 4.0

lnconel cadmlum covered 17.4 26.8
' Cadmlum cover 115.2 26.5

Fractional chonge inre- 12,2 10,6 7.5 £2.6
_dctivity caused by epi-
cudmwm neutrons (%)_

Gold-achvohon codmlum 0.57 0.33
fractlon o

 

 
 

BB

|, it is also bemg consndered as a reacfor

{5\ )..\?‘r"‘.a.,

  

mafenals was therefore mcde in the crmcal as-

sembly. A sample of columbium 4% 7, % 3 % 0.05 in.

 

45

 

(o o33

 
Rkt

 

ANP PROJECT PROGRESS REPORT

 
 
  
 
  

.\.f‘ ‘weighed 98.5 g and was located in the fuel
region adjacent to the ¥ .-in. Inconel separator at
i+ the fuel-reflector interface reduced the reactivify
3.0 cents. A sample of Ifiéonel 43/4 x 3/ X / in.
< that welghed 127.4 g and was located q’r the Sc:me
“position reduced the reactivity 6.6 cents. For

. value must be reduced to 5.3 cents, if it is assumed
© that the effective absorption varies linearly with
i the scmple thickness. The layer of {aconel adja-
‘ cént to the columbium simulated the cladding
”'whlch ‘would probably be required for corrosion
e&:tlon. Some reduction in the critical uranium
":_‘frahon would be effected by the use of
columblum.

" Beryllivm in the Fuel Region. Since the pres-
ence of beryllium in the circulating fuel stream
would be another possnble way of reducing the
:crmcal urcnlum concenfrahon, an experlment was
devised for. evaluahng the effect of beryllium in

~ the fuel | region of the critical assembly, One of

S the cenfrcl fuel subassemblies that consisted of

27 uranium metal sheets (each 0. 004, in. thick) and

. 55 Teflon sheets (each /]4 in. thick) was modified

' -_.‘_m the: followmg manner. Seven uranium sheets

C o and 16 Teflon sheets were removed from the center

o effecf a decrease in reacflwty of 22.7 cents.
. The subsnfuhon of a somple of beryllium, 7/

X 27/ x 1 in., for this uranium and Teflon resulfed

~ in c net reactivity gain of 8,5 cents over the un-

“perturbed value. The addition of a % -in.-thick

 

  
  

f ;'}_’ layer of lnconei cempletely around the beryllium

reduced this net gain to 4.0 cents. An estimate of

_i"i' T‘fhe effect of the |ncone| clad beryllium on the

"concentrahon IS glven ‘by the observation that the
“removal of two of the remaining 20 uranium sheets
" further reduced the reactivity by 4.5 cents, that is,
- “to a value slightly below that of the original array.
- The two sheets removed last were adjacent to the
beryllium sample and were of greater than average
lmporfunce.
~Fuel in the End Ducts. One of the safety features
" of the high-temperature critical experiment presently
5 ""'bemg designed, which is described below, is that
. the liquid fuel can be removed by draining under

HY‘J_,_grav:fy. The degree of safety depen_ds upon the

    

o 5_rcte of removal cnd the sensmv‘f

LN

 
 
 
 

e k{',.{

   

'/:‘4'1"':,upper end duct. A measure of
"*obfclned m the room-tempera 1

D
T
e
0
O"’
B
:J-
¢ e
:3""
(D
U-
S0
m
(.h
...,
0
I
e
o
S0
O_
W
(fl
(Q
-
0
gy ¢
—
|
CD
T
-0
-
,
gy
C
o

 

"j-f':f{_ drmn system.

   

a6

‘comparison with the columbium value, the Inconel -

S o the “fuel Ievel pqrhculorly the’ “Ievel in the
hi: sensmvaty was' )

Three annular rings of fuel were successively
removed in 2%-in. increments from one end duct,
and the resu?’rmg decreases in reactivity were
noted, Each ring contained 825 g of U?33, and
thus a total of 2,475 kg of U235 was removed. The
losses in reactivity accompanying the removal of
the three rings are shown in Fig. 3.2, which also
gives the loss in reactivity per unit displacement,
or fuel sensitivity, averaged over each of the fuel
annuli removed.

SECREL
ORNL—LR—DWG 7378

 

200
180
160
»
c
8 140
>
E
> 120
'_
Q
W
W 400
2 —_
p 80 g
Q &
- o
60 8
>
=
40 >
=
o
20 &
[#2]
o
0 1 2 3 4 5 6 7 8 9

FUEL REMOVED (in}

Fig. 3.2. Effect of Fuel in End Duct on Reac-
'HV“'YG

Power Distributions

The relative fission rate distribution across the
fuel annulus was determined from the fission-
product activities collected on aluminum foils in
contact with the uranium. Exposures were also
made with the foils and the uranium enclosed in

0.02-in.-thick cadmium in order to obtain a measure_
of the energy distribution of the neutrons causing .

fission. The locations of foil traverses within the
fuel section of the assembly are shown at the top
of Fig. 3.3. As may be recalled, the fuel consists
of laminae of vranium (0,004 in. thick) and Teflon
(0.142 in. fhlck), the measurements were made at

" selected positions on lines perpendlcular to these

laminae, The fission rate distribution ocross each
traverse is plotted in the lower part of Fig. 3.3 as

a function of the number of uranium sheets between _

 

 

 
 

 

Ly

RELATIVE ACTIVITY

bl - B RIS b B B BBy By Bogiciail,

40

30

20

 

 

 

Cm

B e SR sl

ORNL-LR-DWG 7379 -

 

  
 
 

BERYLLIUM REFLECTOR

 

 

 

 

 

 

 

FOILS
|—-———— END DUCT
N T b T [T
Nt I 1 -
N ) L
4 4 i
D E F
BERYLLIUM ISLAND
| o
l ® BARE CATCHER FOIL
O CADMIUM-COVERED
’

 

 

¢ CATCHER FOIL
: CADMIUM FRACTION'

 

 

 

 

 

 

 

i B C D ' ® F 9

 

 

 

 

 

NV

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Q o) l{ )l
" \/
27 0 27 0 27 0 8 O 8 o 8

LAYERS OF FUEL FROM ISLAND {1 LAYER =0.004in. OF U AND Q.14 in. OF TEFLON})

Fig. 3.3. Power Distribution Across Fuel Annulus.

 

c.8

0.6

0.4

0.2

CADMIUM FRACTION

§S61 ‘0L ANNr ONIANI QOl¥3d
e i

PR

 

 

 

 

the datum point and the island of the assembly.
~ Values of the cadmium fraction (the fraction of all
~ fissions produced by neutrons having energies less

than ~ 0.5 ev) are also plotted, At the time these
measurements were made, the reflector around the

. center sécfidn ‘of the assembly was only 85/8 in.

'ASSEMBLY FOR HIGH-TEMPERATURE
 CRITICAL EXPERIMENT

" QOperation at high temperature of a critical as-
-~ sembly mockup of the ART is scheduled for late
' _sdmmei', 1955. |n addition to a determination of
“the critical uranium concentration at the elevated
" temperature, the purposes of the experiment are to
“evaluate the temperature coefficient of reactivity,

to investigate the effectiveness of control rod
materials, and to measure the contribution of the
fuel in the upper end duct to the over-all reactivity.

The experiment will be performed at zero nuclear
power and at about 1200°F, Heat will be supplied
by electrical heaters external to the reflector. The

. assembly is to include a beryllium island and a
- beryllium reflector, essentially 12 in, thick, which
~will be in a helium or argon atmosphere and will

contain no sodium.

The ART core configuration, as presently envi-
sioned, will be exactly mocked up in the experiment
between points 18 in. above and below the equatorial
plane of the core. Some minor deviations from the

s

ART design are being made beyond these points to

simplify filling and draining. The fuel will not be
circulated. The core will be filled with molten
fuel from a sump tank by using helium to force the
molten salt into the assembly, The system will be
tilled initially for cleaning and testing with a 50-50
mole % mixture of NaF and ZrF . [ncrements of
Na,UF (with the uranium enriched with 93% U233)
will subsequently be added to the NaF-ZrF, mix-
ture in the sump tank. After each addition of
Na,UF the mixture will be pressurized into the
core and then drained until the critical vranium
concentration is determined, |

A single rod will be located within a 1,50-in.-1D
Inconel thimble along the vertical axis of the
beryllium island. |t will extend from the top of the
reactor tank to 10 in, below the equatorial plane of
the core. The control rod will be a cylindrical
annulus of the mixture of rare-earth elements
described above, enclosed in an |nconel shell.
Annuli of two widths, 1/8 and ]/4 in., will be provided
for comparison. '

The system is being designed to operate iso-
thermally at 1200°F normally, with provision made
for short periods of operation at temperatures up to
1300°F to enable reactivity temperature coefficient
measurements to be made. An attempt will also be
made to measure the fuel temperature coefficient
by inserting the fuel into the reactor assembly at
a temperature different from that of the beryllium.

.y

o,
 

 

 

Part Il

MATERIALS RESEARCH

 

 

 
 

o ¥ o

 
 
 
 
 
 

W

: _accelerated engmeermg test progrom

   
   

- 4. CHEMISTRY OF REACTOR MATERIALS
W. R, Grimes

Materials Chemistry Division

-~ Phase equilibrium studies were made of the
systems LiF-ZrF,, NaF-LiF-ZrF, and NaF-LiF-
ZrF -UF,. Two BeF ,-beafing systems - NaF-
LiF-Berund NaF-LiF-BeF,-UF , — were studied,
and the solubility of UF, in BeF ,-bearing systems
was investigated., A method for zone melting of
fused salts was devised as an aid in phase
equnllbrium “studies. Additional work was done
in mvestlgctlng the equilibrium reduction of FeF
by hydrogen in NaZrF, the reduction of U'F by
structural metals, and fhe stability of chromlum
and iron fluorides in molten fluorides. Vapor
pressure measurements were made on UF in the
temperature range 1270 to 1390°C.

‘The investigation of the variables affecting the
reduction with metallic uranium of UF, dissolved
in alkali fluorides was continued. A study of the
effect of 'rhe nickel filters used in expenments for
determlnmg the stability of UF3 in the NaF-KF-
LiF system showed that disproportionation of the
UF, occurred because of the filter,

Fuel purification and preparation research
included experimental use of electrolysis under a
hydrogen atmosphere fo remove oxides and the
use of metallic zirconium to replace most of the
hydrogen in the stripping operation. In addition,
attempts were made to prepare mixtures containing
very high UF,/UF, ratios., A study of the con-
ditions for the preparahon of BeF ,-bearing melts

was conhnued and producnon operuhons were
resumed to provnde fes? moterlai for the_greatly -

(‘J’\

nrfi e

 

oM Bratcher, V. S. Coleman, and C. J, Barfon,
ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556, p 44.

 

PHASE EQUILIBRIUM STUDIES
C. J. Barton F. F. Blankenship

Materials Chemistry Division
H. Insley, Consultant

The very considerable interest in obtaining
fuel mixtures with physical properties more
favorable than those available in the NaF-ZrF ,-UF
system has led to evaluation of the NaF-LiF-ZrF,
ternary and the NaF-LiF-ZrF ,-UF, quaternary
systems. Largely as a consequence of studies
pursued during the past quarter, the former system
has been reasonably well defined; the latter
requires considerably more effort. While phase
equilibrium data show that quite low melting
points are available in this system and that
adequate melting points are available at ZrF,
concentrations as low as 21 mole %, no compo-
sition with physical properties better than those
available in the NaF-ZrF -UF, system has been
demonstrated.

Study of the analogous NaF-LiF-BeF, system
has been continued. Adequate melting points are
available over wide areas in this system, Whether
the physical properties of proper compositions in
this system can show significant advantages over
those in the NaF-ZrF -UF , system cannot yet be
answered with certainty.

The Binary System LiF-ZrF,
R. E. Moore R. E. Thoma

Materials Chemistry Division
D L __StocktonMMerck & Company

A ?en’raflve | ~diagram f “the LiF- ZrF bmary
system, based prlmorlly on thermal anulys:s data,

  

_was published prewousiy. A re-examination of
... this system by pefrograph[c and x-ray diffraction

studies of quenched and slowly cooled samples

and by dlfferenhal thermal anclysns wos started

A ___durmg the past qucrter becouse of currenf mferest
_in the NaF-LiF- ZrF, system. o
) Alfhough quenchmg of LiF- ZrF4 mixtures is

s : r\oir comp|ete|y sahsfoctory because of erld
phase re!ohonshlps in fuel” systems “of interest.

crystallization of ‘many “mixtures in the system,
some conclusions may be drawn from the results.
A eutectic between LiF and Li;ZrF, at about

51

 

T Y

 
 

 

 

ANP PROJECT PROGRESS REPORT

21 mole % ZrF was confirmed by quenching

"expenmenfs on both sides of this composition.

The eutectic temperature found by quenching is
~ about 600°C instead of the 590°C indicated by
- thermal analysis. The melting point of <Li s ZrF,
"(645°C accordmg to thermal analysis) could nof
be checked by quenching because of too rapid
crystallization, Decomposition of Li,ZrF_ into
Li oZrF ; and LiF occurs somewhere below 480°C.
S Thls decomposmon explains a series of cooling
~ breaks at about 465°C and the apparent discrepancy
between thermcl ano|y51s dota and the results of
" 'solid-phase’ “studies (reported earlier!).  The
melting point of Li ZrF (585°C according to
-recent thermal analys:s dataz) was found to be
- approximately 590°C by quenching. The primary
. phase was found to be Li ,LrF, and thus the

‘belief that 1’]’1!5 compound melts congruently was

, conflrmed _
_ The only other binary compound in the system,
“‘as yet unidentified, was found by examination of
both quenched and slowly cooled samples con-
taining more than 50 mole % ZrF,. No evidence
was found for the existence of LiZrF . A careful
study of samples of compositions with 57.1, 60,
62.5, and 66.7 mole % ZrF , which were held both
above and below the solidus temperature for about

12 hr before quenching indicates that the unidenti-
fied compound is probably Li 2ZrsF . {60%) or
LizZr ,F .y (57.1%) and that it meh‘s incongruently
to ZrF and liquid at about 515°C. The optical
properties of this compound are almost identical
to those of Li,ZrF ; so it was necessary to rely
almost entirely on x-ray diffraction to differentiate
between them. Work on questionable aspects of
the compositions with more than 50 mole % Zrl:4
is continuing.

The differential thermal analysis technique and
the equipment previously used in the study of the
NaF-ZrF, system® were also employed to obtain
more precise thermal data for LiF- Zrl':4 mixtures.
Ten-gram samples were prepared by mixing purified
- Li ZrF with the required amount of ZrF, or
- Lig ZrF in welded nickel capsules fitted with
"fhm-wq(led center thermocouple wells. Manual
‘control of a Variac transformer was found to give

a reosoncbly unlform rate of heating, about 90°C/hr, '

 

2H A, Frredman and F. F Blankenship, unpublished
" data,

. 3R, A Bolomey, ANP Quar. Prog. Rep. Dec. 10, 1953,
_ ORNL 1649, P 54.

52

and of cooling, about 75°C/hr. The base line
obtained on the automatic plot of sample temper-
ature vs differential temperature (difference
between sample and Al O, reference amplified by

a d-c amplifier) seemed to be less erratic when
fhe transformer was controlled manually than when
the automatic temperature controller was used.
The compositions studied to date cover the range
from 31.9 to 66.7 mole % ZrF,. These studies are
incomplete, but the data obtoined indicate a
melting point of 493 + 5°C for the eutectic at
about 50 mole % ZrF and a value of 520 + 5°C
for the incongruent meltlng point of the unidentified
compound containing more than 50 mole % ZrF .
The thermal effect at 462°C was apparent on the
heating curves of all composmons, presumably
because a small amount of LiF present in the
Li,Z¢+F , combined with Li 9ZrF to give LiZrF_,
Thls e?fecf was noted only upon cooling of the
mixture with 31.9 mole % ZrF,, in which it would
be expected that Li 3ZtF ., would be present, The
existence of only one fLermal effect, at 565°C,
for this composition, in addition to the 462°C
solidus transition, appears to verify the existence
of the LizZrF,-Li,ZrF . eutectic near this compo-
sition. These studies are continuing, and it is
expected that a revised diagram will be completed
for this system in the near future.

The Ternary System NaF-LiF-ZrF,

C. M. Blood H. A. Friedman
F. P. Boody F. W. Miles
R. E. Thoma

Materials Chemistry Division

Fuel mixtures based on the NaF-ZtF, system
have suitable melting points only if the compo-
sition lies close to the compound Na Zr6F31,
which is characterized on the phase dlagrom by a
flat maximum at 520°C with 500°C eutectics on
either side. The typical chl:-ZrF"‘-UF4 fuel
mixture therefore contains about 45 mole % ZrF ,,
[t is unfortunate that several of the physical
properties values of this class of fuels are some-
what unsatisfactory because of the high ZrF,
content. The physical properties that are trouble-
some are vapor pressure, kinematic viscosity, and
heat capacity per unit volume, In the L:F-ZrF
system, the low-melting region occurs at 50%
ZrF, and 500°C, and therefore there is little

promise of improved physical properties here.

 

 
S ket

 

i v, Lol s, i e

e s

 

e e kA .

1600 ,

 

However, there was a possibility that low-melting
regions with much lower ZrF, content would be
found in the NaF-LiF-ZrF, ternary system, The
substitution of LiF for NaF was expected to
result in lower melting points, because LigZeF
(mp, 645°C) melts at a much lower temperature
than does Na,ZrF (mp, 850°C), and Li,ZrF ; (mp,
585°C) melts at a much lower temperature than does
Na,ZrF . (incongurent, liquidus at 700°C). The use
of LiF would be expected to result in higher heat
capacities. In order to find low-melting regions
that would have as little ZrF, and as much LiF as
possible, the ternary system NaF-LiF-ZrF, was

942
ZrF,

   

TEMPERATURES
ARE IN °C

PERIOD ENDING JUNE 10, 1955

explored, The results are summarized in Fig. 4.1.
No ternary compounds were encountered,

The liquidus contours were obtained from
cooling curves on melts of at least 500 g con-
tained in 4-in. graphite liners which were enclosed
in nickel pots in a 5-in. pot furnace, The melts
were stirred by belt-driven nickel stirrers; temper-
atures were measured by thermocouples in the
hollow shafts of the stirrers. A flowing atmosphere
of helium was maintained over the melts, The
important features of the ternary system were
determined on mixtures blended from eutectic or
compound compositions which had been purified

CONPIDENTAL
ORNL—LR-DWG 7380

 

 

 

 

  

 
 

 

 

§4‘!
Ry

N
NoF(992) 950 900 850 ~ 800 750 700 €50

 

B ' T Li,ZrFg (585)
- NagZrfg 750 v, N i AR |
e LizZrF(645)
/5?5MN \ 600
[ L A\ 650

 

T NS,
o

S T Heg B0 LiF(842)

Fig. 4.1. Phase Diagram of NaF-LiF-ZrF, System.

 

53

 
 

 

 

o i il ¢ o R

"ANP PROJECT PROGRESS REPORT

All the

solidified samples were examined petrographically

by a high-temperature HF treatment,

to determine the phases present, and representative
samples were examined by x-ray diffraction when-
ever questions arose which could be studied by

~ this technique,

The first objective of the investigation was to

find quasi binaries which would serve to define

the compatibility triangles in the ternary system.
Beginning at the low ZrF, concentrations, the
join between LiF and Na,ZrF_ was found to be a
quasi binary with a simple eutectic at 625°C and

- 38 mole % LiF. The compatibility triangle formed

with this join and NaF as an apex contains a
ternary eutectic at NaF-LiF-ZrF , (51-38-11 mole %)
which melts at 600°C and contains NaF, LiF, and
Na,ZrF ..

The nex’r quasi binary is the Na 3ZrF-LiZrF

- join. Here extensive solid soluflons occur, as

shown in Fig. 4.2, Cooling curves gave uncertain

~indications of the solidus, and therefore the

limits of miscibility at the eutectic are not known.
However, the limit of miscibility of Na,ZrF
Li,ZrF, mixtures at 350°C is clearly defined.
This is due to the existence of a lower limit of
stability of LisZrF, at 490°C, coupled with the
lowering of the limit of stability by increasing
amounts of solid solution with Na,ZrF., and the
large thermal effect associated with the decompo-
sition. At the composition NaF-LiF-ZrF , (26-49-
25 mole %) the lower limit of stob:hfy of the
Li,ZrF-rich solid solution has decreased to
350°C and the saturation limit has been reached.
The x-ray diffraction lines for the Na 2 ZrF o-rich
solid solutions are puzzling in that they show the
expected shifts up to about 30% LiF and unex-
pected, decreased shifts at higher LiF concen-
trations. The eutectic occurs at 50% and 590°C,
The compatibility triangle lying between the
Na,ZrF-LiF join and the Na 3ZrF,-Li ZrF, join
is charccter:zed by three two-phase crystalhzohon
paths with a relatively small temperature gradient;
the paths join to form a eutectic at NaF-LiF- ZrF,
(22-55-23 mole %) that melts at 575°C.
Superficially, the join between NazZrF6 and
Li ,ZrF . appears to be a quasi binary, since only

~ the end6 members are readily identifiable in so-

Ildl_fled melts lying on this join, and a eutectic-
type break in the cooling curves is found at
445°C and 31% LiF. However, Na,ZrF  melts
incongruently at 630°C, and on the Na,ZrF  limb

54

EONHPETTIRL
ORNL-LR-DWG 7381

 

 

 

800 '
\\\ LIQUID

 

 

\

\

\ \ L~ LIQUID AND Nag7ZrFy
\ Al  SOUD SOLUTION

700 \- '
\
\ \
\\ LIQUID AND Ll3ZrF7

X SOLID SOLUTION 2
\ \ \.\ ///

 

 

 

 

 

 

 

 

 

 

— \ * 3
& \ '\ |~ ’/
w 600 “.‘ . '\.. ’.’/ ["
g vt o/
E L LigZrFy SOLID SOLUTION /
o % |
w 1 7K
o \\ I ,I
= NogzE, |\ TWO SOLID ;
"~ 500 o3 A SOLUTIONS /
'SOLID SOLUTION / A
' ,II //
LIF L!aZI’FG, " @’

 

 

SOLID SOLUTION N

 

 

 

\
\
|
\
1
V1 AND LigZreF, , /
1
\
i
1

 

 

 

 

 

 

 

 

 

 

400
LiF, LiyZrFg, i / ¢
— AND LigZrF, —= T —e . N
SOLID SOLUTION ~a 5
300 ,
NasZrF, 20 40 60 80

LiF {mole %)
(ZrF, CONTENT IS 25 mole % THROUGHOUT RANGE)

Fig. 4.2. Phose Diagram of the Quasi-Binary
System Na,ZrF,-Li, ZrF .

of the eutectic, NGSZrF -rich solid solutions are
the primary phase. The eutectic type of halt at
445°C is observable only on the Li,ZrF . limb,
and it corresponds to the appearance of an
Na,ZrF_-rich phase, which must subsequently
react with liquid to give the Na,ZrF . found as a
final product. The true eutectic between Na,ZrF
and Li,ZrF, appears to occur at about 435°C,
The thermal effects accompanying this behavior
are shown in Fig. 4.3. '

These complications have prevented the es-
tablishment of meaningful compatibility triangles
between 25 and 33.3 mole % ZrF,. In this region
at 455°C and the composmon NaF-LiF-ZrF,
(34-35-31 mole %), there is an intersection of the
two-phase crystallization paths involving leszfin

L

 
 

PERIOD ENDING JUNE 10, 1955

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SONTTDETITIT Attempts to define the crystallization path
‘ 650 ORNLZLR—DWG 7382 between Na,ZrF, and Na,ZrF, have been rather
, ' | sketchy qu unfruitful; cooling curves contributed
relatively little toward the definition, Similarly,
:\ essentially nothing is known about the primary
- o phase fleld of Na,Zr,F,;. The region above 45%
600
i , ! ZrF, is of ltttle interest from the standpoint of
/ fuel composition, and has been investigated only
LIQUID Vo sparingly. There was no evidence of melting
/ points below 450°C when the join between
5 5% \ ; /‘ NaZr F,, and LiF-ZrF, (50-50 mole %) was
w \ | NayzrF, . investigated, The composition of the LiF-ZrF,
2 )(SOLID SOLUTION / incongruent compound containing about 60% ZrF,
_ E *  LiguD Li7rF 4+ has not been established, and attempts to find a
= 500 /2 L Lioup ] quasi-binary join between this compound and
= \” \. ‘/ Na,Zr F,, or between Na,Zr,F,, and Li,ZrF,
Nc3ZrF7\ were unsuccessful,
SOLID Na,ZrF, SOLID
et ingILON:\ _ L/ f?()LLl}JITDION*'LingFU The Quaternary System NaF-LiF-ZrF -UF,
Laup \/ -—I—- et C. M. Blood H. A. Friedman
- - »— 0 ° o— : F. P. BOOC[Y F. W. Miles
. u® R. E. Thoma
- : NozszerLizzriFs ;N Materials Chemistry Division
- Mok 1 e 30 a0 0 60 70 The quaternary system NaF-LiF-ZrF -UF, has
. -7 LiF{mote %) been studied as a quasi-ternary system by con-
. _ : (ZrF, CONTENT IS 33.3 mole % THROUGHOUT RANGE) sidering only compositions containing 4 mole %

o o o UF, and by considering the UF, as a substitution
'Fig. 4.3. Thermal Effects in the N“észs' tor 4 fnole % ZLrF . When UF, is added‘to thle
Li,ZrF, System. ' NaF-LiF-ZrF, ternary system, the uranium is
, found in only four phases: Na,Zr(U)F,, Li, Zr(U)F,,

 Na [Zr(U)]6F3,, and Zr(U)F,. In each case ‘rhe

| and Lissz7-rich's'o'|id solutions on the one hand, uranium forms a solid so[utlon. No Zr(U)F, has
o ,C'"d Na,ZrF - and. L' Z"F "“Ch solid S°|Uf'°"5 appeared in composmons of interest as fuel.

o o o 1m0 e o ol G e Ui na N0

. -'t"rreachlng & eutechc. L _‘,_T’d‘ropped 10°C, to 615 C, as a result of the lowermg

“The |om “betwesn Naz 5 F3.1 and;rL|\Zr. . e of the melflng point of Na ZrF by including

"qu05| bmary wnth d\sunpig eui'echc a’r 28% L|F enough uranium to give an over-cll composition of

' ' - - 4 mole % U. On the NaF side of this join, the

~~f ‘quasi-ternary eutectic was found at 590°C, and

gl?;?,;;;; the composition was shifted to NaF-LiF- -LrF ,-UF
e (53-35-8-4 mole %),

ev')suddle between Nc Zr(U)F and Li, ;Zr(U)F,
wos depressed so strongly that the minimum in fhe
1 was yerlmd by the LlF-phase field. The

vest | el’rl‘ng composmon olong this join started
u freezmg a'r 545°C and had the composmon NaF-
" LiF-ZrF -UF, (20-55-21-4 mole %), The viscosity

 
 
  

    

  
  
     
 
 
 
 
   
 
    

 

 

! mole %). The metastable eutectic at 425°C is  of a s:mllar mixture has been measured (cf.,
o more strongly evidenced at a slightly higher LiF  Sec. 7, ‘“Heat Transfer and Physical Properties’’).
concentration, & Contrary to expectations, the reduction of the

55

 

 
i i o i, . i it il

it

 

 

ANP PROJECT PROGRESS REPORT

'Zer content resulted in an increase in the
wviscosh‘y.

When fhls unexpecfed turn of events became

/,opparem‘ cffenhon was shifted to compositions

containing more ZrF,. The composition NaF-LiF-

- ZrF-UF, (32-35-29-4 mole %) melts at 445°C
~and resembies a eutectic containing Na,Zr(U)F
- “andLi Zr(U)F solid solutions and Na [Zr(U)] Fa,
- Most of the uranium is in the NG [Zr(U)] FSI
f'_phase. It is interesting to note fhc'r the melt
" corresponds stoichiometrically to an approximately

50 50 mixture of Nc ,Zr(U)F , and Li Zr(U)F

The quasi bmary befween Nc172r6|:31 and
Lizsz6 was not changed much by the inclusion

of 4 mole % UF,, The eutectic rose only about

2°C (to 457°C), with an almost imperceptible
shift in the dn‘echon of more Li ZrF The

_"’composmon"of this quasn-blnary eutechc is
- NaF-LiF-ZrF - UF4 (31-28-37-4 mole %). Since

this eufec‘nc appears to be a saddle point, a
quasi-ternary eutectic might be expected at some-
what lower ZrF , concentrations. A rather extensive
search has not revealed any lower temperatures
in this region. The important compositions
established in the phase studies described above
are summarized in Table 4.1.

Ber-Bearing. Systems
L. M. Bratcher R. J. Sheil

B. H. Clampitt R. E. Thoma
Materials Chemistry Division

T. N. McVay, Consultant
The Ternary System NaF-LiF-BeF,. A con-

siderable amount of thermal data has been obtained
by means of cooling curves with mixtures in the
NaF-LiF-BeF, system containing 10 to 50 mole %
BeF,. The LiF-Li,BeF ,-Na,BeF ,-NaF section
of the system has been investigated most thoroughly
because it is expected that melts containing less
than 33 mole % BeF, will have low viscosity,
high heat capacity, and other favorable physical
properties, The thermal analysis investigation of
the system has been hampered by undercooling,
most troublesome with melts containing more than
33 mole % BeF,, and by difficulty in obtaining
reproducible data. The filtration technique was
used to determine liquidus temperatures with two
compositions. Mixtures prepared for thermal
analysis were studied by x-ray diffraction and
petrographic techniques. Although some progress
was made in understanding the phase relationships

in this system, which are complicated by the

existence of ternary components, fhey are far
from being completely elucidated. |t appears that

TABLE 4.1. LOW-MELTING MIXTURES IN THE chF-I..iF-ZrF4 AND THE NuFaLiF-ZrF4=UF4.SYSTEMS

 

Composition {mole %) Melting Point

 

Nature of Mixture

 

NaF  LiF  Z+F, UF, (°C)

51 38 | .'IT 600 Ternary eutectic: Nc:F-Lif"‘-l\luszrl""7

53 35 8 4 590 Quasi-ternary evtectic: NaF-LiF-Na3Zr(U)F7

4 B 15 625 Saddle eutectic: LiF-Na ZrF, |

47 38 1 4 615 Saddle eutectic: LiFQNa3Zr(U)F7

22 .' 55 23 575 Ternary eutectic: LiF-[Na(Li)]SZrF.’-[Li(Na)]3ZrF7

25 50 25 590 Saddle eutectic: [Na(Li)]3ZrF7-[Li(Na)]ssz.,

200 55 21 4 545 segrego{es: LiF (primary)—solid sclufion_s'_'Na7[zr(U)]6F3]
32 35 | 29 4 445 Quasi-ternary eutectic(?): solid solufions_Na%[zr(U)]ég]
3 25 ¥ 430 Quasibinary eutectic(?), NagZr F -Li ZeF,

31 1 41 455 Saddle eutectic: Na,Zr F . -Li,ZrF,

31 28 37 4 457 Saddle eutectic(?): Na [Zr(U)|F 4 -Li ZrF

25 27 48 450(2) High ZrF ,; not well established

 

 

56

y )

 
E:

 

 

equilibrium data are difficult to obtain in this
system. |t is expected that the quenching and
differential thermal analysis techniques will be
applied to the study of this system in the near
future,

Thermal data for the LiF-Na,BeF, join are
shown in Fig. 4.4. Solid-phase studies of the
slowly cooled melts showed that only the two
components were present, The minimum-melting
mixture on the join has the composition NaF-LiF-
BeF, (56-16-28 mole %), and the melting point of
480 + 5°C is the lowest thus far observed with
BeF. mixtures containing significantly less than
33 mole % BeF,. Consequently, this mixture is
potentially important as a fuel carrier. The solid-
phase studies have also shown that mixtures
within the LiF-Na,BeF -NaF triangle contain
only these three components, Other compatibility
triangles that have been tentatively postulated on
the basis of incomplete, and sometimes conflicting,
data are LiF-Na,LiBe,F,-Li,BeF, and LiF-
Na,BeF ,-Na,LiBe,F,. The latter compound is

PERIOD ENDING JUNE 10, 1955

the only ternary compound that has been definitely
identified in the present studies, but other phases,
as yet unidentified, have been observed,

The cooling-curve data suggest the existence of
a ternary eutectic that melts at 325 £ 10°C. The
composition of this low-melting eutectic is not
definitely known, but it may be near the compo-
sition NaF-LiF-BeF, (27-35-38 mole %). If
confirmed, this would be one of the lowest melting
fluoride mixtures known that has favorable nuclear
properties and, possibly, other good physical
properties.

The Quaternary System NaF-LiF-BeF,-UF .
Only one composition in the NaF-LiF-BeF ,-UF
system was investigated. The addition of 2,5
mole % UF, to the LiF-Na,BeF , eutectic lowered
the apparent liguidus temperature, determined by

visual observation, to 465 £ 5°C. The liquidus

temperature will be checked by filtration because
of the observation of poorly reproducible thermal
effects on cooling curves at 570 £ 15°C. The
uranium in this mixture was combined as Na,UF

ORNL-LR-DWG 7383

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

850
»\ ,
750
650 \\
= | A
T 550 . )
& "\. &
% @
= [ e e e i [ e v ——-—..f—-——-*
450
350
o B

20 30 40

 

50 60 70 . 80 90 Na,BeF,

L Nagser; (méle %)

Fig. 4.4. Thermal Data for the System LiF-NazBeF4.

57

T T O T R T e

 
 

 

 

ANP PROJECT PROGRESS REPORT

and th_e: remaining compounds were so poorly
- crystallized that they were almost unidentifiable
in the preparations examined to date,

Solubility of UF, in BeF,-Bearing Compositions.
The solubility of UF, in BeF ,-bearing compositions
was determined with six different alkali fluoride—
~BeF, mixtures, some of them at different temper-
~atures. All the determinations were carried out
in nickel filtration apparatus; the melt was

 filtered fhrough nickel filters after an equilibration

period of 2 hr. Some solubility data obtained with
BeF ,-containing melts and reported earlier were
Qbfained in the same type of equipment.? The

 

4L. M. Bratcher ¢t al., ANP Quar. Prog. Rep. Mar. 10,
1955, ORNL-1864, p 51. |

poor reproducibility of the data, shown in Table 4.2,
indicates that equilibrium was not attained in
these experiments, probably because of alloying
of the uranium with the nickel container and filter
with consequent lack of control of the activity of
the uranium, _

The scattered data in Table 4.2 show a trend of
increasing UF, solubility with increasing temper-
ature and decreasing BeF_ content, but in no
case does it appear that the solubility of UF,
at 600°C is sufficiently high to provide more than
a fraction of the concentration needed for an ART
fuel. The data reported by Mound Laboratory?

 

5 Private communication, J. F. Eichelberger to W. R.
Grimes, Feb. 1955,

TABLE 4.2. SOLUBILITY OF UF, IN BeF ,-BEARING COMPOSITIONS

 

: ‘_ 'Com-po.s,irfion'(mo'l‘e %)

 

.Analysis of Filtrate (wt %) ‘

 

 

s Temperature Uranium
~ O NeF ' LiF T BeF, (°C) Added us? Total U
57 43 600 UF, 0.10 0.59
70 30 600 UF, +U 0.54 2.48
600 UF, +U 0.52 4.38
700 UF, + U 1.09 5.47
700 UF, + U 1.54 8.79
800 UF, + U 3.95 11.7
800 UF, +U 4.74 16.0
800 UF, +U 4.46 16.2
75 25 615 UF, + U 5.04* 16.8
700 UF, + U 6.32* 17.5
800 UF, + U 5.49 19.7
77 23 600 UF, + U 4.62 10.1
- 600 UF, +U 3.52 7.6
800 UF, +U 8.15 21.1
69 31 600 UF, + U 3.37 7.30
‘ 600 UF, +U 1.83 3.68
600 UF, + U 2.83 4.39
600 UF, + U 2.61 5.82
56 16 28 600 UF, + U 1.77 3.49
600 UF, + U 3.97 5.46
600 UF, +U 1.54 4.09
700 UF, +U 3.92 8.13
700 UF, + U 4.76 15.4
700 UF, +U 5.63 9.81
800 UF, + U 8.95 18.6
800 UF, + U 10.2 17.3

 

*Sample contained metal, result probably high.

58

 
 

for the solublh’ry of UF, in NaF-BeF, mixtures
were based upon the assumphon that all uranium
dissolved in the melt was in the trivalent form.
This assumption was erroneous, as the data in
Table 4.2 clearly demonstrate. The .data in
Table 4.2 are now supported by recent Mound
data. ®

The low concentration of uranium dissolved in
the NaF-BeF, (57-43 mole %) mixture at 600°C,
which is in line with results reported by Mound
Laboratory,® suggests that a large part of the
tetravalent uranium present in BeF2 melts to
which UF and uranium metal were added was due
to dlspropor'rlonahon rather than to oxidation by
impurities in the melt. The NaF- BeF, (77-23
mole % and 75-25 mole %) mixtures and also the
ternary mixture NaF-LiF-BeF, (56-16-28 mole %)
shown in Table 4.2 were not purified, and it is
quite possible that a part of the tetravalent
uranium present was due to oxidizing impurities,

Reaction of Uranium Metal with BeFé-Bearing
Compositions. Uranium metal was added to all
except one of the mixtures used for UF, solubility
determinations, but only a few expenmem‘s were
performed in which uranium metal was the only
form of uranium added. The analyses of material
filtered through nickel filters after equilibrating
20 g of a purified alkali fluoride—~BeF, mixture
with 4 g of uranium metal for 2 hr are shown in
Table 4.3, The data in Table 4.3 are insufficient
to determine whether the dissolved uranium was
due to oxidation of uranium metal by 'Be'Fz'or by
oxidizing imp'urifies in the melts, and further

experlmentctlon |s plonned to clar:fy thts pomt.

 

 

 

6Prlvcn‘e commumcoflon, J F Elchelberger fo W R
Grimes, April 19, 1955.

M. S. Grim, ANP Quar. Prog. Rep. Sept. 10 1954.
ORNL ]771, P 56

    
 

 

PERIOD ENDING JUNE 10, 1955

Visual Observation of Fluoride Melts
R. J. Sheil

Materials Chemistry Division

A plastic, inert-atmosphere box similar to one
previously described” was used for observation of
a number of fluoride melts. Some of the results
of these observations are given below,

All the alkali fluoride—BeF, mixtures prepared
to date showed a scum when first melted. The
unpurified mixtures were worse than the purified
mixtures in this respect, as might be expected.
The small amount of scum that appeared on melts
of purified mixtures was dissolved when the melt
was heated to about 750 or 800°C, and, in some
instances, a water-clear melt was obtained. When
the mixtures contained UF,, the scum had a
reddish-brown color suggestive of UO,, but this
scum likewise dissolved when the melt was heated
to higher temperatures.

It was found that a melted mixture having the
nominal composition NaF-LiF-ZrF -UF, (20-55-
21-4 mole %) was reddish-brown when only 15%
of the uranium was in the trivalent form. When
all the uvranium was in the tetravalent form, melts
of this composition had the normal green coloration.

Alkali fluotide~BeF, melts appear to have a
rather high surface tension, and they do not wet a

clean nickel surface. These effects could possibly -

be ascribed to small amounts of impurities in the
melts being concentrated in the liquid surface, but
this is not considered to be likely.

Phase Separation by Zone Melting

F. Kertesz H. J. Buttram
No V Smlth

Ma'retzlals Chemlstry Division

Apparatus has been assembled recently to

- determine whether zone melhng of fused salts can

 

AL;MIXTURES

 

 
  

F\2 Mixture

Temperature’

 
 
 
 

* Analysis of Fi ‘i’f‘r&ié:(w’t %)

 

 

us?

  

TotalU

  

 

 
 

NaF -BeF, (57-43 mole %)

   

089
0.0 | 0.85

 

*Sample contained metal, result probably high.

   

59

 

 

 
itk

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

be u'rirli'ze'd. as an aid in phase equilibrium studies.
The zone-melting method consists in alternately
melting and freezing the salt mixture at very slow

‘rates to allow use of the principle of fractional

recrystallization on o multistage basis. Some
preliminary trials with mixtures of LiF and UF,
indicated that, while some equipment modifications
may be required, the method holds promise of being
a useful supplement to present phase-study
procedures,

CHEMICAL REACTIONS IN MOLTEN SALTS
F. F. Blankenship L. G. Overholser

Materials Chemistry Division

Equilibrium Reduction of FeF, by H, in NaZrF,
C. M. Blood
Materials Chemistry Division

Previous measurements® of the partial pressure
of HF at equilibrium during reduction of FeF,
by hydrogen showed higher values than would be
predicted from thermochemical data and ideal
solution behavior. These high values were tento-
tively ascribed to a lowered activity of the
metallic iron because of alloying of this element
with the nickel apparatus. The postulated alloying
has been confirmed by low values for final FeF,
content of the melt and by chemical analyses of
portions of the nickel apparatus,

Attempts have been made to repeat these ex-
periments with a mild-steel liner in the nickel
apparatus, a steel gas entrance tube, and steel
screen baffles. Preliminary data have shown that
the presence of a stainless steel thermocouple
well extending into the melt has been causing a
steady increase in chromium content of the melt
with consequent loss of FeF, from the solution.
This unintended side reaction was taken into
account in the calculations and will be avoided in
the future,

The results again showed somewhat higher
HF values than those predicted. The apparent
equilibrium constant (mole fraction of dissolved

" species used as activity) was 5.7 for this ex-

periment; this is significantly lower than the
value of 8 obtained in the nickel equipment and is
much higher than the calculated value of 1.6 for

~ ideal behavior.

 

 

8C. M. Blood, ANP Quar. Prog. Rep. Mar. 10, 1955,
ORNL-1864, p 57.

6

Reduction of UF by Struct-u-r.al Metals
J. D. Redman C. F. Weaver
Materials Chemistry Division

The reduction of UF, by metallic chromium and
iron has been studied by methods previously
described in ANP quarterly progress reports, and
equilibrium data, obtained by using either NaF-ZrF
(50-50 mole %) or the NaF-KF-LiF eutectic
(11.5-42-46.5 mole %) as the reaction medium,
have been reported. More recently, the reduction
of UF, has been investigated with NaF-ZrF
(53-47 mole %) used as the solvent. Also, the
reaction of UF, with chromium metal and Inconel
has been studied by using mixtures of UF, and
UF, dissolved in NaF-KF-LiF, '

The results of some studies on the reduction of
UF , by chromium metal in NaF-ZrF, (53-47 mole %)
at 600 and 800°C are given in Table 4.4, In
these experiments, 2 g of hydrogen-fired chromium
was reacted with UF, (11.4 wt %) in 40 g of the
NaF-ZrF, mixture contained in nickel.  The
values given for K_ were calculated for the
reaction

2UF, + Cr"v—fZUF3 + CrF,

by wusing concentrations expressed in mole
fractions.

Previous studies with NaF-ZrF, (50-50 mole %)
as the solvent and 4.1 mole % UF , gave a chromium
concentration of 2250 ppm and K_ =4 x 10-4 at
600°C, and 2550 ppm of chromium and K, =5x 104
at 800°C. Thus, an increase from 0.48 to 0.51
in the final mole fraction of NaF in the melt
containing uranium causes a significant decrease
in the equilibrium CrF, concentration.

Data for the reaction of UF, with metallic iron
inNaF-ZrF , (53-47 mole %) are given in Table 4.5.
In these runs, 2 g of hydrogen-fired iron wire and
11.4 wt % UF , were used.

The iron concentrations given in Table 4.5 are
very nearly equal to those found for this reaction
when NaF-ZrF, (50-50 mole %) was used as the
solvent. They also agree very closely with those
obtained when NaF-KF-LiF (11.5-42-46.5 mole %)
was used as the solvent. The iron values were
somewhat higher at 600 than ot 800°C in all three
systems studied.

Data are presented in Table 4.6 for the reaction
of chromium metal with 13.3 wt % UF, plus 1.5
wt % UF, ot 600 and at 800°C when NoF-KF-LiF

 

 
 

i bk A S

i

 

e “taining 15 wi %_“UF

PERIOD ENDING JUNE 10, 1955

TABLE 4.4. EQUILIBRIUM DATA FOR THE REACTION OF UF, WITH CHROMIUM METAL IN
MOLTEN NaF.'sz4 (5347 mole %) AT 600 AND 800°C

 

 

 

 

Conditions of Equilibration Found in Filtrate
Temperature Time Total Uranium Total Chromium* Total Nickel Kx
(°C) (hr) (wt %) (ppm) (ppm)
600 3 8.4 1710 30
. 0
3 8.6 1740 2 1 x 10-4
5 8.3 1710 35
5 8.5 1770 30
800 3 8.5 2160 50
' 3 8.6 ' 2130 40 1 x 10-4
5 8.3 2180 80
5 8.6 2170 25

 

*Blank of 250 ppm of chromium at 800°C.

TABLE 4.5. EQUILIBRIUM DATA FOR THE REACTION OF UF, WITH METALLIC IRON
IN MOLTEN NaF-ZrF 4 (53-47 mole %) AT 600 AND 800°C

 

 

 

 

 
 
 
 
  
 
 
 
  
    

Conditions of Equilibration Found in Filtrate
Temperature ‘ . Time Total Uranium Total lron* Total Nickel Kx
(°C) {hr) {(wt %) {ppm) (ppm)
600 3 8.4 | 435 50 ™\
3 8.5 535 80 .
5 8.6 630 50 b
5 8.6 445 35 (" 2% 10
5 8.4 440 130
5 8.7 460 55 ./
800 455 45 7
o 1385 30
350 40 » 7 x 10~7
445 - 50

 

 

 

"’flval’ues wn‘h ’rhose glven in Table 4. 6 shows the
"effect of a relatively small amount of UF, on the
Mequnhbrlum chromium concentration. [t must be
recognlzed that the UF ; may disproportionate under
3 s stud;ed and that the final UF
showed eqml:bnum “Shromiom’ "“concentration moyfi‘be dlfferem‘ from ‘that calculated.
7 r »f*'I'IOQ und 2700 ppm. cn‘ 600 and Slmlior studles were mcde with lnconel in place
'800°C, “respectively, A ‘comparison of these of the chromium metal.  The results given in
: Table 4.7 were obtained with hydrogen-fired
9). D. Redman and C. F. Weaver, ANP Quar. Prog. inconel and 13.3 wt % UF , plus 1.5 wt % UF, in
Rep., Mar. 10, 1955, ORNL-1864, p 58. the NaF-KF-LiF eutectic,

 

  
 

 

  
 

 

 

61

   

TR

 
"ANP PROJECT PROGRESS REPORT

EQUILIBRIUM DATA FOR THE REACTION OF UF, PLUS UF ; WITH CHROMIUM METAL IN

TABLE 4.6.
L L - MOLTEN NaF-KF-LiF (11.5-42-46.5 mole %) AT 600 AND 800°C

 

Conditions of Equilibration Found in Filtrate

 

 

Total Nickel

 

 

 

Temperature Time Total Uranium Total Chromium

(°C) (hr) (wt %) (ppm) (ppm)

600 5 10.8 255 110
5 10.6 195 40
5 1.7 165 75
12 11.0 195 25
12 11.0 250 105
12 12.4 330 <1

800 5 10.7 1780 60
5 10.7 1740 270
5 10.9 2070 10
5 11.9 1870 35
5 1.1 1730 20
5 11.2 2030 55
5 11.0 1850 65
12 10.8 1920 10
12 10.8 1580 3
12 10.6 1740

<1

 

MOLTEN NaF.KF-LiF (11,5-42-46.5 mole %) AT 600 AND 800°C

TABLE 4.7. DATA FOR THE REACTION OF INCONEL WITH UF , PLUS UF, IN

 

Conditions of Equilibration

 

Found in Filtrate

 

 

Temperature Time Total Uranium Total Chromium Total Nickel Total lron
(°C) (hr) (wt %) (ppm) (ppm) {(ppm)}
600 5 10.6 25 130 105

5 10.56 30 260 115
12 10.8 35 110 80
12 10.7 25 200 50
3800 5 10.7 430 95 20
5 10.5 525 295 110
5 10.8 485 85 75
5 10.7 470 60 55
12 11.1 675 30 60
12 11.4 725 15 80

 

 

The values given in Table 4.7 for 600°C are
much lower than the 360 ppm of chromium reported
earlier? for a mixture containing 15 wt % UF

' f\_n'/i-tfi no UF ) originally present. At 800°C, a

value of 800 ppm for chromium with no UF,

originally present is to be compared with the

62

somewhat lower values given in Table 4.7. Since
the chromium concentration increased when the
heating period at 800°C was extended from 5 to
12 hr, it may be that equilibrium had not been
attained.

 

 
 

P

 

Stability of Chromium and Iron Fluorides
in Molten Fluorides '

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

A number of experiments were run to determine,
if possible, the extent to which some of the
structural metal fluorides are reduced by metallic
iron or chromium. In one experiment, 10.9
wt % Fe3* was added as FeF, to the NaF-KF-LiF
mixture contained in an iron charge bottle and
reacted at 800°C. The filtrate was found to
contain 14.7 wt % total Fe and 13.8 wf % Fe'*®.
Another run, at 400°C, gave 3.9 wt % Fe**? and
4.4 wt % total Fe, which indicated that the Fe3*
was almost completely reduced by the metallic
iron. Some results obtained for several other
combinations are given in Table 4.8. The results
given in Table 4.8 show that NiF, is quantitatively
reduced by metallic chromium and iron at 600 and
800°C in the NaF-KF-LiF mixture and that FeF,

is reduced by chromium under similar conditions.

PERIOD ENDING JUNE 10, 1953

Since previous studies'® had indicated that
FeF, was relatively stable in the NaF-KF-LiF
eutectic at 600 and 800°C, some additional runs
were made for which a longer heating period was
used. The results of these runs are given in
Table 4.9. The data showed that some 90% or
more of the iron was divalent after 12 hr of
heating. = This finding is in agreement with
findings of the earlier studies for which the
heating period was 3 to 5 hr.

The previous reports®:10 also presented dota
which indicated that CrF, is not stable in the
NaF-KF-LiF eutectic but probably disproportionates
according to the reaction

3CrF, == 2CrF, + Cr

The data obtained at 600°C gave a chromium
balance approximating that expected for the
reaction, but the values obtained at 800°C did

 

105, D. Redman and C. F. Weaver, ANP Quar, Prog.
Rep. Dec. 10, 1954, ORNL-1816, p 63.

TABLE 4.8. DATA FOR THE REDUCTION OF STRUCTURAL METAL FLUORIDES BY METALLIC IRON
OR CHROMIUM IN MOLTEN NaF-KF-LiF (11,5-42.46.5 mole %) AT 600 AND 800°C

Equilibration Time: 5 hr

 

Equilibration

Fluoride Added*  Metallic Reducing

Found in Filtrate
++

 

 

 

 

 

 

 

 

 

 

Temperature (wt %) Agent crtt Total Cr Fe Total Fe Total Ni
(°C) (wt %) (wt %) (wt %) (wt %) (ppm)
600 5Nt Fe 1.29 1.40 130

1.5 Ni Tt Fe 1.21 1.10 110

800 15Nt ~ Fe , 1.14 1.24 45

 Fe 1.23 1.03 25

€055 041 105

Cr 089 0.77 45

O os 123 245

a | 052 .._771..-;24, o

cr 003 0.32 © 0.009 25

G000 043 0.012 105

Cr 013 0.54 ' © 0.009 80

" Cr 034 w31 0.007 45

*Njckel added as Nin; iron added as Fer.

63

 
_ Cr++ by disproportionation,
. total chromium in the filtrates and in the residues

 

 

 

ANP PROJECT PROGRESS REPORT

~ not. VA_clid'i:‘rriénd runs were therefore made at
800°C, and the results are given in Table 4.10.
The values given in Table 4.10 are, in general,

" in agreement with the previous data, Most of the
" Cr** changes valence state, but the total chromium

. content found is too high to account for removal of
The similarity of

'7":'5'_Uggés'f§ that finely divided chromium passes the
- filter, but, if the case were this simple, the values

-'.Q'--répbr'te.'d for Crt* should include this metal and

. therefore be higher than those found. No satis-
~factory explanation for the behavior of CrF, under
these conditions can be given at present.

The Disproportionation Pressure of UF,

S. Langer
Materials Chemistry Division

Previous studies!! of the disproportionation of
UF, gave evidence that UF, was more stable at
elevated temperatures than free energy estimates!?
had indicated.  This evidence has been sub-
stantiated by wvapor pressure measurements on

 

Ty, c, Whitley and R, J. Sheil, ANP Quar. Prog. Rep.
Dec. 10, 1954, ORNL-1816, p 60.

12| | Brewer et al., Thermodynamic Properties and
Equilibria at High Temperatures of Uranium Halides,
Oxides, Nitrides, and Carbides, MDDC-1543 (Sept. 10,
1954, rev. Apr. 1, 1947).

TABLE 4.9. STABILITY OF FeF, IN MOLTEN NaF-KF-LiF (11.5-42-46.5 mole %) AT 600 AND 800°C

Equilibration Time: 12 hr

 

Eqfi‘ilibrutibr;

Found in Filtrate

 

 

   

 

- Temperature F:vrjf;id;d Fe't Total Fe Total Nj
. Q) ’ (wt %) (wt %) (ppm)
coe00 13.7 4.8 5.3 | 185
i 13.7 5.6 5.6 255
5.9 4.5 4.6 210
5.9 5.3 5.4 ' 235
13.7 121 13.2 90
13.7 12.0 13.1 60
13.7 12.0 | 12.6 465
13.7 12.5 12.7 1240
5.9 5.2 5.7 25
5.9 5.4 5.7 135
5.9 5.2 5.4 120

 

TABLE 4.10. INSTABILITY OF Cer IN MOLTEN NaF-KF-LiF (11.5-42-46.5 mole %) AT 800°.C

 

Chromium Found (wt %)

 

 

 

 

Time of Heating CrF2 Added In Filtrate In Residse
(hr) {(wt % Cr)
B crtt Total Cr crtt Total Cr
5 5.8 0.65 6.1 0.20 7.5
5.8 0.48 5.3 0.44 5.2
5.8 0.71 5.2 0.43 5.8
S , 5.8 0.57 5.2 |
1 . | 5.0 0.36 43 0.42 4.4
N R 5.0 0.66 4.1 0.44 4.3
: e 3 5.0 0.89 4.6 0.50 5.4

 

 

-

 

 
ks i Gl

 

ar

k]

- 50

UF ; in the temperature range from 1270 to 1390°C.
At these temperatures, the vapor pressure of UF
is negligible, while the vapor pressure of UF4
ranges from 61 to 152 mm Hg, as shown in Fig. 4.5.
Hence, the vapor pressure developed above o
system initially containing only UF; is due to the
UF, resulting from the disproportionation of UF,
according to the equation

4UF, = 3UF, + U°

The condensed phases in the system were
presumed to be solid UF; (mp, 1400°C), liquid
UF, (mp, 1035°C) saturated with UF,, and
uranium metal or an alloy of uranium metal with
the type 316 stainless steel container material,
The vapor pressures were measured by the
method of Rodebush and Dixon,'3 The apparatus
was essentially that described by Moore,'4 except
that the charge was contained in a tantalum liner

SEGRET
ORNL—LR-DWG 7384

~ TEMPERATURE (°C)
1350 1300 1250

1000

500

_ 200

    
   

VAPOR PRESSURE OF UF,

PRESSURE (mm Hg)
S
Q

   

DISPROPORTIONATION PRESSURE OF UFs
. . ’

  
  

   

an 4,5. The Vapor Pressure of UF4'und the

Disproportionation Pressure of UF,.

 

PERIOD ENDING JUNE 10, 1955

in order to minimize alloying of the uranium metal
resulting from the disproportionation of UF..
The low vapor pressure of the system limited the
measurements to temperatures above 1270°C, and
the failure of the container material (type 316
stainless steel) was the limiting factor at temper-
atures near 1400°C.

In all cases, UF‘3 of at least 95% purity was
used as the starting material. Analysis of the
condensate of the vapor phase, collected from the
tubes of the vapor pressure cell, showed that the
vapor phase was pure UF4. The disproportionation
pressures obtained are given in Table 4.11 and
are plotted on Fig. 4.5. The reproducibility of the
measurements was not noticeably influenced by
ascending or descending temperatures. By using
the method of least squares; it was found that a
straight line through the points was represented
by the equation

log P (mm Hg) = (-4187/T) + 3.945

or
—RT In P (mm Hg) = 19,160 ~ 18.05T .

The AH obtained from the slope of this line is
19,160 cal/mole. This AH, however, is a combi-
nation of the heat of vaporization of UF , the heat
of the disproportionation reaction, and the heats
involved in the formation of any solutions existing
in the condensed phases. Several attempts were
made to sample the equilibrium liquid in order to
determine its composition, but this proved to be
unfeasible because of the limitations of materials
at the temperatures involved.

The vapor pressure curve for UF, (Fig. 4.5)
was obtained from the data of Ryon and Twichell!5
(five points at low pressure were excluded as
obviously in error), data of R, E, Moore,'% and
points obtained during the course of this experi-
ment, which extended the temperature range from

]200 to 1300°C. The equation of the line is
logP = (- 8428/T) + 7.250

 

13W. H, Rodebush and A. L. Dixon, Phys. Rev. 26,

851 (1925).

]4R E. Moore, c. J. ‘Barton, and R. E. Traber, Vapor
Pressures of Zirconium Tetrafluorzde and Fused Mixtures

with Sodium Fluoride (to be published).

-19A. D. Ryon and L, P, Twichell, Vapor Pressure and

"‘:":Related ‘Physical Constants of Uramum Tetrafluoride,

H- 53852( uly 25, 1947).

16R, E. Moore, ANP Quar. Prog. Rep. Mar. 10, 1952,
ORNL-1227, p 155.

65

 

 

 

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 4.11. DlSPROPORTIONATION PRESSURES OF UF .

 

 

 

Temperature Temperature ! 4 Pressure
o o x 10
("C) (°K) T(°K) (mm Hg)

1269.8 '1543.0 6.481 | 18.00

1278.6 1551.8 6.444 - 18.95

1286.9 ' 1560.1 6.410 19.10

1287.1 1560.3 6.409 19.50

1290.3 1563.5 6.39 19.30

1304.0 1577.0 6.341 ' 17.50

1304.0 1577.0 6.341 16,007

1316.0 1589.0 6.293 270

1316.3 1589.5 6.291 2035

1317.5 1590.7 6.287 19.60

1325.0 1598.0 6.258 22.70 .
S 13390 1612.3 6.202 "ffzo 95

1342.0 1615.0 6.192 2045
13423 1615.5 6.190 215
 1345.8 1619.0 6.177 120.70

1351.0 1624.0 6.158 | 23.05
13547 1627.9 6.143 22.85

1370.0 1643.0 6.086 23.95

1370.7 1643.9 6.083 - 23.40

1383.6 1656.8 6.036 31.60

1389.0 1662.0 6.017 27.65

1391.0 1664.0 6.010 32.80

 

or
—-RT InP = 38,600 - 33.18T .

The heat of vaporization of pure UF4, as obtained
from the slope of the line, is 38,600 cal/mole.
This value is in fair agreement with the value
of 42,000 cal/mole obtained by R, E. Moore.

It was not surprising to find that the vapor
pressure from UF, was only a small fraction of
that from pure UF ,, but it had not been anticipated
that the fraction would rapidly become increasingly
smaller with increasing temperatures. Indeed, at
sufficiently elevated temperatures, this would not
be the case. The effect of dissolved UF3 is
undoubtedly important; it appears that the solu-
bility of UF, increases markedly with temper-
ature and clso that the UF, is complexed to the
" extent fhof pronounced neganve deviations from
Raoult’s law are observed. There was no definite
“evidence of alloying of uranium metal with the
tantalum liner; no x-ray diffraction data on uranium-
tantalum alloys are available.

66

An extrapolation of the vapor pressure curves
in Fig., 4.5 to the point of intersection might be
expected to provide a rough estimate of the
evtectic temperature of the UF,.UF, system.
The intersection occurs at about 1010°C, whereas
cooling curves have shown thermal arrests, which
were presumed to correspond to the UF,-UF,
eutectic, at 870°C. This discrepancy is probably
due not only to the difficulty of obtaining the
vapor pressure measurements but also to the long
range through which the extrapolation must be
made and the change in AH with temperature
through this range.

The small slope found for the disproportionation
pressure curve means that the disproportionation
of UF, is far from complete under the conditions
prevmhng in the vapor pressure cell and that UF,
is thermodynamically stable with respect to pure
UF4 and uranium metal at temperatures below

1400°C.

 

 

 

 
 

o}

#»

.

 

Reduction of UF , with Uranium in
Alkali Fluorides

C. J. Barton B. H. Clampitt
Materials Chemistry Division

The investigation of the variables affecting the
reduction with metallic uranium of UF, dissolved

“in alkali fluorides was continued ‘during the past

quarter. The technique used in these experiments,
which are carried out in a small, nickel, filtration
apparatus, was described previously.'7 It has
become apparent that alloying of nickel and uranium
can occur at temperatures far below the minimum
nickel-uranium liquidus temperature (732°C) and,
consequently, tha'r dlsproporhonanon of UF, can
be expected to occur at the temperatures of
interest at nickel surfaces or at any metallic
surface with which uranium can alloy, Some effort
was therefore devoted to finding container ma-
terials that will alloy less readily with uranium.
Copper, which has a minimum liquidus temperature
of 950°C with uranium, showed some promise in
this regard. Although the results obtained in
nickel apparatus, and possibly also those obtained
in copper containers, are not to be regarded as
equilibrium data, these experiments did give some
indication of the effect of some of the variables
that are involved in the reaction and thus are of
value., In some experiments, pure UF, was
introduced directly for comparison with results
obtained in reduction experiments. The results
are reported below under the headings of the
variables investigated.

 

, Rep Mar. 10, 1955, ORNL- 1864 p5

17R. 3. Sheil and B. H. Clampltf ANP Quar. Prog.

  
 

PERIOD ENDING JUNE 10, 1955

Effect of Nickel Surface Area, [t was previously
reported that the amount of nickel surface area
exposed to alkali fluoride~UF, melts had a
marked effect on the degree of reduction ob-
tained.'7 This observation is substantiated by

the data shown in Table 4.12.

In these experiments, the normal surface-to-
volume ratio of 2.55 (square centimeters of
surface exposed to melt divided by volume of
melt in cubic centimeters) was decreased by
using a larger charge (40 g) and increased by
adding V-in. nickel balls or nickel powder. The
nickel filter medium surface areq, to which the
melts were exposed for a short time, was neglected
in making these calculations, but it was the same
in all experiments. The nickel surfaces were
exposed to hydrogen at 750°C before the charge
materials were added, and partial sintering of the
nickel powder undoubtedly reduced the surface
area to a smaller value than that calculated. The
marked effect of nickel surface area upon the
degree of reduction of uranium, as shown by the
data in Table 4.12, was probably due to the
previously mentioned effect of alloying of uranium
with the nickel.

Effect of Copper Surface Area, In view of the
results obtained with nickel, it seemed desirable
to determine the effect of another metal on the
reduction of UF, with uranium, Finely divided
copper powder that was hydrogen-fired at 750°C
was added to an LiF-NaF-UF ,.U° mixture (57.6-
38.4-4.0 mole % + 20 wt % excess U°) in sufficient
amount to give a surface-to-volume ratio of about
]000 if the surche area of the mckei contcnner

 

-TO-VOLUME RATIO")ON REACTION OF UF WITH

LKALI FLVUORIDES IN NICKEL CONTAINERS |

 

 

" Ratio of Surface

 

Ratio of ust

emperature

  

   

to Velume

 

 

 

Expéfim'ent.é” o (°C) -1 to Total Uranium
. NaF-LiF (4060 mole %)~ 27 Trrse T 1.89 7 07910 0.82
T e el e Uy '_‘4’ : 7 E0 ) ’ 2.55 . 0.65 to 0.70
2 7500 474 0.43t0 0.51
B 3 £ AT70007 7 0.05 to 0.06
L NaF KF L.F eufecnc' o 5 550 - 2.55 0.55 to 0.65
i1 42'&6‘5’ mole %) Ty S50 T TRM10000T T T 70,04 10 0,05
4 650 2.55 0.55 to 0,65
2 650 ~ 1000 0.04 to 0.05

 

67

 

 

 
ANP PROJECT PROGRESS REPORT

i

was neglected. The filtrate obtained after the
mixture was heated for 2 hr at 750°C contained
19.2 wt % U3* and 26.0 wt % total uranium. These
results compared favorably with those obtained in
the absence of the copper and encouraged an
investigation of copper as a container material
for UF j-containing melts,

.Effect of Excess Uranium, It was reported
earlier'” that there was liftle difference in the
degree of reduction of NaF-KF-LiF- -UF, with
uranivm metal whether the theoretical amount of
uranium requ:red by the equation

AT 3/UF4v—UF

§ was used or whether 1.2 times the theoretical
: amount  was employed It therefore seemed
desuoble to mveshgate this variable with NaF-LiF
(40-60 mole %) as the solvent instead of NaF-KF-
‘LIF ‘This investigation was carried out in nickel
~ apparatus, including nickel filters, and in copper
apparatus equipped with a bronze filter medium,

 

 

The results are given in Tables 4.13 and 4.14.
The theoretical total uranium in the filtrate in «all
these experiments, assuming compiete' reaction of
the UF4 with the uranium metal according to the
above equation, was 27,1 wt %.

The data in Tables 4,13 and 4.14 are somewhat
erratic, but it appears that little is to be gained
by using more than 100% excess uranium (two
times the theoretical amount) under the conditions
of these experiments, Comparison of the results
in the two tables indicates thcn‘ UF, dissolved
in NaF-LiF is more stable in copper thcm in
nickel.

Stability of Binary Alkali Fiuorlde—UF Mixtures,
The preparation of filtered LiF- UF samples
containing 91 to 96% of the uranium in the trivalent
form was reported earlier, '8 but no data on filtered
NaF-UF, samples have been reported. Therefore
a few experiments were performed to determine the

 

18C. M. Blood et al.,, ANP Quar. Prog. Rep. Sept. 10,
1954, ORNL-1771, p 77.

TABLE 4.13. EFFECT OF EXCESS URANIUM ON REDUCTION OF
Nc:F-LiF-UF4 MIXTURES IN NICKEL APPARATUS

 

 

 

 

Amount of. Excess Uranium Found in Filtrate (wt %) R of U3 1o Total U
(%) u3* Total U
Theoretical 18.6 28.4 0.65
17 ’ 18.5 27.2 0.69
80 20.3 25.2 0.79
140 19.7 26.4 0.75
200 21.0 26.3 0.80
500 20.3

26.3 0.77

 

" TABLE 4.14. EFFECT OF EXCESS URANIUM ON REDUCTION OF
NaF-LiF-UF4 MIXTURES IN COPPER APPARATUS

 

Amount of Excess Uranium

Found in Filtrate (wt %)

 

 

Ratio of U3t to Total U

 

(%) y3+ Total U
20 19.1 27.2 0.71
100 24.1 27.1 0.89
| _ 100 23.8 27.8 0.86
1 200 25.2 27.3 0.92

 

 

 

 

68

 

 

 

 

 
 

o

[

stability of UF, in these systems under the same
conditions as those used for more complex systems
except that higher temperatures were necessitated
by the higher liquidus temperatures. The data
are given in Table 4.15. Since the data do not
show any significant differences between the two
systems in degree of reduction and since UF
known to form a stable complex with NaF in the
solid state but not with LiF, it appears that
complex formation has little influence on the
stability of fused cchIi—UFg- mixtures.

Effect of Filter Medium on Stability of UF,
M. B. Panish

Materials Chemistry Division

Previous studies!? on the stability of UF, in
the NaFE-KF-LiF eutectic suggested that the
sintered nickel filters used for the filtrations
probably had an effect on the degree of reduction
found in the filtrates. Since it appeared likely

that the UF3 concentration found in the filtrate -

did not correspond to that present during the
equilibration because the UF, was reduced by
rapid removal of uranium from the melt through
alloying with the nickel filter during filtration,
some experiments were performed to confirm or
disprove this possibility,  These experiments
included runs in which a sintered nickel filter was
used and others in which a perforated nickel
stick was used in place of the filter. The use
of container materials other than nickel has been
investigated, as well as filtration through graphite.

The runs were made by using NaF-KF-LiF
{11.5-42-46.5 mole %) as the solvent and 7 to 14%
l.:'rcmium'pre':'seniL as fluoride. The uranium metal

qdded was 500 to 1000% in excess of thcn‘ reqmred“

 

19R J. Shell and B. H. Clampifi ANP Quar Prog.

. Rep. Mar 10 1955, ORNL 1864 p53

TABLE 4.15.

STABILITY OF UF

PERIOD ENDING JUNE 10, 19535

to reduce all the uranium fluoride to UF, if it
were assumed that all the uranium was present as
UF,. This large excess was used to ensure that
some wuranium metal was present at unit activity
even though a large amount of the uranium present
probably alloyed with the walls of the nickel
container, Data were obtained only from the runs
in which unchanged uranium metal was visible in
the heel

The reaction-filtration apparatus was modlf:ed
to permit rapid interchange of containers of differ-
ent types and materials. Sintered nickel of 0.0015-
in. pore size was used as the filter in all cases
except where otherwise indicated. A nickel stick
with four 1,:’32-in. holes hear the bottom was used
to draw up portions of the melts that were not
filtered, The results given in Table 4.16 were
obtained by using an equilibration period of 2 to
4 hr, except as otherwise indicated.

If it is assumed that there is very little or no
uranium metal present in the samples obtained
in the perforated nickel stick, it may be deduced
from the results in Table 4.16 that the dispro-
portionation of the UF, occurs, at least in a
major part, because of depletion of the uranium
from the melt by the sintered nickel filter. kI
should be noted that the total uranium found was
lower by 1 to 1.5 wt % (absolute) in those cases
in which the sintered nickel filters were used,
whereas the total uranium was, in general, the
same (within experimental error) as that expected
in the cases in which the perforated nickel sticks
were used. This loss of 1 to 1.5% of the uranium

~accounts for the difference in the degree of

reduction observed in the melts. Analyses of.

. .several sintered nickel filters gave results which
showed that there was more uranium in the filters

than would be expecfed if ’rhey were simply

lN L|F UF AND NaF- UF MIXTURES

 

 

 

 

S o 3+
Ne | .Nu?nberrrof o Temperature Raho of U Uranium Added as
Compo:;:hon Experiments . C) to Total Uranium
JLi‘F.'U"Fé”'(?s;z?' mole % 1 | ""8'5"0’" 075 UF,
TR A 5 850 0791082  UF, + U° (20% excess)
NaF-UF, (71-29 mole %) 2 820 "0.61100.67  UF,
4 820 0.72 to 0.82 UF, + U® (20% excess)

 

69

 
 

 

 

 

 

8.8

ANP PROJECT PROGRESS REPORT N .
" "TABLE 4.16. EFFECT OF SINTERED NICKEL FILTERS ON THE DISPROPORTIONATION o
s OF UF, IN NeF-KF-LiF EUTECTIC '
; & & - : .', o S Equilibrqtion Tlme: 2 to 4 hr, excepf as mdxcatéd
' ' o R ' Filter Pore Si;é: 0.0015 in., except as indicated
U : ' : Calculated Total Uranium Totd! Ana.lyzed Uranium 'Rqfio of U3.+to
Temperature
' ©C) Uranium Added as in Charge Materidl After Equilibration Total Analyzed
o (wt %)¢ (wt %) Uranium
Filtered Samples 7
600 UF, 9.2 0.68
; - UF, 13.4 11.4 0.66
i | UF, 13.4 11.8 0.70
- 700 UF, 7.6 " 0.49°
el : UF3 9.9 0.73
UF, 10.1 0.67
- UF, 13.4 12.1 0.70
UF, 13.4 12.0 0.74
UF, 13.4 11.9 0.75°
- 800 ur, 8.8 0.61
UF, 10.2 0.60
UF, 13.4 12.2 0.74
UF, 13.4 12.4 0.62
Unfiltrered Samples
550 UF, 8.3 0.96
600 UF, 7.6 0.99
' UF, 9.1 0.98
UF, 13.4 13.0 1.08°
700 UF, 9.0 0.98
' UF, 13.4 13.2 1.01
UF, 13.4 12.8 0.99%
UF, 13.4 13.2 1.03¢
UF, 13.4 12.1 1.08/
800 UF 7.1 0.97
UF

1.10¢

 

 

2The value given for the calculated total uronium is that expected for total reduction of all UF'4 added. No calcu-
lated vclues are gwen for runs starting wnh UF3, because erratic results were obtained in analyses of the charge
; N muferlals.
4 o Flh‘er used had a pore size of 0.0004 in,

_.CSmall amounts of metallic nickel were observed.

_ Equnllbrohon time 15 min.
. Eqmllbrchon time 30 min,
quuilibrcfion time 60 min.

70

 

 
 

&

-

Y

.

saturated with the melt and that this could account
for 25 to 50% of the uranium lost.

Several runs were made with the NaF-KF-LiF
eutectic plus UF, in gold, platinum, and mo-
lybdenum containers, No uranium metal was
added, and perforated nickel sticks were used for
obtaining samples. The ratio of U3* to total
vranium found at 650°C varied from 0.42 to 0.67,
and evidence of alloying of uranium with gold and
platinum was observed.

Three runs in which the NaF-KF-LiF mixture,
UF ,, excess uranium metal, and a graphite filter
were used were made in the temperature range
600 to 800°C. The degree of reduction varied
from 0.41 to 0.63, and indicated that graphite is
not a suitable medium for filtering melts con-
taining U3*. Analyses of the graphite filters gave
uranium values that accounted for 3% of the missing
uranium at 600°C and 75% at 800°C. Approximately
four times as much potassium was found in the
filters as existed in the NaF-KF-LiF mixture, and
thus it is thought that the potassium may have
reacted with the graphite under these conditions
to give a carbide.

Stability ofinFz-Ber Melts in Inconel
C. J. Barton R. J. Sheil

Materials Chemistry Division

The desirability of determining the stability of
PbF ,-BeF, melts in Inconel was pointed out in
the prevuous progress report.2? To determine the
stability, mixtures containing 50 and 75 mole %
BeF ., were sealed in Inconel capsules and heated
for about 10 hr at 800°C in a helium atmosphere.
When the capsules were opened and the contents
melted out in a helium atmosphere, bright globules,
presumed to be metallic lead, were observed on
the capsule walls, It was concluded that Pb**
ions are not complexed in these mixtures strongly

_enough to prevent reaction with Inconel at 800°C,
“and no further 1est|ng was performed with mixtures
- in thls system.

Solublhiy of Meiuls in Molfen Sults .

- 'R E. Cleary”
- Pra‘rt & Whltney A;rcruft

 

Experimem‘s are under way for determlnlng the
o solublln‘y of chromlum, nlckel and iron in molten

 

20, M. Bratcher et al., ANP Quar. Prog. Rep. Mar.

10, 1955, ORNL-1864, p 52.

PERIOD ENDING JUNE 10, 1955

salts, The systems being investigated consist
of two phases that are thoroughly mixed to ensure
intimate contact. The metais to be studied (Cr,
Ni, Fe) can be alloyed with uranium to form
eutectics with melting points below 900°C, and
therefore at experimental temperatures over 900°C
the alloy may be used as one phase of a two-phase
system, The salt phase of the system must
consist of material not reduced by uranium metal,
and,  consequently, for the initial experiments,
the salt phase is limited to alkali-metal halides
or combinations of alkali-metal halides.

It was expected from engineering data obtained
from thermal-convection loops that the vuranium
alloys could be formed by heating the metals and
the salts in a graphite container under an inert
atmosphere or under vacuum. Consequently, an
apparatus was constructed with readily available
material to carry out the primary part of the experi-
ment. Contrary to expectations, when the experi-
ment was carried out under vacuum with chromium
as the added metal, there was catastrophic
decomposition of the graphite container and
distillation of alkali metals through the vacuum
system to the glass-wool trap. A more thorough
search for information concerning the reactions of
graphite and alkali metals then indicated that
potassium reacts with graphite and causes
decomposition of the crystal structure. The salt
used, NaF<KF-LiF, became green and seeped
through the graphite crucible. There was no
evidence of alloy formation between the uranium
and the chromium., The metal fragments present
had retained their initial shaope and size, even
though the reaction vessel was heated to 1000°C.

In subsequent experiments, ¢ helium atmosphere
was also found to be unsatisfactory for this
reaction. Additional attempts to make a uranium-
chromium alloy in graophite liners in a nickel pot,
under helium or vacuum (50 to 75 p) at a temper-

“ature of 1150°C, have proved unsuccessful. Con-

sequently, it is now planned to use molybdenum

‘and/or tantalum crucible liners in a much smaller

stainless steel reactor. |t is hoped, based on

“prévious work with molybdenum, that these liners

will be inert to the reactants. The smaller volumes
and mass will make it possible to heat the
apparatus to temperatures of over 1200°C in
available equipment,

71

 
 

 

 

ANP PROJECT PROGRE’SS REPORT

PRODUCTION OF PURIFIED FLUORIDES

G J Nessle F. F. Blankenship
oo L. G. Overholser

Materlols Chemlsfry Division
Fuel Punflcohon ond Preparotlon Reseorch

C. M Blood F. L. Daley
- F. P, Boody F. W. Miles
MctermIs Chem!sfry Division

lnvesflgaflons of methods for rapid purlflccn‘lon
of fuel mixtures have included attempts to use
Ve'iectfo]y‘é'i's' “under a “hydrogen atmosphere ‘o
remove oxides in order to avoid the container
corrosion that resuits when HF is used and to use
metallic zirconium fo replace most of the hydrogen
in the stripping 6peration-. {n addition, attempts
have been made to prepare mixtures containing
very h|gh UF /UF4 ratios.

Use of ercomum Mefcl in Fuel Purification,

'The flme-consumlng step in fuel purification -

'processes is the reduction of such materials as
FeF, and NiF2 with hydrogen. Reduction of these
materials by the addition of an active metal should
afford a more rapid process, but it is not possible
to add an excess of active metal to a UF ,-bearing
mixture without the formation of large and variable
quantities of UF,. A process has been demon-
strated on a 5- and 50-1b test scale, however, that
is quite rapid and effective if small UF, concen-
trations are tolerable or desirable,

To prepare NoF-ZrF4-UF4, the NaF and ZrF ,
in proper quantity, are charged to the standard
apparatus and are hydrofluorinated in the usual
manner, At this stage, o considerable excess of
metallic zirconium chips in a nickel basket is
introduced into the reaction vessef, After 3 hr
of stirring with hydrogen, the basket and excess
zirconium metal are removed, The proper quantity
of UF, is then added to the melt, and another
nickel basket containing a small, known quantity
of zirconium metal is introduced. Stirring for 2 hr
suffices for the melt to react with all the zirconium
introduced.

In a 50-lb test preparation of NaF- ZrF -UF,
(50-46-4 mole %), the final mixture confamed ]5%
of the uranium as UF,, with the Fe, Cr, and Ni
contents being 50, 35, and 25 ppm, respectively.
it appears that utilization of some variant of this
procedure might cut the processing time on the
250-Ib production unit to below the 24 hr required,
at present, for each bateh.

72

Electrolytic Purification of Fluoride Mixtures.
Experiments for determining the rate of oxide
removal by hydrofluorination were made during the
preparation of some LiF»Z'rF4 mixtures. It was
found that hydrofluorination of a mixture for 1 hr
at 700°C would not lower the oxide concentration
below the limit of petrographic detection; similar
treatment at 800°C for 2 hr, however, was suf-
ficient. Since the dissolution of NiF, by the melt
and the length of the subsequenf smppmg time
are directly dependent on the hydrofluorination

time and the temperature, attempts were made to

remove oxides completely by electrolysis between
graphite anodes and nickel cathodes under a
hydrogen atmosphere.  Previous experiments?!
had shown that oxides could be removed by
electrolysis and that the anode products were
CO and CO, when a graphite anode was used.
When hydrogen was used in conjunction with
electrolysis in recent experiments, the predominant
anode product was HF. The purpose of the experi-
ment was to test the combined effect of oxide

removal as a primary anode product and hydro-

fluorination by the electrolytically produced HF,

After 0.4 equivalent of HF had been produced by
electrolysis at a current efficiency of 37%, the
HF was stripped from the melt with hydrogen.
No oxide was detected by petrographic examination
of the transferred product, Chemical analyses
showed 1.1 wt % oxide in the sample before
electrolysis as compared with 0.4 wt % in the
finished product. The structural metal contami-
nation was simultaneously reduced to the following
valves: Fe, 55 ppm; Cr, 15 ppm; Ni, 30 ppm.

The earlier experiments with electrolysis?}
had been carried out either in the NaF-KF-LiF

eutectic or in mixtures containing NaF and 40 to

50% ZrF,. The alkali fluoride eutectic was very

prone to an anode effect, while the ZrF ,-bearing
mixtures could be electrolyzed smoothly, Batches
of NaF-LiF eutectic proved to be variable in this
respect, and, in one case that gave a bad anode
effect, the addition of 0.1 mole % ZrF, gave no
significant improvement, On the other hand, an
NaF-ZrF, (81-19 mole %) batch could be electro-
lyzed smoofhly, the transition range must lie

below 19 mole % ZrF4

Mixtures with High UF3/UF4 Ratios., Attempts

have been made to prepare fuel mixtures con-

 

21C. M. Blood et al, ANP Quar. Prog. Rep. Dec. 10,
1954, ORNL-1816, p 66. :

v

 

 

 
 

AF

”»

taining UF, and no UF, for experimental work.
Earlier ottempfs indicated that such mixtures
could be prepared with NaF-ZrF , mixtures as the
base, but not with the NaF-LiF-KF eutectic. In
order to obtain solubilities of the order of 3 mole %
uranium at 600°C, it was found that compositions
containing more than 50 mole % ZrF, would be
necessary, There was a possibility that UF,
was capable of furnishing fluoride ions to complex
the ZrF, so that the physical properties, particu-
larly vapor pressure, might be far different from
those of fuels in current use. Accordingly, a trial
was made with a composition that could be con-
sidered, from a stoichiometric point of view, to be
composed of NaZrF, and U(ZrF,),. The compo-
sition was NaF-ZrF ,-UF, (44-53-3 mole %). After
a thorough purification of the NaF-ZrF, base, the
UF, was added, along with uranium metal (1 wt %
of the amount of UF.) and zirconium metal (0.2 wt %
of the amount of ZrF,), and equilibrated under
hydrogen for 2 hr at 700°C. The uranium and
zirconium were included in an effort to suppress
the formation of UF,. After cooling to 600°C,
the 4-kg melt was filtered; 75% of the uranium had
remained as U3*, There was no insoluble heel.

The preparation was then repeated on a 50-Ib
batch; only 65% of the uranium was U3*,  This
was thought to be a consequence of the previous
history of the preparation vessels for the two
batches. The walls of the 4-kg reaction vessel
had been subjected to melts which deposited
uranium in the form of a nickel alloy, while no
preparations involving UF, had ever been made in
the 50-1b apparatus.

Another trial was made in the small apparatus
by using essen’rmlly the same procedure, except
that a lorge excess of ZIl‘COnIUm in the form of

_' 1/ -in, bars was cllowed to remcm in the melfk‘
_-;;'_-_-_Ci.,:‘}‘fildurlng the equullbranon under hydrogen foHowmg
i ~-1-fhe odditlon of UF '

These prepcranons |||usfroted‘

""This time, 85% of the

    
 
 
  

 

 
 

the questlon of what fmchon of UF3 would

"'*,-*_‘,:;normally survwe ina reacfton vessel which

o 'f’had no prevnous hlsfory of exposure fo UF3 or
©* uranium metal cmd 'ro whlch no uranium or zurcomum_

L had been added.

 

225, M. Watson and C. M. Blood, ANP Quar. Prog.

Rep. June 10, 1954, ORNL-1729, p 51-52.

l:JF dur:ng »prroces‘s:}i‘g of a UF3 fuel and raised

PERIOD ENDING JUNE 10, 1955

An experiment to answer this question showed
that 70% of the uranium remained as U3* in 3 kg
of NaF-ZrF ,-UF, (52-46-2 mole %) after 3 hr of
equilibration with hydrogen bubbling at 700°C.
In the 50-1b apparatus, an attempt to prepare a
composition containing 1.2 mole % UF, and 0.8
mole % UF , in NaZrF by starting with 1.7 mole %
UF, and 0.3 mole % UF, was unsuccessful, since
the resulting product analyzed only 0.5 mole %
UF,. Resampling and reanalysis confirmed the
validity of the initial analysis.

Pilot-Scale Purification Operations

J. P, Blak-ely F. A. Doss
C. R. Croft J. Truitt

Materials Chemistry Division

Fifty-six preparations totaling 630 Ib of material
were produced during the quarter. These prepa-
rations were of various compositions and wetre
consigned to various requesters for small-scale
corrosion studies, for physical property de-
terminations, or, in many cases, for use as purified
intermediates for phase-equilibrium investigations.

Uranium trifluoride was o component of nearly
25% of the materials prepared. These batches
were prepared, in every case, by complete purifi-
cation of the solvent fluorides with the desired
quantity of UF, followed by addition and disso-
lution of the desired quantity of pure UF .

The study of optimum conditions for the prepa-
ration of BeF -bearing melts was continued
without complete success being attained. Hydrogen
reduction of the nickel and iron compounds to
acceptable concentration levels appears to be
more difficult in the BeF,-bearing melts than in

- the ZrF ,-bearing mixtures. When the HF concen-

tration in the hydrogen from a ZtF -bearing fuel

reaches 1 x 10~4 mole per liter, the Fe™¥ + Ni**

concentration is less than 150 ppm (by analysis

~of the filtrate). As the data shown in Table 4.17
~-indicate, this is not the case for the BeF ,-bearing
. fuels even at HF levels of 10~% mole per liter.
. Further study of this system will be attempted as

time permits.

. Production-Scale Operations

J. P. Blakely J. E. Eorgan
* Materials Chemistry Division

Production operations were resumed on March 1,
1955, on a three-shift, five-day-week basis to

73

+

 
 

 

ANP PROJECT PROGRESS REPORT

TABLE 4.17. IMPURITIES IN NaoF -BeF 2-UF, MIXTURES AS A FUNCTION
OF HF CONCENTRATION OF EXIT HYDROGEN

 

Concentration of HF in H,

‘Concentration of Impurities (ppm)

 

 

{mole/liter) Fe Cr Ni ' 5
1 x 10~4 235 12 1 19
- 235 50 120 h 55
155 55 150 45
140 15 65 59
180 40 90 ' 300
140 25 200 ' ns
125 10 85 v
1% 1073 55 17 100 57
65 14 85 77
125 30 10 7
195 37 95 6
130 25 _ 70 44
140 15 - 50 ' 315
90 17 80 -
75 25 125
60

20 15 <5

 

provide test material for the greatly accelerated
ANP engineering program. A total of 4800 Ib of
purified material was prepared during the quarter.
Several small, but significant, changes in
operating techniques have contributed to better
operation, shorter processing times, and, conse-
quently, higher yields of material per man hour.
The three most important changes were the
institution of rigid specifications on the purity of
the raw materials, the use of a 30% larger flow
rate of hydrogen during the final stripping cycle,
and the use of poralle! rather than series flow of
the hydrogen through the receiver and reaction
vessel, It appears that these changes have
reduced the processing time by nearly 50%. With
new or recently cleaned equipment, 250-Ib batches
can be processed to meet specifications in 24 hr.
[t has definitely been established that, by using
two men per shift, 1000 Ib can be produced per

 five-day week; it is believed that this production

figure can be improved slightly in the near future.
Present estimates of ORNL and Pratt & Whitney

requirements during the next several months call
~ for production of 1500 Ib per week. It is believed
. that two men per shift on a six-day work week

could meet this production level. Such an ex-
pansion of effort will necessitate an increase in

74

processing and storage containers, dnd, since
little excess capacity will be available for
emergencies, an increase in spare furnaces and
other maintenance items will be required to mini-
mize downtime for repair.

The life of the processing equipment has
already considerably exceeded expectations. A
number of the reactors have prepared more than
25 charges, and only one reactor has ever failed.
Tests are now under way to determine how many
runs can be made with the *‘specification’ raw
materials before excessive stripping times require
cleaning of the reactor,

In spite of the shorter processing times described
above, product purity is being rigorously main-
tained. Analyses of ten typical 250-Ib batches
of NcF-ZrF4-UF4 mixtures are shown in Table 4.18.

The present supply of ZrF, is not adequate to
meet the accelerated demand for the molten salts,
Commercially available (NoF)erF4, containing
about 30 mole % ZrF4, is being used at present,
but, since this must be blended with 40% ZrF, to
produce ART fuel compositions, nedrly 600 ib
of ZrF, is still required per week for the 1500-ib-
pereweek demand. Also, in making the hafnium-
free fuel for the high-temperature critical experi-

 

 

 

 

 

 

 

 

 
 

a

0

 

TABLE 4.18. ANALYSES OF TYPICAL NoF.Z¢F 4-UF4
MIXTURES FROM PRODUCTION PLANT

 

 

Batch No. Uranium Contfent Impurities (ppm}

i (wt %) Ni Cr Fe
514 8.66 25 85 15
515 8.42 10 50 35
516 8.69 13 28 35
517 8.75 10 43 10
518 8.69 10 48 20
519 8.64 15 34 <1
520 8.51 7 40 35
521 8.83 5 40 105
522 8.28 10 55 45
523 8.70 10 75 50

 

ment, the use of the commercial (NaF) ZrF, will
not be permissible.

Since 250 Ib per week appears to be the limit of
the present capacity of the Y-12 Plant for the
production of ZrF ,, alternative sources are being
urgently sought. A commercial source which
claims to be able to supply the required quantity
has been located, and specification tests of their
material are in progress. |t appears, at present,
that this material may solve the problem; if so,
the Y-12 ZrF -production facility will be used
only part time as needed to provide hafnium-free
material.

If the commercial source does not materialize,
it will be necessary to expand the Y-12 ZtF -
production facilities immediately. Accordingly,
S. H. Smiley at K-25 is using some available

pilof-sccle eqmpment to study convers:on of
"It is anticipated that, should
- oddmonal consfruchon prove necessary, a more
rr-_eff|c1enf conversnon planf than 'rhe presenf one .
“could be bu:lf. o ;

ZrC| to ZrF

 

In order

 
 
 

processing reactor, about 750 1b per week would
be the plant’s maximum output of reclaimed
material. '

S 7 on ‘the use of ZrF , aboutwfiww,
27;";;-},1750 !b of prevrouslyf"“prepored NaF ZrF and
. NaF- ZrF UF mixfures has been con "
ART fuel composmons by refurnlng the maferial
" to the processmg equnpmen’r “and oddmg ZrF o
~and UF4, as required, to 0d|usf “the composmon."
These materials have also been repurlfled From
‘the expenence gamed to date, it appears that, if
the material must be melted for delivery to the

PERIOD ENDING JUNE 10, 1955

Batching and Dispensing Operation

J. P. Blakely F. A. Doss
Materials Chemistry Division

Since production of the purified fluorides is
done in 250-1b batches and since a large fraction
of the tests call for smaller quantities, it is
necessary to batch a considerable fraction of the
material produced into 5-, 10-, or 50-lb containers.
In the recently accelerated program it has been
necessary to operate this equipment virtually
full time on a one-shift, five-day-week basis.
During the quarter a total of 6680 Ib of material
was batched and dispensed by this unit.

During recent weeks, equipment has been
installed to permit hydrogen firing of all 50-1b
transfer cans before re-use. Another unit capable
of handling the smaller containers is to be
installed in the near future, This equipment
should minimize contamination of clean fuel by
poorly handled and corroded containers.

For the first time since production began in
March 1954, usage during a quarter has kept pace
with production. Present estimates of anticipated
needs indicate that the demand for processed
materials will necessitate a six-day production
week after July of this year. The balance sheet
for the quarter is shown below:

Material on hand February 1, 1955 4,987 b

Production during quarter 5,430
Total , 10,417

Material dispensed during quarter 6,682

Material on hand at end of quarter 3,818 Ib

| ’Lo;ding und.v D<rd‘ini/ngr.0perafi’ons. .‘
N. V. Smith

~ Materials Chemlstry Division

Responsnblllty has been ossumed by Materials
Chemistry Division personnel for loading and
draining all test equipment other than the thermai-

_convection loops in Butldmg '9201-3.  This’
responsibility includes the loading of Hiquid

metals, as well as molten fluorides, and includes
the obtaining of reliable samples of the molten
material while the equipment is being filled or

75

Pratt & Whitne)./ Aircraft received 2000 Ib of
_this material, and the rest was dispensed to
__various requesters in the ORNL-ANP program,

 

S e g e

 
 

 

L

 

 

 

ANP PROJECT PROGRESS REPORT

drained. Investigations are presently under way to

ascertain the best procedure for the delivery of
oxygen-free sodium or NaK to test equipment and
to determine the most reliable method of sampling
fh:s materlal for unclysns.

Spec:al Servnces

. P. Blukely "~ F. A. Doss
| -~ J.E. Eorgan

Materials Chem:siry Division

Ennched Fue!s for ln-Plle Loops. Two batches

© of fuel com‘ommg enriched uranium were processed
for use in an in- -pile loop. Smce the size and

nature of the in-pile loop equipment require that
the finished fuel batch contain 600 * 50 g, it has
been agreed that in the fufure the material will be

 prepared in larger porhons and batched into
" containers of proper size. An adequate ‘mechanism
for cnalyhcal quality and accountability conirol

has been established to ensure proper handling

" and dispensmg of the finished material,

As a consequence of increased demands for
fuels containing enriched uranium, the processing
facility has been established in a permanent and
properly equipped location. A new, smaller unit
that is better adapted to meet the demands of the
in-pile loops is being installed, and the larger
unit is being remodeled to accommodate require-
ments for the ART high-temperature critical
experiment. Present known demands for material
will require operation of the facility for one week
per month during the balance of the calendar year.

ART High-Temperature Critical Experiment.
Preparations are being made for the production of
the necessary materials for the high-temperature
critical experiment, for loading the experimental
equipment with the NaF-ZrF, (fuel carrier) and
most of the Na ,UF (fuel concentrate), and for
obtaining relloble samp!es of the mixture as the
experiment proceeds,

Production of the fuel carrier is presently under
way, and production of the fuel concentrate will
proceed when the UF, allocation is received, The
processing time for this material should not exceed
two weeks, but a somewhat longer time will be
required to obtain analyses and to comply with

| a'c(:o‘un'rability procedures.

Fubrlccn‘:on and overhaulmg of equipment to be

'fl";lf;'used are in progress. It is anticipated that all
'-"_fprepcroflons can be completed by July 1 of this

76

year, Since all operations will be similar to those

petformed in the ARE startup, there would seem
to be no doubt of the validity of the procedures
to be used.

Reprocessing Pilot Plant, Plans are under way
to provide charge material and to fill the equipment
required to test operation of the uranium recovery
process developed by personnel of the Chemical
Technology Division (c¢f. Sec. 10). Operations
that each require about 800 Ib of material are
scheduled for July and December of this year.
The material, equipment, and procedure seem to
present no unfamiliar problems.

Experimental Preparation of Simple Fluorides

E. E. Ketchen B. J. Sturm
Materials Chemistry Division

Variables Affecting UF Preparation. - The
preparation of UF, by heahng a stoichiometric
mixture of UF, and finely divided uranium metal
in a sealed ball mill of stainless steel was
described in previous reports.23:24 The require-
ments for UF ;-bearing fuels appear to be most
easily met by the addition of this compound to
carefully purified solvent compositions, A study
of variables in UF3 preparation has been made,
accordingly, to determine the optimum conditions
for production of kilogram quantities of the material.

The variables studied have included number and
size of the steel balls used as milling mediums,
amount of material charged, and the reaction time
and temperature, In order to facilitate loading and
unloading of the 4-kg 15-in. reaction vessel, the
loading tube diameter was increased to 1 in., and
one end plate was welded at 45 deg rather than at
90 deg to the cylinder axis.

In twenty preparations for which various reaction
conditions were used, essentially pure UF, was
obtained when the normal reaction flme-temperafure
program was used. This consisted of heating the
rotating mill and contents at 900°C for three
16-hr periods. The mill was cooled to room temper-
ature between heats and pounded with a hammer to
remove caked material from the walls.

Over the range studied (0.75 to 2.25 kg), charge
size does not appear to affect completeness of
conversion.  Quantity and size of the milling

By. c. Whlflley and C. J. Barton, ANP Quar. Prog.
Rep Sept. 10, 1951, ORNL-1154, p 159.

B J. Sturm and E. E. Ketchen, ANP Quar. Prog.
Rep. Mar. 10, 1955, ORNL- ]864 p66

 

 

 

 
 

o

"

medium appeared to have no measurable ef-
fect on the purity of the resulting UF;, Ali
tests in which 1-in. balls and the sfqndard
heating program were used were successful,
However, when ]41 in, balls were used, the
product, in every case, appeared to be sintered.
This effect is serious because the fine powder is,
in general, desired and because the sintered
material adheres tenaciously to the mill and the
balls, Additional runs with other ball sizes have
been made, and the results will be reported when
available,

Some 30 kg of satisfactory UF, was prepared
during the quarter. Should the need arise, the
present equipment could produce 15 to 20 |b of
this material per week.

Other Preparations.  Approximately 5 b of
(NH4)3CrF6 was prepared by heating an excess
of NH,HF, with CrF,.3.33H,0 at 200°C. This

intermediate product was converted to CrF, by

 

PERIOD ENDING JUNE 10, 1955

FUNDAMENTAL CHEMISTRY OF FUSED SALTS
EMF Measurements
L. E. Topol

Materials Chemistry Division

Potential measurements were made during the
past quarter with combinations of several half
cells consisting of metal electrodes bathed in
solutions of the corresponding metal ion in the
molten salts. The temperature range studied was,
in general, 550 to 700°C, with the system main-
tained under an inert atmosphere. The half cells
were contained in crucibles of Morganite, nickel,
or platinum; electrical contact between them was
achieved by a‘‘salt bridge’’ consisting of a porous
ZrO, rod impregnated with the molten salt solvent.

ZrF, Concentration Cells. Cells consisting of
zirconium rods immersed in various NaF-ZrF
melts yielded constant and reproducible potentials.
The values obtained at various temperatures for a
series of cells of the type

ZrO
Ze [ZeF F Zr
4y (“2) NaF-ZrF (az) zr 4(ay) ‘

 

heating under helium at 500 to 600°C. The 'ower
temperature yields a material containing small
amounts of nitrogen, whereas heating at 600°C
gives a product that is contaminated by divalent
chromium, apparently because of interaction with
the nickel container. Chromic fluoride was also
prepared by hydrofluorination of anhydrous CrCl,.
At 400°C the conversion is mcomplete but ma-
terial prepared at 550°C | is qun‘e pure CrF,.

One batch of AgF was prepared by hydro-

’ ‘7_‘_;1?Huor|nat|on of Ag CO at 150 to 200° C. Some‘:\:: o Z
' ',‘,»,;CeF was synfhesmed by the addition of HF to an
The net electrode reaction is, ac'cbr'dingly,

< ~;._'j::f:—_aqueous ‘i‘s‘oluhonr Of Ce(NO3)3 fO“OWBd bY fll-

 
 
  

(NH"‘)ZSHFJ: The hydrofluorlno’rlon of

 

\...‘._:SnC| 2H20 at 200°C gave a mlxture of chlorlde'

.:.r o

: {';'cmd fluorlde. - i«
About 8 kg of UF was purn‘ied by hydro-
fluorination at 600°C for use as charge material
in the preparation of UF3. |

. lgove ‘o br’oduct contmmng"“'m‘”'
. S’rannous hydromde tha'r -

 
 
 

.1 Nc(a )—-;» 4t Nu+

and

where (@) denotes the activity of the species, are
shown in Table 4.19, In these experiments, three
NaF-ZrF, mixtures were used as the half-cell
electrolytes; the concentrations (c) in mole % ZrF ,
were 50.0, 41.8, and 36.0.

In a cell of the type used the anode reaction is:
—> 7rF + 4e= ,

o =
Zr° + 4F(, | stan)

while at the cathode the reaction is

o+ 4e — 7Zr° + 4F" .
(“]) (az)

+ 4F=
(@))

ZF v arg

— S ZrF
4(61}) (d )

4(a )
‘However, ‘hb'i"g-;"t;c;i"i'dh through the bridge proceeds
according to 4 '

(a))

0, s
(ay)

- z)F(a),

where ¢ is the transference number of the sodium

77

 
"ANP PROJECT PROGRESS REPORT

TABLE 4.19. MEASURED POTENTIALS FOR Z¢F , CONCENTRATION CELLS )

 

 

 

 

 

 

 

() 2)

 

 

_ ":Te_mp'é.rat'ure EMF (v)
;: S Cell No, 1* Cell No. 2% Cell No. 3* . Ey +E, .
o700 0.141 0.120 0.264 o026l
: 650 0.129 0.113 0.252 0.247
} 600 0.114 0.113 0.229 o 27
i 550 0.092 0.108 0200 ) "0,200 B
o *""'.‘C'ellAN-d 1‘ C..l = 41.8 mole % ZrF4; C:: = 50.0 mole % ZrF“1
: Cell No. 2: cy = 36.0 mole%ZrF4;c; = 41.8 mole % ZrF4
Cell No. 3. .c.l = 36.0 mole % ZrF4; c.} = 50.0 mole % ZrF4
lon. The 'ro'rcl cell recctlon is TABLE 4.20. ACTIVITY RATIOS FOR NuF -
o IN NeF-ZrF MIXTURES AT 650°C
' ZrF 4 41 NOF —é ZrF + 4t NaF .
4 4(ay) (al

 

 

 

 

. : F NaF vy NaF
The elecfromoflve force of the cell may, therefore, (cg) (29) (ag)
be expressed as NaF NaF | yNoF |

4¢
RT ZrF4(a ) NaF(al)
E = oI 2 CeliNo, 2 64.0/58.2 2.4 2.1
4F Z'F4(a:) NaF(a,) CellNo.1  58.2/50 3.4 3.0
Cell No, 4*  50/44.8 1.1 1.0

It is reasonable to assume that the fluoride ion |
in such a system is rather completely complexed *Cell No. 4: ¢y = 50.0 mole % ZrF ,; c, c1 =55.2 mole %
by the ZrF, to form complex ions of the type  ZrFi Eg5q0c =0.010 v.

ZrF.~, ZrF '”', etc. Since current transported by
such Icrge ions should be very small compared
with that carried by small simple ions, it seems
likely that the transport number of Na' is near
unity,

The vapor pressure of ZrF, over NaF-ZrF 4
melts has been measured with considerable
precision. |f the ratios of vapor pressures for
ZrF, at the concentrations shown above are
cssumed to represent the activity ratios of ZrF
and if ¢ is assumed to be unity, the activity rahos

for NoF in the various combinations can be ob-

tained directly from the emf equation. The results

of such calculations are shown in Table 4.20.
From the calculations, it appears that the rapid

and uniform addition of ZrF, to molten Na,ZrF

  - reduces the c:chwty of the NaF until the 50-50
"mole % composmon is reached; further additions

of ZrF , do not alter the NaZrF . complex,

Cells with Structural Metal Fluyorides. Cells
consisting of metallic nickel electrodes immersed
in solutions of NiF, in molten NaF-ZrF , mixtures
showed potentials of 1 to 3 mv when the NiF,
concentrations in the half cells were varied; the
Nil'-'2 concentrations were, in all cases, suf-
ficiently high to afford a saturated solution.
Similar Fe/FeF, cells in which the half celis
contained equal concentrations of FeF, showed
potentials of 1 to 10 mv. When the half cells
contained differing FeF, concentrations but more
FeF, than that required for saturation, potentials
were obtained which varied from about 40 mv at
550°C to 10 mv at 700°C. Exact voltages varied
from cell to cell, but the decrease in emf with
temperature seemed quite reproducible.

 

 
 

©

   

A cell of the type

PERIOD ENDING JUNE 10, 1955

 

FeF,(c,) 7:0, FeF3(c:)
+» NaF-ZrF, NaF-ZrF, , Fe ,
FeF (c,) NaF-ZrF, Fer(c;)
where
o . where
¢y = cyand ¢, = Cg + ¢, = 7.0t07.3wt%,

resulted in emf’s of 1 mv. It was believed that
all the FeF, would be reduced to FeF, by the
iron electrodes and that as a result a pure Fe/FeF
cell with etched Fe electrodes would be obtained.
However, analysis of the final melt proved that
most of the FeF, was still present; it is not
likely that equilibrium was established,
Cells of the type ‘

ZrO2

- Zr ZrF NaF

4leqg) ! NaF-ZtF,

have yielded moderately reproducible potentials
at various temperatures, |In these cells, ¢, = 47
mole % and ¢, = 5.0 to 5.6 wt %. From the data
obtained, E° is estimated to be 1.4]1 to 1.42 v by
assuming that the saturating phase is the pure
fluoride and that the activity of the ZrF, can be
established, as before, from vapor pressure data.
From this E°value, it appears that AF® = <131 kecal
for the reaction

Zr® + NiF, —> ZrF, + 2Ni°

This is in excellent agreement with AF® = =127
- kecal based on the thermochemical estimates of'_;___
‘ Brewer

25

   

Na F--ZrF4

it

c, = 5. 1to54wt%,
have yielded the data shown in Table 4.21.

The thermochemical estimates of Brewer suggest
that E° for this cell should have a nearly constant
value of 0.25 v over the 550 to 700°C temperature
interval.  Again, examination of the solidified
melts showed that isomorphous complex compounds

 

NaF-ZrF, , NiF, Ni

€2
sat’d

of NiF2 and FeF2 with the solvent occurred. If

the junction potential of this cell can be assumed
to be negligible, then at 650°C,

 

» RT aFeF2
042v = 0,25V = — In ——
2F  aniF
2
and
AFeF
2
—_ = ~10-2 ,
anie
Nle

" Since the saturation solubility of FeF, is nearly

There is some evndence, “however, from pefl.o,"""“'I"'""fw'i'c"e that of NiF, ot this temperature and since

- graphic and x-ray ‘examination of the cooled melfs'
,__whlch mdlca'res that ’rhe soturctmg phdse |s no'r"' )

the fluoride ion activity should be similar in the

"‘be""s'dlu'rions, it appears that the activity coef-

' f|c1ent for NI is hlgher by '200-1"old than thot for

f‘f_'j;fluonde._ The da’ra presenfed WOuld seem to
. indicate thot thls complex fluorlde is qun‘ew
‘ "Vii-funstable. | C : :
~ Experimental mecsuremenfs of cel!s of the type '

 

25L Brewer et al., p 107 and '110 in Cbemzstry and

— Metallurgy of Miscellaneous Materials; Thermodynamics

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

2 - .
NaF-ZrF,, N:F2< c, ) Ni
sat’d

79

 
s i e Mk B dian e

 

'ANP PROJECT PROGRESS REPORT

TABLE 4.21. POTENTIALS OF CELLS
Fe/Fer/Nin/Ni IN NuF-ZrF4 SOLVENT

 

EMF (v)

 

 

 

Temperroiure |
(°C) Trial No. 1 Trial No. 2
700 0.425 0.423
650 0.2 0.424
600 0.414 0.416
. 550 0.407 0.409
Fett, H appecrs also that the solvent-Fe **

complex is much more stable (about 8 kcal/mole)
thon is ihe comp(ex mvolvmg Ni*

Vapor Pressures of LIF-ZrF Mixtures
| R E. Moore
Materials Chemistry Division
‘The determination of the vapor pressures of a
series of LiF-ZrF, mixtures by the method of
Rodebush and Dixon2% and Fiock and Rodebush27

 

 

 

 

26W H. Rodebush and A. L. Dixen, Phys. Rev. 26,
851 (1925)

E. F. Fiock and W. H. Rodebush, J. Am. Chem.
Soc. 48, 2522 (1926).

was started during the past quarter because the

NaF-LiF-ZrF, system is being considered as «
possible fuel carrier. The vapor pressure work
on mixtures containing 33.3 mole % ZsF, (LiéZrFé)
and 50 mole % ZrF, was completed. The data
for these two mixtures are given in Table 4.22.

~ The vapor pressure equations, which were ob-

tained from the best straight lines on a log
pressure vs reciprocal temperature plot, are, for

the 33.3 mole % ZrF, mixture,

log P {mm Hg) = —(8333/T) + 7 967
and, for the 50 mole % ZrF mixture,

log P (mm Hg) = —-(8848/7‘) £ 9397,

where T is in °K, The heats of vaporization are
38 kcal/mole for the 33.3 mole % ZrF, mixture
and 41 keal/mole for the 50 mole % ZrF, mxxture.
The vapor pressures of the LiF- ZrF4 mixtures
are considerably higher than those of the corre-
sponding NaF-ZrF, mixtures. For example, the
vapor pressures of the 50% LiF mixture are 50
to 100% higher than those of the 50% NaF mixture.
It might have been expected that the small [ithium
ion would produce compounds in the fused state
which were more stable than the sodium com-
pounds. One possible explanation for the higher
vapor pressures is that LiF,~ ions may exist in
the melt, If the lithium ion has a marked tendency

TABLE 4.22. THE VAPOR PRESSURE OF TWO LiF-ZrF4 MIXTURES

 

Temperature

Calculated Pressure

 

Observed Pressure
(°C) (mm Hg) (mm Hg)
LiF-ZtF , (66.7-33.3 mole %)
944 13 13
992 22 24
1014 30 31
1059 54 52
1112 89 89
1178 166 168
LiF-Z¢F , (50-50 mole %)
809 16 16
812 17 17
'330 23 24
849 33 32
897 67 68
941 125 130
208

977 209

 

80

Gl i

 

 

 

Ei

 
 

&

iy

to attract fluoride ions to produce such complexes,
fewer fluoride ions would be available for the
formation of complex ions with ZrF,.

Solubility of Xenon in Fusé.dwlgalts
R. F. Newton

Research Director’s Department

The previously described procedures?® for de-
termining the solubility of xenon in fused salts
were modified so that the source of the spread of
values obtained could be ascertained. Long ex-
posure without stirring gave essentially the same
results and thus indicated that the supposition of
the production of fine gas bubbles by stirring,
which bubbles were then transferred to the stripper
along with the liquid, was unfounded. However,
fong continued cycling of helium through the melt,
without recent exposure to xenon, gave material
which was caught in the liquid nitrogen trap and
was read on the McLeod gage as xenon, This
material may have been SiF, from the reaction of
the glass with the HF liberated from the NaF-
KF-LiF (11.5-42-46.5 mole %). Means for elimi-
"nating this material or correcting for it are now
under s’rudy.

X.Ray lefrochon Study of quuxds

P. C. Sharrah M. D. Danford
H. A. Levy P. Agron
R. D, Ellison M. A. Bredig
Chemistry Division

The construction of an x-ray diffractometer de-
signed specifically for studies on the structure
of liquids was completed recently. The diffraction

pattern from the horizontal surface of the liquid
sample is obtained with a divergent beam tech-

nigue similar to the Bragg -Brentano flat- sample

... 'system frequenfly used on powder somples. The

“‘instrument prowdes for simultaneous angular

   
   
 
 

~: ~surface.

 
 

Pl fdrnoce for work wnth
o in the de5|gn.
“sential to good work “with liquids, is obtained

"*”_}hrough the use of c benf crysfol monochromcn‘or

 

28R F. Newton, ANP Quar. Prog. Rep. Dec. 10,
1954, ORNL-1816, p 72.

PERIOD ENDING JUNE 10, 1955

mounted on the arm with the de’rector.' A scintil-
lation detector and a molybdenum target x-ray
tube are in use in the system now undergoing final
testing.

High-Temperature X-Ray Spectrometer Studies

G. D. White
Metallurgy Division

T. N. McVay, Consultant

A high-temperature attachment for an x-ray
spectrometer is being used to help clarify some

~of the phase relationships in fuel systems of

interest, The apparatus consists of a water-
cooled aluminum cylinder with a vacuum-tight lid
which fits the open end of the cylinder and o
projecting rod on the closed end by which the
attachment is fitted into the goniometer. The
sample holder, tantalum heating element, reflector,
and thermocouple leads are all mounted on the
lid, which also contains an aperture for pumping
a vacuum. Inserted around the center of the
cylinder is a beryllium window through which the
x rays pass., When assembled, the attachment can
be pumped down to a vacuum of less than 1 p by
using a Welch pump and a small diffusion pump.

Initially, the attachment was aligned by using
a Th02 sample, The bracket which holds the
sample can be rotated and can be shifted
vertically or horizontally. It is held in position
by three radial set screws, By manipulating these
set screws, the ThO sample was put in a position
where one of its sfronger x-ray peaks was at the
proper angle and had maximum intensity. Thus
far the alignment has remained true, although
samples have been x-rayed almost dmiy for the
past three months,

In order for the attachment to be used most

" effectively, the sample to be studied must give

an” x-ray diffraction pattern with fairly intense

- motion of the x-ray ‘tube and the defector on arms _"_;l'predksr the diffraction patterns of the polymorphs

| mévmg oboui a horrzontal axis lying in “the hqu:dw
_7 TI‘HS ‘makes it possible to work with
liquids ‘with a free “surface where sample holder
“absorption and sccn‘terlng are ellmlna’red A small" o
' { " and loaded into a nickel sample holder with space
Monochromqflc X mdjgf,on,' es':’"":;"'m'll eft for expansion. After the sample holder has

“must be considerably different, and the inversions
"ot phase changes must be fairly rapid, especially
"if the sample being studied oxidizes readily,

The sample to be x-rayed is ground to a powder

been mounted in the bracket and fhethermocoup|e
bead has been placed on the surface of the

~~"powder, the lid is fastened to the cylinder. The

system is pumped down overnight, and the x-ray
diffraction patterns are obtained the next day.

81

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

Roc_')m-temperafure patterns are always obtained
before and after heating to check on the alignment
and flatness of the surface of the sample.

- Tol_d‘q'fe, work has been done on compositions

in the systems: NaF-ZrF_, LiF-ZrF,, and NaF-
BeF,. Only the study of the composition
‘2NaF.BeF, is discussed here. Two samples of
Na2BeF were found to be different optically,
One of :‘1e samples contained twinned monoclinic
crystals, whereas the parallel extinction of
crystals in the other indicated an orthorhombic
structure.  Room-temperature x-ray ~diffraction
patterns of the samples were very similar, the
only difference being in the intensities of the
fieaks;

- ngh-temperafure X-fcy diffraction patterns were
also obtained on the samples, When heated, both

": . "$“c‘l'mp-le_s' showed inversions at 236 and 336°C,
“and when cooled, showed inversions at 310 and

below 185°C, The temperature could not be
maintained kelow 185°C, and therefore the lower
inversion temperature could not be obtained.

The intensity and 4 values at room temperature,

300, and 340°C are

82

At Room

 

 

At 300°C At 340°C
Temperature
- d I - d I
d i
- Sdrh#le 1
4.18 14 2.71 35 2.78 7
3.90 6 2.65 33 2.65 27_
3.67 10 2.34 30 2.18 27
2.92 35 2.25 7 2.11 8
2.68 10 2.21 22 Sample 2
2.62 20 2.17 22 2.9 12
243 26 2.1 13 2.78 7
2.37 32 2.02 15 2.73 9
2.23 9 ' 2.65 27
2.11 17 2.18 27

1.84 18 a2 8

These data suggest that vth-e.'f samples are
actually a mixture of two phases: g room-temper-

ature phase and the 300°C form. There are

several 4 values at room temperature which are
present at 300°C if allowance is made for ex-
pansion of d values as the temperature is
increased. Therefore, the difference in intensity
values of the two samples at room temperature
could be due to a difference in ratios of the two
phases in the samples.

 

 

 

 
 

"

. PERIOD ENDING JUNE 10, 1955

5. CORROSION RESEARCH

W. D. Manly

G. M. Adamson

Metallurgy Division

W. R. Grimes

F. Kertesz

Materials Chemistry Division

Several Inconel forced-circulation loops that were
“operated with fluoride mixtures and with sodium
as the circulated fluids were examined. Results
of corrosion and mass transfer under dynamic
conditions were obtained. Data were obtained on
an alkali-metal base fluoride mixture containing
UF, and on ZrF ;-base mixtures with UF ~and with
combinations of UF, and UF, circulated at low
(1,000) and high (15,000) Reynolds numbers at a
maximum fluoride mixture temperature of 1500°F.

Further thermal-convection loop studies were
made of alkali-metal base fluoride mixtures with
UF, and UF, and ZrF,-base mixtures with UF,
and UF | in Inconel. The effects of temperature on
mass transfer and of the size and shape of the loop
on corrosion were investigated, and an evaluation
of the erratic results recently obtained with control
loops was made. Several brazing alloys on type
310 stainless steel and ''A’" nickel were tested
in sodium and in a ZrF ,-base fuel mixture in an
attempt to find a brazing alloy that has good cor-
rosion resistance fo both mediums. Cermets that
do not form solid-phase bonds were screened for
suitability for use in valves, bearings, and seals
exposed to liquid metals and fused fluorides.

Mass transfer of sodium in an Inconel thermal-
convection loop ‘and of lithium in a type 347 stain-
less steel loop was studied, as well as the cleanlng
*of berylhum-lnconel systems with Versene.

In the fundamentol corrosion research, further

~ work was done on the mveshgahon “of film for-
‘me’rlon ‘on metals, |ncludmg tests of oxidation
“theories, and oddlfional experlments were conducted

in the lnveshgahon ‘of corrosion and mass transfer
L f"of mefuls by fused sodium hydroxude.
o The chemlcal sfudses of corrosion include in-

  
   
  

.:-.LIF BeF

_ and by LiF-BeF, UF4, ‘the effect of
'-UF UF

mixtures on corrosnon of Inconel by

o fvarious 'solvents and studles of the sodium

:;'fhydromde-—mcke[ reachon o

3 ‘\,;..’_3_,.'_;;1

 

]L. A. Mcnn, w. B. McDonald and W. C, Tunnell,
ANP Quar. Prog. Rep. Dec, 10. 1954, ORNL- ]816
Fige 3.4, p 45.

 found in the first heated section.
< IC > _ ___study of this loop showed that the maximum wall
eshgahons ' of “the corrosmn of Inconel by

“in this area.

'FORCED-CIRCULATION STUDIES

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

Fluoride Mixtures in Inconel

Several forced-circulation loops that were oper-
ated by the Experimental Engineering Department
at the request of the Metallurgy Division were
disassembled and examined. The conditions of
operation of these Inconel loops in which fluoride
mixtures were circulated are summarized in

Table 5.1.

Two of the loops examined (4695-1 and -2} had
circulated NcF-ZrF4-(UF4 + UF3) (50-46-4 mole %)
in which about 2 wt % of the uranium had been
converted to U3Y. The design of loop 4695-1 was
similar to that used previous[y.] Operation of this
loop was terminated after 385 hr because of a leak at
a heating terminal. Examination showed the maxi-
mum attack, which was found in a bend in the first
heating leg, to be 12 mils (Fig. 5.1). In a straight
portion of the heated section, the maximum attack
was to a depth of @ mils. The depths of attack
were thus only slightly less than those found in
loops which circulated a similar mixture containing
no trivalent uranium; however, the number of voids
was reduced by about one-half.

The second Iobp examined (4695-2) had been

~ fabricated with two straight heating sections
rconnected with a U bend (Fig. 5.2), and it had
_ operated for 887 hr before belng terminated because
~ of a pump bearing failure.

The maximum attack
was to a depth of 8 mils (Fig. 5.3), and it was
A temperature

temperature occurred in the first heated section
but that the maximum fluoride mixture temperature
occurred in the second heated section. The loop

was assembled so that no heating occurred in the

bend, and upon exammcmon, no attack was found
‘A thin, as yet unidentified, deposit
was found in the final portion of the cold leg. For
comparative purposes, a portion of this same batch

83

 
 

 

TABLE 5.1. CONDITIONS OF OPERATION OF FORCED-C}I.RC.ULATION_ INCONEL_ LOOPS‘.‘

 

Loop Number
4695-1 '

 

Operating Conditions

4930-A 4935-1 4695-2 4695-3

 

 

Fiuvoride mixture circu- NuF-ZrF4-UF4 NcF-Zer-UF4 NaF-ZrF4—~(UF4+UF3) NoF-ZrF4—(UF4+UF3) NaF-KF-LiF+UF4

(mils)

lated (53.5-40-6.5 mole %) (53.5-40-6.5 mole %) (50-46-4 mole %, (50-46-4 mole %, (11.5-42-46,5 mole % plus
2.2 wt % of U as U3+) 1.7 wt % of U as U3+) 12.3 wt % UF4)
Operating time (hr) 1000 1000 385 887 630
Preliminary operating 24 50 71 43
period at isothermal
temperature (hr)
Maximum fluoride mix- 1500 1500 1500 1500 1500
ture temperature (°F)
Approximate maximum >1710 1680 >1735 > 1670 >1640
tube wall temperature
(°F)
Temperature gradient of 300 300 300 200 300
fluoride mixture (°F)
Reynolds number 1000 1000 10,600 15,000 10,000
Velocity {fps) 0.6 0.7 6.7 10.2 3.3
Length of heated tube (ft) 23.5
First section 3.5 5 5 -5
Second section 3.5 7 7 7
Total length of loop (ft) 14 48 50 50.5 | 50.5
Method of heating Electrical resistance Gas Electrical resistance Electrical resistance Electrical resistance
Shape of heated section Coiled Coiled Coiled Straight Straight
Ratio of hot-leg surface 1.6 3.1 1.5 1.5 1.5
to loop volume
(in.2/in.3)
Cause of termination Scheduled Scheduled |_eak Pump bearing failure Leak
- Maximum depth of attack 25 11 12 8 35

 

 

 

 

LY0dIY SSIHD0¥d LDIr0¥d dNV .
 

 

«}

‘ ”concenfr

   

  

UNCLASSI‘FIED
2 ;

s

 

T

Fig. 5.1. Maximum Attack Found in Inconel Loop
4695-1 After Circulating NaF-ZrF,-(UF, + UF,)
(50-46-4 mole %, 2.2 wt % U as U3?) for 385 hr
at a Maximum Fluoride Mixture Temperature of
1500°F and a Reynolds Number of 10,000, 100X.
Reduced 16%.

of fluoride mixture was circulated for 1000 hr in
an Inconel thermal-convection loop. The hot leg
of the thermal-convection loop was attacked to «
depth of 10 mils, whereas a depth of 13 to 15 mils
was normally found in control thermal-convection
foops in which the fluoride mixture contained UF,
but no trivalent uranium. However, thermal-
convechonr 100ps opemfed prev:ously in wh:ch

trwaent uranium was obfumed by adding zirconium

hydride were attacked to a depfh of only 3 mils.

. The deepest attack found in any forced-circulation
~loop was that found in |06§1"4695‘3 ~ This loop
""rc:rculated the alkoh-metol base mixture NaF KF- LiF
; i‘(” 5.42- 46 5 moie %) wfih 12.3 W’r % Urdmum addedw -
S as UF4. “The ‘maximum attack in this loop (de-
"signed as shown in Flg 5.2) occurred in the
:"\‘second heated sechén as a very heavy mtergranular
RS *"fgice voids to a depth of
35 mlls (Flg 5'_4) The 'reinperature pattern of this
foop showed 'rhcx'r ‘because of the better heat
trcmsfer properhes of fhe ulkah me’ral base mlx-‘:_

 

 

,'_'fures both ‘rhe maximom wall cnd fluorlde mixture
temperatures ~ occurred in the second heated

section. A deposit that was up to 65 mils thick
was found on the lower inside surface of one area

PERIOD ENDING JUNE 10, 1955

in the cold portion of the loop. In other areas,
similar deposits that were up to 5 mils thick were
found. A spectrographic analysis of the deposit
showed the following: >5 wt % Cr, 0.5 wt % Fe,
>10 wt % K, 3wt % Li, 2wt % Na, 0.5 wt % Ni,
> 10 wt % U (balance unidentified}). Another portion
of this batch of fluoride mixture was circulated for
500 hr in an Inconel thermal-convection loop. The

maximum aftack was to a depth of 42 mils, and

was unusual in that it was found only around one
third of the circumference of a sample of the hot
leg. In the remainder of the hot leg the attack was
to a depth of 5 mils. A cold-leg deposit that was
0.3 mil thick was found.

Two other forced-circulation loops that had
completed the scheduled 1000 hr of operation were
also examined. These loops had circulated NaF-
ZrF4-UF4 (53.5-40-6.5 mole %). One of these
loops (4930-A) was heated by the electrical
resistance of a coil,] and the other (4935-1) was
heated in a gas furnace. These loops were similar
except that the gas-fired loop had a heated length
of 23.5 #t, and the heated length of the other loop
was only 7 ft. To transfer the same amount of
power, it was obviously necessary to use a much
higher wall temperature for the short loop. The
maximum attack in the short loop (4930-A) appeared
as a heavy intergranular concentration of sub-
surface voids to a depth of 25 mils and was found
in a bend in the second leg of the heated coil. The
maximum attack in a straight section was 21 mils.
The attack was both deeper and heavier than that
found in loops operated previously with this
fluoride mixture, even when much higher velocities
and Reynolds numbers were used. Examination of
loop 4935-1 showed the maximum attack to be to a
depth of 11 mils; however, this attack was moderate
in intensity and was found for only a short length
at the hottest end of the coil. In this loop no

~differences were found on opposite sides of the

bends. The original purpose of these loops was to
compare gas and electric heating; however, because
of the differences in heater length, it will be
necessary fo repeat the tests with identical loops

-~ -operated under similar conditions.

From the results obtained with these loops and
from those reported previou’sly,:z it is apparent that

‘corrosion and mass transfer in these systems are

not so serious as had been feared. It also appears

 

2G. M. Adamson and R. S. Crouse, ANP Quar. Prog.
Rep. Maf. 10) 1955‘ ORNL']864' P 75.

85

 
 

 

 

 

86

 

ANP PROJECT PROGRESS REPORT

HEATED SECTIONS

  
 

UNCLASSIFIED
ORNL-LR-DWG 7414

     

  

\m

COLD SECTION

 

 
 

 

 

HEATER CONNECTING LUGS

  
 
   

FILL TANK

Fig. 5.2. Schematic Diagram of Forced-Circulation Loop with Straight Heated Sections.

that reducing the tube wall temperature in these
experiments would result in a reduction in depth of
attack. The cold-leg layer found in loop 4695-2
must be investigated further.

The depths of attack found in the forced-
circulation loops are not much, if any, greater
than those found in the low-velocity thermal-

_ convection loops Thus velocity and Reynolds

number appear to be only minor variables. This
conclusion is confirmed by the deep attack found
in the low-velocity forced-circulation loop 4930-A.
The data obtained from the thermal-convection
loops should fherefore be applicable to forced-
circulation systems

The most important variable is the maximum tube
wall temperature or, more exactly, the temperature

of the reaction interface, This is shown by the
~ deepest attack occurring in areas of maximum wall

temperature rather than where the maximum bulk

fluoride mixture temperature occurred. Additional

evidence is afforded by the deepest attack in a
loop having occurred in the hot spots on the inside
of bends.?

The addition of UF, to the ZrF -base mixtures
reduces the attack, and even though the low UF,
concentrations used in these forced-circulation
loops did not reduce the attack as much as had
been hoped, it appears that higher concentrations
would result in a greater decrease in attack.
Unless the addition of UF; can also reduce the
attack and mass transfer found with alkali-metal
base mixtures without disproportionation of the
UF, and the formation of deposits, it appears that
such mixtures will not be useful in large, Inconel
dynamic systems.

Sodium in Inconel

An Inconel forced-circulation loop (4689-4) was
examined in which sodium was circulated at a
maximum temperature of 1500°F for 1000 hr, with

 

 

 

 

 
 

W

”’:“:for 480 he wi

   

  

UNCLASSIFIED
T7643 !

i a -
Sl T
£ .
e
. wrF :
G - LA i
vt ¢ S
J = T
—t St e
s ¥ LW -
. . & . o
. © . . L
? " ¥ w .
. & T
- Y, Lk .
s : g .
L ' . LS . bdl
£ - ., ML .
& P [ ) v
. i cal :
+ . < i . -
- v L
v A . i
o - i ;
-
- —
L w’ ;
- i
. 2 . -
# . LR
- & ¥ .-
{
*

Fig. 5.3. Maximum Attack Found in Inconel Loop
4695-2 After Circulating NaF-ZrF (UF + UF )
(50-46-4 mole %, 1.7 wt % U as US™) for 887 b
at ¢ Maximum Fluoride Mixture Temperature of
1500°F and a Reynolds Number of 15,000. 100X.
Reduced 11%.

a 300°F temperature drop and a Reynolds number

~of greater than 50,000. The loop had the configu-

ration shown in Fig. 5.5. A heavy layer of dendritic
metal crystals was found in all sections of the
economizer and to a lesser extent in the cold loop,
as shown in Fig. 5.6. The crystals were shown by
chemical analysis to be 89.6% Ni and 8.6% Cr and
to contain '50'ppn1 Fe. The layer found in the cold

“end of the economlzer was shown by metollogrqphlci
'exommohon to be 26 mils '|'|"IIC|< The hot leg
‘showed mfergranulor attack to a dep'rh of 2.5 mlls.w |

- Efforts were made to obtain oxygen onolyses on
' ‘_'both the orlgmol ‘and the dmlned sodlum, but bothf
'--f=va|ues are’ quesnonuble. From other work, it

: ‘-..appeors fhat fl1e orlgmof sod:um contolned obouf _'

| --7-'.‘,:50 PPM oxygen. SR

        

A second snmllar'

    

The maximum thickness of the deposn found metal-

 
 

jrl"o‘a'p 'k(}195'l l)mwus opercted_fl_;:
ith a _200° F temperqture drop, and the':f
Cresults conf:rmed 1hose reported above. The moss:
~“fransfer found in this loop is shown in Flg 57.

PERIOD ENDING JUNE _10, 1955

UNCLASSIFIED
T-7727 |

 

. e :
v’ + . o N . T
- e :
“ . ' * o M
- . ‘,. i
3 e
..'. - ., - “\,
,, ; eyl
. . Ly
LT P :
- ..M O
r’; + . [
_ - O
. e e M
> [N 7 o
S . .
AN L p 3
e T O
> - B3
. e e, T
. cw LA
. St
~ Y
w e
¢ ; ‘s i
[Py
' .
. ,

Fig. 5.4. Maximum Attack Found in Inconel Loop
4695-3 After Circulating NaF-KF-LiF + UF, (11.5-
42-46.5 mole % plus 12.3 wt % UF 4) for 630 hr at
a Maximum Fluoride Mixture Temperai‘ure of 1500°F
and ¢ Reynolds Number of 10,000. 100X. Re-
duced 20%.

fographically was 11 mils. The hot leg in this
loop also showed intergranular attack to a depth
of 2.5 mils.

Two additional loops were operated with portions
of the cold legs constructed from type 316 stain-
less steel. Loop 4689-6 had a type 316 stainless
steel economizer and cold leg, .while loop 4689-5
had only a type 316 stainless steel cold leg. Both
these loops also showed mass transfer, but not so
much as was found in loop 4689-4. The maximum
deposit thickness in loop 4689-5 was 15 mils and

—in loop 4689-6 it was 12 mils. The hot legs

showed similar intergranular attack to a depth of

2.5 mils in loop 4689-5 and 1.5 mils in loop 4689-6.
- These data indicate that a type 316 stainless steel

cold leg reduces the mass transfer slightly, but
further study of the variables in the process is

"j__nnecessory Additional IOOps are bemg operated
7 vunder com‘rolied condmons in which' the effects of

oxygen concen’rrohon of 'rhe sodlum the use of a
cold frop, the use of c: Iowe-r maximum sodium
temperature, and ‘the use of an all-stainless-steel
system are being investigated.

87

 
 

 

 

 

| ANP PROJECT PROGRESS REPORT

     

. UNCLASSIFIED _
ORNL-LR-DWG 7415

el

r SURGE TANK

e wlll”
Y n
Y

N
2

 

PUMP

ELECTROMAGNETIC
FLOWMETER

    

Fig. 5.5. Schematic Diagram of Loop Designed for Forced-Circulation of Sodium in Inconel Tubing.

"~ THERMAL-CONVECTION STUDIES

G. M. Adamson
. Metallurgy Division
- T. C. Price V. P. Treciokas
' - Pratt & Whitney Aircraft
'Alkqlri-Mefal Base Mixtures with
UF, and UF in Inconel

The results of metallographic examinations have
 been received for the remaining Inconel thermal-

88

convection loops in which alkali-metal base
fluorides with varying proportions of UF,/UF,
had been circulated. The results for the first
half of this series of tests were reported previ--
ously.® The metallographic data are presented in
Table 5.2. The results for the loops with high
UF,; content confirm those reported previously.
No attack was found in any section, but deposits

 

36. M. Adamson and V. P. Treciokas, ANP Quar.
Prog. Rep. Mar. 10, 1955, ORNL-1864, p 70.

 

 

 

 

 
PERIOD ENDING JUNE 10, 1955

S ' S UNCLASSIFIED ' ' UNCLASSIFIED
. S - o T-7761 - T-7763

 

      

 

 

 

 

 

 

Fig.5.6. Sections from Cold Leg and Economizer Fig. 5.7. Sections from Cold Portions of Forced-
. of Forced-Circulation Inconel Loop 4689-4 After Circulation Inconel Loop 4951-1 After Circulating
Circulating Sodium for 1000 hr at a Maximum Sodium Sodium for 480 hr ot o Maximum Sodium Temper-
Temperature of 1500°F and a Temperature Drop ature of 1500°F and a Temperature Drop of 200°F.
- of 300°F.
TABLE 5.2, RESULTS OF METALLOGRAPH!C EXAMlNATlON OF INCONEL THERMAL.- CONVECTION LOOPS
AFTER CIRCULATING NGF KF- lF (” 5-42-46.5 moie %) CONTAINlNG UF AND UF
| T oo
o Time of , i I,q U3 Content Maximum
L_oop . Uranium { %) .
: . Operation wt % Attack Metallographic Notes
No. h Content : . i1s)
: (l’) L wmw F'”, Fmal‘ (mils
51 s00 ns 17 o 2 Thin cél.d-iég doposit
626 ) 500 .2 _ 1.5. 08 R 1.5 Cold- Ieg deposn 0.3 mll thick
594 1000 126 20 06107 2 Cold- ieg depos:f o s mil thick
| 625 s46 12 30 09 0 Layerto lmil thick inall sections
593 _ ]000 - 138 - 56 - f_ 0.9 ."0 o Layer to 1 rn||'rhlck in all sections
< 592 - 'I-OOOV 109 5.2 - 0.9 -‘ 0 Layer-‘fo i n.}i'! *hick in all sections
P

89

 

 
 

 

ANP PROJECT PROGRESS REPORT

were visible in both the hot and cold tegs. The
trivalent uranium content had been almost com-
pletely lost by disproportionation in all loops.

The results for the three loops with low UF,
contents are encouraging. In loop 594, after
1000 hr, the hot-leg attack was only to a depth of
2 mils, and there was no evidence of a hot-leg
layer (Fig. 5.8). The cold legs of these loops
showed thin, as yet unidentified, layers. Even in
these loops it seems likely that some dispro-

 

fe———————you1 0400

‘,21 ’1'

 

:‘,’;: "?/? ‘/,

Fig. 5.8. Hot-Leg Attack of lnconel Thermal-
Convection Loop After Circulating NaF-KF-LiF
(11.5-42-46.5 mole %) Containing UF; ond UF,
for 1000 hr at a Hot-L.eg Temperature of 1500°F.
250X. Reduced 21%.

UNCLASSIFIED
T.7544

portionation occurred, because one-half the origi-
nal U3Y content was lost, a

Chemical analyses of the fluoride mixtures
before and after operation of the loops are pre-
sented in Table 5.3. The data show that the
impurities react differently in the alkali-metal
base mixtures than they do in the ZrF,-base
mixtures. The chromium content appears to de-
crease slightly, whereas the nickel content shows

~a small increase.

 Zirconium Fluoride Base Mixtures with
UF, and UF, in Inconel

A group of Inconel thermal-convection loops was
operated at a hot-leg temperature of 1500°F with
the ZrF,-base mixture NaF-ZrF4~UF4 (50-46-4
mole %) containing varying amounts of the UF,
converted to UF,. Metallographic results have
been received on only the first half of this series
of tests and are reported in Table 5.4, The chemi-
cal results for these tests are given in Table 5.5,
While the depths of attack found were slightly
lower than the 11 mils now being found after 500 hr
in contro! loops, the results are disappointing in
that, in previous thermal-convection loops operated
with a mixture containing UF, obtained by the
reduction of UF, with ZrH,, attacks as low as
3 mils were found. As in the forced-circulation
loops, the presence of UF, reduces the amount and
intensity of the attack more than it does the depth.
In the loops operated previously, the UF, content
was not known and may have been slightly higher.
1t should be noted that the chromium content was

TABLE 53. RESULTS OF CHEMICAL ANALYSES OF ALKALI-METAL BASE
MIXTURES BEFORE AND AFTER CIRCULATION IN INCONEL LOOPS

 

 

 

 

 

Locop Uro;-aium (wt %) Nicke! (ppm) Chromium (ppm) lron (ppm)
Ne. Before After Before After Before After Before _ Aftgr
591 15 1.5 65 100 &0 30 140 75
626 1.2 2 10 110 4 15 to 50 120 85 | |
594 12.6 | 10.7 60 110 115 15t0 ? 70 " '75 B /
| 625 13.2 12.5 60 40 40 140 ) | .
i 593 s 10.0 45 85 75 15 o 0 | |
' 592 109 10.5 110 95 55 ?

0108 165 0

 

90

 

 
 

"

0"

TABLE 5.4,

PERIOD ENDING JUNE 10, 1955

RESULTS OF METALLOGRAPHIC EXAMINATION OF INCONEL THERMAL-~CONVECTION

LOOPS AFTER CIRCULATING ZrF ~BASE MIXTURES CONTAINING UF, AND'UF4

 

L.oop

Operating Maximum

Metaliographic Notes

 

 

Time Attack
Ne. (hr) (mils) Hot-Leg Appearance Cold-1_eg Appearance
633 500 10 Moderate to heavy general attack and No attack or deposit
o intergranular voids
634 500 8 Moderate to heavy intergranular Few voids to a depth of 1 mil with deposit
voids covering voids
683 500 7 Moderate intergranular voids Few voids to a depth of 1 mil
635 1000 7 Moderate general attack and inter- Few voids to a depth of 0.5 mil and some

_stlll in operatlon.

 
 
 

i‘appears to be rea o

granular voids

thin deposit

 

TABLE 5.5, RESULTS OF CHEMICAL ANALYSES OF ZrF -BASE MIXTURES CONTAINING UF ;, AND UF,
BEFORE AND AFTER CIRCULATING IN INCONEL THERMAL-CONVECTION LOOPS

 

 

 

 

Loop U3t (wt %) Total U (wt %) Nicke! (ppm) Chromium (ppm) Iron (ppm)

Ne. Before After* B—etore After Before After Before After Before After

633 1.37 0.7 1.8 1.4 25 15 100 570 65 50

634 1.28 0.9 8.8 8.9 35 20 65 300 40 70

683 1.43 0.8 9.0 8.5 25 25 140 230 25 30
60

635 1.98 0.7 8.9 9.7

10 100 225 80 40

 

*These values showed considerable variation and it was dlfflcult to estimate an average; the hot-leg value was con-

sistently lower than the cold-leg value,

lower than normal in three of these loops after

circulation. Addltional |oops in this series are

 

data obtamed in these experlments are tabulated
in Table 5 6. |t is not completely understood why“ '
:attack is deeper “at IQOOOF than at 1350°F but
th’lS effect has been noted m prev:ous tests and
: eal. The'r remammg lToops show a
- definite incredse in depth of attack with increasing

:hot Ieg temperaturew In |oops operated previously

 

 

under similar conditions for 500 hr, no effect of
temperature on the depth of attack was found.
Since the mass transfer effect is masked by fhe

effect of impurities and nonequilibrium conditions
during the first 500 hr of operation, it may be

" considered that the increased depths of attack with

L3 o | o7l increased te tures, in the | ted f
- Effect of Temperature on Muss Transfer _, {;00 ?1r weremcf::;osefl by n;:ss :q::fei operated for
L A group Of |nconel ‘°°P5 ‘was operated mthsww
" NaF- ZrF -UF (53 5-40-6.5 mole %) for 1500 hr at

f___,-,‘;,_.;"'varlous hot-leg temperatures m order to determine
- the effect of temperature on mass tronsfer The

Effect of Loop Size and Shape on Corrosion

A series of Inconel loops was operated with

- NciF-Zer-UF4 (50-46-4 mole %) to determine the

" effect on depth of attack of varying the loop size

‘and shape. These loops all operated for 1000 hr

with hot-leg temperatures of 1500°F. The data

‘obtained from these experlments are presented in

“"Table 5.7.

Within the accuracy of the data, it does not
appear that varying the length of the horizontal
leg is any more effective than varying the vertical
leg. For special materials in limited supply, it
appears from the data that the loop size may be

N

 
i e e

 

ANP PROJECT PROGRESS REPORT

'TABLE 5.6. EFFECT OF HOT-LEG TEMPERATURE ON MASS TRANSFER IN INCONEL
THERMAL-CONVECTION LOOPS CIRCULATING NoF-ZF .UF , (53.5-40-6.5 mole %)

Operating time: 1500 hr

 

Meiu!logruphic'Nofe.s.

 

 

 

 

 

 

 

 

 

o L 'Hrcz>‘t-'Leg' - AT Final Chromium Maximum
o NODP Temperature °F) Concentration Attack Hot-Lea Attack Cold-Leg
e (VOF) (ppm) (mils) otr--eg Aftac Appearance
n 578 © 1200 _ 165 610 to 640 8 Heavy witH small voids A -N(‘)'vi‘s ible defiositl
579 1350 ‘ | 190 690 to 815 5 Moderate to heavy No visible deposit
o R : intergranular voids - _ »
583 ' ]350'- | 195 675 to 775 5 Moderate to heavy | " No visible deposit
e T o ' intergranular voids '
580 1500 ~210 615 to 830 12 Heavy intergranlar No visible deposit
o ' : voids o
| 584 ~ 1500 ' _ - 210 955 to 980 15 Heavy intergranular Very Iigh.f'dépc':s it
e voids
585 - 1600 7 240 540 to 810 18 Heavy infergranu‘ar " Metal i'n cold tr..ap
‘ , o voids
TABLE 5.7. EFFECT ON CORROSION DEPTH OF VARYING THE SIZE
AND SHAPE OF INCONEL THERMAL-CONVECTION LOOPS '
Circulated fluid: Nc:F-ZrFA-UF:4 (50-46-4 mole %)
Operating time: 1000 hr
Hot-leg temperature: 1500°F .
_ Length of Length of Maximum Final
Loop Vertical Horizontal AT Depth of Chromium Hot-Leg “
No. Leg Leg (°F) Attack Content Attack
{in.) (in.) {mils) (ppm)
601 8 8 143 10 620 to 680 Heavy
606 8 8 145 10 530 to 675 Heavy
598 15 8 172 11 1030 to 1075 Moderate to .heavy
603 15 8 188 10* 740 to 965 Heavy
600 26 8 220 16.5 980 to 1040 Heavy
605 26 8 218 16 890 to 9215 Ht_ea\fy .
597 s 15 208 10 895 to 1040 Moderate to heavy
o602 15 15 195 13 890 to 960 Heavy
599 15 26 195 15 710 to 1030 Moderate to heavy
614%* 26 17 245 15 835 to 890 Heavy ) .
: *Inr‘t‘h'i's loop:, uryery few boundaries were attacked to a depth of 14 mils. E
**Cofifiéh standard size, shape, and temperature differential. 7 . . 2

92

 
 

1

n

  
  

reduced to 15 by 15 in., at least for screening
experiments. A series of loops of this size will

be operated to get a better statistical picture for

determining whether all the loops should be reduced
to thls size. '

Evoluanon of Control Loop Results

A gradua! increase in depth of attack of the

~standard Inconel thermal-convection loops operated
~with ZrF ,-base mixtures has been noted recently,

and difficulty has been experienced in reproducing
results. As part of a study of this difficulty, three
Inconel loops were filled the same day from the
same batch of fluoride mixture and operated for
500 hr in as nearly an identical manner as possible.
The data obtained from these experiments are
given in Table 5.8. The attack was again deeper
than in previously operated standard loops, and
the data show no reproducibility. It should also
be noted that in each ]oop some cold-leg a’rtack

PERIOD ENDING JUNE 10, 1955

was found, whereas previously the cold legs
appeared to be attack-free.

To - determine whether the increase in attack
could be caused by a leak during operation or by
air blown into the fluoride mixture during filling,
an additional series of three loops was operated.
In one experiment, 1 liter of air was slowly bubbled
through the fluoride mixture in the fill pot before
transfer to the loop. In the second experiment,
the loop was airtight, but it was operated with the
helium atmosphere at atmospheric pressure. In the
third, the helium atmosphere was again at atmos-
pheric pressure, and, in addition, a Swagelok
connection above the loop was loosened slightly
so that there was a slight air leak. The data from
these loops are presented in Table 5.9. A fourth,
standard loop was operated as a control with the
normal helium atmosphere at a pressure of 7 psi.

The data indicate that to get increased attack
from contamination during operation would require

_ TABLE_S.’_B. RESULTS OF OPERATION OF IDENTICAL INCONEL CONTROL LOOPS

 

 

 

. Fi!lifig Maximum Metallographic Notes Average Final
Loop No. Attack Chromium Content
‘ Order (mils) Hot-Leg Attack Cold-Leg Attack (opm)
684 | 3 ' N Moderate to heavy intergranular  Light, general, to a 850
' voids depth of 1 mil
685 2 14 Moderate to heavy intergranular  Very light, general, to - 750
’ - voids ‘ a depth of 1 mil '
] 686 o - 9 Moderate to heavy intergranular  Very light, general, to 770

voids

a 'depfh of 1 mil

 

     

  

Operotlng t:me
L Hot Ieg femperofure 1500 F B

-

500” hr SREERL TR

  

 

 

 

" Maximum Attack

 

 

   

Loop No " Variable Chromium Content
s s T - {mils)
) - ~ SR (pom)
687 10 o 800
- 689 ) 4 t_,ro;Jg oring'mal fluortde mlx’rure ?flfi: A "9- o 950 T
i beFore tronsfer fo loop ‘ ' o ST e
698 : Control {7 psi heliuvm pressure) 10 775

 

93

Average Final

 

k
¢
!
¥

T I T T T

 
 

 

 

ANP PROJECT PROGRESS REPORT

both an air leak and a loss of helium pressure,
‘a combination which does not seem likely to occur.
It also appears that the small amount of air that
could be trapped in a transfer line would be enough
to cause difficulty., One other possible source of
the increased attack may be the cleaning cycle.
Therefore, a group of loops cleaned by various
procedures have been operated, but the results
‘are not yet available, .

| 'G'ENERIAL CORROSION STUDIES

_ E. E. Hoffman 7
W H. Cook _ C. F. Leitten, Jr.
L Meta”urgy DlVlSlon

Bruzmg Alloys on Type 310 Stainless Steel
and "A" Nickel in Sodium and in Fuel Mixtures

quzmg alloys ‘'submitted by the Wall Colmonoy
Corporation have been tested in both static sodium

and in static NaF-ZrF ,-UF, (53.5-40-6.5 mole %).

" These tests were conducted in an effort to find

a brazing alloy that has good corrosion resistance
to both mediums.

The tests were conducted on type 310 stainless
steel T-joints brazed with the brazing alloys listed
in Table 5.10, which also presents the resulits of
exposure to static sodium at 1500°F for 100 hr.
The brazing alloys are listed in order of decreasing

NICKEL PLATE

 

corrosion resistance. The results for tests of
these alloys in static NaF-ZrF ,-UF, (53.5-40-6.5
mole %) are given in Table 5.11, where, as in
Table 5.10, the brazing alloys are listed in order
of decreasing corrosion resistance.

A comparison of the results in Tables 5.10 and
5.11 indicates that brazing alloy B-13 (9% Si-
2.5% P-88.5% Ni) has the best corrosion re-
sistance to both test mediums. Metallographic
examination showed no attack along the surface
of the braze fillet when tested in sodium (Fig. 5.9a).
The cracks shown in Fig. 5.9z are not the results
of corrosion, but rather were caused by the brit-
tleness of the alloy. Similar cracking was" ob-
served in many of the brazed T-joints listed
Tables 5.10 and 5.11. Brazing alloy B-13 is shown
in Fig. 5.95 after being tested in NaF-ZrF ,-UF,
(53.5-40-6.5 mole %) for 100 hr at 1500°F. Surface
attack to a depth of 1 mil can be seen along the
entire brazed fillet. No cracks caon be observed
in this joint.

Several conclusions can be drawn from a study of
the metallographic notes in Tables 5.10 and 5.11.
Brazing alloys containing relatively high per-
centages of phosphorus appear to be inferior in
sodium. However, additions of silicon tend to
improve the corrosion resistance. On the other

hand, brazing alloys having high percentages of

NICKEL PLATE

 Fig. 5.9. Type 310 Stainless Steel T-Joints Brazed with Brazing Alloy B-13 (9 wt % Si, 2.5 wt % P,
88.5 wt % Ni) After Exposure for 100 hr ot 1500°F to () Static Sodium and () Static NuF-ZrF -UF
(53 5-40-6 5 mole %). Etched with aqua regia. 200X. Reduced 26%.

 

 

 

s

 
 

ar

PERIOD ENDING JUNE 10, 1955

TABLE 5.10. BRAZING ALLOYS ON TYPE 310 STAINLESS STEEL
' TESTED IN STATIC SODIUM AT 1500°F FOR 100 hr

 

Alloy Alloy Composition

 

Weight Change
. . : —_— tall hi t
Designation* (wt %) o) ) Metallographic Notes
B-13 9 Si—-2.5 P—-88.5 Ni 0 0 ' No attack along surface of fillet;
. several cracks in fillet
P-11 10.5 Si~7.5 Mn—82 Ni ~ —0.,0002 -0.018 No attack along surface of fillet;
_ : : several cracks in fillet
P-12 9 Si-15 Mn=76 Ni -0.0003 ~0,033 No attack along surface of fillet;
- : several cracks in fillet
P-10 _ . 16.27 Si~5.9 Mn-77.83 Ni +0.0004 +0.043 No evidence of attack; several
' ‘ large cracks throughout fillet
P-13 7.5 $i-22.5 Mn-70 Ni -0.0008 -0.085 Erratic surface attack to a depth of
. o 0.5 mil; several cracks in fillet
B-12 © 10.5 Si—1.25 P—88.25 Ni 0 0 Erratic surface attack along fillet
S i ; . o to a depth of 1 mil
B-14 7.5 Si—-3.75 P—88.75 Ni -0.0001 ~0.01 Uniform surface attack along fillet
. ‘ _ ' S to a depth of 1 mil
P-14 6 Si—30 Mn~64 Ni —0.0002 -0.022 Erratic surface attack to a depth of
' 1 mil; large cracks throughout
o S : R Lo fillet '
G-20 9 P-1 ].49 W-.79.51 Ni -0.0012  -0.14 Maximum attack of 4 mils along
- - : ' ' surface of fillet -
$-10 14, 31 Cr—9.34 $i-2.56 0 0 Subsurface voids to a depth of 3
T Mo=19.32 Fe-54.37 N: mils along fillet surface
B-15 6 Sl—5 P 89 Ni ' 0 0 Maximum attack of 4 mils along
L N S : surface of fillet
B-16 4.5 51—6.25 P—~89.25 Ni ~0.0001 ~0.01 Small subsurface voids to a depth
o o L foT . of 4 mils
G-21 ' 10.98 P—-6.16 W—82.86 Ni ~0.0001 -0.0M Uniform surface attack ciong entire
o o o : fillet to a depth of 4 mils
B-17 38i-7.5 P-89.5 Ni -0.0012 -0.14 Maximum ettack of 4 mils along
CL L o : oo surface of fitlet
L-20 38 Ni-5 Cr-57 Mn —-0.0004 —-0.043 Attack in the form of stringers to a
v . maximum depth of 5 mils; not
uniform

 

*quzing alloys listed in order of decreasing corrosion resistance to sodium.

 
      
 
 
  
   
   

 

CEpT fncke|

  

_’_suhcon in the absence of phosphorus tend to ‘be
mah‘acked in the fluoride mixture, '

 

 

the same bruzmg ollcy after bemg ‘tested

- ';Flg 5,108

No qflfcck can be seen’ along the

  

 

e conducted on

    

  

S V'hove' fair corrosion resnstance fo bofh medla are '..:a hot-zone femperafure&of 1500°F A femperature
" the 90 wt % Ni<10 wt % P alloy, the 8O wt % Ni— dlfferenhal ‘of ‘about 400°F was maintained in all

10 wt % Cr~10 wt % P alloy, and Nicrobraz. A 1.5-  the seesaw tests,

mil layer of small subsurface voids can be seen in All the brazing alloys listed in Table 5.14 had

95

_-|0|m‘ brazed with Nlcrobraz and tested in sodlum,

NaF;ZrF'--UF4 (53.5-40-6.5 mole %) is shown in

and the reéults are glven -

pbged ifor 900 hr ot

i

 
o i

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 5.11. BRAZING ALLOYS ON TYPE 310 STAINLESS STEEL TESTED
IN STATIC NaF-ZrF ~-UF , (53.5-40-6.5 mole %) AT 1500°F FOR 100 hr

 

Alloy Alloy Composition

 

Weight Change
. : ~ Metallographic Notes
Designation* (wt %) (g) (%)
G-20 9 P-11.49 W~79.51 Ni -0.0017 -0.152 No attack on surface of fillet;-
- ' : several cracks in fillet
B-15 6 Si—-5 P—-89 Ni —~0.0046 -0.510 Uniform surface attack along fillet
o _ _ ’ to a depth of 0.5 mil
B-16 4.5 Si—6.25 P-89.25 Ni -0.0057 ~0.562 Surface attack along fillet to a
s A . depth of 0.5 mil
G-21 10.98 P-6.16 W—-82.86 Ni -0.0061 ~0.674 Erratic surface attack to a depth of
e 0.5 mil along fillet
B-17 3 8i-7.5 P-89.5 Ni —~0.0053 —-0.516 Surface attack along fillet to a
' Cee ; depth of 0.5 mil; several cracks
L e e _ in fillet . _ ‘
B-13 ' 9Si-2.5 P—88.5 Ni —0.0027 —0.279  Surface attack along entire fillet to
R o a depth of 1 mil
P-12  95Si-15 Mn—76 Ni ~0.0034 ~0.331 Surface attack in form of small
. , ‘ _ ' voids to a depth of 1 mil; several
e - - cracks in fillet ‘ 7
B-14 7.5 Si-3.75 P~88.75 Ni —~0.0052 ~0.479 Surface attack along entire fillet to
C : : ~a depth of 1 mil
P-11 10.5 Si=7.5 Mn~-82 Ni -0.0042 ~0.401 Surface attack to a depth of 1.5
, mils; several cracks in fillet
B-12 10.5 5i-1,25 P-88.25 Ni -0.0018 -0.171 Surface attack to a depth of 2 mils;
large cracks throughout fillet
P-10 16.27 Si~3.9 Mn—-77.83 Ni —~0.0026 -0, 290 Uniform surface attack along fillet
to a depth of 3 mils
S-10 14.31 Cr—9.34 Si-2.66 —0.0030 -0.344 Surface attack to @ maximum depth

Mo—19.32 Fe-54.37 Ni

of 7.5 mils along entire fillet

P-13 7.5 5i~22.5 Mn=70 Ni —0.6043 -0.455 Complete attack of entire fillet
L-20 38 Ni—5 Cr—=57 Mn -0.0078 -0.74 Complete attack of entire fillet
P-14 6 S$i-30 Mn—64 Ni -0.0099 -1.06 Complete attack of entire fillet

 

*Brazing alloys listed in order of decreasing corrosion resistance to the fluoride mixture.
g Y g

fair corrosion resistance to both sodium and the
fluoride mixture, with the possible exception of
the 65 wt % Ni~25 wt % Ge~10 wi % Cr alloy. The
Coast Metals No. 52 alloy had good corrosion
resistance to the fluoride mixture in the seesaw
test in contrast to its poor resistance in the static
test, A large degree of porosity was present in

Coast Metals No. 50 alloy that hampered the

evaluation of the corrosion data. These two
brazing alloys will be retested in order to check
the results of these seesaw tests,

| Scféénfng Tests of Solid-Phase Bonding

. The a”bys which have been proposed for use in

the fabrication of the ANP reactor fuel and coolant
lines have a tendency to form a solid-phase bond
at elevated temperatures in liquid metals and in

96

fused fluoride salts. Bonding of this type makes

them unsuitable for use in valves, bearings, and

seals; thus cermets and ceramics that do not form
a solid-phase bond are being investigated. The
cermets (metal-bonded ceramics) appear to be the
more promising because of their high corrosion

resistance and other chemical and physical

characteristics.

The test apparatus being used is shown in
Fig. 5.11. This apparatus was adapted from that
designed for lever-arm stress-rupture tests.® The
container for the central components, for the test
specimens, and for the liquid metal or fused fluo-
ride salt serves as a housing for the compression
tube and upper platen. By using this arrangement

 

4R. B. Oliver et al., ANP Quar. Prog. Rep.. Mar. 10,

1955, ORNL"‘864, Fig- 702, P 105,

el iac i

 

 

 
 

°%

O

PERIOD ENDING JUNE 10, 1955

TABLE 5.12, BRAZING ALLOYS ON *‘A" NICKEL T-JOINTS TESTED
IN STATIC NaF-ZrF ~UF (53,5-40-6.5 mole %) AT 1500°F FOR 100 hr

 

Alloy* Composition

Weight Change

(g) (%)

Metallographic Notes

 

(wt %)
80 Ni-10 Cr—-10 P 0 0 No attack on braze fillet
0 0 No attack on braze fillet

50 Ni—-25 Mo—-25 Ge

Nicrobraz

69 Ni-5 B~15 Cr~5 Si~5 Fe-1C

Electroless nickel

90 Ni—10 P
65 Ni—25 Ge—10 Cr

Coast Metals No. 52
90 Ni—-4 B—4 Si-2 Fe

General El..ectric No. 81
66 Ni—10 $i—19 Cr—4 Fe—1 Mn

35 Ni-=55 Mn=10 Cr
60 Mn—40 Ni
68 Ni—32 Sn

-0.0004  -0.016

-0.0004 -0.013

-0.0014  ~0.05
-~0.0003  —0.012

—-0.0111 ~0.48
—-0.0159 -0.59
—0.0998 ~3.49

Mo attack on braze fillet
No attack on braze fillet
Small subsurface voids to a depth of 0.5 mil along

braze fillet

Nonuniform attack to a depth of 6 mils along braze
fillet

Nonuniform attack to a depth of 12 mils along fillet

Complete attack of braze fillet
Complete attack of braze fillet

Joint partially dissolved ot fillet surface

 

*Brazing alloys listed in order of decreasing corrosion resistance to the fluoride mixture,

TABLE 5,13, BRAZING ALLOYS ON *‘A” NICKEL T-JOINTS TESTED

IN STATIC SODIUM AT 1500°F FOR 100 hr

 

Brazing
Alloy* Composition
(wt %)

Weight Change

Metallographic Notes

 

Electroless nickel

90 Ni-10 P

Coast Metals No. 52
90 Ni-—4 8—4 Sl 2 Fe

. 80 Nl-]O Cr—-'IO P
. Gen‘erdl E'[ecfr'lc No 8]
66 Ni-'lO Sl-'|9 Cr--4 Fe—l Mn

"N:crobraz &
~© 69 Ni=5 B-15 Cr-5 5i-5 Fe-1C
[ '50 Nl-25 Mo-25 Ge' -

o 65 Nl_zs c;e..lo Cr

e Mn—40 N. -

s unelo

68 Ni-32 Sn

-0.0171

(g) (%)
~0.0004  —0.018
~0.0019  —0.068
’"°-°.°.°'.‘,_7 ~0.08

S
~0.0022  -0.082
_0.0009  —0.036
\_ ‘"Q‘-__OZ_O_'#-' | ':—o 085
*00005 | _’%0;620 |
~0.540

No attack along fillet surface
Surf;:!ce attack along fillet to a depth of 0.5
mi

Nonuniform attack along fillet to a depth of 1
mll

Attack on surface of hllef to a depth of 1 mil
1.5 mil layer of small subsurface voids along
Tillet edge

Surf?ce aftock along fallet to a depth of 2.5
mils

Unlform surface attack along flllef to a depth
of 3 mlls

Unlform qtrqck ulong entire fliiet to g depth of
9 mlls

Small voids in from surfuce' of fillet to a depth
of 13 mils

Complete attack of whole fillet

 

*Brazing alloys listed in order of decreasing corrosion resistance to sodium.

97

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 5. 14, BRAZING ALLOYS ON INCONEL T~JOINTS EXPOSED IN SEESAW
APPARATUS TO SODIUM AND TO NaF-ZrF UF4 (53,5-406.5 mole %) FOR 100 hr AT 1500° F

 

' Bruzing Alloy*
Composition

_ Bath Materiol
vt %) _

Weight Chonge

Metallographic Notes

(9) (%)

 

" Coast Méfals No.-:'52‘ _ Fluoride mixture
89 Ni-5 Si~4 B-2 Fe

7 Sodium
Low-melhng 'I\:Ii.crol':raz V Fluoride ;.nixture
- 80 Ni-5 Cr-'-_6 Fe—3 B-5Si-1C
Tl Sodium
.Céésf"Méfa'].s l\:.lo.'SO ' Fluoride mixture
] __93 Ni-3.5 $i-2.5 B-1 Fe
Coadln » Sodium
70 Ni--r-}lr;l_ Cr—6 Si—13 Ge .Fluoride mixture
Sodium
Nicrobraz Fluoride mixture
70 Ni—14 Cr—6 Fe~-5 B-4 Si--1 C
Sodium
65 Ni—-25 Ge—10 Cr Fluoride mixture
Sodium

~0.0008 -0.052 N.Vonu‘n‘ifo.rm‘ .s'firfcce'fif.foé-k '
along fillet to a depth of 0.5
mil

~0.0011  ~0.073  No attack along surface of fillet

~0.0008 ~0.063  Nonuniform surface attack to a
depth of 0.5 mil along fillet
Subsurface voids to @ maximum

depth of 1.5 mils along surface
: of flllef

—0.0007 —0.051

~0.0014 ~ —0.085  Uniform surfuce aficlck ulong
f:!lef to a depth of 0.5 mil

~0.0012  ~0.077 'Very erratic surface attack olong
' co flllef to a depth of '| 5 mlls
—0.0011  ~0.067 Nonumform attack to a dep’rh of
: 1.5 mils along surface of fillet

 —0.0023  —0.139 Nonuniform attack along surface

of fillet to a depth of 2.5 mlls
~0.0005 —0.030 Errahc surfcce attack olon?
fillet to a depth of 1.5 mi

Very erratic stringer attack to a
maximum depth of 4 mils along
surface of fillet

~0,0010 —0.056  Stringer-type attack to a maxi-
mum depth of 4 mils in a few
localized areas

-0.0019 -~0.113 intermittent surface attack to a
maximum depth of 4 mils along

fillet

 

*Brazing alloys listed in order of decreasing corrosion resistance to both test mediums.

the test specimens can be brought into mutual
compression between the platens,
Corrosion-resistant cermets and hard-facing alloys
have been tested in an apparatus fabricated from
Inconel for solid-phase bonding in NaF-ZrF ,-UF,
(53.5-40-6.5 mole %) for 100 hr at 1500°F. The
test specimens were dimensionally the same
(+0.0002 in.) and the contacting surfaces had

roughnesses of less than 10 pgin. The test was

. begun by holding the contacting surfaces of the
- test specimens apart while the apparatus was

 flushed with NaF-ZrF -UF, (53.5-40-6.5 mole %)

at 1500°F This was done to ensure that any
surfczce films on the specimens would be removed
prior to the application of the compression load.

A fresh charge of NaF-Z¢F 4-UF, (53.5-40-6.5
"mole %) at 1500°F was ’rhen put into the test

. 98

150,000 psi.

chamber, and the test specimens were pressed
together at the desired contact pressure. At the
conclusion of the 100-hr test period, the NaF-ZrF -
UF, was removed, and the system was cooled to
room temperature and disassembled. '

The contact surfaces of the test specimens were
examined with a low-power microscope for signs
of solid-phase bonding. The seating was not
perfectly uniform in any of the tests, and therefore
the contact pressures between the test specimens
were probably in excess of the calculated values
in certain areas. Table 5.15 is a summary of the
solid-phase-bonding screening tests made at calcu-
lated contact pressures of 6,600, 10,000, and
In the most severe tests, that is,
those in which the calculated contact pressure
was 50,000 psi, no bonding was observed, except

 

 

L
!
i
M
5,
i
i
i
5
g
i
g
i

 
for a slight amount in the tests of K150A vs K152B
- and K162B vs K162B. Since uniform contact was
extremely difficult to obtain, the slight bonding
that occurred in these tests may have been caused

DR
‘ UNCLA_ssu;;ED

Y-14

0.01

PERIOD ENDING JUNE 10, 1955

by contact pressures in excess of 50,000 psi.
This explanation seems even more reasonable
when it is considered that the quantity and compo-
sition of the binder metal of the test specimens,

UNCL ASSIFIED
Y.15144

 

BNICKEL PLATE

 

0.03

 

0.04

 

 

 

Fig. 5.10. *“A” Nickel T-Joints Brazed with

ay

() aqua regia, 100X. Reduced 24%.

TABLE 5.15.

%)

Compositions:

 

Nicrobraz Afief
Static Sodium and (b) Static NaF-ZrF,-UF, (53.5-40-6.5 mole %).

 

 

Exposure for 100 hr at 1500°F to ()
Etched with (a) 10% oxalic acid and

RESULT'S'OF SOL'!D-PHASE'-BON.DING SCREENING TESTS OF VARIOUS CERMETS AND ALLOYS
EXPOSED TO NuF-ZrF +UF, (53.5-40-6.5 mole %) AT 1500°F FOR 100 hr AT VARIOUS CONTACT PRESSURES

K150A (80 wf % TiC=10 wt % NbTaTnC -10 wt % Ni)
K151A (70 wt % TiC—10 wt % NbToT|C ~20 wt % Ni)
_K'|52B (64 wt % T:C 6 wt % NbTuTtC -30 wt % Nl)

 

 

 

   
    

 

 

£

     

50,000 KI50A vs K150A°

 

 

Contact S Contact
‘Pressure on Pressure Contacting Specimens’ Results*
(psh) (psi)
50,000 K150A vs K151A No bonding
K150A vs K152B Some bonding
K150 A vs K162B No bonding
K151A vs K151A No bonding
K151A vs K152B No bonding
K151A vs K162B No bonding
K152B vs K152B Ne bonding
K152B vs K162B Ne bonding

K162B vs K1628 Some bonding

 

 

*There was inconsistent bonding of test specimens to the supporting Inconel platens in all tests.

99

 
0oL

 

 

 

PARTIALLY ASSEMBLED

DISASSEMBLED

 

   

 

 

 

Fig. 5.11. Apparatus for Solid-Phase-Bonding Screening Tests.

COMPRESSION TUBE

PULL ROD

UPPER PLATEN

TEST SPECIMENS

LOWER PLATEN ——— 7]

 

    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

UNCLASSIFIED

12:{ LEVER

BELLOWS

HELIUM ATMOSPHERE IS
MAINTAINED ABOVE LIQUID
METAL. ( ATMOSPHERE
ATTACHMENTS NOT SHOWN )

 

 

~—— MUFFLE OF
ELECTRIC FURNAGE

ORNL-LR-DWG 7514

 

 

 

 

 

 

LEVER LOAD—/

\LIQUID METAL LEVEL

 

 

140d I SSTYD0¥Ud L1I3rodd dNY

 

 

 

 

 
 

 

>

90 .

particularly that for K150A vs K152B, show no
apparent relationship to the occurrence of the
bonding. On the basis of these datq, it is believed
that none of these Kentaniums would form solid-
phase bonds to each other in NaF-ZrF,-UF,

(53.5-40-6.5 mole %) at 1500°F in 100 hr if the

true contact pressure between the cermets did not
exceed 50,000 psi. A recheck of the pairs that
did bond is planned that will conclude the screen
testing of the Kentaniums, K150A, K151A, K152B,
and K162B, for solid-phase bonding in all contact
combinations with each other in NaF-ZrF4-UF4
(53.5-40-6.5 mole %) at 1500°F for 100 hr,

Sodium in Inconel

A 1000-hr test was recently completed in which
sodium was circulated in an Inconel thermai-
convection loop. Samples of the sodium used in
this test were analyzed and found to contain
approximately 0.03% oxygen. The oxygen content
was admittedly high and could have been reduced

   

Fig. 5.12. -‘Sfie;:iméns of Héf and Cold Légs of an Inconel Thermal-Convection Loop After Ciréulating

HOT LEG ( 1500° F)

" PERIOD ENDING JUNE 10, 1955

by a factor of 10 by adequate cold trapping;
however, this test did show how serious that mass
transfer can be even in a thermal-convection loop
when the oxygen content is high. The hot leg of
the loop was held at a temperature of 1500°F, and
the coldest section of the cold leg was held at
1200°F.  The mass transfer that occurred was
concentrated in an area in the cold leg where an
air blast had impinged on the tube wall during the
test, as shown in Fig. 5.12. Very little mass
transfer was detected in other sections of the cold
leg. A similar test will be conducted in which the
oxygen content of the sodium will be held below
0.005% to determine the effect the oxygen content
has on mass transfer in a sodium-Inconel system.

Lithium in Type 347 Stainless Steel

Lithium was circulated in two stainless steel
thermal-convection loops for periods of 1000 and
3000 hr, respectively. The hot- and cold-leg
temperatures were 1000 and 550°F, respectively.

UNCLASSIFIED
Y-15138

Sodium for 1000 hr. Crystal deposition resulting from mass transfer may be seen in the cold leg.

101

 
G

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

The loops operated satisfactorily during the test
periods, and macroscopic examination revealed
no mass-transferred crystals in the loops or in the
lithium drained from the loops.

- The hot zone of the loop that operated for 1000 hr
had subsurface voids and a ferritic surface layer
to a depth of 0.3 to 1.0 mil. Lithium metal had
penetrated to this depth. The weld zone of the
pipe in this area was attacked to o depth of 3 mils.
The cold leg of the loop was unattacked; however,
there were a few small (0.2 mil) crystals attached
to the surfdce, Similar crystals were previously
identified as carbides.

‘The attack in the hot leg of the loop that oper-
ated for 3000 hr was similar to that found in the
1000-hr test specimen; however, the spongy ferritic
surface layer was 1.0 to 1.5 mils thick, as shown
in Fig. 5.13. The attack in the weld zone wos
intergranular and extended to a depth of 4 to 5 mils,

~as shown in Fig. 5.14. The preferential attack in

o

        

the weld zone is due to attack of the grain boundary
carbides by the molten lithium, A deep groove in
the pipe wall may be seen at the weld zone—parent
metal interface.

The cold-leg section (Fig. 5.13) of this loop was
very similar in appearance to that of the loop oper-
ated for 1000 hr. A few small (0.2 mil) carbide
particles were attached to the wall of the tube.

It is believed that the corrosion resistance of
type 347 stainless steel to lithium in this tempera-
ture range would be improved by lowering the
carbon content of the steel and by using seamless
instead of welded pipe. The austenite-to-ferrite
transformation detected in the hot legs of these
loops is attributed to leaching of nickel from the
type 347 stainless steel by the lithium,

Versene Cleaning of Beryllium-Inconel Systems

Versene has been proposed as the cleaning agent
for the ART, and therefore the corrosion resistance

 

"

       

 

 

   

 

 

- . .‘ ?fi&??\g: ;.J.‘ .. ‘ o )
E o UNCLASSIFIED
! NICKEL PLA‘TV , 001 . NIKEL AE Y-14937
T
Q
=
0.02
0.03
0.04

 

 

 

 

 

 

- ';Fig.; 513. "Hot- and Cold-Leg Surfaces of Type 347 Stainless Steel Thermal-Convecfib.fi..Loéb ‘After

Circulating Lithium for 3000 hr at a Hot-Leg Temperature of 1000°F and a Cold-Leg Temperature of
550°F. Note austenite-to-ferrite phase transformations which occurred on hot-leg surface (a) and small

o

" crystals deposited on cold-leg surface (b). Specimens nickel plated after testing. Etched with a_qpc
.- regia. 1000X.

 

 

 

 
 

 

PERIOD ENDING JUNE 10, 1955

S UNCLASSIFIED
Y-14933

 

 

 

S0°0
00
0
=z
O
€I

20’0
YO'O

 

 

 

 

 

Fig. 5. 4. lhei;ie wellfief Hef-Leé Wel“l'.l’ from Type 347 S-t(.:.i-niess Steel Therlr-;fi‘a]‘ (::oev'ecfio'n Loop. After
Clrculutlng Lithium for 3000 hr ot a Hot-Leg Temperature of 1000°F, Note deep (4 mils) attack in weld

zone as compcred wfih 1- m|| of’rock on parem‘ mefol

   

of berylhum 1rodync:mlc Versene was stuched Thefi"s

testing procedure was designed to simulate the

~cleaning procedure used in the ARE, whlch was

clecned wn'rh a Versene solution.

 
 
 
 
 

 

ple was placed within the

insert was 3/ in. OD, / in. 1D, and 3 in. in length.

A ]% Versene‘ soluhon was pumped fhrough each'_

 

pump. " Disodium’ versoncn‘e was used, and the
_concentrafion’ was determined on a welghf-vo[ume

basis. The fesfmg temperature for both loops was
180°F, and the Versene was circulated in each for

a period of 24 hr.

comp!efed with’ fwo loops ‘con-
2| ‘with a berylhum msen‘. “The
_ I : mloop‘-'fl
between ~ two 5|ml|ar lnconel inserts whlch were__jf:_"
held in place by crimpmg’ The tubular berylhum):"'

Efched with qqua regla. TOOOX

Upon complehon of the flrs‘r ‘test fhe Versene
was drained while the loop was still at test
temperature. When the loop was sectioned, a small
amount of Versene was found in a static region of

" the loop located in the annular gap between the
_beryll_:um_ insert and the Inconel sleeve. No trace
of retained Versene was found in this section of

the second |oop, ‘which’ wcs flushed ‘with dls’nlled

““water after the Versene’ was drained.

Macroscoplc examination showed the berylhum
inserts in both loops to be quite similar. Each

insert’ had retained its orlgmal polished ap-

pearance. No effect of the Versene solution was

"found by macroscopic or metaliographic exami-

nation of the inner surface of the beryllium insert
or the Inconel in either test. The outer surfaces
of the beryllium inserts, however, were attacked

103

 

 
 

 

 

 

~ANP PROJECT PROGRESS REPORT

by the Versene solution. The attack was in the
form of erratic pits that varied in depth from 0.5
to 2.5 mils,

UNCLASSIFIED
Y-15010

 

INCH
1
[ ] | J

Fig. 5.15. As-received {(a¢) and As-tested (%)
Beryllium 1Insert from Second Inconel Loop in
Which Versene Cleaning Solution Was Circulated
at 180°F at a Rate of 2 gpm.

I
INCH

0.02

 

003

 

0.04

 

 

 

 

The erratic attack of the Versene on the outer
surface of the beryllium insert used in the second
test is shown macroscopically in Fig. 5.15. The
unattacked inner surface and the erratically at-
tacked outer surface of the beryllium insert used
in the first test are shown in Fig. 5.16. The
2.5-mil outer surface attack shown in Fig. 5.16
is representative of that found by metallographic
examination of the beryllium inserts used in both
tests. In each test the beryllium insert lost
weight, as shown in the following tabulation.

Test No. 1 . Test No. 2
Original weight, g 16.8715 16.7970
Final weight, g 16,8680 16.7941
Weight loss
9 0.0035 0.0029
% 0.021 0.017

A chemical analysis of a portion of the Versene
solution used in the second test revealed a beryl-
fium concentration of 0.0024 mg/ml. Since 1.5
liters of 1% Versene solution was used in this
test, the total amount of beryllium in the solution
was 3.8 mg, which agrees fairly well with the
2.9-mg weight loss of the beryllium insert used
in this test.

UNCLASSIFIED
Y-14755

 

 

.V'Fig. 516 " Inner (@) and Outer (b) Surfaces of Beryllium Insert From First Inconel Loop in WHich
Versene Cleaning Solution Was Circulated at 180°F at o Rate of 2 gpm. Etched with oxalic acid. 100X.

R‘edu ced _24%.‘

104

 

 

 

 

 

 
 

a)

n

"

 
   
    
 
 
 
 
   

FUNDAMENTAL CORROSION RESEARCH

G. P. Smith
Metallurgy Division

Film Formation on Metals

Jo Vo CthCGrf
Metallurgy Division

Most of the work done in the past on the oxida-
tion of metals has been devoted to a study of the
oxidation characteristics of the ‘‘heavy’’ metals
and alloys such as copper, nickel, iron, aluminum,
the stainless steels, etc, These materials either
were structurally important or their physical proper-

‘ties were such as to make them particularly amen-

able to oxidation studies. Comparable investigations
of the alkali and alkaline-earth metals are almost
entirely lacking.

Rather elaborate oxidation theories have been
devised for the heavier metals, but it has been

assumed, in general, that the alkali and alkaline-

earth metals exhibit a linear oxidation rate, in
accordance with the old crack theory of Pilling and
Bedworth.®  Their theory was that a metal, for
which the ratio of the density of the oxide to that
of the parent metal is greater than unity, should
obey a linear oxidation law. It was reasoned that
the oxide film formed would be highly subject to
cracking and that, consequently, there would be a
constant re-exposure of fresh metal surface to
oxygen. Thus the rate of oxidation should be
independent of the thickness of the oxide film, the
quantity of oxide formed being directly proportional
to the time of exposure to oxygen,

In order to test the crack theory of oxidation
directly, as well as to obtain experimental data on

the oxidation of the alkali metals, an mveshgoflon:
. of the omdcfl;on charccterlstlcs of sodium was
'_‘;undertaken. Sodiu

 

h'near oxmfcmon_ _rcfe. "

 

 

   

  
   
    

e tho 'oughi i

 

SN. B. Pilling and R. E. Bedworth, J. Inst. Metals 29,

- 529 (1923); see also, U, R. Evans, Metallic Corrosion,

Passivity, and Protection, Longmans, Green and Co.,
1948, New York, p 102,

 

! mary experlmenfs mdicateJ fhc’r?

estlgo’red ‘

PERIOD ENDING JUNE 10, 1955

oxidation mechanisms of copper, aluminum, and
other structural metals.

On the basis of the results obtained with sodium,
it was concluded that a fundamental error existed
in the currently accepted oxidation concepts for
metals that obey a linear oxidation rate law and
that it was therefore desirable to investigate care-
fully the oxidation of some metal that does exhibit
a linear oxidation rate, Columbium was especially
suitable for this study, Below approximately 400°C,
it obeys a parabolic oxidation rate law, while above
this temperature its oxidation rate is linear. At
400°C the initial stages of oxidation also appear
to follow a parabolic rate, but after several hours

" of oxidation the rate increases and becomes almost

linear.® Thus a careful investigation of the structure
and composition of the oxide films formed at or
near 400°C should provide valuable information as
to the conditions which lead to a linear oxidation
rate.

An added incentive for the study of the oxidation
properties of columbium was that this metal pos-
sesses very desirable high-temperature structural
properties, Its use has been limited largely by its
excessive oxidation at high temperatures, It was
believed that further information concerning its
oxidation mechanism would be helpful in overcoming
this defect.

The experimental procedures used and the results
of the studies of sodium and columblum oxidation
are presented below, 7

Sodium Oxidation. The oxidation rate of sodium
was followed by measuring the change in pressure
in a closed reaction chamber as the reaction pro-
ceeded. A sensitive, differential manometer in
which Octoil-S diffusion pump oil was used as the
manometric fluid served as the pressure-sensing

iy 'I“f.llls complefely the PlH,ng-*-f-:,,_q.dewce.

“and Bedworth crl'rerlon'ifbr a’ metal‘thcf exhlbl'rs a\
Thls stpdy was initiated

The apporcx‘rus used is shown in Fig. 5.17. The

j"‘lsodlum reservoir consisted of a glass bulb closed
~at one end with a thln-walled break-off tip. The
" ‘teéservoir was filled, under vacuum, through a side
“arm, with sodium that had been purified by repeated
“vacuum dnsflllahons at a pressure of 1078 to 1077
“mm Hg. Affer the Slde ‘arm was removed, the
‘reservoir was atfcched to a tube below fhe oxndc-

tion bulb as shown in Fig. 5.17. A small furnace
/a _hlfch__‘could be plcced around the entire
) the sodium reservoir and the

   

 

6H. Inouye, Scaling of Columbium in Air, ORNL-1565
(Sept. 1, 1953). :

105

 
 

 

 

ANP PROJECT PROGRESS REPORT

BALL BEARING —_{ |

BREAK-OFF TIP —{-

2-mm CAPILLARY -]

 

 

 

»

"TO VACUUM
PUMP

OXIDATION BULB

 

" UNCLASSIFIED
ORNL-LR-DWG 7416

   
     
     
   
 
 

 

 

 

 

 

 
 

ocToIL "s"
DIFFERENTIAL MANOMETER

L BALL BEARING

2T /BREAK~OFF TIP

 

PURIFIED SODIUM

| Fig. 5.17. Sodium Oxidation Testing Apparatus.

. manometer protruding from suitable holes. The
" entire apparatus could thus be baked-out under
vacuum before any sodium was admitted to the
~ system. Care was taken to load the manometer by
the vacuum distillation of QOctoil-S from a separate
bulb into the manometer arms. This procedure

106

assured that the manometric fluid would not act
as a source of appreciable gas, '

After a bake-out period of 16 to 20 hr the system
was allowed to cool to room temperature, and the
break-off tip above the sodium reservoir was
crushed. The sodium was then distilled into the

 

 
 

 

H

"

 

tube below the oxidation bulb, and the sodium
reservoir was removed, Finally, the sodium was
distilled into the oxidation bulb and allowed to
condense on the walls, Both these final distil-
lations were made under a pressure of approxi-
mately 2 x 1077 mm Hg. As the last step before
the admission of oxygen, the tube below the oxida-
tion bulb was sealed off, as were the vacuum leads
to the two reference bulbs,

A 50-cc flask served as an oxygen reservoir.
Prior to being attached to the oxidation apparatus,
it was filled with carefully purified oxygen to a
pressure such that when the gos was allowed to
expand into the reference and oxidation bulbs the
resultant pressure was 200 mm Hg. The purification
of the oxygen was accomplished by passing it over
hot copper oxide (to remove hydrogen) and Ascarite
(to remove carbon dioxide) and finally drying it
over magnesium perchlorate and in a liquid-nitrogen
trap. - ' '

The tube leading from the oxygen reservoir to the
break-off tip was made of 2-mm capillary tubing.
By thus minimizing the volume above the reservoir,
it was possible in the low-temperature runs to im-
merse the reservoir in a cold bath to precool the
oxygen to the desired reaction temperature.

When the break-off tip above the oxygen reservoir
was smashed, the oxygen passed through the T
connection above the manometer and simultaneously
filled the reference and oxidation bulbs to identical
pressures. The infersection of the tubes of the T
was then collapsed as quickly as possible with a
hand torch, The three arms of the manometer were
thus separated and any change in pressure in the

oxndo’r!on bulb resultlng from the reccflon of sodlum"

 

25, 35, c:_nd 48°C, an apparatus identical to that

 

     
 
 
 
 
 
 
  
 
 
 
  
  

_ neds "remenfs have been'made at
five temperatures: —79, —20, 25, 35, and 48°C, At

PERIOD ENDING JUNE 10, 1955

shown in Fig. 5.17 was used, except that only one
reference bulb was found to be necessary. For
reasons described below, the second reference
bulb proved very helpful, however, for experiments
at the two lower temperatures.

For all experiments the oxidation and reference
bulbs were immersed in a constant-temperature bath
contained in a 4-liter Dewar flask. Mineral oil
served as the bath liquid for the three higher
temperatures. At -20°C a saturated sodium
chloride—ice bath proved to be satisfactory, and
at —=79°C a slurry consisting of powdered dry ice
and a 50-50 vol % solution of carbon tetrachloride
and chloroform was used to attain the desired
temperature. The densities of the liquid phase and
the dry ice in the latter bath were approximately
equal, and thus a slurry was produced in which
there was little tendency for the segregation of the
dry-ice particles,

At —79°C the sodium oxidized at a very slow
rate. The maximum pressure change which oc-
curred in the system as a result of the oxidation
was 0.2 to 0.3 mm Hg. Therefore it was necessary
to pay particularly careful attention to factors
which could cause spurious pressure readings.
One obvious source of error was a temperature dif-
ference between the reference and oxidation bulbs.
This difficulty was overcome by utilizing a relo-
tively thin slurry in the cold bath and stitring it
very vigorously, A less easily corrected source of
trouble was that, of necessity, the reference and
oxidation bulbs were maintained at dry-ice temper-
ature, whereas the manometer and connecting tubes

 were at room temperature, A simple gas law calcu-

lohonxshows thof, in such a system, any change in
perature of elther the cold bath or the
Ster will produce, in generol a difference in
presswo “between the oxndohon bu[b cnd the refer-

 

 

ence bulb A specnol cose ‘that is an excep’r:on to
th:s rule occurs when the volumes of gas at the
fwo tempero’rures ‘are |denhco| on both s:des of the o

 

he mcgmtude of the pressure dif-
ference is ‘dependenf both on fhe ratio of the volume
Tof ¢ gas “held af 'rhe two temperature extremes “and on

thé dlfference in femperoture. Thus, this effect

  

of no grecn‘ |mportance for experlmenfs ‘close

fo room temperafure, bu'r I'|' beccme sngmflcon'r at '

    

"t “was “impossible to maintain exactly equal

 

‘volumes on both sides of the manometer, and the

attainment of precise temperature control in the

107

 

 
i

 

 

 

ANP PROJECT PROGRESS REPORT

cold bath at ~79°C was almost as difficult, The

equilibrium temperature of any dry-ice bath is
- determined by the sublimation point of carbon
: 7_-di:o$<idé'; however, the constancy of the sublimation
| ~ temperature is dependent upon the constancy of the

partial pressure of carbon dioxide above the bath.
Thus, Unless fhe bcfh is contained in a closed

“vessel, variations in temperature will occur in the
“bath because of changes in the partial pressure of
carbon dloxmle obove it, In addition, the negative

hec’r of soluhon of carbon dioxide in carbon tetra-
chleride and chloroform caused the temperature of

the bath to fall below the equilibrium sublimation
" temperature when the cold mixture was first pre-
~ ‘pared. Since the bath was contained in a Dewar

flask, the lowered temperature tended to persist
for long periods of time unless a heater was immersed

- in fhe bcn‘h

It was decided therefore that precise temperature

: 'confro[ of the apparatus at —79°C was impractical.
'However, it was essential that some reference

point be established for the bath so that the ap-
paratus could be brought back to the initial con-
ditions of the experiment after, for example, each
renewal of the dry-ice charge in the bath. The
addition of a second reference bulb proved satis-
factory for this purpose. The two reference bulbs
represented a ‘‘differential gas thermometer.”’ |t
was observed that within about 10 min after the
start of an experiment a small difference in the
heights of the oil in the reference arms of the
manometer was usually established, This dif-
ference remained constant as long as the temper-
ature of the bath and the room temperature were
unchanged,  When deviations in this difference
were observed, it was possible to make slight
alterations in the bath temperature (for example,
by removing part of the cover of the Dewar flask
and thus decreasing the partial pressure of carbon
dioxide above the bath) and thus re-establish the
desired pressure difference in the reference bulbs.

It was recognized that the constancy of the
pressure difference between the two reference
bulbs did not necessarily indicate a corresponding
constancy in the cold-bath temperature. Frequently,
the adjustments in the bath temperature that were

_.necessary to maintain the pressure difference
 represenfed a counterbalance to changes in room
- temperature. The resultant small changes in the

bath femperature meant, of course, a change in the

: femperafure at which oxidation occurred; but, since

108

the oxidation rate of sodium is not a sensitive
function of temperature at —79°C, this alteration
of the experimental conditions was not considered
to be significant. In any event, as may be seen in
Fig. 5.18, the use of the technique described above
produced excellent agreement of the data for three
separate experiments at —79°C,

The experimental data obtained at the five temper-

atures investigated are shown in Fig., 5.18. The
change in pressure in the system, which is ob-
viously a function of the thickness of the oxide
film formed, is plotted against the time of oxidation
on a log-log scale. Each curve represents the
composite of data taken from three or more experi-
ments, R

Perhaps the most striking feature of the experi-
ments was the slowness with which oxidation
occurred. Very roughly, one unit on the ordinate
scale in Fig. 5.18 is equal to a film thickness of
10 A. Thus at =79°C after 1000 min of oxidation,
the thickness of the oxide film was only about 30
A while at 48°C, it was 500 A,

Since the curves, as plotted, were not linear,
there was an indication that the data did not fit
any of the conventional oxidation rate equations of
the form :

x" = kt , n =123

where x is the thickness of the oxide film after
time ¢, and k£ is the rate constant, The data were
also tested against the logarithmic rate equation

x = klogt ,

but no fit was obtained,

As may be seen from the curves in Fig. 5.18, the
oxidation rates of sodium at =79 and -20°C de-
crease very sharply after the first 150 to 200 min
of oxidation. This behavior is qualitatively similar
to that predicted by Cabrera and Mott? for oxidation
processes leading to the formation of very thin
films. The oxide thickness (30 to 40 A) observed
at —~79°C is very close to the thickness range re-
quired by this theory, The data were checked
against the appropriate rate equation

]

- =A-~Blint,

X
where x and t are oxide thickness and time, re-
spectively, and A and B are constants, but no

 

7N. Cabrera and N. F. Mott, Repts. Progr. in Phys.
12, 163 (1949).

T

 

 

 

 

 

 
 

»

-

AP {mm OCTOIL)

2

PERIOD ENDING JUNE 10, 1955

UNCLASSIFIED
ORNL—LR—DWG 7417

 

5 10° o 5 10% 2 5 10°
TIME (min)

Fig. 5.18. Data on Oxidation of Sodium as a Function of Time and Temperature. In apparatus used, a
change in system pressure is a function of the thickness of the oxide film formed.

agreement was found. However, the qualitative
similarity between the observed rate curves and
those predicted by the theory suggests that the
results obtained at —~79 and —20°C represent an
oxidation process in which the mechanism proposed

mechanism. It is hoped to test this idea by making

oxidation measuremenfs ‘af temperatures below .

~79°C.

. .. The very small amount of oxide which was formed -
”"?._:f—raf all femperafures mveshga'red emphasnzed ’rhe'

 

and aluminum. As further details of the oxidation
characteristics of sodium-and other ‘‘ultra-light”

by Cabrera and Mott is competing with some second

~in ccrefully dried < oxygen. The 1c"equem‘ly cccep'red
',;Idea that sodlum__‘ obeys c lmear oxmicn‘lon rcfe_' h

Wmetols) need not be regardéd as belng'-ef a specml-u" '
' ized nature but 1 may be very SImllar to the oxmlcmon' .
“mechanism for the heavier metals, such as copper

metals become available, it may prove possible to
devise a theory of oxidation which is applicable
to at least the low-temperature oxidation of all
metals,

At the present time the oxidation data obtained
at 25°C are being rechecked, and the time of oxi-
‘dation is being extended. It is felt that these ad-
ditional data are required in order to obtain a
meaningful analytical expression for the rate
curves, The measurements at 48°C will also be
extended in a similar manner, =
~ The completion of these rate studies will mark
the end of what might be termed the ‘‘classical”
phase of the investigation of the oxidation of
“sodium. It is planned that the future work with
) :_'_k,;sjodlium Wi” i:r'ié:ld'de a st'tjdy of the surface topogra-
“phy and” structure of the sodium oxide films.
"Electron(mtcroscopy and diffraction techniques
" will be required for this work. The investigations
" of Harris, Gulbransen, and others have demonstrated

that, contrary to all existing oxidation theory,

oxide films formed on copper, nickel, and iron are

109

 

e

 
 

 

ANP PROJECT PROGRESS REPORT

not of uniform thickness and that the major portion
of the oxide is contained in nuclei scattered over
the metal surfaces. These findings emphasize the

necessity of combining rate measurements with a
detailed examination of the oxide films formed in
any complete oxidation study,

L Columbium QOxidation, The work on columbium is
in the preliminary stage. [t is planned to measure
fh_q rate of oxidation of columbium at several temper-
atures in the neighborhood of 400°C and then to

“correlate the oxidation rate and oxide layer thick-
ness with the structure and porosity of the oxide.
In this manner it should be possible to establish
the condifions which exist in the film, both when a
parabolic law is obeyed and when the rgte becomes
linear, 7 '

A check has been made on the oxidation rate of
columbium at 400°C in highly purified oxygen,

Qualitative verification was obtained of the phe-

nomenon, reported by Inouye,® of the transition of
the oxidation rate from a parabolic to a linear re-
lationship.  The initial stages of oxidation pro-
duced oxide films which exhibited interferences
(colors). As the reaction proceeded, however, the
specimens became coated with a white, opaque
oxide film whose surface was rough. The appear-
ance of the white oxide seemed to coincide ap-
proximately with the transition from the initial
parabolic oxidation rate to a linear rate,

Mass Transfer and Corrosion in Fused Hydroxides

M. E. Steidlitz
Metallurgy Division

The study of corrosion and mass transfer of metal
by fused sodium hydroxide is under way, Over 60

 

8H. Inouye, Scaling of Columbium in Air, ORNL-1565
(Sep'l'o ], 1953)0

TABLE 5.16, MASS TRANSFER OF NICKEL IN SODIUM
HYDROXIDE UNDER A HYDROGEN ATMOSPHERE FOR 100 hr

 

 

Bucket Temperature Number
lTerfiperafure Differential of Amotmf of Mass Trc:lnsfer
°Q) of Sgstem* Tosts (Yisual Observation)
e

450** 100 1 Little
450 100 1 None
575 100 1 Very little |
600 100 4  Very little to little
600 200 1 Very litthe
625 50 1 None
625 100 2 Very little
650 __ 50 1 Barely observable
650 100 2 Moderate
650 200 1 Moderate
675 50 1 Little
675 100 1 Moderate
700 | 50 1 | Moderate
700 100 7 ~ Moderate to heavy

800 50 1 Heavy

L s0 100 1 " Very heavy

 

. *Bucket temperature less the cold-finger temperature.

.. **This test was made under a helium atmosphere.

110

 

 

 

 

 

 
 

WY

individual tests have been conducted in the ‘‘cold-
finger’” apparatus. Although some analyses of

_ these tests are not complete, certain patterns of

hydroxide are being observed.

The cold-finger apparatus, which was described
previously,?¢10 consists of any externally heated
bucket of hydroxide into which dips an air-cooled
cold finger. Both the cold finger and the bucket
are constructed of the metal being tested. Thermal
gradients are observed between the outside of the
bucket and the inside of the cold finger so that
there is never thermocouple metal in the hydroxide
bath, All tests have been run for 100 hr, except
as otherwise noted, at the indicated maximum
temperatures and temperature gradients under a
flowing hydrogen atmosphere.

The sodium hydroxide used in this study is
reagent grade. It is dehydrated in place under

 

M. E. Steidlitz and W. H. Bridges, ANP Quér. Prog.
Rep. Mar. 10, 1955, ORNL-1864, p 91.

0y, H. Bridges, Met, Semiann, Prog. Rep. Apn 10,
1954, ORNL-1727, p 52,

3 s- ;.'«t :< () {
lincHes ¢

 

 

PERIOD ENDING JUNE 10, 1955

vacuum at 300°C for at least 8 hr followed by a
minimum of 16 hr at 400°C. In some instances,
previously dehydrated hydroxide is used, in which
case, arapid pumping at 400°C is deemed sufficient,

The metals studied to data have been nickel,
Inconel, and Hastelloy B. Of these, nickel and
Inconel show some promise of being suitable for
use with the hydroxide at temperatures of the order
of 600°C, Hastelloy B is by far the least resistant
of the three to attack by the hydroxide, belng badly
corroded at 600°C,

The conditions and visual observations of the
tests on nickel are presented in Table 5,16, In no
case was oxidation corrosion observed. Mass
transfer does occur at 600°C, as shown in Fig.
5.19.  The cold fingers seen in Fig. 5.19 are
nickel, with the exception of No. 19, which is
Inconel, Bucket temperatures were 600, 630, 670,
and 700°C for Nos. 17, 18, 20, and 21, respectively.
The cold fingers were kept 100°C cooler than the
corresponding buckets. The sequence shows
clearly the very small amount of mass transfer at

UNCLASSIFIED
Y-14550

 

an. 5,19, Muss-Tronsferred Deposns on Ntckel und Inconel Cold Flngers Exposed for 100 hr to
Sodium Hydroxide. Numbers 17, 18, 20, and 21 are “nickel and were “exposed at bucket femperatures of
600, 630, 670, and 700°C, respectively, with a temperature differential of 100°C. Number 19 is Inconel
and was exposed at a bucket temperature of 700°C with a temperature differential of 100°C.,

111

 
T

il

 

 

 

 

ANP PROJECT PROGRESS REPORT

600°C, as well as an acceleration in the rate of
transfer with increasing temperature in the range
600 to 700°C. The mass-transferred deposit at the
liquid line, as shown in Fig. 5.20, indicates the
type of metal crystal growth which occurs. This
specimen is a duplicate of No. 21, for which the
bucket temperature was 700°C and the cold-finger
'remperdh.ire was 600°C.

Several compounds that were chosen for their
strong reducing properties have been added to the
hydroxide melt, [t was hoped that these materials,
together with the hydrogen atmosphere, will lower
the mass transfer rate in nickel. Results of tests
with sodium or ‘sodium hydride added to the sodium

hydroxide are shown in Fig. 5.21. All the nickel

cold fingers shown were held at 600°C, and the
bucket temperatures were 700°C, No addition was
made to the sodium hydroxide to which No. 37 was

 

 

exposed, and Nos. 38 through 41, respectively,
were exposed to sodium hydroxide containing 5%
Na, 10% Na, 5% NaM, and 10% NaH, respectively.
In no case did the addition decrease the mass
transfer. The same was true for the third and last
additive tried, sodium hydrogen phosphite.

In all the tests with additives, there was evi-
dence that some, if not most, of the added material
distilled out of the bucket, These same compounds
might be found to be effective in a more confined
system such as a closed loop. No tests have yet

been run at the temperature at which mass transfer

is first noflced and it is posmb[e that these ad-
ditives mlghf increase the maximum femperoture
that can be attained without mass transfer occurring.

The results of tests on Inconel in sodium hy-
droxide are shown in Table 5.17. The mass trans-

fer of Inconel was significantly less than that of

S o LR

UNCLASSIFIED |

Y-15166

 

 

92000
N
3
Q
*x
000

 

 

€000

2000
z
O
I

 

 

 

 

F|g. 5.20. Mass=Transferred Deposit on Cold Finger Exposed to Sodium Hydroxule at a Buckef Temper-
uture of 7000C wnh a Temperature leferenhal of 100°C for 100 hr. 250X.

112

T

 

 

 

 
 

1

ar

 

PERIOD ENDING JUNE 10, 1955

UNCLASSIFIED
868

Fig. 5.21. Mass-Transferred Deposits on Nickel Cold Fingers Exposed for 100 hr to Sedium Hydroxide
With and Without Various Additives. Cold finger maintained at 600°C; bucket maintained at 700°C;
No. 37 exposed to sodium hydroxide with no additives; Nos. 38 through 41 exposed to sodium hydroxide
containing 5% Na, 10% Na, 5% NaH, and 10% NaH, respectively.

TABLE 5.17. CORROSI_ON. AND MASS TRANSFER OF
INCONEL IN SODIUM HYDROXIDE UNDER A
HYDROGEN ATMOSPHERE FOR 100 hr

 

 

     

 

   

B Temperature
Bucket Differential Number
Tem;;eroture of System* of Comments
( C) 4 (OC) Test
- 450%% 100 1 Brown deposit; no
. o T mass frcmsfer -
450 '7 100 7 1 Clean; no mass
: o E e trcnsfer )
550 ‘ :2_0‘0 o o 3 Brown stoun, no
e T e - mass fransfer )
600 ) - ]00 o S Brown stain; no
S omnSUm Sh L i mass fronsfer
600 1 'Brown S']‘OII"I, no
S Ao b e mass transfer o
650 _ 100 1 'Brown stcln, bare[y '
SO e e T e e “~ahservable mass
G e D R fransfer R
700 ' 17000 1 Brown sfam, very
B R TR _|1Hle mass
, TR e T S “transfer o
o 800 S ]00 o T Rust eoloron

S lmE e et T e bucket; medium
C ' mass tronsfer

 

*Bucket temperature less the cold-finger temperature.

**This test was made under a helium atmosphere.

nickel, but the Inconel corroded to some extent,
Specimen No. 19 of Fig. 5.19 is an Inconel cold
finger tested under the same conditions as was the
nickel specimen No. 21, The metal deposit on No.
19 is less than that on No. 21, but the tarnish on
No. 19 indicates that there was some corrosion.
Therefore, in the remaining Inconel tests, the
Inconel bucket (that is, the hot zone) was examined
for corrosion. Inconel buckets that were exposed
to sodium hydroxide at bucket temperatures from 450
to 800°C are shown in Figs. 5.22 through 5.26.
It is most interesting to note the change in type
and extent of attack that occurs between 600 and
650°C. At 600°C and below, 1 to 2 mils of grain-

boundary attack may be seen both on the inside

~and outside wall of the bucket, The attack on the

outside wall was caused by hydromde that crept
out over the bucket lip. Above 600°C, an oxidized
layer formed that was 3 to 15 mils in depth. This
temperature sensitivity is, as yet, unexplained.
Other metals are to be studied in an attempt to

raise the temperature |limit at which no mass trans-

fer occurs, It is hoped that questions such as what
causes the change in rate and type of attack in the
600 to 700°C range will also be resolved.

113

T e

 
 

 

 

ANP PROJECT PROGRESS REPORT

 

 

UNGCLASSIFIED
Y5167

 

 

o o o I | o [0 o o |o
S 12 |2 250x R 12 1R |2 2
o (v | l | o s Ny |

 

 

 

 

 

 

 

 

0100
6000
8000
12000
9000

z

T
€000
2000
000

 

 

 

 

 

 

 

 

 

Fig. 5.22. Inconel Bucket Exposed to Sodium Hydroxide for 100 hr. Bucket temperoture; 450°C; cold-

finger temperature, 350°C. 250X.

CHEMICAL STUDIES OF CORROSION

F. Kertesz
Materials Chemistry Division

Corrosion of hiconel by LiF-BeF, and
by LiF-BeF,-UF,

H. J. Buttram R. E. Meadows

Materials Chemistry Division

Inconel capsules containing the BeF,-bearing
mixtures LiF-BeF, (69-31 mole %) and LiF-BeF,-
UF, (67.3-30.2-2.5 mole %) were tested for cor-
rosion in tilting-furnace tests. After 100-hr ex-
posures, metallographic observations of the capsule
walls revealed no attack; chemical analyses con-
firmed these findings. The chromium concentration
was found to be about 250 ppm in the LiF-BeF,
mixture and 400+ ppm in the UF ,-containing

114

analog. These results are of the same order of
magnitude as those obtained for the ZrF-base
mixtures.

Effect of UF:‘-UF4 Mixtures on Corrosion of
Inconel by Yarious Solvents

H. J. Buttram R. E. Meadows

Materials Chemistry Division

Incidental to attempts to study the disproportion-
ation of UF,, confirmation of the beneficial effect
of UF, on corrosion of Inconel by NaF-ZrF,
(53-47 mole %) and NaF-KF-LiF (11.5-42-46.5 mole %)
was obtained. A series of UF,-UF, mixtures in
these solvents was heated in Inconel capsules.
The contents were analyzed for total uranium and
trivalent uranium after testing in the tilting furnace
or under static conditions. Considerable dispro-

 

 

 

 

 

 

 

 

 
e GA

 

Ll

u‘i

@

 

 

 

PERIOD ENDING JUNE 10, 1955

 

OUTSIDE

UN(’LASS!FIED ;
Y-15169

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o o o | o 10 lo lo |lo lo o lo o jo e Jo o

! o | ; o o o o o o o o o o 1©

SR 2 lesoxl2 2 2R IZIZIEIBIB IS |inen 318 18

o ©C im | o i o N~ 1O i o N e | W oo
Fig. 5.23. Inconel Bucket Exposed to Sodwm Hydromde for 100 hr.

finger tempercture 500°C 750X.

portionation of -the added UF, was observed in
;_bofh so_lvenf.:_, especna”y when the UF /UF ratio
"'was |ow, the NraF KF-LiF mixtures showed in

euch cqse more dlsporporhonqhon _thqn dld the
NGF ZrF soluhons. '

 

   

 

   
  

on corrosion of
If a plot is made
fesfmg as a
raho it is seen

5 (UF,/UF, O 2)

 
 
  
  
  

   

 

LN

| wal fs: .

  
     

      

bu'r any “Further mcrease in
ratio apparently had little effect.

‘found to be beneflcm],

the UF3/UF4

 

    

Bucket temperature, 600°C; cold-

Studies of the Sodium Hydroxide=Nickel Reaction
F. Kertesz F. A. Knox

Materials Chemistry Division

  

uring previous work on the equilibrium hydrogen
pressure of the sodium hydroxide~nickel reaction,
it was found that the reproducibility was somewhat
poor, Similar difficulties were observed during
determinations of the equilibrium solubility of the
‘metal in molten hydroxide, In studying the possible
reaction

 

2NaOH + Ni—> Na,0-Ni0 + H,

the hydrogen pressure or the nickel concentration
was determined as a function of time and tempera-
ture. The hydrogen pressure was measured with

115

 

 

 
 

 

 

 

 

 

 

 ANP PROJECT PROGRESS REPORT

OUTSIDE

 

 

80'0

L0°0

900
0

 

 

 

 

0
z
O
X

£0°0
00

1070

 

 

 

 

 

 

Fig. 5.24. Inconel Bucket Exposed to Sodium Hydroxide for 100 hr. Bucket temperature, 650°C; cold-

finger temperature, 550°C. 75X.

the reaction chamber connected to a manometer,
and the nickel concentration was determined by
analyzing the contents of a quartz-jacketed metal
capsule after the desired exposure time. Since
the two sets of data were necessarily determined
on separate systems, there was the possibility of
unknown variables being present, even though
every effort was made to keep the variables
identical.

‘ "_This.experiménml difficulty has recently been
“¢ounferacted by constructing an apparatus that

   

makes possible simultaneous determinations of
- the hydrogen equilibrium pressures and the equi-

librium nickel solubility. A charge of purified

“sodium hydroxide is loaded into a hydrogen-fired

mckel ‘capsule which is sealed under helium and
then placed in a quartz tube that is evacuated and

sealed. The jacketed capsule is heated to the test
" temperature for the desired time period. If the

 

exposure is sufficiently long, the hydrogen pres-
sure developed should be in equilibrium with the
melt, since it should not diffuse through the quartz.
The pressure should be nearly equal inside and
outside the nickel capsule, because hydrogen
diffuses easily through this metal at the tempera-
tures used. Upon completion of the high-tempera-
ture exposure, the capsule is placed in a flanged
metallic cylinder which, after evacuation, is con-
nected to a mercury manometer. A small metal pin
held near the quartz capsule by a metallic bellows
extending through a lateral hole in the cylinder can
then be tapped to break the quartz jacket. The
pressure established in the system can be read on
the manometer, and a simple calculation will give
the pressure inside the quartz capsule before it
was broken. '

Calibration tests with capsules containing
hydrogen at a known pressure showed that the

 

 

 

 

 
 

PERIOD ENDING JUNE 10, 1955

 

i
4 UNCLASSIFIED
1 Y-15171

 

 

OUTSIDE

 

 

o o o o o o o
O O O o O INCH o o) O
w ~ o H w n -

 

 

 

 

 

 

 

 

 

 

Fig. 5.25. Inconel Bucket Exposed to Sodium Hydroxide for 100 hr. Buckef temperature, 700°C; cold-
finger temperature, 600°C. 75X.

method allowed pressures to be determined to of failure of the nickel capsules. It was found
within 1%. After the pressure measurement the  that the nickel content of the hydroxide remained
nickel capsule was opened and the hyd"rc’kid'e was nearly constant when the time of exposure was
leached out. with the total alkalinity and the dis-  varied considerably, which would indicate that

 

solved nlckel bemg “determined on the same 'sample.  equilibrium was reached after a short exposure

 

     

Some of ‘the quartz- |ocketed nickel capsdles “time.” The hydrogen pressures determined by this
Acom‘alnlng sodium hydromde were kep'r at 800°C ~method did not level off as expected, and the
for various ‘lengfhs of time in order to ascertqm\ resultmg pressure was found to be greatly in

 

     

 

the time necessary to reach equ:llbrlum. Runs for excess of that to be expected from consideration
long perlods of time were unsuccessful because of the rec:c'rlon postulated.

    

TR

0

 

"

117

 

 
ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Y-15168 ¢ .

 

   

 

   

 

 

 

 

  

o Jo o Je Jo o [ Jo Je Jo o Jo
o o o O o QINCH|©@ o o o o o ‘
: ; B ~ o o N » ™ o o P B N

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 5.26. Inconel Bucket Exposed to Sodium Hydroxide for 100 hr. Bucket temperature, 800°C; cold-
finger temperature, 700°C. 250X.

 

 

 

     

 

 

 

 

 

    
 

&Y

s

Y

 

% ) Coo
Dec 10 1954 ORNL 1816, o 100

PERIOD ENDING JUNE 10, 1955

6. METALLURGY AND CERAMICS

W, D. Manly

J. M. Warde

Mefcllurgy D:vus:on

Addmoncl fcbr:caflon studles of Hasfelioy B
have mcreased the evidence that the poor hlgh-

'tempercture fcbrlcabtllty is related to the impurity

content from which the strength may be derived.
Several nickel-molybdenum binary and ternary al-
loys were studied in oxidation tests and in room-
and elevated-temperature tensile tests. The ter-

~nary alloys all included 20% molybdenum plus

nickel and a third heavy element. The results of

-additional stress-rupture and tensile property studies

of the nickel-molybdenum alloys are presented.
For comparison with previously obtained data on
the static oxidation of several brazing alloys,
cyclic tests were run; the static and cyclic data
were similar for tests at 1500°F. Cyclic tests at
]700°F are under way. '

Fabrlcqhon was comple'red of a 20-tube Inconel
fuei-fo-NaK heat _exchanger, the fuel-to-NaK inter-
mediate heat exchanger No. 2, two 500-kw NaK-tfo-
air radiators, and a liquid mefala'ro-olr radiator de-
signed by the Cornell Aeronautical Laboratory,

Special extrusions were made of three billets of
vanadium and four high-purity molybdenum billets
containing 0.7% titanium, and studies of flow pat-
terns of duplex and three-ply materials during im-
pact extrusion were continved. The fabrication of
clad columbium was investigated, and diffusion
barrier studies were made. Additional information
was obmmed on the properties of B 6C-Cu mixtures

and a magnesnum -lithium alloy fhaf are belng con-

5|dered os‘ posstble shleldm' "'materlals.

 

 

24, Inouye, J. H. Coobs, and M. R. D'Amore, ANP
Quar. Prog. Rep. Mar. 10, 1955, ORNL.-1864, p 97.

 
 
 
 
 
 
 
 
 
 
 
  
    

e

DEVELOPMENT OF NICKEL-MOLY BD ENUM
ALLOYS

J. H. Coobs H. Inouye
Metallurgy Division

M. R. D’ Amore
Pratt & Whitney Aircraft

Fabrication Studies

Hastelloy B. Two impact extrusions of Hastel-
loy B rod were made at 2000°F with good recovery
of sound material. The billets were prepared from
a vacuumemelted ingot of commercial Hastelloy B
and were extruded in the as-cast condition. During
subsequent hot reduction of the extruded rod to
sheet, the material cracked severely, and no usable.
material was obtained.

Two extruded tube blanks that had been shipped
to the Superior Tube Co. for reduction to small-
diameter tubing were processed; one blank was
made from wrought material and the other from cast
material. The tube blank made from a vacuum-
melted cast billet fractured during the first reduc-
tion operation, but the blank made from a wrought
billet was successfully processed from 1.5-in.-OD,
0.250.in.-wall to 0.187.in,-0D, 0.017-in.-wall seam-
fess tubing. It was found that severe reduction
schedules were permissible and that intermediate
stress-relieving heat treatments were unnecessary.

Previously reported experiments! showed that
higher extrusion ratios were attainable if the
Hastelloy B billet was canned in Inconel. There-
fore wrought Hastelloy B billets with and without
Inconel cans have been prepared for furfher ’rube

: "“extrusmn experiments.

© Attempts to roll extruded Hastelloy B rod at

--2000°F were unsuccessful; the material cracked
”‘:‘S‘évérely after several 5% reduction passes. Since

0.03% cerium additions had previously? been found
to improve the high-temperature fabricability of

Hastelloy B, cdditional'me!fs were prepared with
‘various quanfmes of cerlum aclded in the form of

ni 1 ]% Ce—89% Al master o”oy A small vacuum
melt confaining "0.3% cerium “cracked ‘during hot
reduction, and in subsequent experiments it was
shown that a maximum of 0.1% cerium was useful
in improving the high-temperature fabricability.

119

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

An alloy wi'rh. the nominal Hastelloy B compo-
sition, 4% Fe-28% Mo—-68% Ni, but without the

tramp elements vanadium, silicon, manganese, co-

~ balt, chromium, tungsten, and aluminum, was pre-

pared for evaluating the effects of impurities on
the fabrication properties of Hastelloy B. Three-
pound vacuum melted slab ingots 0.500 in. thick
were prepared and rolled at 2100°F to 0.150-in.-
thick sheet. This material showed no tendency to
crack during hot rolling under moderate reduction

~schedules. The microstructure of the alloy in the
as-cast condition showed random distribution of a
" second phase, which appeared to be an oxide. The
source of this phase might be traced to the starting

material or the melting practice. It is becoming
increasingly evident that the poor high-temperature
fabricability of Hastelloy B is related to the im-
purity content, However, since the superior strength
of commercial Hastelloy B in comparison with that

" of Inconel might be derived from these impurities,

mechanical property tests of the pure 4% Fe-28%
Mo—-68% Ni alloy are under way. The preliminary
room-temperature tensile data, presented in Table
6.1, are encouraging in that they show the tensile
strength and ductility to be comparable to those of
the commercial alloy under these test conditions.
Additional mechanical property tests are under way.

TABLE 6.1. ROOM-TEMPERATURE TENSILE
STRENGTH OF THE 4% Fe-28% Mo—68% Ni ALLOY

 

 

Tensile ,
Elongation
Condition Strength
_ ; (%)
- (psi)
Annealed . 130,000 40
Annealed, aged 500 hr . 137,000 2.5

~ at 1300°F

 

Nickel-Molybdenum Binary Alloys. The previ-
ously reported® work on nickel-molybdenum binary
alloys was extended by a study of an 85% Ni-15%
Mo alloy. The dlloys studied previously were
80% Ni-20% Mo, 76% Ni-24% Mo, and 68% Ni-
32% Mo. An oxidation test of the 85% Ni—15% Mo
alloy at 1500°F in static air indicated a lower rate
of oxidation than that observed for the 80% Ni~
20% Mo alloy, but o higher rate than the rates for
’rhe 76% N|~24% Mo and 68% Ni-32% Mo alloys.

 

,3H. Inouye and J. H. Coobs, ANP Quar. Prog. Rep.
Dec IO 1954, ORNL-18164, p 103.

120

Therefore additional oxidation tests are planned to
supplement the present data. The presently avail-
able data appear to indicate that the rate of oxida-
tion of these binary nickel-molybdenum alloys is
not a function of the molybdenum content, Two
thermal-convection loops for circulating fluoride
mixture in the 85% Ni-15% Mo tubing are being
fabricated. '

Both room- and elevated-temperature (1100 to
1650°F) tensile tests of this alloy have been com-
pleted. Erratic results were obtained in the room-
temperature tests of specimens aged 284 hr at ele-
vated temperatures (1300 to 1650°F), The tensile
strength was low at all test temperatures, and low
elongations were observed at test témpero’rures
above 1300°F. The results of these tests are given
in Table 6.2,

Nickel-Molybdenum Ternary Alloys, A screening
program was initiated in orderto observe theeffects
of additions of alloying elements to nickel-molybde-
num alloys. The alloy compositions investigated
have all included 20% molybdenum plus nickel and
a third element. The ternary systems that have
been evaluated are:

2 to 10% Cb-20% Mo~bal Nij
3 to 10% Cr~20% Mo-bal Ni

2% V-20% Mo-78% Ni
1% Zrw=20% Mo-~-79% Ni

2% Al-20% Mo —78% Ni 1% Ti-20% Mo—79% Ni

The physical property data obtained for alloys con-
taining 2 to 10% columbium were reported previ-
ously.y The aluminum-molybdenum-nickel alloys
were investigated because of the exceptional creep
properties reported in the literature for this alloy
system, Hot-forgeability studies were conducted
on 100-g arc melts containing from 2 to 10% alumi-
num, but only the alloys containing less than 5%

~aluminum were found to be forgeable at 2100°F.

Further studies will be conducted on this alloy
system. ' S

Vanadium, zirconium, and titanium additions were
investigated in an effort to improve the elevated-
temperature ductility of the 20% Mo-~80% Ni alloy.
No difficulties were experienced in fhe hot rolling
of these alloys.

Preliminary room-temperature tensile strength
data for the alloys containing aluminum, vanadium,
zirconium, and titanium in both the annealed and
aged conditions are presented in Table 6.3. All
specimens aged at 1650°F for long periods of time

 

41bid., p 100.

 

 

 

 

 

 

 

 

 
 

»}

©

@)

  

 

PERIOD ENDING JUNE 10, 1955

TABLE 6.2. RESULTS OF TENSILE TESTS OF AN 85% Ni—15% Mo ALLOY

 

Elongation in 2-in.

Yield Point, Tensile

 

iti Test T
Condlflon of | est e;nperc:ture 0.2% Offsot Strength Gage Lengths
Test Specimen* ("F) ] .

. (psi) {psi) (%)

" Annealed 1 hr at 2100°F Room 32,100 60,400 15
Annealed, aged 284 hr ot 1300°F Room 32,700 70,200 17.5

Annealed, aged 284 hr at 1500°F Room 32,000 | 99,700 50

Annealed, aged 284 hr at 1650°F Room 31,500 68,400 21
Annealed 1 hr at 2100°F 1100 18,7007 39,900 16.5
1300 17,850 30,300 7.5
1500 17,000 . 26,000 5.0
1600 16,600 23,200 6.3

1650

16,000 20,700 9.0

 

*Sheet, 0.065 in. thick.

TABLE 6.3, ROOM-TEMPERATURE TENSILE STRENGTHS OF SEVERAL NICKEL-MOLYBDENUM
BASE TERNARY ALLOYS

 

Tensile Strength Elongation

 

Alloy Composition Condition of
o (wt %) Test Specimen {psi) (%)
2 Al—20 Mo—78 Ni Annealed 150,500 32
Annealed, aged 500 hr at 1500°F 118,500 59 .
Annealed, aged 284 hr at 1650°F 70,500 12
2 V20 Mo-78 Ni Annealed 117,000 32
: o Annealed, aged 284 hr at 1300°F 127,000 30
1 Zr—20 Mo~79 Ni Annealed | 101,000 33
- Annealed, aged 284 hr at 1300°F 115,000 58
1 Tie20 Mo..79 Nj Annealed 115,000 60
L e _Aged 284 hr at 1650°

73,700 s

 

 

: show lower s'rreng’rhs cnd ducflllhes ‘than fhose _
unaged or aged at fower temperctures. ‘The speci-
mens ‘aged at 1650°F were exposed to a hydrogen o
'."-atmosphere, ‘while the ofher specnmens were treated
" Under a hehum afmosphere or in evacuatec -~
_‘,.-'f'cqpsules. The possnb|||ty of hydrogen embriih‘le-::"
. meént of 'rhe SpeCImens aged m‘ ]650°F is bemg
Y nves'r:gated .
o The oHoys confomlng 3% 10% chromlum have |
S been evolu 3y -
tensile fes'rs.
“made for corrosion 'restmg ‘has been ilml'red how-
“ever, because of shattering during extrusion.” Al-

    

 
    

  

ed i

 

§|dqt|on, ‘s‘tr' ss-rupture, “and
The quantity “of seamless fubmg

 

 

though sound rod and tube blanks have been made
at an extrusion temperature of 2200°F, consistent

) femperafure range ‘was unsuc

_results could nof be obfc:med A study of these

olloys hos led to the supposmon thc’r the shrink

cavities within the mgot are the cause of the frac-
' turlng The billets are currently being mspecfed

by x-ray and gqmma-radlahon techmques, prior to
exfrusmn, in an ‘effort to correlate the soundness
of the starhng maferucrl with the extrusmn result.

" The hot rollmg of extruded rods of chromlum-_:
'molybdenum-nlckel alloys in the ]95Q to 2200°F_

ul because of

 

‘severe crackmg Subsequem‘;sfud:es showed that
“the hot shor‘ ess was probably caused by oxygen‘
‘contamination of the electrolytic chromium used,
It was found that additions of 0.1% cerium for each
3% chromium in the melt would render these alloys

 

121

 

i
i S

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

hot forgeable. Further experiments have been con-
ducted with 100-g arc melts containing up to 2%
titanium and 2% aluminum as deoxidants. The

aluminum additions were found to be ineffective,

“ but the titanium additions served to reduce the hot
- shortness, It was determined that approximately
" 2% titanium would render a 5% Cr—~20% Mo—-75% Ni

- alloy completely hot forgeable. The extruded rods

of the hot-short alloys could readily be rolled to
sheet at room temperature. Work-hardening data

for a 3% Cr-20% Mo-77% Ni alloy are presented

in Table64 “The alloys containing between 3 and
- 10% chromium work-hardened at the same rate.

TABLE 6.4. WORK-HARDENING OF A
© 3% Cr-20% Mo-77% Ni ALLOY

 

" Reduction (%) ' Hardness (VPN)

 

o 168

104 260
211 ' 312
30.9 337
45.5 409
524 441
63.1 480

 

Room-temperature tensile strength data for chro-

miumemolybdenum-nickel specimens subjected to
various heat freatments are listed in Table 6.5.
The data reported are the average of the results
of two to three tests of each alloy. Included in the
tests were alloys that contained additions of 0.5%
columbium or up to 0.25% cerium. The minor ad-
ditions of columbium appeared to incfe_osé slightly
the tensile strength and ductility of these alloys,
and an average decrease of 5% in elongation and
a slight increase in tensile strength were noted
for the alloys with small amounts of cerium added.

The elevated-temperature tensile strengths of
chromium-molybdenum-nickel alloy specimens after
various aging treatments are given in Table 6.6,
All the specimens were tested at the aging temper-
atures listed. The results are given for single
tests, The elevated-temperature ductilities of these
alloys are low compared with the ductilities at
room temperature. The minor additions of cerium
apparently aided in obtaining higher elevated-
temperature strengths in these alloys, but more
tests are needed to verify this observation, No
relationship between the physical properties and
the chromium content is readily apparent in either
room- or elevated-temperature tests.

In the temperature range above 1300°F the alloys
containing up to 10% chromium appear to be single-

TABLE 6.5. ROOM-TEMPERATURE TENSILE STRENGTH DATA FOR CHROMIUM-MOL YBDENUM-NICKEL
ALLOY SHEET SPECIMENS

Aging Time: minimum, 284 hr; maximum, 500 hr

 

 

 

 

Alloy Composition Condition of Tensile Strength Elongation
(wt %) Test Specimen (psi) (%)
3»Cr—20 M9—77 Ni Annealed 107,000 61
Aged at 1300°F 109,500 60
Aged at 1500°F 111, 100 64
Aged at 1650°F 67,500 19
5 Cr=20 Mo~75 Ni Annealed 120, 000 | 58
Aged at 1300°F 119,000 59
Aged at 1500°F 115,000 | 62
Aged at 1650°F 72,500 14
7 Cr-20 Mo=73 Ni Annealed 116,000 | 63
| | Aged at 1300°F 116,000 | 63
Aged at 1500°F 114,000 59
o o | ~ Aged at 1650°F 69,000 14
10 Cra20 Mo—70Ni  Annealed 110,500 n
: C ' Aged at 1300°F 111,600 63

 

122

 

 

 
 

»t

N

  
   

PERIOD ENDING JUNE 10, 1955

TABLE 6.6, ELEVATED-TEMPERATURE TENSILE STRENGTH DATA FOR
CHROMIUM-MOL_YBDENUM-N!CKEL ALLOY SHEET SPECIMENS

 

Tensile Strength Elongation

 

Alloy Composition A'ging Treatment

(wt %) ~(hr) CF) (psi) (%)

3Cro20 Mo-77 Ni 500 1300 33,200 6.3

o 500 1500 27,300 5.0

_ N 284 1650 26,100 19.0

3 Cr=0.5 Cb—20 Mo—76.5 Ni 362 1300 38,100 6.3

| | 362 1500 30,000 - 5.0
5 Cr-0.5 Cb~20 Mo—74.5 Ni 500 1300 39,500 7.5

500 1500 29,400 5.0

5 Cr-0.25 Ce~20 Mo—74.75 Ni 362 1300 59,700 16.5
' _ 362 1500 43,100 11.0*

7 Cr-20 Mo~73 Ni 362 1300 40,800 9.0
' 362 1500 29,800 3.5%

7 Cr—0.5 Cb-20 Mo~72.5 Ni 500 1300 48,800 12.5

500 1500 38,500 7.5

| - 284 - 1650 29,800 3.8

7 Cr=0.1 Ce—~20 Mo—72.9 Ni 362 1300 54,500 s
, | 362 1500 40,500 8.8*

10 Cr—0.5 Cb—20 Mo—69.5 Ni 500 1300 38,300 8.8

: 500 1500 43,600 9.0

1650 29,900 6.3

284

 

*Specimen fractured outside 2-in. gage length.

phase. Additional studies of the microstructures
will be conducted on alloys made with high-purity
chromium as the starting material, Further evi-
dence of the equdnbnum structure is shown in aged

VSpec:mens. No sngmflcont changes in tensile
' sfrength or duchhfy are obServed in specumens'“'
-~ aged at 1300 and 1500°F The Iow_ sfrength and
~ ductilities of spec:mens aged at 1650° "
'-been cq'used by fhe hydrogen atm _sphere used ‘

  
 
 
 
  
    

   

of ‘stress-ru pture tests

  
 
 
 

N""creasmga ’rhe s'-*rength ‘and’ the duc’rlhty 'rhroUgh '
cerium additions appears to be promising, since
the alloys with cerium additions have shown better

may have_

(presenfed_ _
h n) have shown low sfrengths and‘__

physicd[ properties during sfreés#rupture testing.
It has become apparent that improved deoxidation

practices are necessary in melting these alloys,
and efforts are being dlrected toward the production

of sounder mgo’rs.
0xnduhon Studies

Oxidation studies were continued on nickel-

"“molybdenum alloys containing from 3 to 10% chro-
w‘mium,  The results reporfed prevnouslys for these
““alloys were based on oxidation tests of small

arc-melted specimens that had not been chemlccily:'
inalyzed, The alloys hove therefore been retested
by using specxmens from chemlcally anclyzed

   
 

“heats of material, qnd ‘the results are presented in
fTable 6.7. In comporison “with the previous re-
~sults, the oxidation rate for the 3% chromlum ‘com-
“pesition was’ slightly Iower, ‘and mcreoses ‘were
“noted for the alloys containing” 5, 7, and 10%

 

SH. Inouye and M. R. D*Amore, ANP Quar. Prog. Rep,
Mar. 10, 1955, ORNL-1864, p 104,

123

 
 

L e T

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 6.7. OXIDATION OF CHROMIUM-MOLYBDENUM-NICKEL ALLOY DURING
‘ 168 he AT 1500°F IN STATIC AIR

 

Alloy Composition Weight Gain

 

2 Remarks
(wt %) (mg/cm*)
3 Cr=20 Mo—77 Ni 5.17 Oxide spalled completely during cooling
to room temperature '
5 Cr.20 Mo .75 Ni 5.30 Oxide spalled completely during cooling
7 _ _ to room temperature '
7 Cr=20 Mo~73 Ni 3.14 Oxide spalled badly during cooling to
s room temperature ' ’
10 Cr=20 Mo-70 Ni 0.45 No spalling occurred

 

chromium. It appears that about 10% chromium is
necessary to form a nonspalling protective oxide

‘on the ternary alloy, However, it has been ob-

served '.fh'at the oxidation rate under static condi-
tions can be reduced 50% by the addition of 3%
chromium.

STRESS-RUPTURE STUDIES OF NICKEL-
MOLYBDENUM ALLOYS

D. A. Douglas J. H. DeVan
J. W, Woods
Metallurgy Division

Hastelloy B

A series of creep-rupture tests on Hastelloy B
in the solution-annealed condition has been com-
pleted in an argon atmosphere at 1500 and 1650°F.
The results are summarized in the design curves
presented in Figs., 6.1 and 6.2. A similar series of
tests in fused salts is nearing completion that
provides an interesting comparison with the argon
results, ‘ _ }

Stress-rupture plots are presented in Fig. 6.3
for Hostelloy B in the two environments at 1300,
1500, and 1650°F, and it may be noted that at each
temperature the values obtained in fused salts are
octUaIl_y superior to those obtained in argon. Al-
though no adverse effect on physical properties
was expected, in view of the absence of corrosive
attack of the fused sclts on Hastelloy B under
static conditions, the reason for the increase in
ruptire strength is not entirely clear at present.
The test chamber must be flushed with a cleaning
charge before the fluoride mixture to be used for
the test is admitted, and therefore the specimens
tested in fused salts were at temperature approxi-

124

mately 5 hr longer before stress was applied than
were the specimens tested in argon. Consequently,
the slightly longer aging of the specimens tested
in the fused salt may account for the improvement
in strength, There are objections to this theory,
however. A specimen aged 70 hr prior to stressing
and testing in argon showed inferior properties
relafive to those of a comparable unaged specimen
tested in argon. Also, there is some doubt as to
whether aging can account for the improvement at
1650°F, since the phase diagram, if minor ele-
ments are neglected, indicates a one-phase region
at this temperature. (Photomicrographs, however,
have shown evidence that aging at this temperature
under stress does occur.) Other tests are being
carried out to establish the exact causes for the
apparent anomaly in strengths in the two environ-
ments. S
Tests in hydrogen and air also have been made
in order fo determine the effects of these environ-
ments on the creep-rupture pr0perfie$ of Hastelloy
B at 1500°F, The creep curves obtained at 12,000
psi in hydrogen, argon, air, and fused salt are
shown in Fig. 6.4, In comparison with the effect
of argon, the effect of hydrogen is apparently
negligible, and the effect of air follows closely
the pattern observed for Inconel and *‘A’* nickel
in air, insofar as reduced creep rate and longer
rupture life are concerned. However, the final
elongations of Hastelloy B in air are equivalent
to or lower than those in argon, whereas the elon-
gation of other nickel-base alloys are markedly
greater in air than in argon. A fest program simi-
lar to that carried out on solution-annealed "Hds?erifi
loy B has been initiated on solution-annealed
specimens daged 100 hr at 1300°F to determine

 

 

 

 

 
 

0

e Hastelloy*B from the stondpom’r 6\""

20,000

18,000

16,000

14,000

12,000

STRESS {psi)

10,000

8000

6000

 

PERIOD ENDING JUNE 10, 1955

UNCLASSIFIED
ORNL-LR- DWG 7558

RUPTURE

100 1000 _ 10,000
TIME (hr)

Fig. 6.1. Sfress-Rup?ure Curves for 0, 060-m.-Thzck Sheet Specnmens of Hastelloy B Solution-Annealed

at 2100°F and Tested in Argon at 1500°F.

whether aging treatment results in serious em-
brittlement at service temperatures,

Modified Nickel-Molybdenum Alloys

Creep-rupture tests of modified nicke l-molybdenum

alloys at 1500°F in argon are under way. The
results of tests on these alloys, prepared by

- vacuum meltlng, have been somewhat dlsappomhngm ‘

- with respect to duchllty attained before rupture,

  
  
  
 
  
  
  
  

 

  
 
 

; 20% Mo

 

but, at 8000 psi, it had a final elongation of 28%
compared with 16% for Hastelloy B and 7% for a
similar alloy to which no deoxidant (cerium) was

}{H is currenfly beheved that fhe low duchiity maymi'
“be atfrtbufed to meifmg prachce ‘rather than to
“the anure of thé aHoy sysfems. ‘A heat to whtch”_"
a sma“ amq_un. of deoxtdanf was cdded showed a
;:-is;gnlficqn'r Improvemen‘r in_properties compared_:d _strength data for this material after heating at
~with those of other smgle-phase alloys tested to
_date, The alloy had the composition 74 75%”N1-“ h
Cr~0 25% Ce and was inf rior

- from _these heat treatments have been studied and

added, The results of recent creep-rupture tests
of several modified nickel-molybdenum alloys are
presented in Table 6.8,

TENSILE PROPERTIES OF HASTELLOY B
'P. Patriarca R. E. Clausing
Meta”urgy Division

It has been shown® that the physu:dl properties
of wrought Hastelloy B alloy are directly influenced
by the precipitation which occurs at temperatures

..within the intended operational range of ANP

reactors and heat exchangers.  An extensive pro-
..gram is now under way to obtain short-time tensile

_tempercfures from 1100 to 1600°F for times from
100 to 1000 hr. Typical microstructures obtained

_photographed, and a correlation is being made

 

P, Patriarca et al.,, ANP Quar, Prog Rep, Mar, 10,
1955, ORNL.-1864, p 116.

125

 
 

 

 

 

He work

 

. ANP PROJECT PROGRESS REPORT

UNCLASSIFIED -
~ ORNL—-LR-DWG 755¢

 

14,000 ‘

 

13,000

 

12,000

 

~ 44,000

 

 

RUPTURE

 

40,000 <
: . N
R N 2% A
\

9000 -

 

" 8600 <

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

= 3 N
3z \ \ \\\ \\ \k\
' 0.5% ELONGATION N
W 7000 \\ o~ ™, N NN
i N ™~ ™ NN
& N N NG TR
@ . N ™~ \\
’ GOOO \ \\ \\ N
™
\ N
o \ \\ \\\
5000 N SN
4000
1 10 100 1000 10,000

TIME (hr)

"Fig. 6.2. Stress-Rupture Curves for 0.060-in.-Thick Sheet Specimens of Hasfel.by B Solution-Annealed

at 2100°F and Tested in Argon at 1650°F.

TABLE 6.8. CREEP-RUPTURE TESTS OF MODIFIED NICKEL-MOLYBDENUM ALLOYS

 

 

Alloy Composition Stress Time to Rupture Elongation
(wt %) (psi) (hr) (%)
77 Ni=20 Mo~3 Cr 8000 90 7
5000 675 9
75 Ni_20 Mo~5 Cr _ 8000 80 7
74.75 Ni~20 Mo-5 Cr~0.25 Ce 8000 444 28
73 Ni=20 Mo~7 Cr 8000 380 | 9

 

befween physical properfles and the correspondmg
'mlcros’rruc’rures. As a result of this investigation,
‘more comprehenswe information on the high-tem-
perature properties of this material is being ob-
. tained, The relative merits of various preaging
" heat trectments of Hastelloy B are also being
)sfudled and it is hoped that, as a result of this
a procedure can be developed which will

stabilize the microstructure sufficiently to reduce
the sensitivity to high-temperature aging,

A summary of the short-time tensile data that
have been obtained to date is presented in Flgs.
6.5, 6.6, and 6.7, and in Table 6.9. It may be
noted in Fig. 6.5, which shows the results of the
test in which the variable of aging time was in-
vestigated, that the tensile-strength and yield-point

 

 

 

 

 

 
LT . .

e ol o kil L s

 

L

STRESS (psi)

PERIOD ENDING JUNE 10, 1955

-“EeneT
ORNL—LR—DWG 7560

 

60

 

50

 

40

 

 

 

 

 

 

 

30

 

 

n
o

 

 

 

 

 

 

 

10

 

 

 

 

 

 

 

 

 

 

 

   

| FLUCRIDE FUEL: NaF—ZrF4—UF4 (50-46—4 mole %o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 100 1000 10,000
TIME (hr)

Fig. 6.3. Comparison of Hastelloy B Stress-Rupture Data in Argon and in the Fuel Mixture NoF.ZrF,-
UF, (50»46-4 mole %) at ]300 1500, and '|650°F

TABLE 6 9. ROOM-TEMPERATURE TENSILE PROPERTIES OF HASTELLOY B IN THE
:SOLUTION ANNEALED AND SPHEROIDIZED CONDITIONS

 

  

 

   

‘Yi'e'rld "S_t'i'é-n'gfh ‘Elongation

 

 

- curves for 500—
- max1mum Tat ag

  
 

Y ofSpecimen (p's'i)"" - (%)
- Spheridized 175000 10 185 000 100, ooo 0115000 151020
‘Solution-anneal 125,000 10 130,000 50, 000 to 60,000 55 to 65

 
   

 

   
 
  

       

 

- tempera?ure ’rensule properhes obtamed by a specml
~ spheroidization heat treatment® are compared in
"Figs. 6.6 and 6.7 with those of solution-annealed

‘and 'IOOObeZ agmg ’rlmes. reoch a
e'l)-/' 1300°F, while the

-hes o _minimim at this samer_’

_j‘dUCfl[Ify Curve ré
femperature. “The results of meféllographlc studles,:"
‘discussed prewously, “indicafed that this was |
caused by the ‘exfensive precipifation of a second

 

phase — beta or a combination of two phases, beta
and gamma — throughout the matrix. The high-

 

“material in the unaged and in fhe uged conditions,

"'i"'Although ‘the sphermd:zahon ’rrecn‘ment was in-

" tended to produce a stable microstructure between

1300 and 1600°F, it may be noted that the ductility
is lowered considerably in the 1100 to 1300°F

127

 
 

ANP PROJECT PROGRESS REPORT

O CRt
ORNL—LR—DWG 7561

TESTED IN HYDROGE TESTED [N AIR

TESTED IN ARGON

ELONGATION (%)

TESTED IN thF--ZrI-",_‘—UF4 {50--46--4 mole %)

 

o 200 420 600 800 {000 1200 1400 1600 1800 2000 2200 2400 2800
TIME (hr)

Fig. 6.4. Creep Curves for Hastelloy B Sheet That Was Solution Annealed ot 2100°F and Tested at
1500°F and 12,000 psi in Various Environments.

UNCLASSIFIED

temperature region. Even though very high tensile
ORNL-LR—DWG 7562

and yield strengths were obtained with this treat-
1000 HR ment, the low high-temperature ductility would

"O-0— 500 HR probably prevent useful application of this material

% "o 100 MR above 1000°F. Investigations of other promising
< pretreatments are now in progress, as well as
investigations of the composition variables which

1o

-~ 400

W0
O

o]
o

TENSILE STRENGTH &/‘%\
G
2,

 

M
o
5
S .
€70 SILE STRENGTH & have been shown to be of significance.
o 60 5
— N
g o P PO & DEVELOPMENT OF BRAZING ALLOYS
< ‘{\E\' DO//!/
. ,

E 0 YIELD PO P. Patriarca . G.. M. Slaughter
= *—e Metallurgy Division
© 30 g
& e % R. L. Heestand
Y 20 * 2 Pratt & Whitney Aircraft
2 DUCTILITY | DUCTILITY S
= S

L

DUCTILITY High-Temperature Oxidation Tests

1100 1200 1300 1400 1500 1600 Tests for evaluating the static oxidation re-

TEMPERATURE (°F) sistance of several brazing alloys were conducted

U _ previously at 1500 and 1700°F, and cyclic tests
Fig. 6.5. High-Tempercture Tensile Properties have now been initiated, In the 500-hr cyclic

of Hastelloy B That Wes Solution Annecled at tests the samples are subjected to 190 air cools

2100°F Prior to Aging at the Testing Temperature from 1500°F, The results gbtained thus far are

for Various Times, compared with the results of the static tests in

128
 

a)

»n

UNCLASSIFIED
ORNL-LR~DWG 7563

140

130

120

110

100

20

80

70
= SPHERQIDIZED
- SOLUTION ANNEALED
2 HR AT 2100°F

60
50
40 40

30 30

TENSILE STRENGTH AND YIELD POINT ( psix{0™ 3}

20 20

ELONGATION (%)

10 DUCTILITY 10

 

0 O
1100 1200 1300 1400 1500 1800
‘ TEMPERATURE (°F)

Fig. 6.6. High-Tempefufdre Tensile Properties
of Hastelloy B in the Solution Annealed and
Spheroidized Conditions.

Table 6.10,
alloys tested have good resistance to oxidation
under both static and cyclic conditions.

Tests are now under way for evaluating the
attack resulting from cycling from 1700°F. An
apparatus is also being prepared for testing the
oxidation resistance of brazed joints in moving

It may be seen that many of the

moist air.

Physwal Properfy Tests _4

 

 

An mveshgcn‘lon “has been conducfed of accumu-"

S "_ICmVe effecfs on the physical properhes of InconeI/
tubmg a5 a@” result of fcbrlcatlon of 'rube-to-fmj_
:-_,;':-.'IJOII‘H‘S by hlgh temperature brazmg “with Coast
. Metals cxlloy No. 52 Brazed fube ‘|’O-f!n specnmens'
~with_at least 1 in, “of ‘stacked fins were machine =
o ""ground To 'rhe orlgmcl fubmg dimensions. Threaded
| : "henv brozed to fhese tubes' to

 

Room- cmd hlgh ’remperofure fens;le tests were
“then conducted on The as-brazed specimens and

on specimens that were subsequem‘ly heated in
a vacuum for periods up to 500 hr at 1500°F prior
to testing. The results indicated that no signifi-

- heat treatments.

N g2

  

 
 
 

PERIOD ENDING JUNE 10, 1955

UNCLASSIFIED
ORNL~LR—-DWG 7564

140

430

 

120

¥y
140 o Py
r

"
Q
>
‘@
2100
'_
=
o 90
o,
O
3 80
W
>
g 70
<T
I
I 60
LD
&
W 50
-
[p]
S 40 a0
% | =—=SPHEROIDIZED 5
Z 30— SoLUTION ANNEALED 30 z
— ° =
20 2 HR AT 2{00°F 20 '<_I
o
=2
10 10 o
—~— L
0 0
1400 1200 1300 1400 1500 1600

TEMPERATURE (°F}

Fig. 6.7. High-Temperature Tensile Properties
of Hastelloy B in the Solution Annealed and
Spheroidized Conditions After Aging for 100 hr at
the Test Temperature.

cant decrease in strength or elongation of the
tubing resulted from the brazing and subsequent
Tensile strengths of approxi-
mately 100,000 psi and elongations in 1 in., of
approximately 30% were recorded in the room-
temperature tests. At 1500°F, the samples pos-

-sessed tensile strengths of approximately 27,000
e psi, and they retained good ductility.

Test somples are now being prepared to evaluate

‘“""'the effects of other brazing alloys, such as Nicro-

braz and fhe 90% NI-IO% P a]loy.

FABRICATION OF TEST COMPONENTS

P. Patriarca
R E. Clausing G. M. Slaughter
Metallurgy Division

. R.L. Hees’rond
Pm'rf & Whlfney Alrcraff
Twenfy-Tube Heat Exchanger

A fuel-to-NaK heat exchanger containing 20
Inconel tubes was fabricated for studying the heat
transfer characteristics of fluoride fuel mixtures

129

 

 
~ TABLE 6.10.

.ANP PROJECT PROGRESS REPORT

COMPARISON OF RESULTS OF STATIC AND CYCLIC 500-hr OXIDATION TESTS
OF DRY-HYDROGEN-BRAZED INCONEL T-JOINTS AT 1500°F

 

 

Oxidation Results*

 

Tested f;)l’ 500 hr .

 

 

 

 

Brazing Alloy Composition (wt %) Tested in
, ' Static Air with 190 Air Cools
for 500hr  to Room Temperature
Commercial Alloys
Nicrobraz 70 Ni~14 Cre6 Fe~5B—4 Si—1 C Stight Slight
| ffLow-meif.ng Nlcrobroz 80 Ni—5 Cr—6 Fe~3 B_5Si_1 C Slight Slight
| ;:---_'coqst Metals No 50  93Ni-3.55i-2.5 B_1 Fe Slight Slight
51 92 Ni=4.5 $i -3 B-0.5 Fe Slight Moderate
52 89 Ni~5 Si—4 B_2 Fe Slight Moderate’
53 81Ni—d Si-4 B_g Cr-3 Fe Slight | s‘l‘;giéf |
EEE T C 'NP 50 Ni-12 Si—28 Fe—4 Mo—4.5 P—1 Mn—0.5Cr  Slight ‘Sllght
biomd Ni Co Anoy 64 Ag-33 Pd_3 Mn Severe Severe
"copper - 100 Cu | Complete Complete
Experimental Nickel-Base Alloys
G-E No. 62 69 Ni—20 Cr—11 Si Slight Slight ]
81 66 Ni—19 Cr—10 Si—4 Fe—1 Mn Slight Slight ;
Ni-Cr-Si 73.5 Ni=16.5 Cr=10 Si Slight Slight J
Ni-§i 88 Ni—12 Si Slight Slight i
Ni-Ge 75 Ni—25 Ge Slight Slight
Ni-Ge-Cr 65 Ni_25 Ge-10 Cr Stight Slight
Electroless Ni-P 88 Ni~12 P Slight Moderate :
Ni-P-Cr 80 Ni-10 P_10 Cr Slight Slight
Ni -Mo-Ge 50 Ni~25 Mo—25 Ge Stight Slight
Ni-Sn 68 Ni—32 Sn Moderate Severe
Ni-Mn 40 Ni—60 Mn Complete " Complete
Ni -Mn-Cr 35 Ni~55 Mn—10 Cr Severe Severe
Experimental Precious-Metal Base Alloys
Pd-Ni 60 Pd—40 Ni Slight Slight | i
PANiSi 60 Pd=37 Ni_3 Si Slight Slight A S
Pd-Al 92 Pd_g Al Very slight Moderate
 PdGe 90 Pd—10 Ge Stight Severe
AuNi 82 Au-18 Ni Very slight Sllght
Ad-Co 90 Au—~10 Co Very slight Stight .
B u-Cu 80 Au-20 Cu Complete Complete

 

 

 

 

 

 

 

im0

- *Very sllght less than 1 mil of penetration; slight, 1 tc 2 mils of penetration; moderate, 2 to 5 rn||s of penefratlon,

severe, greofer fhan 5 mils of penetration; complete, fillet completely destroyed.

 
 

'

0

o

over a wide range of Reynold numbers (cf. Sec. 2,
‘*‘Experimental Reactor Engineering’’). The design

- of this heat exchanger specified that the 3/] 6-in.-0D,

0.017-in.-wall Inconel tubes were to be joined to a
dished header at the NaK inlet end and to a radial
header at the NaK outlet end. A thick-walled
Inconel pressure shell was to be used to confine
the molten fuel located outside the tubes.

The fabrication techniques used on this heat
exchanger consisted essentially in manual heliarc
welding and subsequent back brazing of the tube-
to-header joints. All welding was performed by
qualified operators? using specified procedures,?
and complete penetration was utilized, except on
the tube-to-header joints, Complete protection of
the roots of the welds by an inert gas was em-
ployed at all times to minimize scaling and oxida-
tion, The tube-to-header joints were back brazed
to eliminate the ‘‘notch effect’” resulting from
incomplete weld penetration and to ensure against
the development of leaks in the event of corrosion
through an area of shallow weld penetration.

An evaluation of the corrosion resistance of
brazing alloys had shown that low-melting Nicro-
braz? was compatible with both sodium and the
fused fluoride salts, and therefore this alloy was
used in the back-brazing process, which was per-
formed in a dry-hydrogen atmosphere by using the
conventional ‘‘canning’’ procedures. Two brazing
operations were required because the Globar pit
furnace available did not possess a heating zone
of sufficient length to heat both ends of the unit to
the brazing temperature. A ‘‘window’’ was removed
from the pressure shell to permit preplacement of
the brazing alloy on the NaK inlet end of the tube
bundle, as shown in Fig. 6.8. The window was

‘welded shut after plucemenf of the alloy, ondi
helium leak- teshng indicated that all welded and
~ brazed joints were |eol<'r|ghf. The completed l'leoi”'
' exchanger is shown in F:g. 6 9 '

   

lnfermed:mie Heot Exchcmger No., 2

     

The semlautomctlc ellcrc we\

 

 

7T R. Housley cmd P Pfl‘l'l’lcl'ca, Operafor S Qua||f|.
"cohon Tesf QTS.

T R. Housley cmd P Pafrlorco "Procedure Spec:fu-;-m

cohon, PS-1.

A brazing alloy developed by the Wall-Colmonoy
Corp., Detroit, Mich.

 

|‘ng- of fhe tuloe-'” ‘
’ro-heoder V|olnfs of the two “fuel-to- NaK heaf ex-v_":"'
chongers for the intermediate heat exchanger No. 2
ot ‘tesf progrom WGs":'d'"k oussed prewously.l0 The

 

PERIOD ENDING JUNE 10, 1955

UNCL ASSIFIED
Y-14799

 

l=ig. 6.8. Twenty-Tube Fuel-to-NaF Heat Ex-
changer Showing ‘‘Window'' Cut in Pressure Shell

to Permit Preplacement of Brazing Alloy on NaK
Inlet End of the Tube Bundle. '

400 welds were heliu'm lecktighl"before initiation
of the stages of fabrication described here.

Header Welding. In order to minimize cracking

of the bri\‘-'rl\e alloy used in back brazing and to
“avoid contamination of areas to be welded with the
boron and 5|lzcon of the brazing alloy, the header
'componen’rs were heliare welded to each tube
““-bundle before back braz:ng of the tube-to-header
. joints, However, to permit access to the joints,
“_fhe components were not attached until after pre-

plocemenf of the brozmg alloy (low-melhng Nlcro-
braz) had been completed.

The welding is shown at an early stage in Fig.

6.10. Strongbacks were used to minimize distor
_tion of the header and consequent tube:to-header

 

IOP. Patriarca et al., ANP Quar. Prog. Rep. Mar. 10,
1955, ORNL-1864, p 131.

131

TR T

 
ANP PROJECT PROGRESS REPORT

 

s o

 

 

3 Y
UNCLASS)
Y-14480

 

 

 

Flg.6.10. Welding of a Header of the Inter-
~ . ‘mediate Heat Exchanger No. 2, The strongbacks
used to minimize distortion may be seen.

Cowm

 

 

UNCLASSIFIED
Y-14796

 

 

7: -:‘r'Fig. 6.9. Completed 20-Tube Fuel-to-NaK Heat Exchange'.'-v

weld fracture. Joint designs were used that per-
mitted complete penetration,!! and qualified
operators and specified procedures were used at
all times. A photograph of a completed, leaktight,
header section prior to back brazing is shown in

Fig. 6.11.

Back Brazing of Tube-to-Header Joints. The
low-melting Nicrobraz alloy was preplaced on the
headers as a dry powder and then secured firmly
in position with a methylacrylate cement. Since
the heating zone of the Globar-heated brazing
furnace was not long enough for a complete tube
bundle, only one end of the tube bundle could be
brazed at a time. As a result, it was necessary
that the complete welding and brazing of one
header on one end of a tube bundle be done prior
to the brazing alloy preplacement dnd header
welding of the other end so as to avoid spalling
of the preplaced brazing alloy from the underside
of the cold header during the brazing of the first
header. It was thought that hot hydrogen gas
might volatilize the binder on the cold header and
leave the brazing alloy insecurely positioned in

¥

 

Nt R, Housley and P. Patriarca, Joint Design of
Inert-Arc Welded Pressure Vessels for High-Corrosion
Application (to be issuved).

k)

 

 

 
 

€

4N

.y

 

     
  
 

UNCLASSIFIED

Fsg 6.11. Cdrfiflered”fleuder of an Intermediate
Heat Exchanger No. 2 Tube Bundle.

its overhead locations. This important considera-
tion increased the fabrication time to some degree,

~since the welding of both heads and the brazing

of both heads could not be performed in sequence.

 The dry-hydrogen furnace brczmg wds performed‘_h
' w:‘rh the tuue bundle m a horlzontal posn‘lon m}'

  
   
 
   

 

   

?he reforf durlng' bra'z‘mg. CAT hr\nke-'remperatqre
record of every brazmg ‘thermal cycle”was medsured

 

" by means of a Chromel Alumel thermocouple flrmly';"m

I

tqflcched‘ to ’rhe assembly, cmd a plot of a typlccl -
rcycle is shown in | Flg. 6.12. A preheated_brazmg:
o furnace 'w_as usedf’ro ‘shorten the nme'l requnred for

'_rpcge or dis-

 

to be the most practical means for holding the
tube bundle rigid over its free-span length and for

 

and a dry-hydrogen‘ ofmosphere was ’mcun’rmned ln‘“

 

. omblike “Spucers. -
The use . "of wire comblike spacers “was considered

PERIOD ENDING JUNE 10, 1955

UNCLASSIFIED

ORNL-LR-DWG 7565
1200

1400
1000
900
800
700

600
500 FURNACE
COOLING

BRAZIN

TEMPERATURE{°C}

400
IR

300 COOLING
200
100

0

PREHEAT

 

0 10 20 30 40 50 60 70 80 920
TIME (hr)

Fig. 6.12. Brozing Cycle Used in the Fabrication
of Intermediate Heat Exchanger No. 2.

keeping the tubes separated in such a way as to
permit the required mode of fluid flow between
them. The ‘‘combs’’ were placed at 5-in. intervals

_along the bundle, with alternate spacers positioned

90 deg to each other. It was required, for the
heat transfer evaluation, to place the spacers on
one bundle perpendicular to the axis of the tubing,

" whereas on the second bundle they were placed in

‘a plane at a 30-deg angle to the tubing axis. Both
types of spacer arrangement are shown in Fig.6.13.
It may be seen that the fabrication and assembly
of these spacers required the use of precision
jigs and careful determination and control of the
cone-arc plug-welding. conditions.  Meticulous
care also had to be maintained when the holes
were punched in the strip headers, because ex-

,_,_jr_emely close folerances were specified.

Installahon of Tube Bundle unfo Pressure Shell
" The mstallm‘lon of the tube bundles into fhe:r_

‘pressure shells reqmred hgh’r mcchlmng of them

shell to permlt proper flt-up, that is, “each pressure '
“shell had to be precrsuon machmed for its corre~
spondmg tube bundle to accommodate the very
smcll vurlqhons in dlmens:ons.._ ‘The heavy-walled
pressure “shell ‘was beveled accordlng to the

'recommended |0|nt desrgn ‘and, as prescrlbed for

“high- corrosaon appilcchons, the root pass was
' :lded inder of the weldmg was
performed by usmg “the metall:c-arc process to
"minimize the heat input and consequent distortion
of the unit. Figure 6.14 shows the heat exchanger

after completion of the root pass. The large

 

133

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

B NCL ASSIFIED
B v.15047

 

Fig. 6.13. Sample Tube Bundle Showing Two Types of Comblike Spacer Arrangements.

I-beam was used as a strongback for the system
to prevent severe distortion.

Joining of Two Tube Bundles. The two tube
bundles were joined together after being welded
into their respective pressure shells; the com-
pleted unit is shown in Fig. 6.15. It may be seen
that the joining operation required several manual
heliarc welds to connect the NaK circuits and to
attach the inlet and exit nozzles. Leak-testing
 with a helium leak detector indicated no flaws,
. and the heat exchanger was subsequently delivered
" for installation into the test rig. A detailed report

:':t:b\'!ering the fabrication of heat exchangers is
" being prepared. .,
NaK-to-Air ‘Radiufors'

It was reported previously '2 that the fabrication

. of two 500-kw NaK-to-air radiators had been
initiated for use in the intermediate heat exchanger

134

No. 2 test program and that the type 310 stainless-
steel-clad copper high-conductivity fins had been
sheared, degreased, and edge-protected with
aluminum bronze. The construction of these units

has now been completed, and a brief discussion

of the intermediate steps is presented here.
Preplacement of Brazing Alloy. Coast Metals
brazing alloy No. 52 was designated for use on
this radiator because it possesses good flowability
at 1020°C and has good resistance to high-tempera-
ture oxidation. Its good compatibility with sodium
was also of distinct advantage, since the brazing
alloy might come in contact with the NaK during
service if used to seal small pin-hole leaks or if
sufficient weld metal corrosion occurred to expose
a braze metal surface. Also, severe tube-wall

 

126 patriarca et al,, ANP Quar. Prog. Rep. Mar, 10,
1955, ORNL-1864, p 128,

TR

 

 
 

 

LI

  
  

  

  

PERIOD ENDING JUNE 10, 1955

 

Fig. 6.14. Intermediate Heat Exehanger No. 2 After Completion of the Root Pass of the Pressure

Shell Weld. Strongback used to minimize distortion.

dilution durmg brozmg or sohd-sfcte dlffusmn in
servnce.mlght produce a hlgh concen_trano of
bruzmg q||oy consh'ruents near the cwculdt-lng NaK.

The dry-powder method of brazing a”oy pre-

 

 

 

 

- R

placement ‘was_used in thefl fabrication offithese
“tadiators. VA sfamless steel templcte confdmmg
ho[es thf were precusmn drn”ed was, uhhze

 

       
   

   

 

  

"‘,- PN

with Ni i crobrai cemenf

 
 

-m.-thlek

    

‘the radmfoi’ was conduc’re'df with the
down, as is shown in Fig. 6. 16. The cssembly of
the first 4 in. of fins was extremely time-consuming

 

  

 
    

  

  

“The brazing powderflwas secured |

geometry of the fm. Supportlng sheet-me‘ml chcnn_e]s
|mporf over-all’

  

because the tubes were not rigidly in place. A
h{equ metal template was requnred to force each
md:v;dual fin into place. Since some of this diffi-
culty ‘was also derived from the slight curvature
of the stralghf |engths of ’rublng, it is thought
-~ improvements in the bendmg fechnlques
mighf substcnhally assist cssembly of future units.
' A se'r of Luche fmger 1igs was desngned and

 

 

 

     

      
     
 
   
  
 
   

were placed at 4. m.{r intervals fo‘_
§tructural sfobllnfy to the radiator, These channels

   

 

‘were heliarc welded together to form sheet-metal

 

L <

side plates. "Two thin sheets of Inconel were

also placed tightly together against the channels
at these 4-in. intervals. The capillary joint be-

135

 

E.

 

T

R T

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

UNCL ASSIFIED
Y1810

 

Fig. 6.15. Completed Intermediate Heat Exchanger No. 2,

tween the two Inconel sheets acted as a sump to
accommodate any excess brazing alloy.

Assembly of Headers. The original design for
the radiators required that all tubes enter the
cylindrical headers normal to the curvature at the
point of entrance. Sample specimens were prepared
to determine the tube-bending and fitting techniques,
but the complex tube design caused extreme diffi-
culties in placing the headers on the tubes.
Therefore a new design was proposed and accepted
that utilized a split header of 3-in. schedule-40
‘pipe so that all tubes except the two outside rows
 on each radiator half-section entered the headers
- without being bent.

Careful hand polishing of each of the tubes was

- without the use of additional force from a hand

~ press.  Precision _dfiiling and deburring of the
- headers were also employed to prevent bending

qnfljrd'i_st_brt_'ion of the thin-walled Inconel tubes.

 

After assembly of the headers, the tubes were
meticulously hand filed to conform to the exact
curvature of the interior surfaces.  Abrasive
grinding or wet milling was avoided because
entrapped abrasive or lubricant would lead to
inconsistencies during tube-to-header welding.

Welding. The tube-to-header joints were manu-
ally heliarc welded, and a helium backup atmos-
phere was maintained on the underside of the joint
throughout welding. The split headers, nozzles,

and end plates were also welded by using manual

heliarc welding techniques. Extreme care was

taken to ensure complete penetration and good

coverage by the backup gas.

After the manifold welds were completed, the
integrity of the system was determined by helium
leak-testing. Although it was assumed that pin-

hole leaks could be sealed during the subsequent

brazing operation, large leaks were to be avoided.
When questionable welds were found, the headers

 

 

 

 

 

 
 

%)

 

 

'VU*‘;?Ube to-header |omts. -
A ’specml baffle de5|gn was used for the brazing
j_\:_of ’rhese rqc]m’rors,‘ since it wa very important fo _
7 "‘Eflus'h out the copious™ amounfs_\\roffVmethylacryIatew;i'__
:",'bxnder present in 'The fube-to-fin matrix. A sheet-

.,;metai cover was u’rtcchecf to 'rhe fronf of fbe

radiator, “and” clean, “dry linfet’ gas was forced

    

UNCL ASSIFIED

 

Filg.-".6“"|6 Early Stages of Assembly of NaK-
to-Alr Rudmtor.

weréf r'é"m'ol\)e'd bylnsiniply cutting off the tubes at
the underside of the header plates. New headers
were then attached by rebending the tubes and
repolishing them. This procedure was employed
on the radiators in order to obtam leckhghf unn‘s
before bmzmg o -
g. The back’ brazmg of the tube-to-heade

 

  

t,."fdroppe-r was 'y ed to place the broze slurryy;on the

 

 

 

 

between the fins before being exhausted to the
outside of the can. The inside of the sheet-metal

PERIOD ENDING JUNE 10, 1955

cover and inlet tube was designed to permit
relatively even flow throughout the entire frontal
area, and good circulation of hydrogen was ob-
tained around the tube-to-header joints.

The radiator was securely supported on a strong-
back to prevent sagging or distortion during the
high-temperature brazing cycle. The support bars
were made from stainless steel which had been
previously aluminized ot 1500°F. This alumi-
nizing prevented the support bars from being brazed
to the radiator, since an aluminum oxide film
prevents wetting by Coast Metals alloy No. 52.

The radiator was placed in the brazing can with
four thermocouples attached in various positions
around the periphery. This temperature measure-
ment around the radiator was desirable from a
contrel standpoint, even though previous experi-
ments had indicated that all parts of a unit of this
size could be held to within 25°C at the brazing
temperature.  In the preliminary experiment, 12
thermocouples were attached at various points
over the test specimen. This close control was
necessary because the flow temperature of the
brazing alloy is 1020°C, and the copper in the
fins melts at 1083°C, |

The thermal cycle used in brazing these radiators

‘is shown in Fig, 6.17, The unit was placed in the
furnace at room temperature to minimize distortion

and was gradually raised to the brazing temper.
ature. The two completed radiators are shown in
Fig. 6.18; as may be seen, negligible distortion

UNCLASSIFIED
ORNL-LR-DWG 7566

  
  
  
 
  

d the ‘br;azlng rof ihe tube-fo-f:n |omfs

   
  

HEATING FURNACE — AR
. COOLING COOLING

BRAZING

0 05 _4.0 45 2.0 25 30 35 40 45 50 55 60 65 70
' TIME(hr)

Fig. 6.17. Brazing Cycle Used in the Fobrl cation
of a 500-kw NaK.to-Air Radiator.

137

 

et g e

 
ANP PROJECT PROGRESS REPORT .1
UNCLASSIFIED | |
Y-15017

 

 

 

 

 

 

L R : Fig. 6.18. Two Completed 500-kw NuK-fo-Air Radiators. ,
138

 

 
 

+%

a5

 

 

 

occurred. Smce no excess of brazmg alloy was
found, it appears that the sump technique for the
removal of the excess is practical for large as-
semblies. Good flowability of the brazingalloy
was oBfoihed. A cross section of a typical tube-
to-fin joint brazed under conditions simulating
‘those used in the fabrication of the radiators is
shown in Fig. 6.19. It may be seen that good edge
protection of the exposed copper was obtained in
these tube-to-fin joints and that good fillets were
present in the tube-to-header joints.

  
  
  
 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
  
 
 
 

UNCL ASSIFIED |
.. Y-15043

“were therefore dellvered for 1nsta|iahon ‘into the

test rrg.

PERIOD ENDING JUNE 10, 1955

Comell Radiator No. 1

The assembly and partial fdbncatlon of a full-
scale liquid metal-to-air radiator designed by the
Cornell Aeronautical Laboratory were described
previously.!? Since the header plates on this heat
exchanger were 5/8 in. thick, the brazed tube-to-
header joints cracked easily, Severe stresses re-
sulted from the differences in thermal expansion
coefficients between the stainless steel base metal
and the brazing alloy. However, an investigation
of several high-temperature brazing alloys revealed
that Coast Metals alloy No. 50 possessed suffi-
cient ductility to overcome this cracking upon fur-
nace cooling from its brazing temperature of 2050°F,
Other brazing alloys such as Coast Metals alloy
No. 52 and G-E alloy No. 62 cracked to such an
extent that they were not comldered satisfactory
for this application.

Since brazing with Coast Metals alloy No. 50,
which is a nickel-silicon-boron-iron alloy, would
necessarily result in the diffusion of boron into
the type 316 stainless steel base metal, an inves-
tigation was conducted to determine the extent of
penetration and the effects on the mechanical prop-
erties of the base metal. Metallographic examina-
tion of typical brazed tube-to-header joints indi-

‘cated that the diffusion was minor (2 to 3 mils),

even though long-time heating and cooling cycles
were used. Samples examined after subsequent
heating for 24 hr at 1650°F also exhibited diffusion
to the same extent. After 100 hr at 1650°F, the
diffusion had progressed to 4 to 5 mils. Several
brazed tensile-test samples were prepared so that
a study could be made of the effects of brazing on
the mechanical properties, Tests at room temper-

ature ‘and at 1600°F at the Cornell Aeronautical

Laborafory indicated no appreciable reductlon in

"the tensile strength of the base metal.

The back brazing was carrled out by placmg the
in the” furnace at room temperature and

) ‘hectm‘ggfldt a rqfe\of ]50°C/hr unh[ the brazmg tems
} __percture was reoched The radlator was held at
" 2050°F for /2 hr and then furnace cooled at a-max-

 

of 150°C/hr, The complefed heat ex-
changer is 'shown in Flg. “6.20. Testing with a
mass-Spectrometer helium leak detector indicated

   
 

‘that all welded and brazed joints were |eok'rtght.

 

3P Patrlarca et al., ANP Quar. Prog. Rep Mar. 10
1955, ORNL-1864, p 132.

139

o

 
PR

Lt

 

 

 

 

ANP PROJECT PROGRES'S REPORT

UNCL ASSIFIED
Y+14668

Fig. 6.20, Completed Cornell Aeronautical Labo-
ratory Sodiumdo-Air Radiator No. 1.

The radiator has therefore been delivered to the
Cornell Aeronautical Laboratory.

SPECIAL MATERIALS STUDIES
Special Alloy Extrusions

J. H. Coobs H. Inouye
Metallurgy Division

M. R D Amore Pratt & Whitney Aircraft

Three blllets of vanadium were extruded at
2000°F, The pressure required for an extrusion

yrroho of 9: 1 was approximately 45 tsi. The billets
were ccmned in mild steel to protect the vanadium

from oxndohori durmg extrusion. The canning
materml was removed from the extruded tube
blonks by p:ckling in hot HCl. The tube blanks
have been shipped to the Superior Tube Co. for
further reduchon and cladding on the outside with

‘stcmless steel. The clad tubing will be used in

cofrosion studies.

o

Four high-purity molybdenum billets containing
0.7% titanium were exiruded into rod to be used
for welding studies at Battelle Memorial Institute.
The extrusions were made at 2600°F, and it was
demonstrated that the pressure required could be
reduced by using a salt-type lubricant. These
billets failed to extrude in the first two attempts
when Fiberglas and Necrolene were used as the
lubricating materials. Since the salt lubricant
thermally decomposes at 2600°F, it was swabbed
onto the container walls. The use of salt in the
container reduced the pressure required from 700
to 427 tons on a 3-in. ram for an extrusion ratio

of 4.5:1. Similarly, at an extrusion ratio of 6.25:1,

the pressure required was reduced from 700 to
508 tons.

The studies of flow patterns of duplex and
three-ply materials during impact extrusion were
continued. - The flow pattern studies of three-ply
composites have been directed toward the pro-
duction of stainless steel seamless tubular fuel
elements. The experiments on duplex materials
are designed for studying the fabrication problems
involved in the cladding of tube materials. Six
additional extrusions of composite stainless steel-
mild steel billets have been completed. The
results of the extrusions have not been completely
evaluated, but it appears that material recovery
may be improved by tapering the billet nose to
conform with the cone of the die.

Clad-Columbium Fabrication

J. H. Coobs H. Inouye
Metallurgy Division

M. R. D’Amore, Pratt & WHitney Aircraft

Roll-Cladding Experiments. A series of tests
have been made of various combinations of heat-
resistant alloys roll clad onto columbium. - The
experiments were designed for determining the
lowest temperature at which bonding could be
achieved to minimize intermetallic formation and
contamination (the reactivity of columbium with
gases increases with temperature). In addition,
attempts were made to bond columbium to columbium
in order to develop a method for obtaining large
sheets of the metal. Rolling temperatures of 900,
1050, and 1250°C were investigated. Metallo-
graphic examinations of the columbium to columbium
combinations prepared showed that the interfaces
between the layers were not discernible at high

Tk

 

 

 

 

 
 

3

 

)

magnifications after hot rolling at all temperatures
investigated. Separation did not occur upon severe
bending. I, therefore, appears feasible to obtain
large sheets of columbium by this method.

In the investigation of heat-resistant alloys as
roll claddings for columbium, maintaining the
metallurgical bonds necessary for heat transfer
was expected to be difficult because of the differ-
ences in thermal expansion, as shown in the
following:

Mea.n Coefficient
~ of Thermal Ekpansion
: at 0 to 815°C

(,_Lin./OC) |
Columbium 7.1
Type 446 stainless steel 1.2
Type 310 stainless steel 18.0
Inconel 16.0
Hastelloy B 15.3
Nickel 13.1

 

PERIOD ENDING JUNE 10, 1955

It was further expected that brittle intermetallic
compounds would form between nickel and co-
jumbium and between iron and columbium.

Service tests of columbium clad with these heat-
resistant alloys were made at 815 and 1000°C for
500 hr. The columbium was found to be adequately
protected, and the effects on ductility were slight.
Interface reactions were, however, evident in all
combinations, and there was, as a result, sepa-
ration at the interfaces.

The interface of the Inconel-columbium composite
after rolling at 1050°C is shown in Fig. 6.21. A
discontinuous layer about 0.0001 in. thick is
evident., The reaction layers increased in thick-
ness with the rolling temperature. The same com-
posite is shown in Fig. 6.22 after heating at 815°C
for 500 hr. The reaction layer is only about
0.0005 in. thick, but it is sufficiently thick to
cause separation upon cooling of the composite
from the test temperature to room temperature.
Voids caused by unequal diffusion rates of the
elements are evident in the Inconel layer.

 

 

0.003

 

0.004

 

C.005

 

c.0086

 

0.607

 

0.008

 

0.009

 

 

 

Fié. 6.21. lnterfot.;e of ihéone.-CéluniBium':Compo.site After Rolling at 1050°C, 500X.

  

141

 

 
 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

  

 

UNCLASSIFIED
[ Y-14870

INC H

0.002

 

0.003

 

0.004

 

0.005

 

0.006

 

0.007

 

 

0.008

 

0.009

 

 

 

 

Fig. 6.22. The Inconel-Columbium Interface Shown in Fig. 6.21 After Exposure at 815°C for 500 hr.

Inconel-clad and type 310 stainless-steel-clad
columbium composites tested at 1000°C for 500 hr
are shown in Figs. 6.23 and 6.24. The reaction
layer in these composites is about 0.0015 in,
thick, and the brittle nature of the reaction layer
is exhibited by the transverse cracks. Similar
results were obtained for other composites of heat-
resistant alloys and columbium.

Diffusion Barrier Studies. The use of a diffusion
barrier for preventing separation at the interface
of the cladding and the columbium was investi-
gated. Copper was selected for this study on the
basis of the results of some brazing work reported
by the Fansteel Metallurgical Corp. A 0.001-in.
layer of copper was roll clad onto columbium at
1000°C from an evacuated capsule. Examination
of specimens tested for 150 hr at 815°C showed
‘two reaction layers totaling about 0.0001 in. in
t_hic_;knes'rs,' and separation at the bonded interface
was slight or nonexistent.

142

If, after additional experimentation, copper is
found to be a suitable barrier for columbium, an
additional barrier for the copper will be necessary
if the external cladding is to be Inconel. Previous

work has shown, however, that iron or the 300

series stainless steels are satisfactory barriers
between copper and Inconel.

B,C-Cu Shielding Material
J. H. Coobs H. tnouye
Metallurgy Division
M. R. D’Amore, Pratt & Whitney Aircraft

The studies of boron-containing shielding ma-
terials, reported previously,M were continued. A
program has been initiated for evaluating the
compatibility of Inconel and a boron carbide—~copper
mixture at 1650°F, since, for reactor use, it will

 

4H. Inouye, J. H. Coobs, and M. R. D'Amore, ANP
Quar. Prog. Rep. Mar. 10, 1955, ORNL-1864, p 136.

o

 

 

 

 

 

 

 

 
 

%y

“@)

@

UNCL ASSIFIED
¥-13033 )

 

_ Fig. 6.23. Interface of Inconel-Col;finbium Com--
posite After Testing ot 1000°C for 500 hr. Un-
etched. 200X. Reduced 30.5%.

be necessary for the boron-containing material to
be clad with inconel. For the initial studies, the
B,C-Cu mixture was prepared by pressing powder
mixtures into compacts 1x 3 x 0.337 in. containing
40, 50, and 60 vol % B,C, respectively. The
boron content and the density achieved after pres-
sing, sintering, and coining of these compacts
are listed below:

G fB
B6C (vol %) % Theoretical ramsa et Soron
per cm® of Mixture

40 89.3 0.737
50 . 86.3 ' 0.890

60 . 838 1.03_8' o

: On the bas:s of prewous work a duplex dlffusnon
barrier con5|shng of foils of iron and copper ‘was
inserted befween ’rhe Inconel and B C-Cu mixture.

The componem‘s were assembled in a picture frame

- ‘and rolled at ]800°F to bond the |ncone! clcddmg,

diffusion’ bamers cnd B6C -Cu core. The quality

of the bond qumed has not been eva[uated but

i x-rqy ana|ys:s after a hot reduction of 50% revea[ed_
- that no crucklng of 'rhe B6C -Cu had occurred in
_any of the three composn‘es studled These ma-

terials W||| be sub;ected to room- temperaturé bend“
tests and to hehum permecblhty and Vmaterlcl'

- compohb:hty tests at ]650°F "

 

 

155, H. Jackson et al., J. Metals 185, 149 (Feb. 1949).

PERIOD ENDING JUNE 10, 1955

© UNCLASSTFIED ©
Y3034

Fig. 6.24. Interface of Type 310 Sfoinléss Steel=
Columbium Composite After Testing at 1000°C for
500 hr. Unetched. 200X. Reduced 30.5%.

Magnesium-Lithium Alloy

D. A. Douglas J. H. DeVan
J. W. Woods
Metallurgy Division

A magnesium-lithium alloy is being considered
as a possible shield material for use in the con-
struction of an aircraft crew compartment. Ad-
ditions of lithium to magnesium reportedly bring
about a reduction in tensile strength while greatly
increasing tensile ductility, particularly in the
case of alloys containing more than 10.3% lithium. 15
Such alloys have a body-centered cubic lattice
structure. 4

Tensnle Properhes. In order to evaluate the

properties of alloys containing relatively large

percentages of lithium, a slab ingot having the
composition 80% Mg-20% Li was obtained from
Dow Chemical Company and was rolled at 200°F
to. a sheet 0.060 in. in thickness. Specimens
having a ‘/2-in. reduced gage section were machined
from the sheet, and tensile and creep tests were
conducted on the alloy in air. Results of tensile

~ tests completed at room temperature and at 200°F

are listed in Table 6.11. A comparison of these

_values with the room-temperature properties of

pure magnesium, given in Table 6.12, points out
the decrease in strength and the increase in
ductility brought about by the alloy addition.

There is very little capacity for work-hardening
in the alloy, as indicated by the small difference

143

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 6.11. TENSILE PROPERTIES OF 80% Mg—20% Li ALLOY

 

Yield Strength,

Elongation

Modulus of

 

 

 

 

Tefi’nperdturé . o Tensile Strength ) .
©F) 0.2% Offset (psi) in 3 in. Elasticity
o (psi) (%)
‘Room 11,230 12,470 35 4.5 x 108
200 2,970 3,010 45 3.2 x 108
TABLE 6,12, TENSILE PROPERTIES OF PURE (99.80%) MAGNESIUM
' S Yield Strength, Tensile Strength Elongation Modulus of
Condition 0.2% Offset ; ‘ ' . .
' o (psi) (%) Elasticity
(psi)
Annealed 14,000 27,000 16 6.5 x 10
27,000 37,000 9

‘Hardrolled

6.5 x 108

 

between yield and tensile strengths. Also, these
strengths are quite temperature sensitive, since
at 200°F they have been reduced by a factor of 4
below the strengths at room temperature.

Creep-Rupture Properties. Creep-rupture tests on
the alloy were initiated at room temperature at
7350 and 1500 psi. At the higher stress the test
was discontinued after 7 hr when the presence
of numerous cracks indicated impending failure
to be only a matter of minutes away. Final elonga-
tion for the specimen, measured over a 3-in. gage
length, was 20%. The test of the specimen stressed
at 1500 psi is still in progress after 3200 hr.

Welding of Columbium Thermal-Convection Loops

P. Patriarca G. M. Slaughter
Metallurgy Division

Several type 310 stainless-steel-clad columbium
thermal-convection loops were fabricated for test-
ing to obtain information on resistance to dynamic
corrosion. In preparation for the fabrication of
these loops, test specimens of 0,020-in. columbium
sheet were machine-heliarc-welded with both argon
and helium as the shielding gases. |t was found
that when excellent shielding was utilized, welds
‘made wi.th'blbth gases were ductile, However,
manual welds on the same material were found to
~ be quite brittle, and thus it appeared that the high

‘welding speed possible with machine welding was
beneficial. The high rate of travel minimized the
time interval over which the weld metal was molten,

144

with a consequent decrease in the quantity of
atmosphere contamination,

Since most of the welds could not be made by
machine - welding, dry-box manual welding of the
clad-columbium loops was initiated with a helium
atmosphere in the chamber. A loop was fabricated
which was pressure-tight to 35 psi helium; how-
ever, several extremely small pin holes were
found upon leak testing of the loop with a mass-
spectrometer helium leak detector. Since these
pin holes were attributed to the difficulty of torch
manipulation inherent in the use of helium, argon
was used in the dry-box during the fabrication of
four subsequent loops. These were all found to be

leaktight,  The welding of the stainless steel
cladding was completed according to normal
practice,

Dimensional Stability Test
on an Inconel Spun Core

P. Patriarca
Metallurgy Division

A proposed design for the core shell of the ART
calls for the use of a spun Inconel configuration,
and therefore an investigation was conducted to
determine the dimensional stability of a small-
scale specimen (Fig. 6.25) obtained from Pratt &
Whitney Aircraft. The stability of this configura-
tion was tested by measuring the distances be-
tween several reference points both before and
after thermally cycling from 1500°F to ambient

 

 

 

 

 

 

 
 

¥)

a0

UNCLASSIFIED
Y-14465

 

INCHES u

Fig. 6.25. Inconel Spun Core Shell.

TABLE 6.13. CONDITIONS OF THERMAL CYCLING
OF INCONEL SPUN CORE

 

Heating Cooling

 

Cycle ] Time at 1500°F .
No Time (hr) Time
’ (hr) (hr)
1
1 4 /2 4
]
2 4 /2 4
1
3 1 /2 4
1, 3
4 1 /2 4

 

 

__'._under the condmons hsted in Tab[e 6 13 Thel
_' ,-;cychng process was performed in a pro’rectlve"
- helium atmosphere. ‘At the completion of the
~ “experiment, no recordable diametrical, axial, o
"thlckness mstcbmty was observed

  

 

Bruzmg of Cermets fo Inconel

R E. Clausmg
Metallurgy D|v15|on

 

The use of cermef vclve secfs has been proposed\'
“for certain high-temperature fused Fluoride fuel

applications. Cermets composed of titanium carbide
with a nickel binder (Kentanium type) are among

PERIOD ENDING JUNE 10, 1955

those most likely to be used (cf. Sec. 5, ‘“Cor-
rosion Research’), If these valve seats are to be
brazed to an |nconel valve body, the brazed joint
must be sufficiently ductile to withstand the stres-
ses created by unequal thermal expansion co-
efficients (10.2 x 107¢ in./in./°F for Inconel and
4.0 x 1076 t0 5.7 x 107¢ in./in./°F for the cermets),
In addition the brazed joint must possess adequate
strength and corrosion resistance at temperatures
to 1500°F.

Good wetting of the cermet has been obtained by
using Electroless-plated nickel-phosphorus brazing
alloy.1®  Satisfactory joints have been obtained
with a 0.0001-in. layer of nickel-phosphorus followed
with a 0.003- to 0.004-in. layer of electroplated
copper. The addition of a nickel shim between the
plated cermet and the |nconel is desirable, since
it serves as a cushion to dissipate the thermally
induced stresses in the joint. During the brazing
operation, the nickel-phosphorus melts at approxi-
mately 1740°F and wets the Kentanium surface;
continually increasing the temperature then melts
the copper, which alloys with the nickel-phosphorus
to form a ductile joint.

A test sample was prepared that satisfactorily
withstood two air cools from 1700°F. [t was com-
posed of @ V. x ]/ x 2 ine Kenfonlum 150A cermet
(90% TIC 10% Nl) brazed 'ro a /2-|n. Inconel plate
along a /4 x 2 in. face. A ¥ -in. nickel shim was
incorporated in the joint, The specimen was
severely cracked, but, as can be seen in Fig.
6.26, the joint remained sound. Sample joints are
now being prepared for tensile tests.

CERAMIC RESEARCH 7
Jo A, Griffin J. R. Johnson
C. E. Curtis A. J. Taylor
© Metallurgy Division
Rare-Earth Ceramics

Ceramics composed of rare-earth oxides should
combine the property of high absorption cross
section (for thermal neutrons) and the usual ceramic

~ properties of high density, strength, corrosion
_resistance, and high melting temperature. Since
“these properties are those required for a control

rod for the ART, enough Sm 5,0 and Gd,0, was

' obfc:ned for feshng. These oxs es, in the form of

bars 3 x ¥ oy % /16 in. thick (pressed from the powdered

 

¥p_ Patriarca et al., ANP Quar. Prog. Rep. Mar. 10,
1954, ORNL-1692, p 90.

145

 
 

ANP PROJECT PROGRESS REPORT

FELST T Aty e

:.UNCLA SIFIED oxides), were fired at 1300 and 1500°C in air. The
‘Y-ifng . results of physical property measurements on these .
L, B specimens are presented in Table 6.14,

Both Sm,0, and Gd,0, sintered at 1500°C to
form ceramics of good strength, High resistance
to thermal shock would not be expected because
the coefficients of thermal expansion are rather
high., The oxide Gd,0, sintered at both 1300 and
1500°C was stable in  boiling water; however,
Sm,0, sintered at 1500°C was not stable.

A mixture of the oxides (available commercially)
containing 63.8% Sm,0, and 26.3% Gd,0, (balance,
other rare-earth oxicfes) was fabricated into cylin-
ders, and these specimens have been tested in the
critical experiment (cf. Sec. 3 ‘‘Critical Experi-
ment’'),  After heat treatment at 1500°C, the
cylinders had a density of 10.94 and a porosity of
8%, and they were stable in boiling water. They
were machined to the following dimensions: ¥ -in.-
oD, ‘/2-in.-lD, 0.8-in.-long hollow cylinders, and
0.45-in.-0OD, ¥,-in.-long solid cylinders to fit into
hollow cylinders. Similar ceramics of the mixed
oxides are being tested for radiation damage and
thermal conductivity. 4

 
    

 

Exploratory tests for possible new compounds
of these oxides were made by combining each in
e ST . equimolecular proportions with each of the follow- -
_ Fig. 6.26. Brazed Joint of Kentanium 150A ing: MgO, Ca0O, SrQ, CdO, BaO, A|203, Fe, O,/
Cermet to Nickel. Note sound appearance of joint Si0,, Sn0,, Zr0Q,, HfO,, and ThO,. After a heat
despite severe crack in the cermet. ' treatment at 1500°C for 2 hr, x-ray examination »

   

TABLE 6.14. RESULTS OF PHYSICAL PROPERTY'MEASUREMENTS OF RARE-EARTH OXIDES
' FIRED AT 1300 AND 1500°C IN AIR

 

 

 

Property Measured ‘ . VSm203 ‘ G6203

Sintering ée‘rfiperg}u;e °C) | 1300 1500 1300 1500

Density (bulk) | 6.0 7.4 Cos97 7146

~ Per cent of theoretical* 81 100 93 o2

" Apparent porosity (%) 23 4 0.4 12

Modulus of rupture (psi) 1800 2000 2740 2840

Modulus of elasticity (psi) 4.4 x 108 26.5 x 108 14.5 x 10° 18.0 x 10°
" Coefficient of linear thermal expansion 9.9 x 10~8 105 % 107¢

(in./in./°C) .
Melting temperature (°C) 2350 % 20 23_50 +'20 -

 

_ *Théorefi_'cql densities: Sm203, 7.43; Gd203, 7.407.
**Higher than theoretical density, probably due to 2% terbium present,

146

 
 

)

showed that both rare-earth oxides formed com-
pounds with A1203, Fe203, Si0,, and SrO, The
compounds with Al 0, and Fe,0, have the perov-
skite structure, and their possible use as dielectrics
is being investigated. In each case the Fe O
compound was magnetic, Both Sm,0, and Gd,0,
form solid solutions with CaO, MgO, HfO,, ZrO,,
and ThO.. In each case the CdO volohllzed out of
the mixture at 1500°C.

Several rare earths were available only in very
small quantities, and hence they were studied only
with respect to their x-ray diffraction patterns,
Their structures, as received and after calcination

at 1400°C, are given in Table 6.15.

TABLE 6.15. STRUCTURES OF RARE-EARTH
OXIDES

 

After Calcination
As Received

 

at 1400°C

(:eo2 | fee fce
Pr2U3 U= U
Nd203 U U**
Sm203 U+ c U
Eu203 c U
Gd203 c U
Dy203 c c
H0203 c .
Er202 c c
Tm203 c c
Yb203 c c

 

  

*U = undefermmed c = cublc.
**2 phases presenf, bofh omsoiroplc. '

***2 phases present one cublc, one dnlSOfl‘OplCo

      

The rare-earth ox\lde

5 in. are being

  

Shevlm at the OhIO State Experiment Station. This
cermet consists of 30 wt % rare-earth oxides and

  

- Eu O 4 Nd 03, Pr ‘03, and
s 1‘}ny20 are now on hand in sufflcnent quantities for
o __‘defermlnmg some of their sintered properties and
"3___':whether their reachon products wn‘h A|203 and
N -Fe 03 have valuable ‘properties. Specimens of
" Eu,0, in the form of disks ¥, in. in diameter and
R 003 in. thick an bars 3 X /4 x ¥
"r_',_prepcred o
"-‘"‘radlohon dcmuge studles.
(Sm, Gd) O -Fe cermet was prepared by

property measurements ond

PERIOD ENDING JUNE 10, 1955

70 wt % iron. Powders of these two materials were
milled for 24 hr in l-gt-capacity steel mills with
tungsten carbide slugs and methyl alcohol. The
mill charge was dried and granulated, and the
batch of material was pressed isostatically at
35,000 psi. The blank thus formed was sintered in
hydrogen at 2725°F for 'Il/2 hr. The specimen was
machined to a cylinder about 0.8 in, in diameter
and 3 in. long by using a conventional lathe and
steel tip tools. A porosity of 14% was measured
by determining kerosene absorption, and a porosity
of 25.9% was calculated on the basis of theoretical
density; these values indicate the close pore
volume to be 11,9%,

UO, Particles Coated with Zr0,

In anticipation of a possible need for protecting
particles of UO, from reaction with molten silicon
in the development of an SiC-Si fuel element, a
technique was developed by Shevlin, at the Ohio
State Experiment Station, for coating particles of

0, (20 to 50 mesh) with a thin layer of ZrO,.
This was done by heating the particles to about
500°F in an open-topped container on a hot plate.
While this container was vibrated mechanically to
agitate the particles, they were sprayed with a
solution of ammonium zirconyl carbonate. Under
these experimental conditions, droplets of the
solution were dried, the salt was decomposed, and
the oxide (ZrO,) product was deposited on the
UO, surface to form a continuous film, which was

_dn‘flculf to remove by rubbing., It appears possible

to use this process to form diffusion barriers which

‘would prevent UO,, from reacting with aluminum and

other metals, Attempts will be made to deposit

'"”""“"’A|203 and NiO coatings on UQ, by this process,
“and it is planned to heat-treat ZrO, and Al,Q, to
“explore the possibility of consolidating the coat-
" ings. Heat treatment of NiQO under reducing con-

ditions will also be investigated to determine the
practicability of producing nickel coatings in this

~manner.,

Graphite-Hydrogen Reactions and
Erosion [nvestigation

For some ANP rocket concepts it is proposed to

“Use hydrogen as a propellant that would pass
- through a graphite-UO, core at high velocity.

Because it was reported that hydrogen reacted
seriously with graphite, attempts were made to

147

 
 

 

 

' ANP PROJECT PROGRESS REPORT

coat graphite with ZrC. All methods so far tried

~ have resulted in unsatisfactory coatings. An

investigation is under way to determine the re-
action of graphite with hydrogen at 2300°C,

« . A static gas test in which a plate of graphite is
- heated in hydrogen to 2300°C has been developed.
= thtie, it any, reaction was noted when the graphite
~was exposed at 1800°C for 2 min and 2300°C for
30 sec. Gas analyses are being made of the
hydrogen which surrounded the specimen.

~ A dynamic test has been set up at Graphite

 

Specialties Corporation, Niagara Falls.

An elec-
trically heated high-density graphite tube surrounded
by carbon insulation is the heat source and speci-
men. Hydrogen that had been processed to remove
most of the water impurities was blown through the
tube at several hundred feet per second while the
tube was heated to 2300°C. There was no con-
clusive evidence that any significant reaction took
place between hydrogen and graphite in this test;
however, there was evidence of reaction between
water impurity in the gas and graphitg.

 

 

 

 

 

 
 

"

¥

-_f.?empercfure errors.” _
of the mlxmgf chcmbers is ‘now in progréss t6 be
sure ’rhcn‘ good mixed mean fluid temperofures;
" were ob'rcmed A visual exammahon of fhe fube'
'showed no surfaf‘e c{epos:ts for fhls system ‘

PERIOD ENDING JUNE 10, 1955

7. HEAT TRANSFER AND PHYSICAL PROPERTIES
H. F. Poppendiek

Reactor Experimental Engineering Division

Fused salt heat transfer studies have included
an experimental determination of the friction factor
as a function of Reynolds modulus for the case of
turbulently  flowing Nc:F'-'ZrFA‘-UF4 in Inconel
tubes. A full-scale ART fuel-to-NaK heat ex-
changer was studied as a water-to-water heat
exchanger. Some velocity profile data were ob-
tained for the 18-in. ART core, and a study was
made of the temperatures and corresponding
tensile stress fluctuations in the Inconel walls of
the ART core for momentary periods of flow
stagnation near the wall. The results of a theo-
retical and experimental study of a free-convection
system containing a volume heat source are
presented

The enthalpies, heat capacities, and viscosities
of several fluoride mixtures were obtained. Also,
some preliminary thermal conductivity data were
obtained with a new conductivity cell.

FUSED SALT HEAT TRANSFER

H. W. Hoffman
Reactor Experimental Engineering Division
S, J. Pias

Pratt & Whitney Aircraft

Previously reported heat transfer experiments
with the fuel mixture NaF-ZrF,-UF, (53.5-40-
6.5 mole %) flowing in an Inconel tube gave resulis
which were 24% below the general heat transfer
corre[q’hon." Two possable explanations for this

discrepancy are that there may have been errors
in fempercture mecsurement or that an cddmonal'
.therma! re5|s'rance may "have been present at the
fluid- meml mterfcce. Calibration of the mlxmg-' '
chamber nd fube wa|| fhermocouples revealed no
| A sfudy of the effectiveness

  

tube surface suggests a nonwet’rmg condmon "
To check on the’ genercl possub||n‘y of some 1ype
of two-phase flow, a study of the pressure-drop

characteristics of NaF-ZrF,-UF, in forced con-
vection through nearly isothermal circular tubes

was undertaken. The preliminary date cre shown
in Fig. 7.1. The results suggest that two-phase
flow does not occur in this system, because the
friction factor measurements fall near the general
single-phase curve. The influence of a nonwetting
fluoride on forced-convection heat transfer is not
yet known.

The heat transfer characteristics of the mixture
I\'ch-KF-LiI"'-UF4 (11.2-41-45.3-2.5 mole %) are
currently being investigated. Preliminary results
for an Inconel system, shown in Fig. 7.2, lie 40%
below the general heat transfer correlation. A
fragile green film was found on the inside surface
of the tube. X-ray and petrographic examination
revealed only the constituents of the salt. Further
heat transfer studies are planned, for which «
stainless steel test section will be used. Pressure:
drop measurements are also to be made.

A small-scale loop that is to include a cen-
trifugal sump pump has been designed and is being
fabricated. The proposed system is shown sche-
matically in Fig. 7.3. The fluid flow rate will be
obtained by weighing the salt as it empties into
a weigh tank suspended within a weigh chamber.
The weigh tank will be arranged so that it can be
rapidly emptied. The sump well will be used to
maintain the liquid level in the pump during the
weighing operation.

ART FUEL-TO-NaoK HEAT EXCHANGER
J. L. Wantland

Reactor Experimental Engineering Division

A full-scale ART fuel-to-NaK heat exchanger tube

" hundle (100 Inconel tubes, 6 ft long, %, in. OD,

'0 017-in. wollr fhickness) was assembled instru-

.'chonge system (Fig. 7.4). Cold water was circu-

lated through the tubes (NaK side) at Reynolds

"'_'fnumbers of sufficient ‘magnitude to produce low and

caleulable  thermal resnsfcnces hot water was

circulated around the tube bundle in a parallel
“direction (fuel side).” The inlet and outlet tempera-
tures  of both streams were measured by two
thermocouples located in each of the four mixing

chambers. The ‘‘fuel’’ side was instrumented with
static pressure taps. Both flow rates were measured

149

 

 

 
 

 

ANP PROJECT PROGRESS REPORT

Fig. 7.1. Friction Factor for lsothermal Flow of INluF-ZrF4-UF4 (53.5-40-6.5 mole %) in a Smooth

FRICTION FACTOR, f/2 - -

0.01C

£/2 = 8/No,

£/2 =0.023 NRe_O'z
0.005
0.002
0.001
1000 2000 5000 10,000

REYNOLDS MODULUS, Mie

Circular Tube.

Fi.g.'7.2-. Comparison of Heat Transfer Measurements on NaF-KF-LiF-UF, J(11.2=41-45.3-2.5 mole %)

/3

o
-
a_

NSt'N

COLBURN FUNCTION, /

0.040

0.005

0.002

0.001 _
1000 ' 2000 ' 5000 10,000
' REYNOLDS MODULUS, ¥

in Inconel with the General Correlation for Ordinary Fluids.

150

 

 

SETRTY
ORNL-LR-DWG 7319

20,000 i 50,000

SEORET
ORNL-LR-DWG 7320

GENERAL GORRELATION,
j=0.023 Re 92

20,000 ' 50,000

 

a#
e+ i, i

 

PERIOD ENDING JUNE 10, 1955

UNCLASSIFIED
ORNL-LR-DWG 7321

 

[ 7 h

 

    
 
   
    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

¥
' WEIGH TANK
WEIGH CHAMBER
CENTRIFUGAL SUMP PUMP
SCALE "
AIR COOLING com\>
ELECTRODES .
MIXING CHAMBER | |
i \ I I
* TEST SECTION
FREEZE VALVE

 

 
 
  

MIXING CHAMBER

SUMP FILL TANK

 

 

Fig. 7.3; ‘{.S.c'l;é'r.na-fii:'.'Iji'agrom of Small-Scale Pumpfrigl Sysfemfot Heat Transfer Studies with Fused
Sst. e o ' o s

   

E HYDR'O'DYNAMICS

0%

   
  
 

   
  
 
   
   
 
   

J,Q Brgdfufe"”f“‘ " F. E.Lynch
s rer curves 1c o L.D. Palmer o
r-;-f_r_:exchcnger are shown -il;.l '."fFlgs 7-5 "dhd 7.6-_,_‘_,;_ N Reoctor Expenmem‘ol Englneerlng D|v15|on
_/Fvlgure 75 is based on 18 data points that fall - G. L. Muller

o

Prafl & Whn‘ney AII'CI'G‘H'

7 Furfher flow studles ‘_were conducted w_n‘h ’rhe
quar’rer scole model_;'of thé'_ ]8~|n. ART core

 

::rfhrough fl'me core w:fhouf ro’rqtlon have been tdken,
“and " quantitafive veloc:ty “profiles have been

»«‘i-’ \’,’\4-’ L

“mined by u’rlhzmg the tube bundle as an electrical-
resistance heater. Results of the sfudy should  obtained.
further substantiate the curve shown in Fig. 7.5. A vaned section was added at the inlet to the

151

 

 

TR T

I I N R T T

 

R OEST mEe

 
 

 

 

Ro pot ALTE

ANP PROJECT PROGRESS REPORT

TS
wwn»‘unnw
e,

=
z
S
STorisiinal

d‘tlo.fctcoooo;,:ccfsv.'fl
e A iorrereYse
TSt

T Tas ol T et

TR
i ST T ot 3 : <3

ey : % b -
T ) o Tae 2 % D
B st 2 ) & eIy amIstTaes
e tOowooHNN“oc a3 s e e T e e

5
> v e red.

oz
o e
2335

SITeTAEL. N
sl

AveELriT

 

 

152

iment.

-to-NaK Heat Exchanger Exper

ART Fuel

4

7

Fig.

 

 
e M w i ey

e R S

 

' | PERIOD ENDING JUNE 10, 1955

core model to give the fluid a rotational component
(approximately o 45-deg helical angle ot the

o ’ ORNL~LR-DWG 7322
7 40

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' o b entrance). The velocity profiles were then ob-
2 A served qualitatively with the flow-visualization
" : Nu/PrO9 0 025 Reo,;_w/ Technique. Separation was observed next. to the
' L7 island wall rather than the shell wall, which was
: 20 /,/ W% the case with sfrdight-throug!h flow. The axial
7 / component profiles are shown in Fig. 7.7.
3. //' In addition, a series of 16-mesh screens was
; //’ ' / added at the inlet, both with and without the
10 / ' rotation-producing vanes. Little effect attributable
// to the screens was noted in the straight-through
P case, except that the separation region appeared
/’ to be much less turbulent; indeed, the negative
/_ - profiles were easy to see. They appeared to be
5 / approximately parabolic, and thus they suggested
. _
000 2000 5000 10,000

LEUREY
REYNOLDS MODULUS, A, ORNL-LR-DWG 7324

 

 

 

an. 7.5. Prehmmary Heat Transfer Churacier-
istics of the Fuel Side of the ART Fuel-to-NaK : m\U]
T Heat Exchanger as Determined by Water-to-Water '
Tests.
| Rl

 

. . : = ORNL—LR=DWG 7323 .
02 -

>~ / / W

 

 

 

 

 

 

 

 

 

».

 

 

Qi S /—/ -
“ . REVERSE FLOW
o N | W ,
& - .
2 N
2 ‘ _
g > g
E - S~
2 005 — -~
o ) : .
SMOOTH PIPEL_| :

 

U

 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    
 

     

  
 

! " Fig. 7.6.
of the Fuel ‘Slde of the ART Fuel-fo-NaK Heat

Exchanger.

Flow “Through o Model of the ART Core af a
Reynolds Number of 3000 with a Rotational Ve-
locity Component at the Inlet.

153

 

Fig.'7/7.‘ Qualltaflve Axial Yelocity Profiles of’ -

 

 
sl

 

 

Qo na? AL

ANP PROJECT PROGRESS REPORT

a laminar type of flow, No effect of the screens
could be observed with rotational flow.

A two-dimensional diffuser has been designed
which permits a variation of diffuser geometries
and cross-sectional area ratios. The entire diffuser
section is fo be made of Plexiglas so that the flow
can be studied by the flow-visualization system.
The assembly is now being fabricated.

A preliminary feasibility study is at present
being made on the control of the boundary layer
in the core for the purpose of preventing separation
of the boundary layer. A boundary layer suction

“technique is being considered.

A plastic 10/44-scale model of the 21-in. core
has been designed and is now being fabricated.
The odd scale was chosen to make the model fit
the available testing facility.

REACTOR CORE HEAT TRANSFER
H. F. Poppendiek N. D. Greene
L. D. Palmer

Reactor Experimental Engineering Division

The hydrodynamic studies being made for the
ART core indicate that when separation regions
exist they are often characterized by repetitive,
short periods of flow stagnation. A study was
made of the temperature and tensile stress fluctu-
ations in the Inconel core wall that result from
these periods of flow stagnation. The following
idealized system depicting the ART core wall
during a short period of flow stagnation was con-
sidered. It was postulated that a layer of fuel
contiguous to the Inconel wall suddenly stagnated
for a period of 0.1 sec under the volume heat flux
condition of 5 kw_/cms. This flux is representative
of the conditions in the ART because of the high
flux peak at the wall. After a 0.1l-sec time interval
the fuel and wall temperature would rise about
260°F if no heat transfer were present; transient
heat conduction into the Inconel wall, however,
would reduce this temperature rise. A numerical
composite-slab heat transfer analysis was made
for this problem, and the results indicate that the

" Inconel-fuel interface temperature would rise about

60°F in 0.1 sec. This temperature rise would, of
course, be much smaller if the thermal conductivity
of Inconel were not so poor. The corresponding
eicshc thermal stress for this temperature fluctu-

-'? ation was found to be about —15,000 psi. These
results suggest that the fluid flow in the core

should not be allowed to fluctuate,

154

All mechanical components, including the power
supply, of the volume heat source experiment for
reactor cores with nonuniform flow cross section
have been constructed and installed. System leak
testing and flow calibration are now being carried
out. An “infinitely’’ adjustable power input (that
is, 15 to 100 kw) to the test section will be pos-
sible by means of several saturable reactors that
have been shown to possess excellent voltage

regulation characteristics under the desired load.

The electrical and power instrumentation is nearing
completion. The temperature structure within an
insulated, divergent test channel will soon be
studied for low power densities.

Several heat transfer analyses that are useful in
predicting the thermal structure within a circulating-
fuel reactor core have been completed. A report
is being prepared which tabulates some of the
detailed temperature profiles derived for the forced-
convection volume-heat-source systems described
previously.l'g These temperature data will make
it possible to determine rapidly the complicated,
radial, fuel-temperature profiles in circulating-fuel
reactors whose pipe or channel ducts are being
cooled at the walls. Also, o transient temperature
solution was derived for the case in which fuel
stagnates momentarily next to the Inconel core
shell.  Another analysis was concerned with a
boundary layer temperature solution where a
volumetric heat source exists in the fluid; thermal
and hydrodynamic boundary layers are not presumed
to be equal. '

FREE CONVECTION IN FLUIDS HAVING
A VOLUME HEAT SOURCE

D. C. Hamilton F. E. Lynch

Reactor Experimental Engineering Division

The objectives of the free-convection reseorch
and the progress were previously reported Both
the theoretical and experimental analyses of the
three-parallel-plates system have been completed,
and a report is to be issued.? The three-parallel-

 

]H. F. Poppendiek and L., D. Palmer, Forced Con-
vection Heat Transfer in Pipes with Volume Heat Sources
Wztbm the Fluids, ORNL-1395 (Nov. 5, 1953).

4. E. Poppendiek and L. D, Palmer, Forced Con-
vection Heat Transfer Between Parallel Plates and in
Annuli with Volume Heat Sources Within the Fluids,
ORNL. 1701 (May 11, 1954),

3D. C. Hamilton and E. E. Lynch, ANP Quar. Prog.
Rep. Dec. 10, 1954, ORNL-1816, p 113,

4D, C. Hamilton and F. E, Lynch, Free Convection
Theory and Experiment in Fluids Having a Volume Heat
Source, ORNL-1888 (to be issued).

sl

 
 

  
 

Do st AL

plates system is shown schematically in the upper
inset of Fig. 7.8; it consists of three parallel and
equally spaced vertical copper plates, 8 in. wide
and 36in. high, with a channel width, x4, 0f 0.5 in.
Plastic bottoms and ends make leakproof containers
of the two free-convection channels, which are
filled with a dilute electrolyte (HC! in HZO)'
With the two outer plates grounded and the center
plate maintained at a given a-c voltage, heat is
generated uniformly within the electrolyte. The
heat thus generated is transferred by free con-
vection to the outer walls and then through the
outer walls fo the coolant channels in which
cooling water flows vertically upward. Thermo-
couple probe wells were provided so that the
temperature within each wall could be measured
at various vertical levels in the system., The
variables measured were coolant flow rate and
temperature rise, power input to the apparatus,
and the three wall temperatures at various levels.

A theoretical analysis for laminar flow was made
that was based on the postulates that, in the region
awdy from the top or bottom, the velocity profile
becomes fully established and unidimensional and
that the temperature becomes linear with the
vertical coordinate, z, and has the same vertical
gradient, A, as the coolant mixed mean tempera-
ture.  Both velocity and temperature solutions
were obtained. The maximum temperature at a

given level occurs at the center wall. The

equation for the dimensionless center wall tempera-

ture, ®(0), follows, and (D‘(O), Ny, and N, are

defined on Fig. 7.8:

sin?2 A + sin A sinh A ~ A sin A cosh A — A cos A sinh A

PERIOD ENDING JUNE 10, 1955

Semiquantitative visual observations of the paths
of suspended droplets indicated that the free-
convection circuit existed as one long cell and
that the velocity profile became established in the
middle region and was similar to that predicted by
the theoretical analysis. The onset of turbulence
occurred at a Grashof modulus of 5 x 107, the
same as that for ordinary free convection.

A curve is given in the lower inset of Fig. 7.8
that shows the typical vertical variation of the
experimental ®(0). |n comparing the laminar regime
minimum ®(0) data with the theory, it is seen that,
for N, less than 1300, the data are about 20%
lower than the theory; this is good agreement. For
these data, 4 was uniform and approximately equal
to A , the coolant mixed mean temperature gradient.
For the data (not shown here) for which Ny was
greater than 1300, A was definitely not uniform;
it was as much as ten times greater than A and
the ®(0) data were 50% lower than the theory.

From the turbulent regime data, it is seen that
free convection can reduce the maximum tempera-
ture difference in such systems to at least one
tenth that which would be present if the heat were
transferred by conduction alone.

Variables not defined on Fig. 7.8 are: %, thermal
conductivity; g, acceleration due to gravity; 3,
temperature - coefficient of expansion; a, thermal

_ diffusivity; v, kinematic viscosity; and §, volume

heat source term.

 

o ®(0) = )\fi’ [sin A éi_nh A - c-z_(_]!-—_ cos A cosh)\) + c(cos)\ si_nh_}t ~ sin A cosh M,

 

sinh2 A — sin? A

 

s (cosh)\ — cos A) (smh X+ sin A) — 2X sin Asinh

    
 

 

c = -
sinh2 A — sin? A

 

155

 

 

b
!

 

 
ANP PROJECT PROGRESS REPORT

" UNCLASSIFIED
. ORNL—LR-DWG 7325

T — T T T ]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LAMINAR REGIME e TURBULENT REGIME CENTER WALL
N =
~— T LEFT OUTER WALL
Ot I — T~
4 | s Sl AT D
[ ¢ .ZS.""--..__ T T
0.5 — —
EQUATION 1+ o K s LEFT COOLANT PASSAGE
S et
~ 1, y 7
d ""--.\‘. A — -
~= T~
~ -~
w5 oz . 07§ J—‘\—-.. N ~
Jol = O MINIMUM $(0) EXPERIMENTAL DATA o U‘-"“&L 5 \\"13
i ® DARKENED POINTS GO WITH THEORY T of
—— T—
S oy A MAXIMUM $(0) EXPERIMENTAL DATA e
& —— AVERAGE LINE DRAWN THROUGH DATA
—— MINIMUM ®(0) FROM LAMINAR REGIME
THEORETICAL ANALYSIS (EQUATION 4) 1.0
0.05 T J ‘ I i { H l - | < MAXIMUM ${0)
| MINIMUM &(0
T T I $(0) @(/7
| ——— 7, CENTER WALL TEMPERATURE 1653 Bg 4
? #, OUTER WALL TEMPERATURE ' av 0
| : TOP BOTTOM
0.02 —— Sx2 HEIGHT
' §.(0)= —= 810 TYPICAL VERTICAL VARIATION
: 24 M= 5354 OF EXPERIMENTAL &{O0)
" a’c
oo L L | L L ||
10° 2 5 10 2 5 10° 2 5 10° 2
W
I
Fig. 7.8. Dimensionless Temperature Function for N\, < 1300.
HEAT CAPACITY Liquid (546 to 899°C)
W. D. Powers Ho —Hyso o =~9.8 + 0.3508T - (5.39 x 10~°)72
Reactor Experimental Engineering Division c =0.3508 — (10.79 x 10~5)T
The enthalpies and heat capacities of four P o
AH = 63 at 510°C

fluoride compositions and of lithium hydride have fusion

been determined by using the copper-block calori- NaF-ZrF ,-UF , (56-39-5 mole %)

meter. The results are listed below: Solid (]37 to 503°C)
| H ~ Hygo =—4.3 +0.1596T + (5.15 x 107 °)12
NaF-ZrF ,-UF , (50-46-4 mole %) | c, = 0.1596 + (10.29 x 107°)T
Liquid (540 to 894°C) ieuid (567 10 892°C)
—ye? iqui to
Hp ~Hygoo ==3.3+0.3178T - (4.28 x 107°)T Hy ~ Hygo o = 0.6+ 0.3033T — (3.24 x 10-%)1?
=0.3178 — (8.56 x 10~5)T -5
“p c, = 0.3033 - (6.47 x 10~>)T
AH, . =57 at 530°C
NaF-ZrF, (50-50 mole %) | NaF-LiF-ZeF -UF , (20-55:21-4 mole %)
Solid (54 o 488°C) ~ Liquid (582 to 900°C)
Hp - Hygo o = ~44+ 0.1798T + (2.69 x 10512 Hop = Hygo o ==25.9 + 0.43WT ~ (7.42 x 10-5)7?2
c,= 0.1798 + (5.38 x 10=3)T c,= 0.4314 — (14.85 x 10~3)T

156

 
 

"f‘"\/f‘scosity measurer '
fluoride mixtures, and the results are tabulated’

/QAMM

Lithium Hydride
Solid (100 to 490°C)

Hyp ~ Hypo =~38.5+ 0.939T + (6.6 x 1074)72

c, =0.939 + (0.132 x 10T

In the above expressions,

H; = enthalpy in cal/g,
c, = heat capacity in cal/g-°C,

p
T = temperature in °C,
AH, . = heat of fusion in cal/g.

The enthalpies and heat capacities of the first
two mixtures had been previously determined by
using the Bunsen ice calorimeters. In view of the
present importance of the zirconium fluoride-base
fuels, however, it was felt desirable to determine
these properties with the more precise copper-
block calorimeters.

The heat capacities of liquid fluoride mixtures
may be predicted within about 15% on the basis of
their chemical composition. It has been found that
the product of the heat capacity (cal/g:°C) and
a function M/N is remarkably constant. This
function is defined as:

M =S M,
?
N =2xN;,
—— z
M = average molecular weight,

. = molecular weight of component,
. = mole fraction of component,
= average number of ions,

= number of ions in componenf

funchon M/N For the ]7 fluorlde mixtures being

*VLS*Ud'ed at bfesent ‘the “average of the préduct of:"'
“and M/N is 9.0. The corresponding average

producf for the six fluoride mixtures containing

'approxlmctely “equal ‘molar amounts of NaF and
o IRy and from O to 7 mole % UF was found to
S _'be 8.1 '

 

  
 
 

 

in Table 7.1. The data are expressed in the form

nts were made on seven

PERIOD ENDING JUNE 10, 1955

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ST
ORNL-LR-DWG 7326
1.0
° NQF-ZrF4 MIXTURES WITH O TO 7 mole % Ul"'4
® ALL OTHER FLUORIDE MIXTURES
. ®

05 N\
p N 7
2 v::;,= 8.0 Hfi_u
©
A 9
>—
k=
NN
g et TN .
- M %
o N
T N\ *
u& ‘ \

0.2 e\

o.1

10 20 30 40 50 60
MIN

Fig. 7.9. c, Vs M/N for Molten Fluoride Mixtures.

where T is in °K. These viscosity data have been
plotted in Fig. 7.10. No equation is listed in
Table 7.1 for salt a because of the slight curvature
in the data, which may be noted in the figure.
Salts b, ¢, and 4 are mixtures which had been
studied prior to the recently completed viscometry
refinement program. Salts a, e, f, and g are

Flé;Jre 79 shows s p[of of heaf cupacny T theff*-"mi)émres ‘which have been formulated recently.

Measurements were made on all the mixtures

- (except salt g) by using both the Brookfield and

capillary viscometers; the results obtained by
these two completely different instruments were
in satisfactory agreement (deviations from the
average line through the data were within 112%),
Measurements on salt g, which contained BeF_,
e made in a separate beryllium facility. Data
5 taken wnh 'rwo cuplllary wscometers to

“furnish a check.

=:Figure 7.11 presents a plot of the viscosities of
seven mixtures containing BeF2 Mixtures g, i,
and & were studied at ORNL. Mound Laboratory
investigated mixtures » and j, and mixtures  and
m were studied at KAPL. The formulas of these
mixtures in mole percentages, as well as the

157

 

 
 

 

D

sl AT

_ANP PROJECT PROGRESS REPORT

TABLE 7.1. SUMMARY OF CURRENT VISCOSITY MEASUREMENTS

 

 

a Vis&osity
Miqufe Coniposition (cp) B Reference

a - NaF-LiF-ZeF -UF .UF, At 550°C, 13.5 (b)

o (20.9-38.4-35.7-4-1 mole %) At 800°C, 4.3 | _

b NaF-ZF 2 UF, At 570°C, 11 0.1307 3730/T ()
. (50-46-4 mole %) At 870°C, 3.4

¢ NaF-UF, At 700°C, 10.25 0.1715 &>784/T (d)
(66.7-33.3 mole %) At 900°C, 5.1

d NaF-KF-UF, At 600°C, 17 0.0866 ¢4611/T (e)
(46.5-26.0-27.5 mole %) At 900°C, 4.4

e NaF-LiF-ZrF ;-UF ,.UF, At 600°C, 12 0.061 481V/T (N
(20-55-21-3.6-0.4 mole %) At 850°C, 3.7

/ NaF.LiF-ZrF, At 600°C, 12 0.061 4617 ()
(22-55-23 mole %) At 900°C, 3.1

g NaF-LiF-BeF, A+575°C, 7 0.105 3560/T (g)
(56-16-28 mole %) At 800°C, 2.9

 

VISCOSITY (cp)

4See Fig. 7.10.

bS. I. Cohen and T. N. Jones, Measurement of the Viscosity of Composition 72, ORNL CF-.55-3-61 (Mar. 8, 1955).
€5, I. Cohen and T. N, Jones, Measurement of the Viscosity of Composition 30, ORNL CF.55-3-62 (Mar. 9,,1955),
dS. f. Cohen and T. N. Jones, Measurement of the Viscosity of Composition 43, ORNL CF.55-3-137 (Mar. 16, 1955).
95, 1. Cohen and T. N, Jones, Measurement of the Viscosity of Composition 2, ORNL CF-55.4-32 (Apr. 1, 1955).
/s.1. Cohen and T. N, Jones, Measurement of the Viscosities of Composition 81 and Composition 82, ORNL CF.55-5.58
(May 16, 1955).
83, I. Cohen and T. N. Jones, Measurement of the Viscosity of Composition 78, ORNL. CF-55-5-59 (May 16, 1955).

UNCLASSIFIED
ORNL-LR-DWG 7327

' TEMPERATURE (°K)

600 700 800 900

1000

1200

50

20

 

{

300 400 500 600

'TEMPERATURE {(°C}

700 800 {000

'_ Fig. 710 .V‘is'c:osi;ies of Fluoride Mixtures
Currently Being Studied. '

158

VISCOSITY {cp)

600

300

Fig. 7.11. Viscosities of Some Fi.uc.ariae_Mik)-&br.‘esi ;

UNCLASSIFIED

ORNL-L.R-DWG 7328

TEMPERATURE {°K)
700 800 900

400 500

Containing BeF,.

1000

m

 

600 700 800
TEMPERATURE (°C) o

1200

1000
. G oieode o i

g A ok e

 

Qo pod dlts

PERIOD ENDING JUNE 10, 1955

TABLE 7.2, THE VISCOSITIES AT 700°C OF SEVERAL MIXTURES CONTAINING Be'F:Z

 

 

- a' . o Viscosity BeF, Content
\ Mixture Composition (cp) (wt %) Reference
~ g Ber-LiF-NaF 4,1 32.2 (&)
1 . ' (28-16-56 mole %) : '
¥ b ' BeF NaF ' ‘ 4.25 40.0 (¢)
o (38361 7 mole %) ’ '
; . BeFyLiF 5.0 44.9 (@)
 (31-69 mole %) |
i BeF2-NqF 6.25 47.0 {c)
(44.4-55.6 mole %)
k ' BeF -NaF | 6.75 45.8 ()
o (43-57 mole %) '
1 BeF.z—NuF-U F4 ' 9.6 53.9 {e)
(51.9-47.9-0.25 mole %)
m Ber—LiF-ThFd-UF4 17 58.6 ()

(49.43-49.43-1.03-0. 12 mole %)

 

%See Figs. 7.11 and 7.12.

S, 1. Cohen and T. N. Jones, Measurement of the Viscosity of Composztzon 78, ORNL CF.55-5-59 (May 16, 1955).
“Communication from J. F. Eichelberger, Mound Laboratory

45, 1. Cohen, ANP Quar. Prog. Rep. Mar, 10, 1955, ORNL-1864, p 143.

. ?J. K. Davidsen, R. J. Herbert, and B. T. Morecroft, Fused Fluoride Homogeneous Reactor System for Submarine
“ . Propulsion, (SAR Phase Il Study), KAPL-992. '

Personal communication from J. K. Davidson, Knolls Atomic Power Laboratory.

: UNCLASSIFIED 1 T
oriCLASSIFIED BeF, contents in weight percentages and the

20 e viscosities at 700°C, are listed in Table 7.2.

Figure 7.12 shows a plot of the viscosity of
each of these mixtures at 700°C vs the weight
percentage of BeF, in each of the mixtures; note
that the viscosity decreases as the BeF2 content
decreases. Mixture g was formulated because it
appeared from this tfrend that the kinematic
viscosity might compete with the viscosities of
salts in the NaF-ZrF ,- -UF, system. The kinematic
viscosity of mixture g, BeF -LiF-NaF (28-16-56
mole %), was found to be ubout 15% higher than
that of NaF-ZrF, UF (50-46-4 mole %), the salt
presently being consudered as 'rhe fuel for the ART.

"VISCOSITY (cp)

+

THERMAL CONDUCTIV!TY

 

 

    

3 - 0 0 200 30 40 50  e0 70 o ) W. D. Powers
- - U Ber, CONTENT OF MIXTURE (wi %)~ Reactor Experlmentcl Englneermg Dw:snon
) T ?:Two methods for ‘the defermmdtton of fhermal
’ -~ Fig. 7.12. Viscosities at 700°C vs B‘éF2 Content conductivities of liquid fluoride mixtures are being
i of Seven Fluoride Mixtures. Compositions of the investigated. A radial thermal conductivity cell
mixtures are given in Table 7.2, has been redesigned and is being fabricated. The

159

 

T T TErT—

 

 
 

 

i i, i s

 

QDo it Al

ANP PROJECT PROGRESS REPORT

major difficulty in using the apparatus as originaily
designed was the lack of sufficient guard heating
to assure pure radial heat flow,

A flat-plate system is now being investigated.
The liquid is contained in cells consisting of two
parallel flat plates of metal joined at the edges
with 10-mil-thick sheet metal. The conductivity
of the liquid in the cell can be calculated by
knowing the area of the plates and the distance
between them and by measuring the amount of heat
passing from one plate to the other and the tempera-
ture differences between them. Some preliminary
data on a ZrF ;-base fluoride mixture are currently
being obtained that are in agreement with previous
conductivity data obtained with the variable-gap
device.

160

ELECTRICAL CONDUCTIVITY
N. D. Greene

Reactor Experimental Engineering Pivision

Conductivity studies of several aqueous electro-
lytes have indicated the presence of a considerable
amount of polarization within the experimental,
platinum conductivity cell. Accordingly, a second
potential electrode will be installed, which, when
used in conjunction with an electrometer having a
10-megohm input, should reduce the effects of
polarization. Similar conductivity measurements of
some nonaqueous electrolytes (molten salts) will
be made to compare the differences in the degree
of polarization between the two types of electro-
lytes.
2

 

 

s

had been irradi

  

 

. :;L',‘."_-};‘ANP Quar Prog. Rep Dec 10, 1954 ORNL 1816, p 120
2Chemical ‘analyses by the General Analysis Labo-

 

PERIOD ENDING JUNE 10, 1955

8. RADIATION DAMAGE

D. S. Billington

J. B. Trice

Solid State Division

Preliminary results from a series of MTR irradi-
ations of Inconel capsules designed to compare
UF;- and UF ,-base fuels show no corrosion, no
significant uranium segregation, and no changes
in concentrations of impurities in the fluoride
mixture. The loop in which a fuel mixture was

. circulated in the LITR has been sectioned, and

preliminary examinations have been made. The
miniature loop designed for operation in a vertical
hole .in the LITRis in the final stages of assembly,
and charcoal traps to delay fission gases in the
event of a rupture have been tested and incorporated
in the cooling-air off-gas line.

A tube-burst stress-corrosion apparatus has been
assembled for insertion in the LITR, and an LITR-

irradiated stress-corrosion rig is ready for exami-

nation. The creep apparatus inserted in the MTR
has completed six weeks of irradiation and is to
be returned to ORNL for measurements. A theo-
retical study has been completed which indicates
that xenon will be sparged from the ART, as
presently designed, in sufficient quantity to
alleviate the control problem,

MTR STATIC CORROSION TESTS

W. E. Browning G. W. Keilholtz
Solid State Division

H. L. Hemph.iil
Analytical Chemistry Division

Two more of a series of paired 'c'apsules are

contained 2 mole % UF3, and in the other capsule
the fluoride mixture contained 2 mole % UF,. A
control capsule for each mixture was also opened
and analyzed. No corrosion could be found in any
of these five capsules. One irradiated capsule
containing UF that was examined earlier had
shown slight corrosion. The next series of capsules
that will become available for analysis consists
of four capsules that will have been irradiated
three times as long (six weeks) as those analyzed
thus far. The capsules in this series will be more
likely to reveal any difference that might exist
between the two fuel materials.

The starting fuel mixtures were analyzed? for
trivalent uranium, total uranium, iron, chromium,
nickel, sodium, and zirconium. The UFB-becring
batch had more than 96% of its vranium in the
trivalent condition, whereas no trivalent uranium
was found in the UF ,-bearing batch when cnalyzed
by a method which is sensitive to 0.005% ust,
The results of the chemical analyses for uranium,
iron, chromium, and nickel in the starting mixture,
the control mixtures, and the irradiated mixtures
were averaged, and ratios were computed to deter-
mine whether there were changes in distribution
or composition. The results of these statistical
analyses, which are presented in Table 8.1,
indicate that there was no significant uranium
segregc’non and that there were no chonges in
concentration of urqmum, iron, chromium, and
nickel.

being ‘irradiated for two weeks each in the MTR o

These Inconel capsules are bemg wradmted

- order to compore the effects of UF3 and UF,

- NaF- ZrF -base fuel mixtures, as described prevu-
e ously Three addltlonal Inconel capsules that
ha ted for two weeks were openedw
‘and anaiyzed chemlcafiliy and metqllogrcphlcaily
if,}.;_ln two of fhese cop5u|es the fluorlde mlxture”

  

 
 
  
 
 

 

1W E Browmng, G. W. Kelrlhohz, und H. L. Hemphili,

ratory, Analytical Chemistry Division,

3Metc|lographic analyses by the Remote Metallography
Group, Solid State Division.

   

Chemical analyses were also obtained of four
additional irradiated inconel capsules in which the
fluoride fuel mixture initially contained 4 mole %

. UF,. _The uranium analyses of inner cores and

outer cores of the samples agreed to within 6%.
The iron, chromium, and nickel concentrations

.. were between 0.01 cmd 0.1%. Additional capsules

f this type are belng prepared for :rradmhon in

.,,.__:th‘et MTR at 1500°F for control runs, and for
thermal cyclmg tests. '

 

4Anczlyses performed by the Y-12 Analytical Labo-
ratory.

161

 

b

L

ke

F
'

 

 
 

oL

 

TABLE 8.1, SUMMARY OF ANALYTICAL RESULTS FOR SERIES OF CAPSULLES TESTED TO COMPARE EFFECTS OF

UI"'3 AND UF, IN N.uF-ZrF4-BASE FUEL MIXTURES

 

-

 

 

LH0dIY SSIYO0Yd LDIF0dd INV

 

Number of Uraniom Distribution {a) Ratio of Final Ratio of Final to Initial
Capsule Category Capsules U (av)/U, (av) to lnitial Impurity Concentrufion(c)
a b Uranium Content( ) Fe Cr Ni
All capsules in series 7 1.003 *+ 0.028 1.021 % 0.022 1.47 +>1.47 1.08 +>1.08 (d)
All irradiated capsules 5 1.021 t 0.033 1.038 £ 0.024 1.45 £ >1.45 0.48 + 1.5 1.48 + 0.56
All control capsules 2 0.96 = 0.0240 0.989 * 0.02(e) 1.53 +>1.53 3.0 + 1.2 (d)
Irradiated capsules containing 2 1.026 £ 0.044 1.031 = 0.04(e) .67 + 1.47 0.0 + 1(2) 3.9 + 1.1
2 mole % UF4
Control capsule containing 1 0.947 + 0.028 0.975 + 0.024 (D {d) (d)
2 mole % UF'4
lrradiated capsules containing 3 1.001 £ 0.048 L051 £ 0..05(6) 1.31 +>1.31 0.54 + 0.6 0.95 £ 0.05
2 mole % UF3
Control capsule containing | 0.972 1 0.020 0.994 + 0.02(e) 1.53 £>1.53 3.0 + L2 {(d)
2 mole % UF3 '
(a)Ua is the uranium content of the inner core of a sample, and Ub is the uranium content of the outer core of a sample.

(b)Correcfed for burnup.

(C)Confidence limits are ot 95% level.

(d)Vcflues obtained were influenced by a sample which was probably contaminated with Inconel and therefore are not significant.

(e)Esfimufed.
 

.y

-

 

Lo sl

LITR HORIZONTAL-BEAM-HOLE
FLUORIDE-FUEL LOOP

0. Sisman J. G. Morgcn
C. D. Baumann M. T. Morgan
W. E. Brundage A. S. Olson

R. M. Carroll W. W. Parkinson

Solid State Division

The operation of the loop for circulating fluoride-
base fuel in the LITR horizontal beam hole to
study the effect of reactor irradiation on corrosion
and fuel stability was described in the previous
report.> The fuel mixture was composed of 62.5
mole % NaF, 12,5 mole % "ZrFA, and 25 mole % UF,.
The loop was cleaned by running it at 1500°F for
16 hr with a non-uranium-bearing salt. The actual
fuel mixture was circulated in the loop for 645 hr,
including 475 hr during which the LITR was at
full power (3000 kw). The power generated in the
loop -was about 2.8 kw, and the linear flow rate of
the fuel mixture was 8 to 10 fps, which corresponds

to a Reynolds number of 5000 to 6500. The dis-

assembly of the loop to provide specimens for
metallurgical examination and chemical analysis
and to reduce the loop to pieces of convenient
size for fuel recovery has been completed.

After perotion of the loop in the LITR, the
section between the linear seal flange and the
pump was sheared with remotely controlled equip-
ment to permit removal of the pump in its shield.
The loop in its water jacket was then fransporfed

fo the hot celis of the Sol:d qu’re Dwuszon for the

remote dlsassemb!/ operation. The out-of—plle end
of the jacket was drawn from the withdrawal shield

with the spec:al dlsassembiy appcsrc:fus show

 

 

30. Sisman et al., ANP Quar. Prog. Rep. Mar. 10,
1955, ORNL-1864, p 150,

  

- system with nonradloactlve material.
-~ pump enciosure was removed, the insulation and
- the various wires and heoters were taken off to
z= provide access fo ‘the pump bowl and the fuel
tubes. After fhe_ fuel tubes were severed and the
nds were clamped fhe pump “bowl was cut ‘usf’
above fhe fuel leve] by usmg the bcnd saw (Fi ig. . 8.5).
‘There : all” 51

through a hole in fhe rear wall of a hot cell equ;pped

   
 
  

“-':ffi'ws;ble on the upp _ _
= the cct:vnty ‘was high enough fo |nd|cate fhe_‘
presence of a cons:derable quam‘l'ry of fnssmni '

 

. prov:ded access to the loop for str:pplng of thermal ™~ the upper porhon of the

A
v e T ’
o S *

PERIOD ENDING JUNE 10, 1955

insulation, heaters, wires, and graphite shield
blocks. The loop is shown partially stripped in
Fig. 8.3, as photographed through a window of the
hot cell. '

Tube sections about 2 in. long were cut from the
fuel tubes for metallographic specimens. These
specimens were taken at about 6-in. intervals from
the part outside the jacket seal flange and at 2-ft
intervals inside the flange, between the heat ex-
changer and the irradiated section. Additional
specimens were taken from the heat exchanger and
the venturi for examination of a longitudinal
section. The U bend of the irradiated section was
cut intfo five specimens from ]/2 to 1 in. long, and
additional specimens were cut from the remaining
18 in. of the irradiated section at about 6-in.
intervals,  The fuel was removed from these
specimens by melting in an argon atmosphere.
Four 2-in. sections of tube were also taken from
the portion of the loop between the jacket flange
and the pump, and two more were taken 6 in. back

- from the ‘irradiated section to furnish samples for

chemical analysis of the fuel.

Disassembly of the pump was complicated by
its size and by the large stainless steel enclosure
around the bowl (Fig. 8.4). A 300-amp, d-¢ welder
equipped with a high-frequency-arc starter was
used to cut the enclosure successfully. It was

-possab[e to guide the cutting electrode with a
- manipulator while cutting the sheet metal in a

manner that prevented contamination of the fuel

After the

 

o’f fuel (~1 g)
ofrthe pump “baffles, buf

     
 
   
  

 
 

too high to permit unshlelded dlsmanflmg and
examination. !t was assumed that fission products
which had vaporized from the fuel were the sources

163

 

S

"

 
 

 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIF [ED
PHOTO 13766

R

RENCH

FOR MANIPULATOR

s

YDRAULIC
MANIPULATOR

CANS FOR SHIE

B L | L

VACUUM CLEANERS
FOR SAW

“SHIEL

T R

FRTPITR S LN

|
|
k4

D W% ) FURNACE POT
WASTE CAN [ T = ' il

i

         
 

 

N Fig.» 8.1. Top View

 

of Hot Cell and Equipment for Dfsassémbly of LITR Fluor.i'de-.FuAei. Loop.

 

 

 

 
o AR AR B, i MG R SR

s

T

 

o

el . A

 

 

 

sl s Sl

 

 

 

 

 

o

FUEL TUBES

AND HEATERS=

 

 

 

o

 

165

il

E

 

 
 

 

 

 

e i e

Do so? LM

ANP PROJECT PROGRESS REPORT

 

 

UNCLASSIFIED
PHOTQ 13889

FUEL TUBE gl

’ ~ Fig. 8.3. Partly Disassembled LITR Fluoride-Fuel Loop.

of the activity, and, since an analysis of the
condition of bearing and seals was precluded,
this portion of the pump was discarded. The fuel-
containing parts were stored for uranium recovery.

The metallurgical specimens have been prepared
and photomlcrogrcphed A preliminary analysis
indicates that subsurface void formation was
generally uniform throughout the loop to a depth of
0.8 mil. The uniformity of attack possibly resulted
from ihe Icck of a femperature differential in the
loop. The temperqture record shows that the fuel
temperature could not have been more than 20°F
lower in the out-of-pile parts of the loop than in
the in- p||e parts. Some sections showed grain
growth, which is an indication that some tempera-
tures were h\gher than those recorded, but even
the areas of grain growth showed no significant

__chq\nge in depfh of corrosion. ,
" The fuel in the nose of the loop was in a maxi-

166

mum perturbed flux of approximately 0.8 x 10"
neutrons/cm?+sec, which corresponds to a fission
power density of 0.5 to 0.6 kw/cm3. The power
density gives a measure of the severity of con-
ditions imposed upon at least one section of the
loop tubing by radiation effects. No difference in
the corrosion could be observed between this
region and other, less severely irradiated, regions.
Also, about the same depth of attack was found in

the venturi section of the loop as was present in

the rest of the loop, even though there was a large
velocity gradient present. No intergranular attack
was noted. N _
The six sections of tubing tcken for chemical
analysis were sampled by a drilling method de-
scribed previously,® After the end of a tube section

 

8¢, C. Webster and J. G. Morgan, Solid State Semiann.
Prog. Rep. Feb. 28, 1954, ORNL-1677, p 27.

 

 
A RS . e o o i A G e, SR il

 

L o o

 

“w .

.“.‘.

-

wrrh a

 
 

 

Flg 8.4. LITR Fluoride-Fuel Loop Pump Priot

‘to Removal of Enclosure for Providing @ Helium

Atmosphere.

“had been cleaned by drlllmg a /-|n. hole abouf
l/ in. deep in the fuel, each sample was bored out
./16"”‘ bit. Tungsten carbide—~tipped bits
used to avoid conrammahng the fuel with

 
  

   

 

‘rhe place where ‘this llne passed through the pump

“shield. The chemical analyses of the samples arewr
“almost complete, “and radiochemical analyses for

- Cs 71_37__, and Zr 95 have been “made__”_of two_ of the

 

_. amples.”

 

elecfrlcal measuremem‘smade durmg 'operaflon of

the loop The resul'rs fromfl‘fhe Cs _ 7_analy5|s

      

indicate escape of the Xe 137 oarent from the fuel.
In orderfo measure the fission power distribution
along the irradiated section of the loop, an experi-

 

nonrad:o‘achve _iron, chromium, or nickel, = Un-
' lrradlared control(samples were obtalned from the
flll lme between the charge ‘tank and the loop at“ '

: tamed markedly more Ry
taken upstream. This difference is exactly that

“The frssron power of 2.7 ancl 55 kw

R - \m'dtcared for the two samples by ‘the Zr®3 aont_entm

 

co pareslfqvombly with fl'ie ? 8 kw_‘vesfjimaiecl from, “

2.1 and 1.5 kw, “are cons;derably lower,fl and they removal rate of Ru

PERIOD ENDING JUNE 10, 1955

ment was conducted in the LITR with a mockup of
the loop. The mockup was designed to duplicate

both the- arrangement of structural materials ‘and

the macroscopic cross section of the fuel so that
effective thermal-neutron flux values could be

‘measured inside the fuel and then could be used

to calculate the power generated by fissioning.

‘The results are being compared with the activation

of cobalt monitor foils attached to the loop and
with additional data from the activation of borings
taken ~from the metallurgical specimens. The
results will be reported when all the data have

been assembled and analyzed.

DEPOSITION OF Ru'%3 IN LITR
FLUORIDE-FUEL LOOP

M. T. Robinson
Solid State Division,

T. H. Handley
Analy’rlcal Chemistry Division

The deposn‘:on of the fission- producr 40 day

Rulq3 on the surfaces of the fuel circuit of the
" ARE was reported prevrously

Supporhng evidence
for this phenomenon has now been obtained from

a siudy of samples taken from the loop described

above. Two samples of tubing from the loop were
examined by gamma-ray scintillation spectrometry.
One sample was taken upstream from the high-flux
region of the loop, and the other was taken at an
equal distance downstream.

The ratio of Ru'®3 to Zr?3-Nb%5 should have
been 0.6 if the Ru'®® had remained uniformly
distributed throughout the loop. However, the ratio
for the sample taken upstream was 5.33, and that
for the sample taken downstream was 8.07. Within

- the experimental error, the Ze?5-Nb93 activity

analyzed the same for both samples. The' sample

""""" 103 fhan did the sample

to be expected from rapid reaction of the ruthenium
in the fuel with 1he container metals. These

‘numbers, comblned wn'rh operahonal and dimensional

data, yielded a hczlrc llfe of 0.5 min for the rate

“constant for removal of l?uw3 from the fuel in the

loop. It was assumed for “the. calculatlon that the
103 was a first-order function
of the concentration.

 

M. T Robinson, S, A, Reynolds, and H. W, Wright,
ANP Quar. Prog. Rep. Mar. 10, 1955, ORNL-1864, p 13.

167

 

 

 

 

 

 

 
e

 

 

 

 

No st delle

ANP PROJECT PROGRESS REPORT

 

 

UNCLASSIFIED
PEOTO 14134
%( 23

Fig. 8.5. Bottom View of Pump Showing Accumulation of Fuel on Upper Parts of the Baffles.

After a delay of 53 days, the two pipe samples
were re-examined to determine the apparent half
lives of the two components of the gamma-ray
spectrum. For the Ru'%3 peak (0.50 Mev), the
half lives of two samples were 42 and 43 days,
respectively, and amply confirmed the assignment
of that activity. For the Zr?5.Nb?5 pedk, however,
the results were 40 and 43 days, respectively,
instead of the nearly 65 days expected. The
contribution to this gamma-ray energy (0.78 Mev)
must be largely from 37-day Nb%3, the 65-day Zr’3
being very low in relative amount (a few per cent).
It appears very likely that niobium deposits on the
Inconel pipes along with ruthenium. The deposition
rate for Nb?3 cannot be estimated, at present, since
it is produced very slowly from its Zr?5 parent,
rather than rapidly from fission as is the case
with Ru 193,

Radiochemical analysis of a fuel sample for
Ru'®® indicated that the amount was below the

168

limit of detection by that technique.® This demon-
strates that the efficiency of ruthenium removal
was comparable to that attained in the ARE.” No
direct neutron activation of the elements of Inconel
was observed.

MINIATURE IN-PILE LOOP

W. R. Willis M. F. Osborne
H. E. Robertson G. W. Keilholtz
Solid State Division

The miniature fluoride-fuel loop for insertion in
a vertical hole in the LITR (described previously®)
was charged with fuel and operated on the bench.
It is now in the final stages of assembly for in-pile
testing. An exact mockup of the experiment has
been insertedin the reactor and has been withdrawn

 

8 . . .
W. W. Parkinson, personal communication,

9W. R. Willis et al., ANP Quar. Prog. Rep. Mar. 10,
1955, ORNL-1864, p 147.
aliainie o

i ol

 

W

 

by using the special withdrawal cask. The operation
proceeded smoothly

DELAY OF FISSION GASES BY
CHARCOAL TRAPS

D. E. Guss, United States Air Force
W, R. Willis, Solid State Division

In the event of a leak in the miniature in-pile
loop mentioned above, the fission gases released
into the cooling-air stream must be contained, and
at the same time, the cooling air must continue to
flow until the temperature of the system has fallen
below the control set point. To accomplish this,
two traps, 24 in. long and 10 in. in diameter, filled
with 6-14 mesh activated cocoanut charcoal, have
been mcorpora’red :ri the cooling-air off-gas line.
These traps, whlch are connected in series and
operate at room temperature can be expected to
hold krypton activity in the system at the design
air flow rate of 20 cfm for 30 sec before the release
of appreciable activity. If the cooling air then
continued to flow at the design rate, the krypton

" activity, which would be released over a period of

about 5 min, would reach a peak in about 2 min.
However, since the air flow would stop in less
than 1 min after a leak occurred, virtually all the
krypton activity would be contained in the traps
and would trickle out over a Iong period of time.
Xenon, the other active gas present in quantity,
will be held back for a much longer period of time;
hence, the trap deS|gn was based on the krypton
holdup time.

Since the action of the charcoal is to increase
the effective volume of the trap space by a factor

VWthh depends dlrecfly upon the amount of chcrcoal_
e used, it was possnble to eshmafe ‘the frap size
S neededwforrfhe_m p||e experlment sumply by scahng N

   

expenmental_' Japparatus

mcheolly in Fig. 8.6. Kryp'ron was |n'rroduced |n'ro"xi_(:_'”

a‘reservoir beneath the frap ond c'r time zero, was

drwen ‘through the” Yrap with air. "It then passed

through a flowmeter and over a counter, where the
activity as a function of time was observed. All
the runs were made at 23°C.

 

   

N

PERIOD ENDING JUNE 10, 1955

The two small traps were tested at flow rates in
the region of interest, and all the elution curves
obtained were similar, the activity rising quite
rapidly and almost linearly to a peak and falling
off more gradually. A typical elution curve is
shown in Fig. 8.7, in which the time coordinate is

UNCLASSIFIED
ORNL-LR-DWG 7330

    
 

RADIATION
DETECTOR

CHARCOAL
TRAP

Fig. 8.6. Schemahc Diagram of Charcoal-Trap
Experiment.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i
ORNL-LR-DWG 7334
10
* o
0.9 / \\
/ ® AIRFLOW RATE: 6600 cm>/min
08 ] \
> L
= | \
»n 0.7
=
. \
= o/ .
z 06 \
o
- \
o / ®
< 05 \
o
) | \
=
< 0.4 4 ®
w
\
/ e :
//0 o
- T e e \..
| J
, | \.
5 | 4
0 20 40 60

‘TRAP VOLUMES OF GAS ELUTED (cm?)

Fig. 8.7. Elution Curve for 334.cm® Charcoal
Trap Containing Krypton.

169

 

"

 

 
 

 

 

&OMW

ANP PROJECT PROGRESS REPORT

v;ekpres.sed' in trcp vol'umes of gas eluted. A plof of
- the inverse of the flow rate vs time to peak activity
for the 23.5-in.- -long, 1.05-in.-dia trap (shown in

“,_',’VFlg 8. 8) indicates a linear relationship. The point

. on ‘|’|‘|IS curve corresponding to 'rhe design flow rate

' -*57:;1'15 shown. _

The e[u‘rlon curve obtained at the in- -pile design
. flow rate is shown in Fig. 8.9. Since the effect
~ of two traps in series is to double the time to peck
“activity and to broaden the elution curve, it is
planned to rely upon two charcoal traps 24 in.
long and 10 in. in diameter to delay any burst of
fission gases from the in-pile loop for 30 sec before
an appreciable amount of activity is released to
the sfcck.

CREEP AND STRESS CORROSION TESTS

W. E. Dav:s J. C. Wilson
N. E. Hinkle J. C. Zukas
Solid State Division
Heat transfer experiments have shown that heat
dissipation in the longitudinally finned, vertical

cylinders in the water-jacketed, helium-filled
enclosure of the tube-burst, stress-corrosion appa-
ratus, as presently designed, % will limit fhe
fluoride salt power densities to about 500 w/cm°.
Therefore fins have. been added between the salt
container and a surroundmg water |acket and the
heat transfer has been increased sufficiently to
permit opercmon at power densities greater than
1000 w/cm®. The first complete apparatus has
been assembled and awaits f||||ng before irradi-
ation in the LITR. Five more rlgs are bemg
fabricated. | -

Tests are under way to determine the suitability
of pressure-regulating valves for supplying gas to
stress the tube-burst specimen. Two pressu're'-
volume fransformers with 1:1 and é: I ratios were
designed and are being fabricated.  These units
are intended to isolate from the remainder of the
system the gas used to stress the specimen so that
rupture of a specimen will result in a known pres-

 

195, C. Wilson et al., Solid State Semiann. Prog. Rep.
Feb, 28, 1955, ORNL- ]851 p 3.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PeRET
. ORNL—LR—-DWG 7332
20

. /
2 15 »—0—7'/

>
o 1

£ ' /

<

<

E ¢ /

ot

u )P

£ 10 -

1

=z / o .

3 re-1FEB 1,1955

- ' A MAR 91955

u DESIGN POINT

a FOR LITR TRAP

s 5 | g

= /—0—4/

’I-.-l
b;é-f_'_'
.fi
0
0 100 200 300

400 50C 600 700

DELAY TIME TO PEAK OF ELUTION CURVE (sec)

) FlgSS Inverse of the Air Flow Rate vs Time to Peak Activity for o 23.5-ifln:-'l._.<;‘n§", 1.05-in.-dia

~ Charcoal Trap Containing Krypton.

 

 
 

 

s R D iR S
o

 

&MW

- SECRTT
ORNL-LR-DWG 7333

 

C 850

 

40

 

30

|

|

|

|

| | |
l | \ AR FLOW RATE: 20 cfm
|
|
|
|

 

RELATIVE RADIATION INTENSITY

 

 

 

 

 

 

20 ][ \-
10 } %
0 100 200 300

TIME (sec)

" Fig. 8.9. Eluhon CU;Ve for 30:,900-cm3.Char<.:ou|
Trap Containing Krypton.

  
  
 
  

: _to 0.0001 in. or better, so that the stress may be

accurate to 1%, several methods of checking wall

~ thicknesses are being studied.

 

PERIOD ENDING JUNE 10, 1955

The stress-corrosion rig irradiated in the LITR
previously ! will be opened when hot-cell arrange-
ments have been completed. A transfer carrier for
bringing subsequent rigs from the LITR to the hot
cells in Building 3025 was designed and is being
fabricated.

The MTR creep apparatus has just completed
six weeks of irradiation in the MTR and will be
returned to ORNL for measurements.

A THEORETICAL TREATMENT OF Xe!33
POISONING IN THE ARE AND THE ART

M. T. Robinson
Solid State Division

A theoretical study of xenon poisoning in a
circulating-fuel reflector-moderated reactor was
made in an attempt to understand the behavior of
the ARE and to exirapolate this experience real-
istically to the ART. The system was assumed to
consist of two phases: the liquid fuel and the
sparging gas (helium), The theory deals only
with volume-averaged concentrations and neutron
fluxes. Turbulent motion of the two fluids was
assumed to assure thorough mixing within each
phase.  With these assumptions, the differential
equations describing the behavior of the poisoning
were derived and solved. The various processes
are iliustrated schematically in Fig. 8.10. The
rates of all processes, except the two Xe'S3
production rates, were assumed to be first order;
that is, the rates are proportional to the Xe 135
concentration in the starting phase. The rate
constants for the two phase-transfer operations
are related by application of the law of mass action,
cm:di thus the eqwhbrlum solubilify of xenon in the

.. fuel is mfroduced

The Xe 135 pouséhmg of a fluid- fueled reactor

'\__under steady-state condmons is gwen by

(a r a,)k

°°,.-~ 2
-mf&?ifffif_

Xw= sfeady state Xe]35 -poi'sbni.ng (%),
Ty = 'IOnyeaf T, .
ay = 00y /o,,

 

”W. W. Davis et al., ANP Quar. Prog. Rep. Mar. 10,

1955, ORNL-1864, p 155,

171

 

 

 
 

 

 

Do et AL

ANP PROJECT PROGRESS REPORT

 

 

SECREY
SSD-A-1167
ORNL—LR—DWG 6430A
PRODUCTION FROM

 

 

 

 

 

 

 

 

 

 

 

 

DECAY OF 1'°°
LOSS BY FLOW OF TRANSFER OF xe'3° ‘ LOSS BY N
o GAS FROM SYSTEM Xe' 29N LIQUID TO GAS xe'3d N THERMAL-NEUTRON ABSORPTION
o ' SPARGING -GAS 135 | LIQUID-FUEL
— . lLOsSBY PHASE TRANSFER OF Xe™> PHASE _
RADIOACTIVE DECAY GAS TO LIQUID * LOSS BY RADIOACTIVE DECAY

DIRECT PRODUCTION
IN FISSION

Fig. 8.10. Processes Governing Xel35 Poisoning in Liquid-Fueled Reactors.

k]=a4+)\f+)\L,
ky=a, + azfi)\f+ Ag s
o, = RTS,
B = VL/VG !
A'I_. = O-Xe(’l‘S
/\g = vG/VG ,
)‘f = rate constant for transfer of Xe 135

from liquid to gas,
)/xe, y, = fission yields of Xel35 1135

f' Xe = thermal-neutron cross sections for
Y235 tission, u23s absorption,
S “and Xe 135 absorption,
e d4' e radioactive decay constant of Xe
" ¢ = volume-averaged thermal-neutron
0 flux,

135

VL’VG = volumes of liquid, gas phases,
ve = sparging gas volumetric flow rate,

S = séiub”ity of xenon in fuel at l-atm
~ Xe pressure,
. T = absolute temperature,

'R = universal gas constant.

- The results of the calculations made by the
equation and the ART data given in Table 8.2 are
presented in Fig. 8.11, which is o design chart for
the estimation of the sparging-gas flow rate for
conceivable levels of Xe'3® poisoning.  The
- parameter 7, (Fig.8.11) is the phase-transfer mean
life from the liquid to the gas:
1

/
R A

Thé \-r;:iue of ;rf 1"ol Ee chosen is unknéwn, but ARE

72

experience indicates a value of about 20 min for
that reactor. |t is to be expected that the value
for the ART will be considerably smalier.

A study of the poisoning behavior of the ART was
made for two different types of shutdown. |n each

case, it was assumed that 7, = 5 min and that
= 1000 STP liters/day. At stead¥ state during
full nuclear power operation, the Xe'>> poisoning

would be about 0.40%. If the nuclear power was
reduced to zero but sparging was continued, the
poisoning would not rise by more than 1 or 2% of
the steady-state value before starting to decrease.
If sparging was discontinued, the poisoning would
rise in about 11 hr to a maximum value of about
12%. If at this point sparging was resumed, about
36 min would be required to reduce the poisoning to
below the steady-state value. Intermediate values
of 10, 5, 2, and 1% poisoning would be reached in
times of about 1, 7, 17, and 25 min, respectively.
It appears highly probable that no difficulties will
be encountered from Xe'3® during short shutdowns,
since removal of the xenon to acceptable levels by
sparging can be made within the time necessary to
force the fuel from the dump tanks into the reactor.

A series of calculations has been performed on
the Oracle to study the approach of the Xe 135
poisoning to its steady-state value. In all cases
of practical interest, this approach is controlied
by the rate of 1'35 production. In fact, except for
a change in ordinate, the poisoning is found to
follow the iodine growth curve.

It is concluded that the removal of Xel3% by
sparging with helium (or other inert gas) appears
to be a satisfactory means of controlling poisoning
in the ART. No difficulties seem to exist in
connection with shutdown behavior of the poisoning
at the sparging rates selected for ART operation.

£

‘s
PERIOD ENDING JUNE 10, 1955

TABLE 8.2. DATA USED TO CALCULATE POISONING IN ART

 

; o 7 - ' : A " Numerical Data
* - ag = 0.254% R = 82,0567 cm3-01'm/moie-°K
: ay = 4.74% | T = 1033°K (1400°F)
; 0 a, = 0.0509 S = 6 x 1077 moles/cm>+atm
Q, = 2.09 x 10 s sec-] Oyo = 1.7 % 106 barns*
) e
_ . Reactor Data™*
- ARE ' ART
v, 535 5.64 f1°
' .3 | 3
VG 1 .fl' | 0.31 ft
v 0.25 em>/sec 1000 STP liters/day
¢. 8 x 10! neutrons/cm? sec | 1 x 10" neutrons/cm?-sec
B 5.35 | 18.2
A’L 1.4 x 10;6 sec“] 1.7 x 10_4 sec—]
- A K_. Ergén:-c;na{ HW Bé}fi‘ni',‘A'NP indr.:t;P'r?o"g. Rep. Ma;. IO, 1955, ORNL- 1864, p 16.
"'-,‘_;' **ARE datae from ARE Nuclear Log Book and from J. L. Meem; ART data from J. L.. Meem and W. T. Furgerson.
ECRE™
~ SSD-B-1168
¢ 5 ORNL—LR—DWG 64 31A

 

A

e

T 1 Ty = 20 min

 

 

“xe'3% POISONING (%)

;= 10 min

 

 

=5 mlin

 

Tf=2m‘in
L [
“Tf=Om‘in"

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

P .
T 3000 0 40000 5000

" SPARGING-GAS FLOW RATE (STP liters/doy)

Fig. 8.11. Stéu'dy-Sfafe Xe'®3 Poisoning in_ART as a Function of Sparging-Gas Flow Rate for Various
Assumed Values of the Phase-Transfer Mean Life, Ty

173

 

  

W

 

 

 
 

 

 

' Do ot bl

- ANP PROJECT PROGRESS REPORT

9. ANALYTICAL CHEMISTRY OF REACTOR MATERIALS

C. D. Susano

J. C. White

Analytical Chemistry Division

Modifications of the method for the determination
of uranium metal in fluoride-base reactor fuels
were completed. Further studies were made on the
evaluation of the methylene-blue method for the
determination of trivalent uranium in fluoride-base
reactor fuels. Investigations were continued on
methods for the determination of oxygen as metallic
oxides in fluoride salts. An improved separation
method invelving anion-exchange resins was de-
veloped for the determination of alkali metals in
fluoride salts. A volumetric method for the deter-
mination of zirconium in fluoride-base fuels was
proposed. o

DETERMINATION OF URANIUM METAL IN
FLUORIDE SALT MIXTURES
A. S. Meyer, Jr. B. L. McDowell
Analytical Chemistry Division
The apparatus for the determination of uranium

metal in fluoride-base fuels according to a method
based on the decomposition of the hydride in an

CLOSED-END MERCURY-SEALED
I \ OIL MANOMETER

TO VACUUM Cu0 TUBE
PUMP

" LIQUID NITROGEN
COLD TRAP

MERCURY PRESSURE
INDICATOR

 

COMBUSTICON
TUBE

FURNACE

atmosphere of oxygen at reduced pressure! is
shown in Fig. 9.1. This apparatus has been cali-
brated and has been applied to the determination of
uranium metal in UF; and KF-UF .

Samples of the uranium metal are placed in a
platinum boat in the combustion tube and heated to
250°C for 1 hr in an atmosphere of hydrogen in
order to convert any uranium metal present in the
sample to the hydride. The excess hydrogen is
evacuated from the system after the sample has
been cooled to room temperature in an atmosphere
of hydrogen. The UH, produced from the uranium
metal is oxidized by heating at 400°C for 20 min
in an atmosphere of oxygen: '

AUH, + 70,—> 4U0, + 6H,0

The combustion gases are passed through a copper
oxide tube at 500°C to ensure complete conversion
of the hydrogento water. The water vapor produced

 

IA. S. Meyer, Jr., and B. L. McDowell, ANP Quar.
Prog. Rep. Mar. 10, 1955, ORNL-1864, p 158.

UNCLASSIFIED
ORNL-LR-DWG 7334

TO KIPP GENERATOR

 

 

 

 

 

 

 

ASCARITE
SAMPLE
y 7 LIQUID NITROGEN
’\\ COLD TRAP
N

FURNACE

0z

ALCOHOL AND MgC102

DRY ICE

" Fig. 9.1. Diagram of Apparatus for the Determination of Uranium Metal.

174

 
B el oo ik

 

L

"

  
 
     
  
 

Do et

by the oxidation of the combined hydrogen is
collected in the cold trap, and, after the oxygen
has been removed by evacuation, the water is
allowed to expand into a system of known volume.
The pressure of the water vapor is measured on
a closed-end mercury-sealed oil manometer similar
to that described by Naughton and Frodyma.? The
pressure registered on the manometer is related to
the weight of the uranium present in the sample by

 empirical calibration of the apparatus.

A preliminary calibration of the apparatys with
BaCl,-2H,0 as.a standard for hydrogen in the form
of water indicated a linear relationship between
scale reading and weight of hydrogen for the range
0.18 to 0.8 mg of hydrogen. This linear relation-
ship does not apply to larger amounts of hydrogen
because the pressure in the system is approxi-
mately equal to the vapor pressure of water.
Results which have been obtained for the cali-
bration of the apparatus with samples of pure
uranium metal are in agreement with the BaCl,-2H,0

" calibration, and they indicate that a scale reading

of 1 mm represents 0.36mg of uranium (Fig. 9.2).
The results of the calibration with uranium metal
have a coefficient of variation of 7% for the range
6 to 60 mg of uranium. ' '

UNCLASSIFIED
ORNL—LR-DWG 7335

 

 

 

 

 

25 T T T T
e OBTAINED FROM CALIBRATION
WITH BaCly- 2 H,0 AS A STANDARD
= 20 —— FOR HYDROGEN (N THE FORM OF WATER -——8
= O OBTAINED FROM CALIBRATION
€ 45 L. WITH URANIUM METAL |
a i ‘ =
o
i :
w 10 /‘/
- :
< - .
2 & . .
o/d :

 

 

 

 

 

 

 

 

 

 

   

Flg. 9.2. Calibration of Apparotus For Determl-
ation of Uranium Metal T

 

 

e R

samples represented portions of the material

 

J. J. Naughton and M. M. Frodyma, Anal. Chem. 22,
711 (1950).

 

 

   
  
 
  
 
 

2 i . L- 9.
r'hICh were onrolyzed for uranium metal were fcken VPrOg Rep. Dec: 10, 1954, ORNL-1816, p 12

by drillmg with” |ncreasmg|y |crger bl’rs so that the e

PERIOD ENDING JUNE 10, 1955

successively nearer the wall of the reaction tube.
Samples of UF; and fluoride salt mixtures which
were to be analyzed for uranium metal were ground
in an atmosphere of hydrogen to prevent oxidation
of the uranium metal. The samples, along with the
necessary equipment for grinding, were placed in
a plastic bag, which could be alternately evacu-
ated and filled with hydrogen. The samples were
ground to a particle size that would pass a
standard 325-mesh sieve.

Several samples of UF; which had been produced
by the reduction of UF, with uranium metal have
been analyzed for uranium metal in the new appa-
ratus. The results for the uranium metal content
paralleled those obtained by petrographic analysis
of the material. The standard deviation for dupli-
cate determinations on the samples of UF, was
approximately the same as that obtained for the
calibration of the apparatus with uranium metal.

The difficulties encountered earlier in the decom-
position of the UH, in certain gaseous atmos-
pheres" have been completely eliminated in the
present method. The period of several hours which
was required for the decomposition of the hydride
in an atmosphere of carbon dioxide was reduced to
approximately 20 min when the ignition was carried
out in an atmosphere of oxygen. Tests indicate
that no interference is introduced by the presence
of fluoride salts or trivalent uranium in the
samples, and therefore the procedure should be
applicable to analyses of all types of fluoride
fuels. A topical report on this investigation is
being written.

DETERMINATION OF TRIVALENT URANIUM
INFLUORIDE FUELHS
A. S. Meyer, Jr. W. J. Ross
Analytical Chemistry Division

Oxidation of Trivalent Uranium by
Mefhy|ene Blue '

jtohiparison ‘of the methyle'\e-blue “and the
hy'drogen evolution® methods for the determination

 

    

A, s Meyer, Jr., and B, L. McDowell,

A S. Meyer, Jr., D. L Mannmg, and W J. Ross
{151\}1’ Quar Prog. Rep. Mar. 10, 1955, ORNI -1864, p

D L. Meanning, W. K. Miller, and R. Rowan, Jr.,
Methods of Determination of Uranium Trifluoride, ORNL.-
1279 (April 25, 1952).

175

 

of trivalent uranium in fluoride-base fueljs has been

ANP”Quar |

T T o

 
R,

 

 

 

 

LD»MM

ANP PROJECT PROGRESS REPORT

completed. Portions of eutectic mixtures, which
had been pulverizedand sampled under inert atmos-
vheres, were analyzed simultaneously by both
methods in an effort to minimize the effects of the
heterogeneity and the instability of such samples.
An evaluation of the results showed that satis-
factory agreement, comparable to that reported
previously? for other eutectic mixtures, was
achieved in the analytical results for samples of
LiF-BeF,, NaF-LiF, and NaF-BeF,. The results
obtained by the methylene-blue method in the
determination of trivalent vranium in KF-UF ,-UF,
eutectics continue to exhibit negative bias when
compared with the results from the hydrogen-
evolution method. No further studies were made of
NaF-ZrF -UF ; eutectics. All the mixtures studied
could be dissolved in methylene-blue solutions
that were 1.5 to 6 M in HCl and saturated with
AlCIS. No hydrogen is evolved during the dis-
solution of the most reactive trivalent wuranium

‘compounds, such as U'Ci-s, in 1.5 to 3 M HCI

solutions of methylene blue if efficient agitation
is maintained while the solvent is being added
to the flask.

Methylene-blue sclutions that are 1.5 to 6 M in
HCI are reduced, during dissolution at room
temperature, to methylene white by finely divided
metallic chromium, nickel, iron, and uranium-nickel
alloy. Zirconium is only very slightly soluble in
such solutions over a 2-hr period. The presence
of metallic impurities in fluoride-base fuels causes
highly erratic results when ftrivalent uranium is
determined by the methylene-blue method, but such
impurities have an even greater adverse effect on
the hydrogen-evolution method. Dissolution of the
metallic impurities in the hot acid solvent results
in quantitative evolution of hydrogen. A more
comprehensive description of the theoretical and
practical aspects of the methylene-blue method for
determining trivalent uranium will be issued in the
form of a topical report.

Simultaneous Determination of Trivalent Uranium
and Total Uranium

The inherent advantage of the methylene-blue
method over the hydrogen-evolution method, that
is, the possibility of determining trivalent uranium
and total uranium in the same sample, is still being
investigated. The studies have led to the postu-
lation of an interaction species of pentavalent

‘uranium and methylene white.* A detailed study

176

of this postulation has been initiated. The method
of Vosburgh and Cooper® will be applied to es-
tablish the stoichiometry of the complex. Attempts
to prepare standard solutions of methylene white
through the reduction of methylene biue by zinc
amalgam have been unsuccessful because of the
ease with which zinc ions react with both the
oxidized and reduced forms of methylene blue.  An

attempt is being made to prepare pure methylene

white through electrolytic reduction of methylene
blue.

DETERMINATION OF LITHIUM IN
LiF-BeF, AND LiF-.ZrF4--UF4

A. S. Meyer, Jr. D. L. Manning
Analytical Chemistry Division

A method for the determination of lithium in
LiF-BeF, and LiF-ZrF ,-UF, has been developed
that is based on a separation of lithium by an
anion-exchange resin, Dowex-1. In this procedure
a sulfate solution of the fluorides that is approxi-
mately 1 N with respect to H,50, is passed
through a column of Dowex-1 in the hydroxide form.
In order to avoid the depletion of the resin in the
column by the H,50,, the solution is equilibrated
with about 20 ml of a water slurry of the resin in
the hydroxide form in a beaker before being placed
onto the column. Beryilium, uranium, and zirconium
are retained on the column as the beryllate,
zirconate, and diuranate anion, respectively,
while the lithium passes into the effluent as LiOH,
The following equations represent the probable
reactions, where ROH is the hydroxide form of the
anion-exchange resin:

H,50, + 2ROH —> R,50, + 2H,0

Zr(SO,), + 8ROH — R,Zr0, + 2R,SO, + 4H,0

BeSO, + 4ROH —> R2Be02v+ R,50, + 2H,0

2U0,80, + 6ROH— R,U,0, + 2R,30, + 3H,0

27277
Li,50, + 2ROH —> R,SO, + 2LiOH

It may be seen from the equations that LiOH and

water are the only species which are found in the
effluent. The lithium is determined by titrating’
the hydroxide with a solution of standard HCI.

If the fluoride mixtures contain sodium fluoride
in addition to lithium fluoride, the sodium will
react in the same manner as the lithium and will

 

6W. C. Vosburgh and G. R. Cooper, . Am Chem. Soc.
63, 437 (1941).

 
i

i

 

B

L

 

 
 
 
 
  

 

Do pa? Ao

be found in the effluent as NaOH. A subsequent
titration of the effluent with standard HCI will give
the sum of sodium and lithium as the hydroxides.
Lithium is then determined by extracting the LiCl,
after titration with HCI, with 2-ethyl-1-hexanol and
titrating the chloride ion in the nonaqueous medium
according to the procedure developed by White
and Goldberg.”

If a single alkali metal is present in the fluoride
mixture, it can be determined directly by titration
of the effluent with standard HCl.  When more than
one alkali metal is present, the sum of the alkali-
metal hydroxide concentration in the effluent is
obtained by the method just described. Lithium is
then determined by the 2-ethyl- 1-hexanol procedure,7
po'rastm, by the tetraphenyl boron gravimetric
method;® and sodium, by the difference.

W||||ams‘c:nd Vaughan (cf. *“ANP Service Labo-
ratory”’) have utilized a two-stage anion-exchange
resin column for the determination of the alkali
metal fluorides in the presence of beryllium,
zirconium, ‘and uranium fluorides. The first column

¢ontains the anion-exchange resin, Dowex-1, in

the citrate form, whereas the second column
contains the resin in the hydroxide form. Beryl-
lium, zirconium, and uranium remain in the citrate
column as the anionic citrate complexes, and the
alkali metals pass into the effluent as the alkali
metal hydroxides. The alkali metal concentration
is then determined by titrating the effluent with
standard HCi

In the method described above, a single column
of anion-exchange resin in the hydroxide form is
used, and thus the need for the citrate form of the

anlon-exchange ‘resin s’ completely eliminated.
© The method has been tested in the ioborafory ‘and
.‘found' to be schsfactory in every ‘respect. The ™
T , no of the mefhod is of ’rhe"
" order of 1% in the rangémlo to 30 mg “of ]n‘h:um A &
"','}"(:'toplcul reporf on the method |s belng wrlt’ren '

 

‘preserice of sulfate has been repor’red by Fritz S

Johnson.? Excess disodium dihydrogen ethylene-
diaminetetraacetate (EDTA) forms o very stable

   

 

  
 
  

PERIOD ENDING JUNE 10, 1955

complex with zirconium in acidic solutions. The
excess EDTA is then back-titrated with bismuth
nitrate; thiourea is used as the indicator. The end
point is noted by the formation of the yellow
bismuth-thiourea complex. "An advantage of this
method in comparison with previous methods for
using EDTA in the determination of zirconium is
that anions which form complexes with zirconium,
such as fluoride, sulfate, phosphate, thiocyanate,
and tartrate, do not interfere,

This volumetric method for the zirconium determi-
nation appeared to have several advantages over
the gravimetric mandelic acid method now being
used, but it was found that the yellow thiourea end
point was obscure, especially in the presence of
hexavalent uranium, which is also yellow. The
precision was consequently poor.

The possibility of modifying the method by back-
titrating the excess EDTA with a solution of
iron{[1l), with disodium- 1, 2-dihydroxybenzene-3, 5-di-
sulfonate (Tiron) as the indicator, is now being
investigated.  This reagent forms an intensely
purple complex with iron(ltl) in an acetate-buffered
solution having a pH of about 5.0. Two moles of
Tiron combine with 1 mole of iron under these
conditions. This compiex, which has a pK of about
10, is relatively stable. !

The pK of the iron-EDTA complex’ is about 25,
which is approximately the same as that of the
zirconium-EDTA complex. It is possible, therefore,
to add a solution of iron{lll) to a solution of EDTA
and Tiron at a pH of 5,0 so that when the EDTA is
completely complexed the color of the solution will
change from yellow to purple. Qualitative, pre-
Ilmmory tests revealed this change in color at the
“end pomf ‘to be well defined. The end point was

Cvery shorp when a solution of 0.05 M EDTA
" “containing about 50 mg of Tiron was titrated with a
“solution of 0.05 M iron(lll}. The end point corre-
""" sponded to the correct stoichiometry of the iron-

EDTA reachon in which 1 mole of iron cemplexes

»_,wflh '| mole of“Er_D‘TA Future 'wcrk w:” lnclude

 

 

 

The volume‘rrlc determmahon of zirconium in the -

"fi}'{—'i‘.’J. C. White und G. Goldberg, Applz'cation of the

Volbard Titration to the 2-Ethyl-1-Hexanol Separation
Method for the Determination of Lithium, ORNL-1806
(Nov 4, 1954).

sy 8¢. R. Williams, ANP Quar Prog Rep. Mar. 10, 1955,
. ORNL 1864, p 162,

J S. Fritz and M, Johnson, Volumeinc Determmatwn
of Zirconium and EDTA Method Involving Back-Titration
with Bismuth, 15C-571 (Feb. 1, 1955).

1OA E. Harvey, Jr., and D. L.. Manning, J. Am. Chem.
Soc. 72, 4488 (1950).

177

 
- ik

bopoi e

 

 

 

ANP PROJECT PROGRESS REPORT

the determination of zirconium by the modified
method of back-titrating the excess EDTA and a
study of the possible interferences, particularly
fluoride and tetravalent uranium. It is expected
that tetravalent uranium would interfere because
it would complex with the EDTA in the same
manner as the zirconium would.

DETERMINATION OF OXYGEN IN
FLUORIDE FUELS

“A. S. Meyer, Jr. J. M. Peele
Analytical Chemistry Division

Further tests were carried out on the determi-
nation of oxygen as oxide in fluoride fuels'! by
electrolysis of solutions of the samples in fused
KHF,. Quantities of oxygen in excess of 90% of
the theoretical value were recovered when known
amounts of oxygen were introduced as water by the
addition of samples of Na 2CO,4, which react with
the fused blfluorlde in accordance with the following
equcmon

Na,CO, + 2KHF ,— 2NaF + 2KF + H,0 + CO,

The yields were not quantitative, because the rate
of oxygen evolution decreased as the concentration
of water in the electrolyte was reduced. Since
traces of hydrogen were deiected in the insoluble
gases, even though a high concentration of AgF
was added to the electrolyte, methods were adapted
for the direct determination of oxygen in the
effluent gases.

For samples containing large quantities of
oxygen the Orsat'? method for the determination
of oxygen in gases was applied by passing the
insoluble gas, which was collected over KOH,
into a solution of alkaline pyrogallate and meas-
uring the decrease in volume of the gas.

A modification of the Winkler method '3 for the
determination of oxygen in water was adapted for
samples which contained smaller concentrations
of oxygen. The apparatus was modified by con-
verting the sweep gas to purified helium and
bubbling the effluent cell gases directly into an
alkaline solution which contained Kl and a sus-
pension of Mn(OH), in which the oxygen was

 

A s, Meyer Jr., and J. M. Peele, ANP Quar. Prog.
Rep Mar. 10, 1955, ORNL-1864, p 159.

2w, w, Scott, Standard Methods of Chemical Analysis,

51‘1’1 ed., ll, 2349, Van Nostrand, New York, 1939

13| . W. Winkler, Ber. 21, 2843 (1888).

178

absorbed according to the reaction
4Mn(OH), + O, + 2H,0—> 4Mn(OH),

The absorbed oxygen was determined by acidi-
fying the absorber solution and thus liberating an
equivalent quantity of iodine, which was then
titrated with standard Na,5,0, solution. If only
microgram quantities of oxygen are present, the
iodine is measured by extracting it info orthoxylene
and determining the concentration of iodine in the
organic phase spectrophotometrically according to
the procedure of Silverman, Bradshaw, and Taylor. 14
Experiments are now being carried out in an attempt
to reduce the time required for the quantitative
evolution of oxygen. No significant increase in
the rate of oxygen generation was obtained by
carrying out the electrolysis at a temperature of
250°C rather than at 100°C. The effect of the
design and current density of the anode on the
efficiency of the generation of oxygen is now being
studied. It has been found that the use of o
mercury cathode, which is introduced by placing
a Teflon cup in the electrolytic cell, simplifies
the electrolysis, Fluctuations of the current are
decreased, and the evolution of fluorine is greatly
reduced. By introducing the sweep gas below the
surface of the mercury, the problem of slugging of
the entrance line is eliminated. Since AgF is
incompatible with mercury under these conditions,
it cannot be used to reduce the rate of generation
of hydrogen.

DETERMINATION OF OXYGEN IN 7
METALLIC OXIDES BY BROMINATION

J. P. Young G. Goldberg
Analytical Chemistry Division

Extension of the method of Codell and Norwitz '3

for the determination of oxygen in titanium to the
determination in fluoride-base fuels was continued
during this period. Several modifications of the
apparatus described previously '® were made. New
traps of smaller dimensions were designed, and, in
addition, an ice-salt trap was placed between the
ignition tube and the first dry-ice~alcohol trap to

 

4[... Silverman ond W. Bradshaw, Determination of

Oxygen in Certain Gases, NAA-SR-892 (April 15, 1954).

M. Codell and G. Norwitz, Chem. Eng. News 32,
4565 (1954).

6;. c. White, G. Goidberg, and J. P. Young, ANP

Quar. Prog. Rep. Mar. 10, 1955, ORNL-1864, p 161.

R’

 
i G, A i

 

DY

 
  
 
 
   

-

- oxygen con’rammchon were found.

=+ In ‘several attempt

Bo not Alel

‘remove more completely the excess bromine vapor.

[+ was found that if the flow rate of the mixture of
helium and bromine vapor were too rapid, the
second dry-ice—alcohol trap became plugged with
solid bromine during the course of a determination.

In the previous application of the bromination
method to the de’rermlnohon of oxygen in BeO,
considerable oxide confcmlncmon was found in the
NaF FeF flux that was used. A small quantity of
the lmpure flux was therefore treated with bromine
vapor at 750°C, and then the prefreated flux was
used in the determination. Essentially complete
recovery of the oxygen in the BeO sample was
obtained =when this mixture was treated with
bromine vapor at 950°C.

Several samples of CrF, were analyzed for oxide
contommahon, and reasoncble values were found
when the samples were treated with bromine vapor
at 950°C The accuracy and optimum conditions

for this determ!noflon will be investigated with

known mixtures of CrF3 and Cr,0O

Also, several samples of Nazsz were analyzed
for oxide contamingtion. The bromination was
performed at 950°C, _and reasonable values of
During the
analysis of two scmp]es of Li,ZrF,, a precipitate
formed on the surface of the Bc:(OH)2 bubbler and
caused inefficient flow of gas through the appa-
ratus. It is believed that the occurrence of this
precipitate was due to a reaction mvoivmg the
hydrogen fluoride or fluorine which was present or
was formed in the samples being analyzed. A trap
of NoF and KBr is belng pre pared for removal of
these gases. The accuracy ‘and op'rlmum condl’rlons
for the defermmaflon of oxygen contamma'rlon in

- ZrF W||| fhen be mveshgated

    

;wcs found that the occurrence of various side

reactions made the method impractical.

   

  

s aoply-" fhe flrst reaction as
a method for determmlng omde present as BeO, it

PERIOD ENDING JUNE 10, 1955

It was impossible to secure reasonable blanks in
the absence of BeO, and furthermore, when BeQO
was present, no evidence of stoichiometry wds
found. In these attempts, samples of BeO and
BeF,-BeO were placed in contact with solutions
of KF, as concentrated as 20% (W/V), at tempera-
tures of 80°C. The concentration of hydroxyl ions
presumed to have been formed during the dissolution
of the sample was then determined by titration with
standard acid. Values from 7.5 to 9.0 pH units
were- chosen as arbitrary end points for these
titrations. Dissolution of either BeO or BeF, (with
oxide contamination) was more rapid in higher
concentrations of KF and at pH values of 7.5, A
fine crystalline precipitate was formed during the
dissolution of all samples containing BeF,. From
the results of x-ray diffraction analysis, it was
found that this precipitate contained K,BeF,.

DETERMINATION OF TRACE AMOUNTS OF
NICKEL IN FLUORIDE FUELS WITH
SODIUM DIETHYLDITHIOCARBAMATE

J. P. Young M. A. Marler
Analytical Chemistry Division

The use of sodium diethyldithiocarbamate in the
colorimetric  determination of nickel in trace
amounts was investigated. The work of Chilton'’
served as a basis for these studies.

Sodium diethyldithiocarbamate is a very sensitive
colorimetric reagent for nickel; however, this re-
agent forms colored complexes with a wide variety
of cations. Many of the metal complexes of sodium
diethyldithiocarbamate are extractible into organic
solvents.

The molar absorbancy index for nickel-diethyl-
dithiocarbamate extracted into carbon tetrachloride
or" 1,2-dichlorobenzene from an aqueous solution

“~ whose pH was 9 was found to be about 34,000 at
‘7 a wavelength of 328 mu. The molar absorbancy
-. index of nickel dimethylglyox'ime is about 11,000
.. at the wavelength of its maximum absorbancy.
'.'/_L(:hllton'|8 subsequently reported that a more ef-
_:W_flment exfrcctlon of the mckel cpmplex occurred |
i the pH of the ‘aqueous phase were first basic
.+: to bromeresol green (pH ~5). Without the addition
_of the complexing agents described by Chilton, it
“Was found that uranium interfered in fhis method,

e STl

 

7). M. Chilton, Anal. Chem. 25, 1274 (1953),
18 M. Chilton, Anal. Chem. 26, 940 (1954),

179

 

 

 
 

 

 

ANP PROJECT PROGRESS REPORT

as did iron and molybdenum to a very slight degree.

-With the addition of the complexing agents de-
scribed by Chilton, it was found that the inter-
“ference of uranium was still present, although
“somewhat decreased. Of the other ions present in
fluoride-base fuels, chromium and the alkali ions
presented no interference, and evidence indicates
that zirconium does not interfere. Absorbancy
“spectra of carbon tetrachloride extracts were
—~determined for aqueous solutions containing sodium
~ diethyldithiocarbamate and all the cations mentioned
-in this discussion.

ANP SERVICE LABORATORY.

W. F. Vaughan C. R. Williams
Analytical Chemistry Division

The number of determinations of the oxygen
content of metallic sodium increased sharply during
‘the quarter. Efforts are being made to increase
the accuracy of this determination, and two steps
have been taken to achieve this goal. More ef-
ficient removal of the thin film of adsorbed moisture
on the outer surface of the glass bulbs used for
the determination is being achieved by immersing
the bulbs in acetone and then quickly rinsing them
in dry ether before placing them in the hexane-butyl
bromide mixture. Also, the reagents, both hexane
and butyl bromide, are being desiccated over
phosphorus pentoxide so that the water content is
less than 5 ppm.

The major portion of the work in the service
laboratory continued to be the analysis of fluoride
salts, with the emphasis being on the determination

180

of the following components: Na, Zr, Li, K, F,
U3, total U, Ni, Cr, Fe, and Mo. A new procedure
for the determination of lithium was proposed and
tested in which lithium and other alkali metals
are first separated from zirconium, beryllium,
uranivm, and sulfate ions by means of two anion-
exchange resin columns in series. The first column
is prepared in the citrate form and the second in
the hydroxide form. The effluent from the second
column is a solution of lithium or alkali metal
hydroxides, the concentration of which is deter-
mined by titration with standard acid. Work by
Manning (cf. ‘“'Determination of Lithium in LiF-
BeF, and LiF-ZrF4-UF4”) has subsequently
shown that only the hydroxide form of the anion
resin is required for the separation. '

A total of 1673 samples was analyzed, on which
8678 determinations were made. The backlog
consists of 151 samples. A breakdown of the work
is given in Table 9.1,

TABLE 9.1. SUMMARY OF SERVICE ANALYSES

 

 

 

REPORTED '
Number of Number of

Samples  Determinations
Reactor Chemistry 1139 6143
Experimental Engineering 525 2515
Miscellaneous 9 20
Total 1673 8678

 

 
 

i

¥

!

Y ‘which

—

PERIOD ENDING JUNE 10, 1955

10. RECOVERY AND REPROCESSING OF REACTOR FUEL

F R Bruce
D. E. Ferguson W. K. Eister H. E. Goeller
M. R. Bennett J. T. Long
G. I, Cathers R. P. Milford

S. H. Stainker

Chemical Technology Division

E. E. Hoffman

C. F. Leitten, Jr.

Metallurgy Division

PILOT PLANT DESIGN

The design of the pilot plant for recovering ANP
fuel by a fused salt—fluoride volatility process
is expected to be completed by August 15. The
engineering flowsheet is 85% complete. Of the
29 pieces of process equipment now contemplated,
10 are completely designed or specified and on
order; 2 of the 10 have been received. The
scheduled construction completion date is still
December 31, 1955,

The plant will be located in cells 1 and 2 of
Building 3019. Equipment for highly radioactive
materials — the fluorinator, the ARE fuel dump
tank, and the vessel for melting other salts to be
processed (for example, salts from the in-pile
loops) — will be in cell 1. Equipment for less
radicactive materials will be in cell 2. The
remainder of the equipment will be in the operating
gallery and on the roof above cells 1 and 2,

The present plan of operation is to lower the
ARE dump tank into the cell, melt the contents,
and pass the molten salt info the fluorinator by
means of'm"f'rogen pressure. |he process differs
from thc’r prevnously descnbed] in the following

~ details: “only two cold traps are provided, one

at -40°C qnc! one at ~62°C; fhe UF will be

w"‘recycled “together “with fluorine gas, through the
- absorber ‘and cold traps to effect additional decon-
-~ tamination nc necessary to meef product spec:fl-'
| "'.':":“"ca'r:ons. o s
. The bcrren molfen sqlf will be removed from the
i fluorlnm‘or with pressure and pu'r mfo metal cans
_ezzhburled The sodium fluoride ab-
" sorber 'con’ral'nhx'ng the volqtlle Ffission- product floe-
rides will be transported to the burial ground,
- where it will be dumped and the contents buried.
Excess fluorine in the off-gas will be reacted with

 

 

'p, E. Ferguson et al., ANP Quar. Prog. Rep. Mar.

5 to 10% aqueous potassium hydroxide in a spray
tower, '

Two refrigeration units, each consisting of a«
Freon F-22 and F-13 cascade system, will be used
to chill Freon F-11 to —40 and —-62°C for recircu-
lation through the cold traps.

PROCESS DEVELOPMENT

In further studies on an absorption bed for

removing volatile radicactive material from the

Ui:é-l:2 gas stream of the fused salt—volatility
process, sodium fluoride was found to be much
more effective than calcium fluoride,  Decon-
tamination was poor in second runs with the same
absorbent bed. |t is believed that the femperature
of the bed is important, but control of the bed
temperature was difficult,

The absorbent beds were 1-in.-dia tubes that
contained 90 g of either calcium fluoride or sodium
fluoride in a 9-in. length or 180 g in an 18-in.
length, The CaF_ used was made by fluorinating
CaSO,. Both the CaF, and the NaF were 12 to
40 mesh. The gas from a fluorinator was passed
directly into the absorbent bed. The fluorination

‘reaction was carried out with 365-g charges of the

ARE-type fuel NaF-ZrF, -UF,

mole %) containing 30 g of uranlum.

(53.5-43.0-3.5

Previous

“work? had been done with similar material on a

sccle of a 67-g charge of fuel
"The NaF was ten times as effective as reported

‘previously., The amount of fluorine used for the
fluorination ‘was much less than the ninefold
"_:excess used prewously. In some of the new runs
“the flucrine was passed info the molten fuel very
rapidly during the induction’period and then slowly

during the volatilization of the UF  to make the
residence time of the off-gas in the a%sorbe-r long.

 

2D, E, Ferguson et al., ANP Quar. Prog. Kep. Dec.

181

 

 

 
ANP PROJECT PROGRESS REPORT

Calcium fluoride at either 200 or 650°C was
about one-third as effective as NaF in removing
volatile ruthenium and niobium fluoride from the
UF(S-F2 gas stream. Tests were made with both
9- and 18-in.-long beds (Tables 10.1, 10.2, and
10.3). '

Good decontamination was obtained with NaF
in either a 9- or an 18-in.-long bed at 650°C when
the NaF /U weight ratio was 3/1 (Table 10.1, runs
1, 2, 3) or 6/1 (Table 10.2, runs 5 and 7). When
the 9-in.-long bed was re-used (run 4, Table 10.1),
so that the over-all NaF /U weight ratio was 1.5/1
for the two runs, decontamination factors for gross
beta, gross gamma, and ruthenium beta decreased
sharply. With the 18-in.-long bed the same effect
was observed, although here the over-all NaF/U
weight ratio for the two runs was 3/1 (Table 10.2,
runs 6 and 8).

The temperature of the NaF absorbent bed was

difficult to maintain at 650°C. The temperature
profile over the 18-in.-long bed varied 90°C. The
data in Table 10.2 were obtained with the hottest
point at 670°C; this temperature may be a little
high for efficient operation, but operation at a
[ower temperature would probably result in too
much uranium retention in the cooler section,

CORROSION STUDIES

In 20 laboratory-scale fluorination runs at 650°C,
corrosion of nickel test coupons and of the nickel
reaction vessel was fairly low. Since conditions

changed continually during the runs and since the

various components of the vessel were attacked
to different degrees, a calculated over-all cor-
rosion rate would have no significance, However,
it appears that a large number of ARE fuel fluori-
nation runs can be made in one reaction vessel
before corrosion interferes with the process.

 

TABLE 10.1. DECONTAMINATION IN A 9-in.-LONG NaF ABSORBENT BED

UF6-F2 gas stream from fluorination of ARE-type fuel at 600 to 650°C passed through 1-in.-dia bed with
temperature of 650°C in hoflesf_portion; same bed used in runs 3 and 4; F"2 flow rate about 300 ml/min
initially, then aobout 150 ml/min during remainder of run

F2/U mole ratio: run 1 3.7 run 3 8.2

run 2 3.6 run 4 4.8
NaF/U weight ratio in absorber: 3/1 for runs 1, 2, and 3; over-all ratio for runs 3 and 4 = 1.5/1

 

Decontamination Factors

 

 

 

Activity

Run 1 Run 2 Run 3 Run 4

Over-all
Gross 3 1.3 x 104 5800 1.0 x 10* 2900
Gross y 3.2 x 104 2.1 x 104 2.4 x 104 4500
Ry y 1700 1600 1.0 x 104 200
Zr-Nb y 3.2 x 105 7.4 x 104 7.0 x 10% L0 x 10°
TRE* 5.2 x 105 5.0 x 104 4.9 x 104 1.1 x 107
Across Absorbent**

Gross 3 340 200 ' 35
Gross y 930 4100 220
Ruy 940 1400 62
ZeNb y 900 1.2 x 104 2000
TRE B 8 17 8

 

*Total rare earths.

**Caleulated on basis of activity found in absorbent and final product.

182

 

 
»”

)

w

PERIOD ENDING JUNE 10, 19535

TABLE 10.2. DECONTAMINATION IN AN 18-in.-LONG NaF ABSORBENT BED

UF F gas stream from fluorination of ARE-type fuel at 600°C passed through 1-in.-dia bed with
hoflest point at 670°C; same bed used in runs 5 and 6 and in runs 7 cnd 8; F flow rate cbou'r
200 mi/min
2/U mole ratio: run 5 8.1 run 7 8.4
run 6 9.9 run 8 10.2
NaF /U weight ratio in absorber: 6/1 for runs 5 and 7; over-all ratio for runs 5 and 6 and 6 and 7 = 3N

 

Decontamination Factors

 

 

Activity

: Run § ’ Run 6 Run 7 Run 8

Over-all
Gross 3 3900 1600 4300 2400 ;
Gross y 9700 2700 2.0 x 104 4000 5
Ru y 1500 140 6700 250
ZeNby 3.5 x 104 2.9 x 104 9.8 x 10% 5.2 x 10%
TRE 3 3.7 x 104 1.0 x 10°
Across Absorbent*

Gross 3 2 14
Gross y 40 | 82
Ru y 5 31
Zr-Nb y 250 700

 

*Calculated on basis of activity found in absorbent and final product,

 

- _’;‘Q",'Nl, | 0 01% 7C"),‘ was also cut |ong:fudma|ly and

 

The A’ nickel reaction vessel was 2 in. in
diameter, The three test coupons were mounted
in an upright position at the bottom of the reaction
vessel, as shown in Fig. 10.1, in such a way that
one-third the surface area of ‘each coupon extended

from the quUICI mfo fhe gas phase. The coupons'
were 3 in, long, % in. wide, and ¥ in, thick. Two
of ’rhe cdupons were”A"n:ckel?nommal compo-
| 'smon 99.4% Ni, 0, 05% C), and one of them wcs' ‘
“cut longn‘udlnqily and welded The thlrd coupon,

WhIC]’I wa _ mckel (nomlncl composmon 99.4%

 

 
   

to pro‘é'évss condlhons for a 'rotal of 30 hr. The:w
' fluorme flow rcte varled from 50 to 300 ml/mm“
“and was regulated so that 9.4 moles of fluorine

was used per mole of uranium in each run,

7.5 mils in depth.

Corrosion of the welded coupons {both ‘'A’" and
“I."" nickel) was greater than that of unwelded ones,
but in both cases the corrosion was of the solution
type (Fig. 10.2), and there was fairly uniform
surface removal, Dimensional and weight-change

analyses also showed that corrosion was greater

in welded than in unwelded coupons (Table 10.4).
Thé most severe attack was on the outer surface
of the fluorine gas inlet tube in the vapor zone
(Fig. ]0 3). The attack on this tube in the liquid

zone was more uniform and varied from 4.0 to

region in contact with molten salt showed that
the attack there was of a solution nature (Fig.

10.4).

Recovery of uranium was high in all runs (Table

10.5). The uranium loss in the waste salt was

consistently lowest in the 50-min runs at the

183

Dimensional analysis of the
__‘;_,E,;;:',reachqn vessel indicated nonuniform attack of

"5 to 9 mils in both the liquid and gas zones. A
NaF - 'C'metallogrcph:c examination of the vessel in the"
.r.,_foi' an md:vrdual run varled from 4, 58 to 0.83 hr,

- the reaction’ vessel cmd fhe coupons bemg exposed f

 

 

n L T A

 
A

ek sk . st i o

i o o e

- ~ANP PROJECT PROGRESS REPORT

TABLE 10.3. DECONTAMINATION IN A CaF , ABSORBENT BED

U\F6:-F-2 gas stream from fluorination of ARE-type fuel at 600°C passed Through. Ir-rir‘l‘.-diu bed;
__F2 flow rate about 200 ml/min

Run 9: 9_iin7-|ong bed at 650°C, F2/U mele ratio 6.5

‘Run 10: ]8-in'.-long bed at 2000(:, F2/U mole ratio 7.5

Run 11: 18-in.-long bed at 200°C, F2/U mole ratio 8.7

 

Decontamination Factors

 

 

 

 

Activity
o Run 9 Run 10 Run 11
Over-u” |

Gross 3 1900 960 1700

Gross y 1800 1200 3600

Ru y 130 | 80 200
CZrNby 5300 6400 | 47 x 104
" TREB | 2.3 x 10°

Across Absorbent*

Gross B ‘ 4 3 5

Gr'oss y 18 20 65

Ru y 5 3 8

ZrNby 40 110 1000

TRE 3 13

 

*Calculated on basis of activity found in absorbent and final product.

TABLE 10.4. WEIGHT LOSS OF NICKEL CORROSION COUPONS TESTED IN
LABORATORY-SCALE FLUORINATION RUNS

 

 

 

. Original Weight Final Weight Weight Change
Type of Coupon
T (9) (g) (g) (%)
Welded *“L** nickel 83.9878 ~ 80.3760 3.6118 4.3 o
Welded "*A’* nickel 86.3445 82.7515 3.5930 42
Unwelded **A’* nickel 82.6071 80.2160 2.3911 2.9

 

 

 
 

o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

¥

 

Do not lelels

highest fluorine flow rate. This result was pos-
sibly due to a smaller loss of fluorine in corrosion
in the short runs than in the long runs. Out of
3935 g of salt, 341 g of uranium was actually
recovered as UFé' which corresponds to an initial

PERIOD ENDING JUNE 10, 1955

uranium content of 8.66%. Analyses of this par-
ticular batch of fuel indicated a urenium concen-
tration ranging from 8.30 to 8.76%. Even if the
higher value is assumed to be correct, the total
recovery was 99.0%.

TABLE 10.5. URANIUM LOSSES IN LABORATORY-SCALE FLUORINATION RUNS

 

Uranium Loss

 

Duration Fluotine Flow Rate
Number of Runs (hr) (ml/min) in Waste
" (% of total)
1 4.58 55 0.11
4 2.50 100 0.02 to 0.16
5 1.35 200 ' 0.06 to 0.23
9 0.83 - 300 ' 0.01 to 0.04

 

e
" S 7 et

Fig. 10.1. Cross Section of
the fluorine gas inlet tube may be seen at point A.

 

      

Nickel Reaction Vessel, The pitting type of aitack on

UNCLASSIFIED
Y- 14847

    

185

 

 

 

 
 

 

 

 

 

Sl B

b, v Sl ¢

 

 

Lo

o Aelele

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
¥-15140

 

Fig. 10.2. Cross Section of Welded “‘L’* Nickel Test Coupon Exposed to Molten Sult |h a Nickel

Reaction Vessel. Note uniformity of attack. Etched with KCN plus (NH )25208' 12X.

- AS RECEIVED

 

Fig. 10.3. Outer Surface Attack of Fluorine Gas Inlet Tube in Vapor Zone of Reaction Vessel. S'eéfioh

taken at point A of Fig. 10.1. Etched with KCN plus (NH )

186

,Og 20X.

»

 
e

5ot

 

 

 

 

¥

n

  

 

 

 

PERIOD ENDING JUNE 10, 1955

 

 

 

0.004

 

0.002

 

0.003

 

0.004

 

0.C05

 

0.006

 

0.007

 

0.C08

 

0.009

 

0.040

 

0.044

 

INCH

 

 

 

 

Fig. 16940 Inner Surface of Specimen of ‘‘A’" Nickel Reaction Vessel Taken from Region Exposed to
ARE-.Type Fuel. Note nonuniform surface attack. Etched with KCN plus (NH4)25208' 250X.

 

  

187

         

T

T

 

 

ek

 

 

 

 

 
 

  

 
 
 
 

 

 

 

 

 

 

-rern - s n v T T R N T A

 
i G et o i LA

 

*®

 

Part i

SHIELDING RESEARCH

 

 

 

 

 

 

 
 

 

 
Atk

K sk o RS

 

-t

"

g

11. SHIELDING ANALYSIS
E. P, Blizard

F. H. Murray

C. D. Zerby

' ~ R. B, Murray
Applied Nuclear Physics Division

S. Auslender
J. B. Dee

C. A, Goetz

J. Smolen

Pratt & Whitney Aircraft

A calculational method has been developed for
tracing all the gamma radiation born within a
circulating-fuel reactor; for this calculation it is
assumed that the reactor is spherically symmetric
and that all regions are homogeneous. in addition,
the codings of two Monte Carlo problems - the
calculation of the heat generation resulting from
the absorption of gamma radiation in laminated
shields and the calculation of the energy and
angular distribution of air-scattered neutrons from
a monoenergetic source — are nearing completion,
The analysis of the constant-velocity transport
equation has also been extended.

GAMMA-RAY DISTRIBUTION IN A __
CIRCULATING-FUEL REACTOR AND SHIELD

C. D. Zerby M. D. Pearson!

A semianalytical Monte Carlo calculation has
been initiated to determine the history of all
gamma radiation born within a circulating-fuel
reactor of the ART type. The calculation is a
joint effort of the Boeing Airplane Company and
ORNL and will be coded for the IBM 701 automatic
digital cdmpm‘gr. The problems of interest will
be compufed at Boeihg. The calculation w1|l be

' smphhed by ’rhe assumphons thc’r the recctor has
spherlcolly symmetrlc geomefry and " that the =
regions w:fhout homogene:fy (for example, the
- ,heat exchanger) ure homogeneous. However, these
~ will be the only snmphflcahons, and all shells of
lr'TheAcore, refiecfor, and “shield, |nc|ud|ng the
_ Inconel rlconfamer she_lls, w:” be ’raken ' m’ro'
_account,” T T e
: The problem' is deS|gned to de'rermme the hlsfory
of “all gamma radiation born within cmy one
‘spherncol shell of the confugurc:tlon and ’ro account_ '
. for the rodlal disfrlbu’non of the source., The‘:_ -

results w:” mc]ude u "determination of the energy @ —————

 

 

]Boeing Airplane Company.

 

. experimental data,

absorption resulting from the transport of the
gamma radiation. These data will be presented as
a function of radius. The results will also include
determinations of the energy spectrum and angular
distribution of the gamma radiation penetrating
the shield., The data obtained will be normalized
per watt of power generated in the core.

The sources of gamma radiation within the core
will be determined from neutron flux distribution
data obtained from one of the existing multigroup
neutron diffusion calculations. The power of the
reactor will be determined from these same data.

The energy and density normalization of the
total prompt-fission .and fission-product decay
gamma rays born in the active fuel region of the
core will be obtained from the relationship

(1) N(E) = 13,7 e~1:08E |
where
N(E) = total photons/fission/Mev,
E = energy (Mev).

The components of this total are given by the
following equations:

(2) Ny(E) = 7.7 e=1-02F |
@) Ny = 6en12E
where -
N,(E) = prompt photons/fission/Mev,

~ N,(E) = decay photons/fission/Mev.
Equation 2 is an empirical fit to published datq,
while Eq. 3 is a best guess inferred from some
3 Equation 1 is the sum of
Eqs. 2 and 3. Better data will be used as they
become available.

In the heat exchanger region, only the fission-

i

2

.. product decay gamma rays will be considered,

 

2, L. Gamble, Pbys. Semiann. Prog. Rep. Sept. 10,
1953, ORNL-1620, p 15,

3R, W. Peelle, private communication.

191

 

 

 

 
ik opes S A

 

 

- data on heating are obtained in a similar manner.

 

ANP PROJECT PROGRESS REPORT

since the thermal-neufron flux, and thus the

- fissioning, is depressed by the boron curtains.
It will be assumed that these decay gamma rays
- will be uniformly distributed in the heat exchanger.
The Vener'gy and density normalization of this
~radiation will be obtained from Eq. 3. In all

other regions of the core the energy and density

~of the capture gamma rays will be determined from

published data.4
The capture gamma-ray sources in the shield

“will be mcluded when data on the neutron flux in

this region become available. It is hoped that the
numerical integration of the transport equation for
spherical shields will be completed so that the
results will be available to complete this problem.

A neglected source of radiation is that resulting
from inelastic neutron scattering. However, this
radiation can be included without modification of
the code when the information becomes available,
It is intended that each region of the assembly
be treated separately, and the results will be
added to complete the treatment of the reactor as
a whole, By treating each region separately, it
wullrbe ‘possible to make parameter studies to
determine  optimum configurations to shield
against radiation from any patticular region.

For this calculation it was necessary to deviate
from the straightforward Monte Carlo methods
because of the many mean free paths of material
to be penetrated by the core-region gamma rays;
for example, there are approximately 16 mean free
paths from the center of the core to the outside of
the shield for 3-Mev gamma rays in a typical
300-Mw design. Many mean free paths of at-
tenuation for radiation usually result in poor
statistics in the solution when the problem is

~ treated by Monte Carlo methods. It was therefore

necessary to use a semianalytical method, which
is a considerable improvement over the straight-
forward procedure. The procedure is to generate

 the spatial energy and angular collision density

of gamma rays born in any one region by standard

~ Monte Carlo techniques and from this to calculate

analytically the energy spectrum and angular
distribution® of the penetrating radiation. The
' 6
In practice the analytical and Monte Carlo parts of
the problem are carried on simultaneously, each
modifying the other.

 

4P. Mittelman, Gamma Rays Resulting from Thermal

.‘ Neutron Capture, NDA 10-99 (Oct. 6, 1953).

192

ENERGY ABSORPTION RESULTING FROM
GAMMA RADIATION INCIDENT ON A
MULTIREGION SHIELD WITH SLAB GEOMETRY

C. D. Zerby S. Auslender

The coding of a Monte Carlo calculation of heat
generation resulting from the fransport gamma
radiation through shields with ' laminated-slab
geometry is nearing completion,” The data re-
sulting from the calculation will include the
energy absorption as a function of depth, as well
as the energy reflected and penetrating the slab,

A revised code makes it possible to consider
the gamma-ray heating in a circulating-fuel reactor
when it is idealized to slab geometry (see pre-
ceding paper). As coded, 1000 histories can be
calculated by the Oracle with standard Monte Carlo
methods in approximately 5 min, This number of
histories should be sufficient to provide good
statistics for gamma rays incident on a shield
with a thickness of approximately seven mean free
paths, The code should be applicable to other
programs as cases of interest arise,

ENERGY AND ANGULAR DISTRIBUTION OF
AIR-SCATTERED NEUTRONS FROM A
MONOENERGETIC SOURCE

C. D. Zerby

A general description of a Monte Carlo calcu-
lation of the energy and angular distribution of
air-scattered neutrons from a monoenergetic source
was given in a previous report.®. Coding of the
problem for the Oracle is continuing, and slight
modifications in the original problem have been
made for greater utility, The program now makes
possible a determination of the energy and angular

distribution of air-scattered neutrons from a unit

surface source on the surface of a sphere,

 

SFor an example of the type of data to be obtained
see C. D, Zerby, ANP Quar. Prog. Rep. Dec. 10, 1954,
ORNL-1816, p 144 C. D. Zerby, *‘Energy Absorption
Resulting from Incident Gamma Radiation as a Funchon
of Thickness of Materials with Slab Geometry,"’ Reactor
Shielding Information Meeting, Nov. 15-16, 1954, General
Electric Co., Aircraft Nuclear Propulsion Dept., Cincin-
natz, Ohio, WASH-185 (Part I), p 26 (Feb, 1955).

®Eor an exomple of the type of data to be obtained
see C. D. Zerby, ‘*The Penetration of Composite Slabs
by Slanf Incident Gamma Radiation (Monte Carlo So-
luhon), WASH-185, op. cit., p 15.

7c.p. Zerby and S, Auslender, ANP Quar. Prog. Rep.
Mar. 10, 1955, ORNL-1864, p 173,

8¢. p. Zerby, ANP Quar. Prog. Rep. Mar. 10, 1955,
ORNL-1864, p 173,

£l
i, Rk, i idiai A i o ey e e MR i R

 

e

 

P

The surface source can have an angular distri-
bution about the normal given by

P(f) = Acos™ 8,

where

P(6)

probability of emission per unit solid
angle in a direction inclined at an
angle 6 with the normal,

f = angle with respect to the normal,

A = normalizing constant,

n o= 0,1,2,0. .
The sphere on which the source is located will be
treated as a black body to radiation returning to
the sphere.

With these modifications it is possible to calcu-
late the answers for several sources having
different powers of the cosine distribution and
then to combine them to duplicate the distribution
from the surface of a circulating-fuel reactor,
Detailed information about the neutron radiation
reaching the crew compartment shield can thus be
obtained.

A study of the effect of neutrons originally
emitted in the half space to the rear of the aircraft
on the total flux at the crew compartment shield
can also be made with these modifications,
Economical shaping of the reactor shield is
dependent on a thorough knowledge of behavior of
this radiation.

 

9 Murray, ANP Quar. Prog. Rep. Mar. 10, 1955,
ORNL 1864, p 173.

 

 

 

PERIOD ENDING JUNE 10, 1955

ANALYSIS OF THE CONSTANT-VELOCITY
TRANSPORT EQUATION

F. H. Murray

The analysis of the constant-velocity transport
equation previously developed9'm has been
extended.!! The inversion of the Fourier trans-
forms computed in the first paper may be avoided
by the use of eigenfunctions for all the separate
strata or shells of a stratified medium. The
course of the calculations is the same os for a
system of nonhomogeneous differential equations.
To any solution of the nonhomogeneous equation
is added the sum of all eigenfunctions, each
multiplied by a constant factor, and these factors
are determined by the conditions at the boundaries
between different media. In this case the con-
ditions are that the product (n,v)f(S,) be con-
tinvous at each surface and that the flux vanish
at infinity for finite sources. The eigenfunctions
for cylindrical and spherical source distributions
may be calculated from those for plane source
distributions with the aid of matrix representations
of rotations in three-dimensional space. The
solution of the transport equation with finite
sources but for unbounded space represents the
solution of the nonhomogeneous equation,

 

mF. H. Murray, Anisotropic Scattering of Neutrons
in a Uniform Medium with Beam Sources, ORNL CF-
54-11-83 (Dec. 3, 1954).

Ne w Murray, Analysis of the Constant Velocity
Transport Equation with the Aid of Eigenfunctions of
the Various Medza:, ORNL CF-55-5-2 (to be published).

193

 
 

 

 

ANP PROJECT PROGRESS REPORT

12, LID TANK SHIELDING FACILITY

G. T. Chapman

J. M. Miller

Applied Nuclear Physics Division

W. J. McCool

H. C. Woodsum

Pratt & Whitney Aircraft

~ The second series of experiments with mockups

- of 'fl-jé”<:_"'_i_fculbti'r'1§->fu‘e|' reflector-moderated reactor
' ___(RMR‘)‘ @:nd shield have been continued at the Lid
Tank Shielding Facility (LTSF). This series con-
‘sists of two sets of fests: the static source tests
“and the dynamic source tests. The measurements

fcken to date have been concerned w:Th the static
source fests

REFLECTOR MODERATED REACTOR AND
' ' SHIELD MOCKUP TESTS

G T Chcpman J. B. Dee
: H. C Woodsum

Mcmy of ‘rhe measurements in the static source

 

‘tests of the reflector-moderated reactor and shield

mockup experiment at the LTSF have been com-
pleted, As reported previous!y,] these tests were
designed to determine the primary and secondary
sources of radiation that reach the outside of the
shield. The tests completed to date have included
(1) gamma-ray dose rate measurements behind
various configurations of the mockup (the standard
configuration is shown in Fig. 12.1), {2) thermal-
neutron flux distribution measurements within the
beryllium reflector, (3) neutron measurements be-
hind the mockups as a function of the thickness of
the lead gamma-ray shield region, and (4) sodium
activation measurements within the heat exchanger.

Before any mockup tests were performed, a series
of gamma-ray dose rate, fast-neutron dose rate, and
thermal-neutron flux measurements were made both
in plain water and in borated water contained in the
iron tank that holds all the components of the
mockup in the experiment. These measurements
were made in order to determine the effect of the
borated water in which all the mockup measure-
mén"fé' 6re made. The mockup fcnk which is placed
agclnst the source nlate, has a / in.-thick alumi-

num window onthe source side. Rodm’rlon measure-
‘ménts were also made in the plain water of the

LTSF (that is, no mockup tank used).

 

1J .B. Dee, ANP Quar. Prog. Rep. Mar. 10, 1955,

: ORNL 1864, p 176

194

Gamma-Ray Dose Rate Meos-urem’e‘nfs

The gamma-ray dose rate measurements in plain
water and in borated water (Fig. 12.2) were used
to determine the source strength and approximate
energy of the gamma rays from the LTSF source
plate. In order to separate the sources, three
simultaneous equations with three unknowns were
solved:

- +T. .
&

 

D')’(HZO,LT) H Capture
D'y(Hzo,RMRT) =1
+ FH Capture + FAI Capture !
D')’(B-H20,RMRT) =T,
r
H Capture Al Capture
+ + ,
66 10
Sremr
2-0$-057-66-117
3pin, Al N e R \\\\\\\

    

 

 

 

 

 

 

 

~25em AR
Ygin. INCONEL DRy e 4-in. DRY
—- . o el TANK WALL
1 (Fe)
Ya-in Al \f— - rel
TANK WALLS <] fl\\ / {/ ALl -
T TR D
_ ,,‘ N WOODY 5 i1 BoRaL
T i \ T (4 SHEETS)
2in BORAL | WOOD|i720% L HEAT EXCHANGER (2 TANKS] - -
(4 SHEETS) ~—{ | || L | - 1lin. NICKEL

 

 

 

 

4% in. LEAD (3 SLABS)

 

 

 

e — = BORATED WATER
(195wt % BY _

 

 

 

 

Fig. 12.1. Sfandor'd‘ Confijfiraiion- (Top View)
for LTSF RMR.Shield Tests.

 

 

 
e e

 

  

" the mockus of the RMR and shisld was
in the LTSF, and the effects of varying dlfferenf“'
' reglons were studied. Measurements behind various

T weEenE T
2-01—057~-66-~11

2 —Fa PLAIN WATER, RMR TANK
o PLAIN WATER, LID TANK
5  BORATED WATER, RMR TANK

GAMMA-RAY DOSE RATE {mr/hr)

 

10
50 60 7O 80 920 {00 {10 120 {30 140 {50 {60

z, DISTANGE FROM SOURGE PLATE {cm)

' Fig. 12.2. Gdfim&fi&ybofié Rate Measurements

in Plain and Borated Water at the LTSF.

where
'D,), = total gamma-ray dose rate meas-
ured,
" = gamma-ray dose rate from LTSF
- source plate,
H Capture = gamma-ray dose rate from hydrogen
. capture of thermal neutrons,
I, Capture = 9amma-ray dose rate from alumi-

num capture of thermal neutrons,
ratio of the integral of the thermal-
neutron flux in pure ond boro’red
woter,

66

il

 
  
  
       
 
 
 
 

“and bérated woter

:ro’no of the fhermol neufron fluxes'__'_‘ .
the alumlnum wuncfow In pure

lun_form Ieadr thlck” t F
“the curves is due to”rhe 0 5 Mev‘gammo ray re-

PERIOD ENDING JUNE 10, 1955

PAECHET.
2-01-057-66-116
WATER CONTAINED IN TANK WITH M WINDOW

TOTAL GAMMA PLAIN WATER

TOTAL GAMMA DOSE, BORATED WATER
DOSE FROM SOURCE PLATE

GAMMA DOSE FROM HYDROGEN
CAPTURE, PLAIN WATER

GAMMA-RAY DOSE RATE {mr/hr)

 

10
50 60 70 80 90 100 HO 120 130 140 150 160 {v0 180

z, DISTANCE FROM SOURGE PLATE (cm)

Fig. 12.3. Contribution of Various Sources to
Gamma-Ray Dose Rate in Plain or Borated Water

Against LTSF Source Plate.

beryllium and the heat exchanger showed that, for
thicknesses of boral greater than / in., there was
no change in the resulting gamma-ray dose rate
(Fig. 12.4). This simply means that gamma rays
resulting from epithermal neutron capture in the
heat exchanger are not an important contribution
to the total gamma-ray dose rate (see further
discussion on the sodium activation below).
Figure 12.5 shows a decrease in total gamma-ray
dose rate with an increase in heat exchanger
thickness. This decrease was investigated and
was found to be just an attenuation of the primary
gamma-ray dose.

A lead thickness variation study was carried out

“ not only to determine the resulting effect on the

- gamma-ray dose rate but also to learn something
econdcry gamma-ray produchon in and

Results of ’rhls sfudy (Fag 12.6)

   

 
   

sulhng “from fhermal-nelfron capture by the boron

 
   
   
   

estobhshgd\

 

fhlcknesses of the boron curtain between the

 

 

 

'qf the borofed wa’rer |mmedLate|y behmd ‘rhe iecd

|nsfa|1édfw\

   

gamma wrcy n r fl'le back of ‘the tonk Ana!yses of
these curves md:cote ‘to dc're, “that most of this
secondary dose is caused by boron capture gamma
rays, hydrogen capture gamma rays, and thermal

195

 

 

 

 

 

 

T+

R Py
ANP PROJECT PROGRESS REPORT

SEL R
-66-112

2™ e CONFIGURATION 2A, 13 in. BORAL
4 CONFIGURATION 2C, Y% in. BORAL -
107 "0 GONFIGURATION 2B, NO BORAL

GAMMA-RAY DOSE RATE {mr/hr)

 

50 60 70 80 90 {00 #HO {20 {30 440 {15C 180
z, DISTANGE FROM SOURGE PLATE {cm)

" Fig. 12.4. Gammo-Ray Dose Rate Beyond RMR-
* Shield Mockups: Effect of Boral Thickness.

 

PECRET
2-01-057-66-110

STANDARD CONFIGURATION
(4-in. HEAT EXCHANGER)

v

n

STANDARD CONFIGURATION
PLUS 2-in, HEAT EXCHANGER

GAMMA-RAY DOSE RATE (mr/hr)
3

o

 

10—2 o
50 60 7O 80 90 00 O {20 130 40 50 160
: z, DISTANCE FROM SOURCE PLATE (em)

Fig. 12.5, 'lGhr-fimui-'R;:y Dose Rate Beyond RMR-
Shield Mockups: Effect of Heat Exchanger Thick-

ness.

 

and epithermal neutron captures in the lead. For

large lfead and water thicknesses (for example,
.9 in. of lead and 80 cm of water), the dose rate
_ from epithermal neutron capture in lead seems to
"f:‘[S:"_%:"dominate'.' More experiments are planned for
",inyes'f'igating these results. The placement of the
"l_éd'd will also be investigated in a later test to

196

 

o
[

o
N N

N . OO,

GAMMA-RAY DOSE RATE {mr/hr)

10

 

5
2 7Y, in.
~2

0
50 60 7C 80 90 100 #O {20 430 140 150 160 180
2, DISTANGE FROM SOURGE PLATE {cm}

Fig. 12.6. Gamma-Ray Dose Rate BeyondRMR- |

Shield Mockups: Effect of Lead Thickness.

| SEORr™
2-01-057-66-113

STANDARD CONFIGURATION
{tin. Ni).

STANDARD CONFIGURATION PLUS 1 in. Ni

GAMMA-RAY DOSE RATE {mr/hr)

 

102
50 60 7O 80 90 100 110 120 {30 {40 150 160
z, DISTANGE FROM SOURGE PLATE (cm)

Fig. 12.7. Gamma-Ray Dose Rate Beyond RMR-
Shield Mockups: Effect of Pressure Shell Thick-

ness.

enable shield optimization and to further study
the secondary gamma-ray dose in and beyond the

fead.

A study of the effect on the total gamma-ray
dose rate of adding 1 in. of nickel to the pressure
shell of the standard configuration indicated that
the extra thickness of nickel merely attenuated

the primary gamma-ray dose (Fig. 12.7). A study

-

 
e Motk s e ik

 

."f_,é:‘ .

‘"'d'-were"maéi‘e be;fh in bora’recf wqfer and in plcun waer

of the effect of adding an Inconel cladding on the
boron curtain at the rear of the beryllium (Fig. 12.8)
showed a negligible increase in the gamma-ray
dose rate with 0.020 in. of Inconel and an increase
of approximately 10% with 0.125 in. of Inconel.

SEOMERee
2-01-057-66-45

e CONFIGURATION 2A, STANDARD GONFIGURATION
(NG INGONEL )
o GONFIGURATION 1G, STANDARD CONFIGURATION
+ 0.020in. INGONEL
2 4 GONFIGURATION 1H, STANDARD CONFIGURATION
+0.125 in. INCONEL

GAMMA-RAY DOSE RATE {mir/hr)

 

~2
10 :

50 60 70 8> SO {00 {0 120 {30 {40 {50 ({60
2, DISTANGE FROM SOURGCE PLATE (cm)

Fig. 12.8. Gomma-Ray Dose Rate Beyond RMR-
Shield Mockups: Effect of Inconel Cladding on
Boral.

Additional gamma-ray dose rate studies in this
static source series will include variations of the
thickness of the beryllium reflector region and the
addition of heavy metals in the reflector region.
An investigation of the gamma-ray captures in the

beryllium reflector and Inconel core shell is

   
 

reported below.

 

Neutron Measuremenfs

 

“As was mentioned V_qbove,r neu’rron e UFSIaATS

  

ockup tank A comparason

was nof cffecfed‘;fo nof|'cecxb|e degree. The

dnsplacemem‘ of ’rhe L-zdvTank plcnn water CUrVBs
. _cnd the mockup tunk”

In order to determme the distribution of the
thermal-neutron flux in the beryllium of the mock-

 

   
    
    
   

 

PERIOD ENDING JUNE 10, 1955

SEemET
2-01-057-66-18

 

 

 

 

BORATED WATER IN RMR TANK

|
| ——PLAIN WATER IN LID TANK

|

/PLAIN WATER IN RMR TANK
N\
40 N % 10

PLAIN OR BORATED . \
WATER IN RMR TANK— \" \:\
1 i N A 1

PLAIN WATER IN LID TANK .
10 < ot

 

5. 5,
./.
N //f
L~
/j

 

2
4

 

THERMAL-NEUTRON FLUX {nv)
o
W
% g
Tt /\. /
"
3
o

 

 

 

 

 

 

FAST-NEUTRON DOSE RATE (mrep/hr)

 

 

 

 

 

 

 

 

1G°2 12
0 20 40 60 80 100 120 140
z, DISTANGE FROM SOURGE (cm)

 

Fig. 12.9. Neutron Measurements in Plain and

Borated Water (1.95 wt % B) ot LTSF.

ups, gold foils were exposed throughout the re-
flector region. The foils were placed on the faces
of each beryllium slab from 5 in. above the source
center line in intervals of 4 in. to the top of the
beryllium. Plots of the results as a function of
the beryllium thickness (Fig. 12.10) are uniform
and peak just after the first face of the second

4 __Lberylhum slab (that is, at about 4 / in. of beryl-
“#4% [jum),  These data have not yet been corrected
;- for self-absorption and flux depression by the foils.

Plots of the gold foil measurements as a function

...of the vertical distance above the source center
... line (Fig. 12.11) again are uniform, with a relatively
.. flat region across the source center line. As would
. be expected, the flux falls off rapidly as the upper

- edge of the beryllium is approached.

Neutron measurements (Fig. 12.12) were also

'spoce between the source <+ made behind the mockups as the lead region of

.~ the shield was increased from 0 to 9 ]/2 in. to study
«the contribution to the gamma-ray dose rate by

secondary gamma rays from the lead. There were
a great many low-energy neutrons in the region

197

 

 

 

 

 
 

 

 

 

 

 ANP PROJECT PROGRESS REPORT

DEOREF )
2-01-057-66-119

108

THERMAL-NEUTRON FLUX {nv)

<%

STANDARD GONFIGURATION

5 x = DISTANCE ABOVE SOURCE CENTERLINE

 

0P
0 2 4 6 8 0 12 14
#, BERYLLIUM THICKNESS (in}

Fig. 12.10. Thermal-Neutron Flux Distribution
in the Beryllium of the RMR-Shield Mockups:

Horizontal Traverses.

Seenss
8 2-01-057-66-120

3
2,

o,

THERMAL-NEUTRON FLUX (nv)
N

OO)

CONFIGURATION
#=BERYLLIUM THICKNESS

 

10° :
5 9 13 17 24 25 29 33

#, DISTANCE ABQVE SOURCE CENTERLINE ({in.}

F.‘ig‘./"'IZU.‘“. The;mal-Nefitron Flux Distribution

in ihe Bery|li--um of the RMR-Shield Mockups:
" Vertical Traverses.

 

- T
4 2-01-057-66-122

4lin. Po

&in. Pb
1Y in Pb

T¥2in Pb

3in. Pb

THERMAL-NEUTRON FLUX {nv)

in.Pb
4Y%in. Pb
3in. Pb

FAST-NEUTRON DOSE RATE (mrep/hr}

 

e0 80 10C 120 140
DISTANCE FROM SOURCE {cm)

Fig. 12.12. Neutron Measurements Beyond RMR-
Shield Mockups: Effect of Lead Thickness.

immediately behind the lead, and the increase in
fead thickness, in effect, only moved the thermal
and intermediate flux out. The slopes of the fast-
neutron dose rate curves (A = 4.7 cm) indicate that
these neutrons were predominantly of intermediate
energies.

Sodium Activation in Heat Exchanger Region

A study of the activation by core neutrons of the
sodium in the heat exchanger region was made as
a function of the boral curtain thickness and the
heat exchanger thickness. In each of the six con-
figurations tested (Fig. 12.13), 12 in. of beryllium
preceded the first boral slab. The last boral slab
behind the heat exchanger region was followed by
1 in. of nickel, 4 l/2 in. of lead, and borated water,
The measurements were made by placing small
pellets of sodium flyoride (p = 2.1 g/cm®) in the
center of each heat exchanger tank at the source
center line. The plots in Fig. 12.13 show the
measured specific activity of the pellets corrected
to saturation for each case. For the first three

cases the effect of decreasing the thickness of

"w
P

 

i

i

 

-

o oL

 

F|g. ]2 14. RMR-Shield Mockup Experiment:

PERIOD ENDING JUNE 10, 1955

SEGRET-
2-01-057-66-121

 

 

 

 

 

 

     
  

     
 
     

    
    
     
 

 

   
  

     
  

 
 

R

INRUANNRNRLRN
NN

N

R~

N

DEENNRNN
A

D
T
B N A RN

N

10° -
— 7 =
- ; . —
5 - ’ | D N
- - 5-in-THICK %-in-THICK {%-in-THICK -
~ 2 el
= I~ BORAL SLAB Ni WALL NaF SLAB ]
E
% (F=1.6g/cmd)
s 2 N
B
o
@ 4 \ -
= 10 [~ \ e, .. _
% — ’ N ™~
- E | \ : \ =
— \ _
% 5 . \ \ ]
o — Y Y \ . |
9 \\ \\ \ A
N -
[t N, \
o ® \
e N\
= 2 \
<[ N\
x .\
o
\
g 10° v9%% 552% P ) 7 %4
$7%% 2972 7% %25 L 7
o 7h%7 %7 2 gies Vo v
= 7% Vo a7 74 o 7
E 5 2277 i 747 ) 0
a v w7z 7 b 477 77
= v 9957 7 74%% 7257 V
= 7277 9247 7% B4 9% 7%
= v 9972 27 7254 P 2%
o ] = . 0 v 4 5
=4 7254 %77 %% 5227 o Z
v gl ,
7z %4%% 7 2979
%7 o 1 e
7 L % 2742
7% 9772 4 w42
7! [ ] Lo
% [ 4 Loy
7 o Z 7742
4 ) A 4%

 

 

 

 

 

     

 

 

 

 

 

 

  

e A NN,
B R AN

NN AR AN NN
R N
&\\\\\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\
‘\\\\\\\\\\\k\%k\

  

ANNTINNNR N
S

   

BN

     

“

N

 

 

 

 

STANDARD CONFIGURATION

Flg. 12.13. Effect of Heat Exchanger and Boron Curfam Thicknesses on Sochum Activation in RMR-
Shield Mockups.

the firs*_r boral curtain is apparent. The results of
BORAL THICKNESS (in) the last three cases show the effect of a 6-in.-thick
© \ 2 3 4 5 heat exchanger with various thicknesses of the

Sowrr
2-01-057-66-1234

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4-end boral curtai in.

A sample calculahon of 'rhe sodlum activation in
the NaK in the heat exchanger of a 300-Mw alrplane
reactor gave 2000 curies. |f the NaK-to-alr radiator

were unshielded, this would give a dose rate of
16 rem/hr at a distance of 60 ft,

10% i | | [ |
,,iaooo | k - first and second boral curtain.
g
5 1-2-4’,/1— i . . . p.
= i-3-1q B £ oo ST e For each configuration an average specific
£ ] 23 | N . , . .
< TR activation was determined by integrating numeri-
£ i—2-2 e . :
£ N85 | cally; the results are shown in Fig. 12.14. Each
2 \ ©g . . . .
3 \\ & & 2000 soea) | point on the curves is labeled with a code in
5 o 7 . . .
3 \\ "% which the first and last numbers designate the
Q A ] g
3 N % 2 v e s number of boral sheets preceding and following
8 N, HEAT EXCHANGER THICKNESS (in.)
N 1 the heat exchanger, respectively, and the center
I
s 2-2-4% N_ 9 Y
% N number des:gnates the number of heat exchanger
« ~,
& Y\ tanks in the conf:guraflon._ The mserf of the figure
= . ~
o ' o _ is a cross plot showing an mcrease in the average
o
£ RGN specific activation with increasing heat exchanger
| e p
o 2 . thickness for two dlfferent ’rhlcknesses of the first
&
&
v
=
<L
o
[
2

 
  

 

 

 

 

 

 

 

 

 

{00~ 200 300 400" 500 < 600 700"
“ U FIRST SORON CURTAIN DENSITY (mg of 8'%/em?)

 
       
        

Sodium Activation Caused by Neutrons from LTSF
Source.

199

 

 

TR

 

 

 

 

TS
™

 
ANP PROJECT PROGRESS REPORT

13. BULK SHIELDING FACILITY
F. C. Maienschein

T. V. Blosser K. M., Henry
G. M. Estabrook E. B. Johnson
J. D. Flynn T.A.Love
M. P. Haydon! R. W, Peelle

W. Zobel
Applied Nuclear Physics Division

K. M. Johnson
Pratt & Whitney Aircraft

 

 

The duct portion of the GE-ANP reactor shield
was tested in the ORNL Bulk Shielding Facility
(BSF) during the past quarter, and arrangements
are being made for three experiments to measure
the spectrum of fission-product gamma rays for a
circulc’ring-fuei reflécfor-moder'ated reactor.

GE-ANP AIR-DUCT MOCKUP EXPERIMENT

F. C. Maienschein J.D. Flynn
J. S. Arterburn? E. B. Johnson

A mockup of the air-duct system for the GE-ANP
reactor was tested at the BSF. The mockup con-
sisted of a pair of annular ducts placed against
the BSR, which was modified with GE-type tran-
sition sections,

One objective of the experiment was to deter-
mine the optimum spacing, if such existed, between
the reactor face and the inlet air duct, Another
objective was to obtain information about the
angular distribution of radiation, particularly fast
neutrons, leaving the reactor shield. The latter
data were needed in order to calculate the scattered
radiation dose arriving at the crew compartment.

The extensive data obtained from the measure-
ments, which included more than 80 traverses, have
-not been completely analyzed by ORNL, and it is
contemplated that the remaining work will be done
by the GE group at Lockland. A description of
" the experiment and some tentative conclusions are
presented here.

- The fuel loading of the BSR was arranged to
correspond as closely as possible to that of the
'GE-ANP aircraft reactor. This necessitated the

. use of a long, narrow reactor (Fig. 13.1). Trane
sition sections were mounted on either side of the

reactor (Fig. 13.2) to simulate the two transition

~ sections of the GE-ANP aircraft engine. The two

 

1 Part time,

2General Electric Company.

200

duct mockups, labeled “‘inlet’’ and “‘outlet,’’ are

shown outside the transition sections. The ducts
were mounted on a large plate (Fig. 13.3), and the
entire assembly was arranged as a“*package shield
mockup,”’ the package being lowered as a unit into
the reactor pool. This method was employed be-

“cause the BSF reactor pool is now filled with de-

mineralized water, and refilling the pool after
draining to assemble equipment in it would be very
time-consuming. All materials of construction were
corrosion resistant, being either aluminum or stain-
less steel. The duct mockups were essentially
two-dimensional analogs of the actual three-
dimensional annular ducts. In order to determine
the optimum position of the inlet air duct, the

UNCLASSIFIED
ORNL-(LR-DWG 5898

BeO REFLECTOR REACTOR GRID PLATE

STANDARD FUEL SAFETY RODS T

ASSEMBLY %
u :1 SNl i:))i )

DD
lfim W EDIEIIED
EDEE) DD DD D IED

e @ k) © @ )
CEEEEEEII)

% FUEL ASSEMBLY

 

 

  

 

 

 

3 FUEL ASSEMBLY

REGULATING ROD

Fig. 13.1. Bulk Shielding Reactor Loading 38.
 

PERIOD ENDING JUNE 10, 1955

SEeRER
ORNL-LR-DWG 6175

  
 
  
 

INLET AIR DUCT
{DUCT NO.1)

R -~ T PES Je - v

 

DUCGT NO. 5

 

 

.
i b
i
{l“ila
i
+
i
|
fll :J

TRANSITION SECTION
~(DUCT NO.3) “l i
 BSF REACTOR- | lF;H

o ' . : | QUTLET AIR DUGT
- | TRANSITION SECTION / (DUGT NO.2)

et
L4

(DUCT NO. 4)

T T

Fig. 13.2. GE-ANP Duct Mockup in the BSF.

spacing between reactor and duct was varied by  to drive the instrument carriage laterally across
‘means of an electric motor mounted in a box on the  the BSF pool. Measurements by the detectors were
stainless steel base plate (shown in lower right  automatically plotted as a function of position on a
“corner of Fig. 13.3). o “two-dimensional recorder.  This mechanization
| | used in this experi- permlfieétheqccumuiaflon of much more data than

“The instrumentatior that vias used in this experic permit ulation of much more.
consisted conventional gamma-ray and  couldhave beenobtained by hand in the two months

    
 

     
 
 

   
   

s, however,

d not operate in the large
he ducts. Col h

“data were taken with an essentially

 

  
    
    

 

corresponding flux data for reactors of known power. | A

   
 

      
 

   

rument, the Hornyak button, which was ~ Since the data of most importance from this experi-
SR I S L ] s e o S ang e S AR e L T SN fo e S i T e g s =

     

    

 

nsidered

only relative, | s
-consuming measure-

worth while to make the time-consuming me
ments necessary for mapping the flux over the

1ld be required f

      
 
 
   
 
  

  

ely that this”

dosimeter, in general, but all

    
   

i

            

 

 

d s ig,_“_,‘- & RN R e e e R P S e R ST L
wer determination. The neutron leakage fluxes .
- t of the two sides and the front of the reactor ‘

 

, | 7 e investigated by means of the wusual foil
o Ancther innovation in instrumentation for this  exposures. Eg
experiment was the use of a variable-speed motor Of the many traverses obtained, two are included

201

 

 

  
 

 

gt

 

 

 

ANP PROJECT PROGRESS REPORT

 

—iE

SEGRER
~ PHOTO 13848

Fig. 13.3, GE-ANP Duct Mockup Assembly.

in this summary in order to provide an illustration
of the data. The first traverse (Fig, 13.4) shows
what happens to the leakage out the end of the large
inlet duct as its position is varied, The detectors
were actually placed on the end of the duct. This
plot shows the counting rates for each of the three
detectors as a function of separation distance be-
tween the inlet and outlet ducts., Since there is
simply a smooth decrease in counting rate as the
separation is increased, there is no evidence of
streaming down the entire length of the duct.
Figure 13.5 shows traverses lengthwise behind
the large air duct as a function of the reactor-duct
spacing. The fast-neutron traverse was made, in
each case, at the same location; only the inters
vening inlet air duct was moved. For the spacings
shown, there is little difference in the fast-neutron
dose rate along the traverse behind the duct. It
appears that as far as fast neutrons are concerned
the duct simply reduces the density of the shield,

202

JDmem
2-01-058-38-2

108

TRAVERSE ¢ (COUNTER RIDES ON END OF

MOVING DUCT){DUCT NO.5).
= RELATIVE DISTANCE DUCT NO.2 IS OFFSET
FROM TRANSITION SECTION.

103

2

 

20 6 12 8 49 0
DISTANCE BETWEEN INLET AND QUTLET DUCTS (in.)

RELATIVE FAST-NEUTRON DOSE RATE {cpm}, HORNYAK BUTTON
RELATIVE THERMAL NEUTRON FLUX (cpm), FISSION CHAMBER

Fig. 13.4. GE-ANP Air-Duct Mockup Experiment;
Radiation Measurements Along Traverse c.

REACTOR POWER = 1 kw {NOMINAL) 150

FISSION 100

50

GAMMA-RAY DOSE RATE {r/hr), ION CHAMBER

>

 
 

"

the best availabl
Slot of thaf
‘f Tom

 

 

Lo
2-01-058-38—1 Rt

x=20"gin.

5 x =15 '/8 in.

x=10 Y% in.

x=2% in.

02

RELATIVE FAST -NEUTRON DOSE RATE (cpm)

HORNYAK BUTTON COVERED WITH Y in. LEAD.

5 L4 TRAVERSE ¢ (PARALLEL TO AX!S OF PLANE).
¥ =DISTANGE BETWEEN INLET AND QUTLET
AIR DUCTS
REACTOR AND DUCTS 2, 3 AND 4 IN FIXED
> POSITION

REACTOR POWER 4 kw (NOMINAL)

 

100 50 0 50 00 150
" " DISTANCE FROM REACTOR CENTERLINE (cm)

F:g 135 GE-ANPA"-Duct Mockup E xperlmenf

Fast-Neutron Dose Rate Along Traverse 4.

and its position has no great effect on that portion
of the radiation which penetrates the shield.
In order to determine some of the characteristics

of _i{he angular distribution of radiation escaping

  
 
 
 
 
  
    
 
 
    
 

| of ‘a”5-ft-long aluminum fube

U

ast-

 

e

 

“on_the intersection of the reactor center line and

‘the ‘second leg of the large inlet duct. The curve

 

shows that the escaping radiation peaks quite
heavily in the forward direction. There is no evi-

  

 

   
  
     
 

"PERIOD ENDING JUNE 10, 1955

Wttt
2-01-058-38-3R1

 
  

 

 
  

o
& 2
& w0
2 Q
o 0
2 b 3
(0925 0 15 20 25 (o)

/sec)

 

 

Fig. 13.6. GE-ANP Air-Duct Mockup Experiment:

Angular Distribution of Fast Neutrons.

dence of fast neutrons escaping through the large
duct, since the curve is symmetrical in the region
around 50 deg. As soon as time is available, fur-
ther investigations will be made of the efficiency
of this type of fast-neutron collimation and of angu-
lar distributions for other shields.,

The results of this experiment seem to indicate
that the problem of shielding the air ducts of an
air-cooled aircraft engine may not be so serious as
was once thought. This conclusion applies spe-
cifically to the annular air duct.

THE SPECTRUM OF FISSION-PRODUCT
GAMMA RAYS

_ R. W. Peelle
T. A. Love F. C. Maienschein

One of ‘ 'fh em osfdlfftculf ‘ éi;o blems in the des ign

| of a o-ft-long aluminum tube  “of q divided aircraft shield arises from the lack of
which acted as a collimator when | '

informati gr&\:_gg_wthg__epergy ‘spectrum and angular dis-
.\'rr_iblqt:i‘c}_nrbf'gamxficjlr;jys at the outside of the primary
reactor shield, These data are needed for an air-
‘<eattoring calcslation because of the rapid variation

.. with gamma-ray energy of the Compton scattering
_ process, This variation makes the results of the

air-s ca'r'rerm g calculation extre mely sensitive to
the . g‘g_,_lkg_' of the gamma-ray flux at the surface of
the roadtor shisld, e

_In the case of a circulating-fuel reflector-moder-

“ated reactor a large fraction of the gamma-ray flux
outside the reactor shield originates from fission-
product gamma rays in the region of the primary

203

 

 

 
 

 

'ANP PROJECT PROGRESS REPORT

heat exchanger. The fission-product gamma-ray
emitters are entrained in the fuel mixture and reach
the heat exchanger very shortly after the corre-
sponding fissions have occurred. The energy
spectrum of these short-lived fission-product gamma

~ rays has not been measured, and, since most of the
~ gammasray energy is given off with short periods,

a complete lack of knowledge of this important

“spectrum now exists. All estimates® which have
_ been mode of the spectral shape must be con-
" sideredtobe quite unreliable, and even the average
‘gamma-ray energy release is in some doubt, It is
" felt® thatthis lack of spectral information for decay
- gamma rays ¢could cause an error of a few thousand

pounds in the supposedly optimized shield weight,
Three experiments relative to the fission-product

_gamma-ray spectrum are presently under way. All

three experiments will make use of the multiple.

‘erystal gamma-ray spectrometer, which allows
- 'measurements of continuous gamma-ray spectra to
be analyzed with reasonable certainty.

One of the experiments will be performed at the
LTSF and will make use of the RMR mockup (see
Sec, 12). In this test the circulating-fuel belt will

 

 

i S

3Reporz of the 1953 Summer Shielding Session, ORNL-
1575, p 233.

J. B. Dee, private communication.

be placed near the LTSF source plate, and the
spectrometer will be exposed to gamma radiation
from the belt. Thus the spectrometer will measure
the gamma-ray spectrum that would be encountered
from a heat exchanger operating with the same type
of circulating-fuel parameters. These results should
shed some light upon the accuracy of the more
general experiments described below.

The other two experiments will be based upon
the brief bombardment of small samples of enriched
U238 jn the interior of the ORNL Graphite Reactor.
These samples will be handled with the aid of the
fast pneumatic probe assembly available there.
One experiment will involve the measurement of
the disintegration rate of the sample for only a few
energy groups. In this case many points in time
will be studied for each energy group. The other
experiment will be similar except that many energy
groups will be studied for only a few different times
after bombardment. 1t should be possible to analyze
the results of these latter experiments in such a
way that the observed spectrum from the heat ex-
changer of an aircraft reactor could be derived for
any set of circulation parameters, Times after
fissionas short as a few tenths of a second will be
studied in the above experiments. Preliminary re-
sults should be available during the next quarter.

 

 
 

AL Al e Rl

o i

(R

il

 

AR s o B i MR i, B s o i B . 3

 

" PERIOD ENDING JUNE 10, 1953

o L | 14. TOWER SHIELDING FACILITY
e - - C. E. Clifford
. ’ Jc Lo HU“ Fo Jo MUCkenfhaier
¥ | ' - ~ F. N. Watson
Applied Nuclear Physics Division
M. F. Valerino, NACA, Cleveland

Js E. Van Hoomissen, Boeing Airplane Co.

A major porhon of the expenmental results to study the penetration to be expected in the side
necessary for the optimization of the neutron of a crew shield as related to the angle of neutron
shielding in a divided shield has been obtained at emission from the reactor tank (Figs. 14.2, 14.3,

- the Tower Shielding Facility (TSF). These results and 14.4). As a cross check, the fast-neutron dose

concern the fast-neutron dose rate distribution in rates were measured with the detector held station-
the detector tank and in a crew shield mockup as a ary successively at the right side, left side, and
function of a variation in the reactor shield geometry  bottom of the tank and with the angle of radiation
and are identified as Phase | of the differential emission, 8, being changed in steps from 360 to

experiment. A technique has been devised for pre- 180 deg (Fig. 14.5).
dicting the dose in an arbitrary shield design from Studies were also made of the penetration into
the differential experiment information as the first the front and rear of the detector tank for various
step in obtaining a procedure for optimizing a values of & (Figs. 14.6 and 14,7). Figure 14.8
shield design. The technique has been applied to shows a cross check measured directly as «

U previous mockup experiments, and the results agree  function of 6,

* to within experimental error. Measurements were made of the direct beam
' ' , G reaching the rear of the detector tank for three
t THE DIFFERENTIAL SHIELDING values of reactor shield thickness, p. When these
ot EXPERIMENTS AT THE TSF: PHASE | ' measurements were plotted as functions of the
F. N. Watson total water thickness, that is, y plus p, it was

" found that the flux was a unique function of this
sum. This indicates that the boration of the water
made no difference in the dose for this configuration

(Fig. 14.9).

The optimization -of the neutron shielding
necessary in a nuclear-powered aircraft depends
primarily upon the variation of the fast-neutron
dose rate in the crew compartment with changes in
such variables as reactor shield thickness, ongle

_of radiation emission, and thickness of crew com-
~ partrent” shleldmg. ’ Mequremenfs of the fast- The fast-neutron dose rate distribution within the

" ‘neutron dose rates’ m$|de the detector tan andi crew compartment (Fig. 14.10) of the GE-ANP

T AN Craw eopsEir ot “** divided shield was also determined,  The crew
“"“‘"compartmenf shielding consisted primarily of water
“‘containing apprommately 0.65 wt % boron and lead,
»The dose rate measuremenfs along the y axis
(fo}'e and aff) of ’rhe crew compartment (Fig. 14.11)
show that, for 6 =0 deg, neutrons entering the rear

Measurements in the GE-ANP Crew Compartment

   
 
 
  
 
  
 
 
 
  
   

.7}80 deg, the conmbuhon through the rear. face was

o "qune small :

. Meosurements from fhe left ’ro the rlghi of the
i Afl.hcomplete s : : | “erew compartment (Fig. 14. 12) md:cate that the

the TSF geometry conventlon is given by C. E. Clifford dose rate is symmetrical for 6 = 0 = 360 deg. The

 

 

 

 
 

. . et al,, Preliminary Study of Fast-Neutron Ground and Air
Scattering at the Tower Shielding Facility, ORNL CF- curve for § = 270 c‘:leg Show_s the .e'Hed of most of
54-8-95 (Aug. 23, 1954). the neutrons entering the right side. Because of

205

   

.,,_.'fcce of the crew sh:eld made a s:zabie contribution

""fo ’rhe to'rol dose ms:de 'rhe shueld For 6 270 and’

T

- [

T P

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

 

REACTOR TANK

 

" UNCLASSIFIED
ORNL-LR-DWG 843

  
 

DETECTOR TANK

Fig. 14.1. Experimental Arrangement for TSF Diffelrential Shielding Experiments.

a low counting rate, only one point was obtained
for 6 = 180 deg.

The dose rates at the center of the crew com-
partment were measured as a function of 6 for two

thicknesses (p) of water shielding the reactor. The

shapes of the two curves, p = 17 and 45 cm, are
roug hly the same (Fig. 14.13),

. The dose rate at the center of the crew com-
partment as a function of p was also investigated

for various values of @ {Fig. 14.14). By observing
that the relaxation lengths of the curves are slowly
varying, it can be seen that either the angular dis-
tribution of the flux from the reactor tank or its

VSp'e'c‘:frum is changing.

206

ANALYSIS OF THE DIFFERENTIAL
SHIELDING EXPERIMENTS

M. F. Yalerino

An idealized configuration of an aircraft divided
shield is illustrated in Fig. 14,15. The fast-neutron
dose rate in the crew compartment can be considered
to be a weighted integration of the dose rate radiated
into the air from the primary reactor shield surface.
The weighting factors for this integration must
account for the probability of the air scattering of
the neutrons into the various sides of the crew
shield and for the probability of neutron penetration
through the crew shield into the crew compartment.

 

  
 

s B i, A . s A, S

 

v

uabacai

 

)

 

 

srener
2-01-056~-7-25-123A

10-3
5 CENTER OF TRIPLET NEUTRON DOSIMETER AT:
X VARIABLE
- CONTRIBUTION SUBTRACTED OUT
2 0.0¢m
~4
10 g = 360 deg
5
2 195 #
64 ft
1073 45 ¢m

VARIOUS

10-6

FAST-NEUTRON DCSE RATE (mrep/hr/watt)

 

10-7
=180 deg
5
§=270 deg
2
1078
80 68 56 44 32 20

£, HORIZONTAL DISTANCE FROM & AXIS TO DETECTOR CENTER (¢m)

Fig. 14.2, Fast-Neuviron Dose Rates Near the
Left Side of the Detector Tank for Various Angles
of 0; x Variable.

| eGaney
2-04-056—7=-3t-116A

   

A 195 £t
0 641t
p 45¢em
é

 
   
    
 

* FAST~NEUTRON DOSE RATE (mrep /hr/watt}
5

- CENTER OF TRIPLET NEUTRON DOSIMETER AT:

200 cm ) )
» CONTRIBUTION SUBTRACTED OUT
7 VARIABLE

 
   
     
   
 

 

-es T =56 —'»44 -3
A VERTICAL. D|STANC 'R’OM'&A‘x]s

     
 

Frg. 14.4. Fast-Neutron Dose Rutes Neur the
Bottom of the Detector Tank for Various Angles
of §; z Variable.

 

’F/ Of 64 ffc

DETECTORCENTER(C R
- e factors in more precise terms to show how they can

 

PERIOD ENDING JUNE 10. 1955

SECTREF
2-01-056-7-10-~1224

 

10
5 b
CENTER OF TRIPLET NEUTRON DOSIMETER AT: -
X VARIABLE
y CONTRIBUTION SUBTRACTED OUT .
Z 0.0 cm b
10 ?
£t s
3 9=330 deg
= 5 195 ft
= 2 g 64 it
[+
€ p 45¢m
o 1g 3 8 VARIOUS
5 a —-90 deg
o
N
3 8=300 d
a - €9
= 2
g
i =360 d
5 106 8 0 deg
w
=
L 5
v
&
2
167 *
5
&=270 deg
2
.-
-80 -68 -56 ~-44 -32 -20

¥, HORIZONTAL DISTANCE FROM & AXISTO DETECTOR CENTER (em)

Fig. 14.3. Fast-Neutron Dose Rates Near the
Right Side of the Detector Tank for Various
Angles of 6; x Variable.

The primary purpose of the fast-neutron differential
shielding experiments (reported above) was to ob-
tain these weighting factors as a function of the
angle 4 which a fast-neutron beam from the reactor
shield surface makes with respect to the axis
joining the reactor and crew compartment. in the
experiment the fast-neutrons from the reactor were
filtered through 45 cm of water prior to beaming
into the air. Some data were also obtained for
neutrons filtered through only 15 cm of water; for
this case, the energy spectrum and the angular
distribution? are somewhat different than those for
the 45-cm case, and therefore comparison of the
results for these two cases provides an indication
of the effect of a change in neutron _energy spectrum
on the integration weughfmg factors. All the data
were obtained at a constant separahon distance,

"1t is necessary to define the mtegrahon weighting

 

2Thea change in the angular distribution for the two
cases was small and was corrected for in the interpre-
tation of the data,

207

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

H 4195 f+
d 641t

p 45c¢cm
8 VARIABLE
a APPROPRIATE

348 336 324 31z 300 288

 

276

erty

2-04-056-7-43+30+37-443A

RIGHT SIDE
X=-68cm

BOTTOM
Z=-68cm

LEFT SIDE
X= + 68 c¢m

CENTER OF TRIPLET NEUTRON DOSIMETER AT:
x 0.0cm, OR AS SHOWN

» CONTRIBUTION SUBTRACTED OUT

7 0.0 cm, OR AS SHOWN

264 252 240 228 216 204 492 180

8, HORIZONTAL ANGLE BETWEEN p AND & AXES (deq)

Fig. 14.5. Fast-Neutron Dose Rates Near fhe Right Side, Left Slde and Boflom of fhe Detecior Tonk-

6 Yariable.

be obtained from the differential experiments, For
this purpose, the reactor shield surface source can
be replaced by a point source because the reactor
shield dimensions are small compared with the
separation distance. The strength of the equivalent
point source per unit solid angle in a given direction
must then equal the integral over the reactor shield
surface of the surface source strength per unit
area per unit solid angle in the same direction,

Definition of Dose Scattering Probability

The probability of the dose from a line beam of
fast neutrons scattering into the sides of a cylinder
is defined in Fig. 14,16, The radiation intensity
in direction {a,¢,) is conveniently given by the
dose rate on the surface of a unit sphere at co-
ordinates (a,qS ). This dose is denoted by D _(a,¢5, )
Consider the radlahon passing through the ele-
mental area d§) about the point (a,gé ) on the unit
sphere. If all this radiation unlformly intercepted
. the unit area at coordinate ¢, on the side of the
‘Cylmder, the dose rate at coordmcfie ¢, would

208

equal D (a,¢,)dQ. The dose scattering probability
for a line beam is defined as the ratio of the actual
dose at ¢, due to the emission from dQ} to the dose
rate whi %1 would have been obtained if all the
radiation from J) had arrived at the unit area about
coordinate ¢,. The dose scattering probability for
a line beam is denoted as PHa ) ~ o)

By symmetry, the line-beam dose scattering
probability is a function of the difference in the
azimuthal angles (by ~ &, )e Assume that PS(a,gs
- ¢2) is known. Lef the angulor distribution of fhe
dose rate on the surface of the unit sphere about
the point source § be given by D o ,qS ). Then the
dose rate at coordinate ¢, on the side of the cylin-
der is given by the integral shown in Fig. 14.16.
This integral weights the differential dose con-
tribution over the unit sphere by its proper scattering
probability.

in general, primary reactor shield configurations
are symmetric about the axis joining the reactor
and crew compartment, and so the radiation from
the primary shield is also symmetric about this

 

 
il M,

skl

 

S
2-01-056-7-18-1104

 

1074
CENTER OF TRIPLET NEUTRON DOSIMETER AT:
5 ¥ 0.0cm
¥ VARIABLE
o
-z 0.0cm
- 2
. -5
= 10 195 ft

d 84ft

p 45cm

g VARIOUS
0

2 =330

FAST-NEUTRON DOSE RATE (mrep/hr/watt)

-7 8=300 deg

8=180deg

 

00 106 (12 g  t24 {30 136 142 148
y, DISTANCE FROM REAR FACE OF DETECTOR TANK TO
o DETECTOR CENTER (cm)

 

: _ 7 i'Fi'g.- 14.6. FééiQNéut‘rofi. Dose Rates Near the
' " Front of the Detector Tank for Various Angles
of 6; y Variable.

  

 

REAR; y =10 ¢m

RATEHmrep/hi/woit):

    
 

£UTRON DOSE
S
1.
n

x 0.0 cm

 

360 330 . 7300 "

 

PERIOD ENDING JUNE 10, 1955

SEfP
2-01-056~7—-39-{{5A

10
5 o R
WITH 5 in, OF LEAD BLOCKING
8=330 deg DIRECT BEAM
2
8=360 deg
1973
® DATA TAKEN AT LOW POWERS
5 WITH NO LEAD IN DETECTCR
£ TANK
2
C \
i‘ 2
g . o
E4g? 19=300 deg h 195 1
o 8=270 deg g 64ft
;1: 5 p 45cm
W 8 VARIOUS
3 @ O DEG
a 2
=2
o 8=360 deg
& .5
E 40
L
| 5
L
o
<
|
2
6=300 deg
T

 

5 LCENTER OF TRIPLET NEUTRON
DOSIMETER AT:

x¥ 00 cm
2 ¥ VARIABLE
z 00 cm

o) 12 24 36 48 60 T2 84 96 {08 {20
y, DISTANCE FROM REAR FACE OF DETECTOR TANK TO
DETECTOR CENTER {cm)

-Fig. i4.7. Fast-Neutron Dose.Rufes. Near the
Rear of the Detector Tank for Various Angles of G;
y Variable.

2-01-056-7-19+28+ 40-44A

A 195 ft

d 64 ft

p 45 cm

8§ VARIABLE
o O deg

FRONT; y =445 cm

CENTER OF TRIPLET NEUTRON DOSIMETER AT:

¥ SEE ABOVE
come 2 00 6m

270 T 240 0 240 7 180

. B, HORIZONTAL ANGLE BETWEEN p AND o AXES (deg)

- Fig. 14.8. F_c-:'s:t-fléu-fron Dose Rates Near the
Front and Rear of the Detector Tank; 0 Variable,

  

209

 

 

 

b T

 

 

 
ANP PROJECT PROGRESS REPORT

EempT
2-04—056—-7-39-134A

h 195 ft
d 64 ft

P VARIOUS
é Odeg
a Odeg

FAST—~NFUTRON DOSE RATE (mrep/hr/watt)

CENTER OF TRIPLET NEUTRON DOSIMETER AT:
x . 00em

¥ VARIABLE
z 00cm

 

0 12 24 38 48 60 T2 84 96 108 120
f, TOTAL THICKNESS OF WATER BETWEEN REACTOR FACE AND GEOMETRIC CENTER OF DETECTOR (cm)

 

Fig. 14.9. Fast-Neutron Dose Rates Near the Rear of the Detector Tank for Three Valves of ; ¥ Variable.

SGRET
2-01-056-3-T24

THIS PLATE LEFT OUT

 

ALUMINUM

 

 

4
; REMOVABLE PLUG FCR
DETECTOR INSERTION

LEAD

 

LUMINUM
ALL DIMENSIONS ARE IN INCHES

Fig. 14.10. Schematic Diagram of the GE-ANP Crew Compartment.

210

 

 
e

Sddiil

 

.5

-

-
v

 
 

 

  
 

FAST-NEUTRON DOSE. RATE' {mrep/hr/watt)

Fig. 14.11.

1673

1078

. SPeRET
2-04-056-3-28-124A

 

 

 

 

 

 

g =360 deg

 

 

 

 

 

 

 

 

 

 

 

 

QUTSIDE REAR AT y = 00 cm
INSIDE REAR AT y = 899 ¢m
INSIDE FRONT AT y = 2480 cm

 

 

 

 

 

 

 

 

 

 

OUTSIDE FRONT AT y = 280.4 ¢m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

| CURVE SHAPE OBTAINED AT p=152 cm, § = 270 deg |
5 195 fi P 45 cm
d 64 ft 8 VARIOUS
L4t
| i 1 I ]
11—
[ CURVE SHAPE OBTAINED AT p={7 cm, & =180 deg —|
C _
O e ?
CENTER OF TRIPLET 4
NEUTRON DOSIMETER AT:
X 0.0 cm
¥ VARIABLE
z 0.0 cm
Lol 1 1
100 #0120 130 140 150 160 170

- ¥, DISTANCE FROM REAR FACE OF CREW COMPARTMENT
TO DETECTOR CENTER (cm)

of 4; y Variable,

 

RATEE", (mrep/he/watt) -

“FAST-NEUTRON' DOSE;

1078

FushNéuirbn Dose Rates Inside the
GE-ANP Crew Compartment for Various Angles

h 195 ft
d 64 ft
p SEE ABOVE
8 vaRIABLE

300

 

 

FAST-NEUTRON-DOSE RATE {(mrep/hr/watt}

Tom]

108

10~7

108

PERIOD ENDING JUNE 10, 1955

RIGHT SIDE WALL

-48
x, HORIZONTAL DISTANCE FROM & AXiS TO DETECTOR CENTER (om)

 

~-36

§=360 deg

SRGRET
2-01-056-3-29-125A

195 ft
64 ft

4% ¢m
VARIOUS

8= 180 deg

CENTER OF TRIPLET NEUTRON DOSIMETER AT:

X VARIABLE
¥ 169 cm
z 00 ¢cm

—-24 —12 0 12 24 36 48

Fig. 14.'52. Fast-Neutron Dose Rates Inside the
GE-ANP Crew Compartment for Various Angles
of 9; x Variable.

 

 

 

  

CENTER OF TRIPLET NEUTRON DOSIMETER AT:

JECRET
2-01-056-3-34 +37-427A  _
o

 

 
   

 

        
   
   
     

  

x 0.0 e¢m ?é
¥y 169 em 2
Zz 0.0 ¢m £
—g ™
10 g
£
5 ¢
<t
o
w
3
2 a
=
7

-5
10 ’5
w
=
1
5 &
<
w

2
107¢
240

0, HORIZONTAL ANGLE BETWEEN p 'AND J Axes (deg)' '

 

180

Fig. 14.13. Fasi-Neuiron Dose Rutes lnsnde the GE ANP Crew Compartmeni for Two Values of p;
0 Variable.

 

 

211

 
 

 

 

FAST-NEUTRON DOSE RATE (mrep/hr/watt) . .

 

 

ANP PROJECT PROGRESS REPORT

~axis, Flgure 14.17 defines the dose scattering
~ probability for a conical shell beam. Again, a unit
sphere is drawn about the point source, The radia-

GEORET
T 2-01-056-3-38-126A

5 CENTER OF TRIPLET
NEUTRON DOSIMETER AT:
x 0.0 ¢m
. y 169 ¢m
) Z 00 cm

1072

 

5
360 deg
2
1078
330 deg
5
A 195 ft  p VARIABLE 300 deg
¢ 64 ft G SEE ABOVE
2
270 deg
1077
0 10 20 30 40 50 60 70

p, THICKNESS OF WATER BETWEEN REACTOR FACE
AND REACTOR TANK WALL (cm)

Fig. 14.14, Fast-Neutron Dose Rates Inside the
GE-ANP Crew Compartment for Various Angles
of 9; p Yariable,

DO .
2-01-056-0-T9

 
  
 

REACTOR

REACTOR | shiglDING

CREW
COMPARTMENT

 

. .Fig."|4'.15.. | Radiation Attenuation in the Ideal-

ized Aircraft Divided Shield,

212

tion is emitted from the annular shell contained be-

tween two cones which have their common vertex
at S, the point source, The conical angle of the
shell is a. For this symmetrical case, the dose
rate on the unit sphere is independent of the
azimuthal angle ¢,. Cotrespondingly, the dose
rate received at the side of the cylinder is inde-
pendent of the azimuthal angle ¢,. The dose scat-
tering probability P¥(a) for a conical shell beam
is defined by the integral over the azimuthal angle
¢, of the dose scattering probability for a line
beam. Assume P®(a) is known. [et the angular
distribution of dose rate on the unit sphere be
D (a). Then, on the basis of the definition of
P*(a), the dose rate at the side of the cylinder is
given by the integration shown in Fig. 14.17,

UNCLASSIFIED
2-01-088~-T-D134

 

 

  

(052, 4,)]

s e 1 P2

P ( a, ¢1 ¢2)— Da(a"';")"dn -

FOR DOSE EMISSION 0O (a,¢,) FROM POINT SOURCE S,

Ds(¢2)=f P*la, d=$,) 05 la,4)d0
[t}

Fig. 14.16. Probability of the Neutron Dose
Scattering.

UNCLASSIFIED
2-01-056-7-0132

 

 

  

2w
Sy o 5 _
Pilars [ A (a,4,-4,) 09,
O
FOR CONICAL SHELL BEAM EMISSION Dé(a),

Dg () = fq;(a) P3(a) du
W

Fig. 14.17. Probability of thé Neutron Dose
Scattering for a Conical Shell Beam.

 
 

o e v
R

S

i

ey

-}

  
   
 
  
  

i

“

This integral weights the differential dose rate
contribution over the unit sphere by the dose
scattering probability for a conical beam. The
integration is taken with respect to the cosine of
the angle o that each conical shell in the beam
makes with the reactor crew compartment axis

" (denoted by symbol u én Fig. 14.17),

-’ Evolfiéfi§fi _.of- Prd;dbflity from
o TSF Experiments

Figu;ur‘e 14;]8 “shows how the dose :-"s.‘cafiering

probability is obtainable from the TSF differential
shielding experiments. The experimental fast-
nevtron beam has an angular spread about its axis
of symmetry, as indicated in the figure. Hence,
there is obtained from the experiments the line-
beam dose scattering probability averaged in some
way over the given experimental beam. For the
present purpose the TSF beam geometry can best
be characterized by the polar angle @ measured
between the reactor crew compartment axis and the

Po(a, -9, Dpla,,) 0 5

PERIOD ENDING JUNE 10, 1955

line of symmetry of the beam and by the azimuthal
angle ¢y which the axis of symmetry of the beam
makes with the horizontal plane., By definition,
the line-beam dose scattering probability averaged
over the experimental beam, denoted by Ps (6, - ¢,),
is simply the dose scattering probability weighted
over the dose rate contribution of each differential
element of the beam divided by the integrated
dose over the experimental beam. The numerator
is simply the measured scattered dose rate at the
side of the cylinder; this rate is a function of the
beam angle @ and also of the difference between
the azimuthal angles of the beam and of the unit
area under consideration at the side of the cylinder,
that is, of (¢ - ¢.) This is converted to the
average dose scattering probability for a conical
shell beam by integrating over the azimuthal angle
s of the beam. Because the dose scattering
probability is a function of the difference in angles
(Y —~ ¢>2), this is equivalent to integrating over
angle ¢, for a fixed angle . Substitution for

UNCLASSIFIED
2-01-056-7-D133

 

 

 

 

 

 

- 8,y -
PS(0,y - )= A _ S(MEASURED)( ¥4,
e [ bta,4)a0 [ Deta,g) a0
‘BEAM BEAM
. /
\\"&u BEAM

   

 

DS(MEASURED)(Q’q’ 9_62) 2

 

O, u;"w a0

 

Fig. 14.18. Relaiionr;h‘ips -fof Dért'er.miri'ing Dose Scattering Probability from TSF Experiments.

 

213

 

TR T

 
 

 

 

 

s G Dt ©

- ANP PROJECT PROGRESS REPORT

Ps(6,4 Z 952) of the equation at the top of Fig.

- 14.18 results in the final relation for obtaining,

from the experiments, the average dose scattering
probobllny PS(@) for a conical shell beam. The
numerator in this relation is the integral of the

S scch‘ered dose, around the side of the cylinder,
“with respect to angle ¢,. The denominator is the

integral of the direct-beam dose rate emitted from
the equivalent point source.
This relationship for conversion of the experi-

~ mental dose rate measurements to a fundamentally
‘basic probability function is general. |t applies
“not only to the scattered dose rate at the outside

surface of the cylinder but also to the dose rate at
various depths into the cylinder. [n this case the
probability function is also a function of the shield-
ing thickness. This relationship is also applicable

for the scattered dose rate at the front or rear face
. of the cylmder or for the scattered dose rate within
. ;:fhe crew compartment cavity. The method of
- evaluation of the scattered dose integral [numerator
‘in the PS(6) relationship] from the detector tank

measurements is presented in the following paper.
The method of evaluation from the experimental
results of the direct-beam dose rate integral
[denominator in the PS(4) relationship] is given
below,

Evaluation of Direct-Beam Integral

A plot of the relaxation length of the direct-beam
fast-neutron dose rate vs total water thickness be-
tween reactor face and detector is given in Fig.
14.19. Because of the large separation distance
between the dosimeter and the reactor face in the
TSF experiment, the relaxation lengths measured
correspond to the point attenuation kernel for the
fast-neutron dose rate. For comparison, relaxation
lengths obtained from BSF data and corrected for
geometry to correspond to a point attenuation kernel
are also plotted in Fig., 14.19. The TSF and cor-
rected BSF data agree well in the overlapping
region, both as to magnitude and to trend.

The variation of the dose rate measured at the
rear of the detector tank as a function of beam

. angle @ is shown in Fig. 14.20. The ordinate is
~ 'the ratio of the dose rate for beam angle 6 to the
. .dose rate for § = 0 deg, that is, when the reactor
" is pointed directly at the detector tank. For a given

“““dngle @, the dosimeter in the detector tank reads

" the direct-beam dose rate coming off at angle 6
“with respect to the line of symmetry of the beam

214

2—-01-056-T-Ai28

 

o—eo—5 **

 

o

O TSF DATA
® BSF DATA CORRECTED FOR GEOMETRY

rlem)

 

 

 

 

 

0

 

0 50 ' 100 150
TOTAL WATER THICKNESS BETWEEN REACTOR FACE AND DETECTOR (cm)

Fig. 14.19. Relaxation Length (\) for Fast-

Neutron Dose.

serRet
2-04-056-7- 129

5 DIRECT BEAM INTEGRAL*

J' RECTED (ppan)

2
(9)-!2- 2 sin 8 00 =167 x10"2x ¢

0.5

CALCULATED DIRECT

on

3 02 PLUS ESTIMATED

9 SCATTERED
Bl od
‘—-
g MEASURED DIRECT
wlul 0,05 PLUS SCATTERED
Qo
oo

0.02

CALCULATED DIRECT
0.04

0.005

ME ASURED
CONTRIBUTION = 12.5% SCATTERED DOSE
0.002 AT SIDE

 

0.004 :
0 20 40 60 80 100 420 440
8, BEAM ANGLE {deg)

Fig. 14.20. Angurlur Distribution of Dose at Rear
of Detector Tank (p = 45 cm).

 
 

e e o ki v

bt

 

" “verse R? relationship, D_

o ‘p‘|u‘s an air-scattered dose rate from the other por-

tions of the beam. At § = 0 deg the direct-beam
dose rate is large, and therefore the scattered dose
rate is negligible. As @ increases, the direct-beam
dose rate decreases rapidly, owing to the beam

- collimation, and the scattered dose rate becomes

proportionately larger. In order to evaluate the
direct-beam dose rate integral, it is necessary to
subtract out the scattered dose rate. The method
used was, briefly, the following: The neutron

~angular emission from the faces of the reactor was

assumed to have a cosine distribution. The direct
beam from the reactor faces was attenuated along a
straight-line path by using the relaxation lengths
shown in Fig. 14.19. The results of this procedure
for the various angles @ are shown in Fig. 14.20.
The calculated direct-beam curve and the measured
direct- plus scattered-beam curve agree well for ¢
below about 45 deg, where it is to be expected that
the scattering contribution will be relatively small.

“For @ greater than about 45 deg the calculated

direct-beam dose rate drops rapldly, and, at 90 deg,
the direct-beam dose rate is down by a factor of
10%. As expected from the geometry, the dose rate
measured at § = 90 deg is an almost totally scattered
dose rate. The scattered dose rates measured at
the side of the detector tank for @ = 90 and 135 deg
are also given in Fig. 14.20. The ratio of rear- to
side-scattered dose rates for these two angles is
about 1.6, If this ratio is assumed to be valid to
6 = 60 deg, an estimate of the rear-scattered dose
rate based on the measured sco'rtered dose rate
at the side can be made for 6 = 60 deg. If this
is done and fhe esflmated rear-scofiered dose
rate at @ = 60 deg is added fo the calculcted direct-

‘beam dose ra're, the result IS fhat gwen by fhe”,

    

 
 

’ o corrected
the dose rate which would have been obtained on

 

 

PERIOD ENDING JUNE 10, 1955

the surface of a unit sphere drawn about the
equivalent point source. The element of solid
angle dQ is 2% sin 6 d6. The direct-beam dose
rate integral for the TSF experimental beam (at

=45 ¢m) is equal to 1,67 x 1072 x {2,

Calculation of Scattered Dose

Figure 14.21 presents a typical plot of £2p5(6)
vs the angle 0 of the conical shell beam. Note
that the ordinate is the product of the square of
the separation distance by the dose scattering
probability for a conical shell beam for this sepa-
ration distance. The dose scattering probability
presented in Fig., 14.21 is for a water thickness
p = 45 cm at the reactor, a water thickness ¢ into
the side of the cylinder of 5.2 cm, and a separation

PG
2-01-056-7-A136
O T T I T I

 

| LET D; (8,0) = DIRECT-BEAM DOSE OBTAINED AT )
DISTANCE ¢ FOR A GIVEN REACTOR SHIELD DESIGN

| THEN: 0= [ 0 (8,0} a2 P(8) dp

,u

0.05 - o4, e | s e

‘ pszffi_zué__ {12/:3(9)] I
@ £

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.02 \
&
w

N
Qoo N
= N\
\\
0.005
\\
30 0 780 9ol U427 T 450 180

8, CONICAL SHELL BEAM ANGLE (deg)

 

  

Fig. 14.21. Dose Scattering Probability for Coni-
cal Shell Beam as Obtained from TSF Experiments
(o = 45¢cm, t = 5.2 cm, { = 64 #).

215

 

 
   

   

 
 

 

e

 

- ANP PROJECT PROGRESS REPORT

diétqni_:e ffiof 64 ft. The functional dependence of
P*(6) on p, t, and 1 should be accounted for in

o sccfl‘ered dose rate caleulations, In the present

expenmenfs a range “of values of t is covered for a

" single sepcrahon distance £ and for a single value
- of p; however, some data were obtained with p
- changed from 45 to 15 cm, and the effect of this
- change on the dose scattering probability is dis-
. cussed later. To illustrate the use of the dose
" scattering probability for a conical shell beam in

the prediction of scattered dose rate, assume that

~the angular distribution of the dose rate emission

from a prlmary reactor shield is obtained by dose
rate measurements af a distance d from the reactor.
Thls dose rate d:strlbuhon obtained at distance 4
is denoted as Dd(é? d). Then D6, d)d2 is the dose

‘rate which would be obtained on the surface of a
“ “unit sphere drawn about the equivalent point source.
" The scattered dose rate is obtained by weighting
"~ Dj(6, ,4)d? by the scattering probability PS(6) de-

fermmed from the differential experiments and
integrating with respect to y = cos @ {first integral
in Fig. 14.21). The second integral in Fig. 14,21,
which is equivalent to the first, is in terms of the
£2p<(6) used in the plots.

Effects of Direct-Beam Collimation

Figure 14.22 shows the effect of beam collimation
on the dose scattering probability averaged over
the beam. These results were obtained from single
air-scattering calculations. The beam in the TSF
experiments had approximately a cos? distribution.
The cos® and cos'® distributions represent an
increasing degree of collimation of the beam. |t
must be remembered that in the interpretation of
the TSF experiments the assumption was made that
the dose scattering probability averaged over the
TSF experimental beam was equal to the dose
scattering probability for a differential beam having

- an angle @ equal to the angle of the axis of symme-

try of the experimental beam. These results indi-

-cate that this assumption is valid within 10 to 15%

- for beam angles greater than 30 deg. For beam
" angles between 0 to 30 deg the averaging of the
 dose scattering probability is quite sensitive to
beam collimation, and therefore the experimental
results obtained are not valid if any sharp well-
jdefmed beams are emitted from the reactor shield

- in this range of beam angles. Consideration of
multiply scattered neutrons in the calculations
"~ would have tended to reduce the effect of beam

 

UNCLASSIFIED
2-04-056-7-A430

0.3

0.2

04

0.05

2
LB 18

0.02
0014

0.005

0.002

 

0.001 '
0 30 60 90 120 150 i80

8, BEAM ANGLE (deg)

Fig. 14.22, Effect of Beam Collimation on
Scattering Function Based on Single-Scatter Calcy-
lation.

collimation shown on this figure because the

sources of multiply scattered neutrons are of
necessity more diffuse than the sources of singly
scattered neutrons,

Effect of Neutron Energy Spectrum

A small amount of data was also obtained for a
water thickness of 15 em at the reactor to obtain
an indication of the effect of change in neutron
spectrum on the dose scattering probability.
Figure 14.23 presents the angular distribution of
the dose rate at the rear of the detector tank for
p = 15 cm of water. Again, to obtain the direct-
beam dose rate, it is necessary to subtract out the
scattered dose rate. The procedure descnbed for
the p = 45 cm case was used, but the results are
not nearly so satisfying. The calculated direct-
beam dose rate has a flatter distribution than the
measured direct plus scattered dose rate in the
region of O =0 to 45 deg, where the scattered dose
rate should be a relatively small contribution.
For all the data obtained at p = 45 cm, both for the
direct-beam and scattered dose rates, a plot of the
dose rate vs the beam angle always showed a
slight dip in the curve in the region of 6 = 30 deg.

‘This dip has been attributed to an out-of-roundness

 

 
i o i deaails Gm i

 

ol i, i

 

 

ca

Fig.

ST
2-01-056-7~A137

~1 2
RECT DOSE INTEGRAL= 9.6x10 'x.J

0.5

CALCULATED DIRECT
PLUS ESTIMATED
SCATTERED

0.2

04

MEASURED DIRECT
0.05 PLUS SCATTERED

O deg

0.02

DOSE AT &
DOSE AT 8

CALCULATED DIRECT
.01

0.005

CONTRIBUTION= {8
©.002

 

0.004
0 20 40 60 80 100 120 140
8, BEAM ANGLE {deg)

Fig. 14.23. 'Afigijlur 'Distribufion of Dose ‘u;"Rear
of Detector Tank (p = 15 ¢m).

of the reactor tank. For the 15-cm case, the ef-
fect of an out-of-roundness of the reactor tank
~ would be much more pronounced owing to much
shorter relaoxation |engfhs. The discrepancy ob-
tained here is believed to be due, in part, to this
effect, The direct-beam dose rate integral was
obtained from the curve drawn through the data
points from 0 to 45 deg and faired into the calcu-
lated curve from 60 to 90 deg. The contribution to

the direct-beam dose lntegral of the dose rafe from _
- scaflered neutrons into the side of the detector
“tank is qu:te chfferent for the p = 15 cm case com-
‘pared with the p = 45 cm case. Examination of the
“"relaxation lengths for the two cases shows that they
’are approxlmotely proporhonal to the relaxchonr
Ieng'rhs for the dlrect-beom dose. More dafa are

| _50 to 90 deg is only about 18%.

   

o N b

 

Frgure 14, 2% compores\’rthe Ime-beam dose scatter-

mg probobsl n‘y obfcmed fl"om the TSF me

 

=z S XA

l5|de_6f the detectoi' ank.

   

: 14, 24 e that the effect of this dif-
“ference in energy’ Spectfbm’on the air-scattering
process is not large, the difference in the relative
air scattering between the two cases being de-

 

 

      
 
    
   
 
  
  
 
  
      

‘PERIOD ENDING JUNE 10, 1955

SUERES
2-01-056-7-A135
0.05

0.02

12 % DIFFERENCE

0.01

0.005

0.002

Dy, £)d0

0.004

BEAM

7

0.0005

0.0002

 

0.0001
o) 4 8 12 16 20 249
/, WATER THICKNESS AT SIDE OF DETECTOR TANK (cm)

Fig. 14.24. Effect of p (Water Thickness at Re-
actor) on
D

s

 

2P0 =0, f ~ p, =0) =
| L

edam

D d(Q,{) dQ

' termined as only 12%. However, a rough estimate

of the uncertainties in arriving at this percentage
difference adds up to the order of 20%. Hence, at
least a £20% figure should be placed alongside the
12% difference indicated. More data for larger

‘water thicknesses at the reactor are needed to

definitely tie down this point.
The plot shows that the relaxation length of the

“needed to definitely tie down this apparently simple

relahonshlp.

iiiiee SHIELDING EXPERIMENTS

J. E. Yan Hoomissen

differential experiments (see preceding discussion,
*“The Differential Shielding Experiments at the
TSF: Phase I’’) has made possible a more detailed

217

 

APPL!CATION OF THE DlFFERENTIAL -

_The hknowyledge gumed “from the first data of the

il

R e

 
 

| S

 

rflro cm
20¢cm
__.‘

Kmm = = Ym = X
40 ¢m
Y= ———— == X —

n

0.38In. STEE

0.50in. STEEL

L

0.62in, STEEL

FRONT

COMPARTMENT A

2.05in. LEAD
(3 PIECES)

COMPARTMENT B

22.7¢cm

 

Jd2in. STEEL

.78in. STEEL

3in. STEEL

in. STEEL

3in x 3in.
LEAD

 

J——x— XX K e e X

Fig. 14.25.

38in. STEEL!

 

 

C2-01-05%6-~3-T6A

35.25in.

50-in. GASKET

38in. STEEL
X
o
I_ 40 ¢cm
';.] ; _ 20cm r-
|
TSF ' -~ REAR ——->|<—:—x————
REACTOR COMPARTMENT F
15in. STEEL _1.5in. LEAD
COMPARTMENT D
0.62in. STEEL
0.42in. ALUMINUM
COMPARTMENT € VoD
0.62in.STEEL COMPARTMENT E
i
1
X
[ 10cm
X
{' -
20¢cm _
; X - POSITIONS OF GEOMETRICAL
g | , CENTERS OF COUNTERS
Mockup of G-E R-1 Shield.
§ * - - »

1¥0d3Y SSTUO0U LDIrOYd dNY

 
 

A

"

 

-
w

j:cm for TSQ(__deg. Sl_n

 

-1816, p 158; ORNL-1864, p 178.

analysis of the measurements taken in the divided-
shield mockup experiment reported earlier.® It also
serves as a basis for a prediction of the results of
future mockup experiments. It is the purpose of
this paper to correlate the results, to date, of the
mockup experiment and the differential experiments
and to outline the procedures for future predictions.

In the mockup experiment the reactor was encased
in the R-T shield (Fig. 14.25), and measurements
were made with a fast-neutron dosimeter along the
x axis in the detector tank, The experimental

~values obtained in that experiment are compared

with values predicted by applying the data differ-
ential and the procedures given in the two preceding
papers.

Since the differential experiment includes x
traverses across the detector tank and z traverses
up and down in the tank, the extrapolation of the
dose rate measurements to the surface of the four
faces of the tank _WQUId give dose rates (Fig.
14.18) from a line beam? emitted at a given g as a
function of the angle ¢, (Fig. 14.26). The integra-
tion of the area under the curve in Fig, 14.26
would then be equivalent to the dose rate at a
point on the surface of the cylindrical crew com-
partment from a conical shell beam at the same
value of 6. A plot of a series of these integrated
dose rates as a function of water thickness of a
cyhndrlcal crew comportmen‘r side shield can then

be made for # = 0, 30, 60, 90, 135, and 180 deg,

for which experlmen‘ral data are available. A

composite plot (Flg. 14.27) shows the actual
attenuation curves that should be used in designing
a cylmdrlcal crew compartment,

The average relaxation length of the curves in

e

PERIOD ENDING JUNE 10, 1955

oo
Z2-01-056-3-A149
20 2

 

 

o

 

FAST-NEUTRON DOSE RATE (mrep/hr/watt x 10°)

 

 

 

 

 

 

 

10 /
5 \ /
0
0 90 180 270 360

4>2, AZIMUTHAL ANGLE AT CREW COMPARTMENT (deg)

Fig. 14.26. Fast-Neutron Dose Rate from a Line
Beam as a Function of ¢, for 0 = 60 deg, p = 45

cm, t = 0,

Flg- 14,27 varies from 64 7ch for 9 0 deg fo__75 9 oo

   
  
 
 

 

 

 

    

3c E. Clifford i"al., ANP Ouar, Prog. Reps.,

4The concept of a line beam is discussed in the pre-
ceding poper,

 

 

‘ Slnce the( dose rate emn‘fed at eoch‘cmgiua 0 from

   

ORNL-"

cylmdrlca! crew. comparfment ‘or the equivalent

* thickness :n the s:de of fhe defector tank, for a
‘unit dose rate emltfed in a comcal shell beam at

- any angle @ can be deduced from Flg. 14,27,

  

      

A comprehens‘n}e eXper‘lmenft“ is now under way at
the TSF to determine this angular distribution,
that is, the fast-neutron dose rate leaving the

219

 

 
 

 

 

 

 

 

_ANP PROJECT PROGRESS REPORT

~SECRET
1 2-01-056-7-A—{359A
407

10-3

- ZZPS(B) x1.67
ny

 

2
o- 8=135 deg
f=180d
5
2
10-6

52 12 {72 232 292 352 412 472

DISTANCE IN FROM SIDE SURFACE
OF A CYLINDRICAL CREW BOX (cm)

Fig. 14.27. £2P5(6) for a Conical Shell Beam as
a Function of Water Thickness at the Side of
Cylindrical Crew Compartment for Yarious Angles

of 0 and for p = 45 deg and £ = 64 ft.

shield as a function of angle 6; in the meantime, a
preiiminory measurement of this distribution (Fig.
14.28) is used in thls analysis as the best available
data,

 Ata given water thickness the values of dose
“rate as a function of 6 can be read from Fig. 14.27.

. -__;These pomts — an example for a given water thick-
~niess is shown in Fig. 14.29 — are plotted, and a

“smooth curve is drawn between the points. This
shows the ‘dos’e_rrafe_fo be expected at agiven water

: _,220.

2-01~056-7-Al46A

 

»
o

 

W
o

 

o
™

N\
[\

 

0o

@
——
]

 

 

 

FAST-NEUTRON DQSE RATE
AT A DISTANCE OF 64 ft FROM REACTOR (mrep/hr/waft x10%
rn ~ ’
O £
\
-—/

o
o

0
P
]
/

 

o

J o
TN

0 20 40 60 80 100 120 140 160 180
REACTOR ANGLE & {deq)

 

o
©

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 14.28. Estimation of Fast-Neutron Dose
Rate at 64 ft from Reactor Encased in R-1 Mockup
as a Function of Reactor Angle 0.

thickness from a conical shell beam at any angle 6.
This distribution is weighted by the angular dis-
tribution of the mockup and integrated to determine
the total dose rate to be expected at a given water
thickness in from the side of a detector tank. A
plot of the total dose rate at a given water thick-
ness as a function of @ for the R-1 mockup is given
in Fig. 14.30. This process was repeated for
several different water thicknesses at the detector
tank, and the results, along with the experimentally
measured values, are shown in Fig. 14,31, |In
general, the agreement is good. The apparent dis-
agreement can be attributed to three factors, the
first of which is the uncertainty of the angular
distribution of the fast-neutron dose rate emitted
from the R-1 mockup. This uncertainty, possibly
the largest contributor to the disagreement, will be
cleared up by future experimentation, The second
factor contributing to the d:sagreement is the low
count rate encountered in the actual R-1 experi-
ment. The third factor is the lack of complete

collimation of the direct or unsccttered becm in the ]

differential experiment, as pointed out in ‘the
preceding paper.

 

L

 
 

FOFPERIPny

A g i, R A M

 

»

"

coss dlstrlbuhon. _,;‘The dnfferem‘lalfi"exp’eraments in
whlch ‘the TSF reactor was placed in the reactor

2%P°(0) x 167

O 30 60

- PERIOD ENDING JUNE 10, 1955

eFereT
2-04-056-7~A-140A

 

90 120 150 180

8, ANGLE BETWEEN CONICAL SURFACE BEAM AND REACTOR-CREW COMPARTMENT AXIS (deg)

Fig. 14.29. 12P%(6) ot a Point 37.5 cm in from Surfucekof Cylinarical Crew Compartment as a Function

of Reactor Angle 0 for p 45 deg und 13 = 64 ft

The method to be used in reachmg a predlcf:on
of the dose rate to be expected in the crew com-
partment with the reactor in the R-1 mockup can
now be illustrated. Up to this pomt the dISCUS-

snon has’ beenulrlmited to measurements made in the

 
 
  
 
 
 
 
 

 

3). E. Faulkner, Focusing of Radzatzon ina Cylmdrzcal
Crew Compartment, ORNL. CF-54-8-100 (Aug. 18, 1954)

 

   
 
 
 

the _dose rate ’measureé_ gf th_e center of the cclylfy'
1 trengfh c’r the surface of th _ cavutym

 

  

_at the ang|e 9 hove /been determme&? The lcmo of )
these values for different ‘angles of 6 (Table 14.1)

tank and measurements were made in the cavity of
the GE-ANP crew compartment mockup allow an
estimate of this ratio to be made experimentaily.
The dose rate at a given water thickness at the
s:de of fhe cy||ndrlco[ crew compartment (Fig.

ed and IS qssumed to be the source
From the experlment (F:g. 14.12) the

‘f"‘dose rafe at the center ‘of the cov|ty from a line
_,_\_.beam qt ‘the angle 6 is known. This experimental
" value |'s 'cfllawsted to the correspondlng value (that
s, it is multlplled by 27) at ‘the center of the
_cavity for a comcal shell beam based on the pro-
) v'_':__‘";cedure Voutllned in the precedmg paper. Ncnw, both

 
 
      

   

  

is shown to be effectively a constant with an
average value of 5,1. This indicates that the

221

 

28) from a comcal shell ‘beam at the angle 6

 

 
 

 

o

FAST~NEUTRON DOSE RATE SCATTERED IN SIDE OF DETECTOR TANK

 

. ANP PROJECT PROGRESS REPORT

SECRET
2—04—-056-7-A147A

 

" 150

..

n

o
/

 

 

o
S
—

 

o
~
a

 

0.50 \

0.25 |

 

 

 

 

 

 

 

 

 

 

 

 

 

0 20 40 &0 80 100 {20 {40 160 180
‘ REACTOR ANGLE 8 {deq)

Fig. 14.30. Fast-Neutron Dose Rate Scattered
in Side of Detector Tank as a Function of Reactor

Angle 6 (Reactor in R-1 Shield, ¢t = 10.2 ¢m).

angular distribution of the neutrons leaving the
inside surface of the crew compartment can be
assumed to be the cos3 of the angle measured from
the normal to the surface. This picture, along with
Faulkner’s work,® neglects any neutrons that
might go across the cavity and be reflected back
into it, This should be a small contribution to the
source strength, on the order of 10%, if an albedo of
]/10 is assumed, -

The dose rate at the center of the cavity of the

‘cylindrical crew compartment with the reactor en-

cased in the R-1 shield mockup can now be pre-
dicted. The source strength at the surface of the
cavity is found by making a calculation as out-
lined above except that a water thickness equal to
the side shielding of the crew compartment mockup

. is chosen. This source strength is integrated

(multiplied by a factor of 5.1) to give the con-

tribution to the dose rate at the center of the crew

222

TABLE 14.1. RATIO OF DOSE RATE AT CENTER OF
CREW COMPARTMENT CAVITY TO DOSE RATE
AT INSIDE SURFACE AS A FUNCTION OF @

 

Dose Rate at

 

9 Corrected Dose Rate Surface of
(deg) at Center of Cavity Cavity Ratio™*
(mrep/ hr/w) (mrep/he/w)

0 1.55 % 1079 3.14x 1078 485
30 .08 x 1073 2.10x10°%  5.14
60 5.30x 1078 9.70 x 1077 5.46
90 1.86 x 1076 3.80 x 107 4.89
180 3,15 % 1077 6.00x10°8  5.25

 

* Average value of ratio = 5,12,

compartment mockup from neutrons scattered through
the side of the compartment.

The dose rate contribution at the center of the
cavity resulting from neutrons entering the rear
and the front of the crew compartment can also be
found from the differential data by using the y
traverses for a line beam at a given @ (Figs. 14.3
and 14.4). This is converted to data for a conical

“shell beam by multiplying by 2m. This gives the

dose rate at a given water thickness in the front or
the rear of the crew compartment as a function of
the angle 6. Again, by weighting these distribu-
tions by the angular distribution of the R-1 mockup,
the source strength at the inner surface at the
front and rear of the crew compartment can be
found. An angular distribution of emission from
the inner surface must be assumed and an integro-
tion performed to find the dose rate at the center
of the cavity, A good approximation is to again
use a cos® distribution for the front of the crew
compartment, but for the rear the distribution will
be closer to a cosine distribution because of the

existence of thick lead. The three component

parts of the dose rate at the center of the cavity
(the neutrons scattered in the side, the rear, and
the front) are then added. This prediction will be

completed when a more exact measurement of the

angular distribution of neutrons leaving the R-1
mockup has been made,

N

1.

 
 

e e Y £

[T
- o

 

ki

 

QQAI

»

»

FAST-NEUTRON DOSE (mr‘ep/hr/wafi)

 

PERIOD ENDING JUNE 10, 1955

SELRET

-5 : ’ ' 2-04{~056-3-22-85R1A

hH=1495 f+
d= 64 ft
a= 90 de

CALCULATED CALCULATED

MEASURED MEASURED

CENTER OF TRIPLET FAST-NEUTRCN DOSIMETER AT:

x = VARIABLE
y=90c¢m
2= 0Ocm

~-78 —-66 -—54 -42 -30 -8 -6 0 6 18 30 42 54
‘ x, HORIZONTAL DISTANCE FROM & AXIS TO DETECTOR CENTER (cm)

Fig. 14.31. Fast-Neutron Dose Rate Along x Axis of Detector Tank.

 

66

 

78

223

 
3

3

 

 

 

 

 

 
   
 
 

 

 

 
 
 

 

"

T

Part IV

APPENDIX

 

 

 
 

 

 

 

 

 

 

g

'Lyl

()

>

 
 

 

o
-

© CFi55-2-93

CF-55-484

15. LIST OF REPORTS ISSUED FROM FEBRUARY 1955 THROUGH MAY 11, 1955

 

 

 

 
 
  

: :_. Crn‘:cal Assembly wnh /8 in. Incone!
Core Shells

e IS1’IC$ of Confrol Rods and Mcferlals
o Potenhally Suncble for Use in fhe o
ART F’arf || | o

  

CF-55.4-137 tio ontrol Rod
terials. Part 11]: The Effect of
Neutron frradiation on Some Rare

Earth Samples

 

bLEvdluotion' of ART ControlRod Ma- S

AUTHOR(s)

H. C. Hopkins
L.. T. Anderson
W, T. Furgerson,
H. C. Hopkins

C. S. Burtnette

A. S. Thompson
A. S. Thompson

V. J. Kelleghan
E. 5. Bettis
B. Cottrell

A. S. Thompson

A. S, Thompsen

L. T. Anderson

Tc Jo B(:I”es

L. A. Mann

REPORT NO. TITLE OF REPORT
. Reflector-Moderated Reactor
CF-55-2-142 Vertical Component of Fuel Forces on
Reflector and Island
CF-55-3-161 Gamma and Neutron Heating of the ART
Fue! Pump Assembly
CF-55-3-167 Circulating-Fuel-Reactor-Powered
Ramjet
CF-55-3-191 Fission Product Heating in the Off-Gas
System of the ART
CF-55-4-34 Empirical Correlation for Fatigue
Stresses
CF-55-4-44 Allowable Operating Conditions
~ CF-55-4-83 High Temperature Valve Information
Summary
CF-55-4-8.7 ART Design Data
CF-55-4-116 The ART Off-Gas System
CF-55-4-124 Flexible Mounting Systems
CF-55-4-159 Thermal Stresses in Tube-Header
Joints
CF-55-5-76 Calculation of the Beryllium Contri-
bution to the ART Temperature
Coefficient of Reactivity
CF-55-5-93 Surface-Yolume Ratios for Five
Different Fluoride Fuel Systems
H. Ex_prerrimeptal Engineering
 CF-55-2-100 ART Reactor Accidents Hazards Tests
" Critical Experiments
Three Region 'Ref‘lecfa{Mader&éJ T

R. M. Spencer

,Eyaluohon of Reachvny Chdracfer_ AT

" J. W, Noaks

J. W. Noaks

DATE OF DOCUMENT

3-2-55

3-28-55
3-24-55
3-28-55

4-5-55
4-11-55

4-5-55

4-18-55
4-21-55
4-11-55

4-25-55

5-11-55

5-12-55

2-11-55

2-14-55

4-13-55

4-25-55

227

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

REPORT NO,

CF-55-4-178

CF-55-4-18

- CF-55-2-79
CF-55-3-157

CF.55.3-179

- CF-55-4-167

CF-55-4-181

CF-55-2-89

CF-55-2-148

CF-55-3-15
CF-55-3-47
CF-55-3-61
CF-55-3-62

: C.F-55-3-137

CF-55-3-174

' .CF-55-4-32

8

Examination of First Three Large

TITLE OF REPORT

Evaluation of ART Control Rod Ma-
terials, Part IV: The Variation of
Reactivity with Control Rod Diameter

AUTHOR(s}

J. W, Noaks

IV. Chemistry

Solubility of Composition 30 in Water

V. Corrosion

Data and Results of ARE Corrosion

Capsules

Fluoride Pump Loops

Examination of Inconel-Flueride 30-D

'Pu_mp Loop Number 4695-1

Examination of Sedium-Incone! Pump
Loop 4689-4

Examination of Fluoride Pump Loops

4930-A and 4935-1

J. C. White

R Baldock

| G. M. Adamson,
R. S. Crouse

G. M. Adamson,
R. §. Crouse,
P. G. Smith

G. M. Adamson,
R. S, Crouse

G. M, Adamseon,
R. S. Crouse

Vl. Heat Transfer and Physical Properties

Measurement of the Viscosities of

Composition 35 and Composition 74

Electrical Heating and Flow in Tube
Bends

Qualitative Velocity Profiles with
Rotation in 18 Inch ART Core

Heat Capacity of Lithium Hydride

Measurement of the Viscosity of

Composition 72

Measurement of the Viscosity of

Composition 30

Measurement of the Viscosity of

' Composition 43

Status Report on Forced Convection
Experimental Work in Converging and
Diverging Channels with Volume Heat

Sources in the Fluids

Measurement of the Viscosity of

Composition 2

S. I. Cohen,
T. N, Jones

H. W. Hoffman,
L. D. Palmer,
N, D, Greene

G. L. Muller,
J. 0. Bradfute

W. D. Powers,
G. C. Blalock

S. l. Cohen,
T. N. Jones

$. I, Cohen,
T. N. Jones

3. I, Cohen,
T. N. Jones

H. F. Poppendiék,
N. D. Greene

S. l. Cohen,
T. N. Jones

DATE OF DOCUMENT

| 4-29-55

4-4+55

 2-15-55

3-24-55

3-28-55
4-21-55

4-26-55

2-15-55

2-22-55
3-1-55
3f7'55 .
3'8'55_
39:55
_3,-] 6-55

3-24-55

4-1-55

 

 

e

)

 
ot s it

 

 

%uh%n%.m‘ i nbi ol o

 

 

¥,
o

 

PERIOD ENDING JUNE 10, 1955

REPORT NO, TITLE OF REPORT AUTHOR(s) DATE OF DOCUMENT

Yili. Radiation Damage

_t CF-55-2-36 The Fate of Certain Fission Products M. T. Robinson,
in the ARE S. A. Reynolds,
_ H. W. Wright 2-7-55
F
CF-55-4-16 Ru Deposition in In-Pile Loop M. T. Robinson 4-5-55
CF-55-5-22 A Theoretical Treatment of Xe 135 .
Poisoning in the ARE and the ART M. T. Robinson 5-2-55

Vill. Shielding

CF-55-2-111 Calibration of the Revalet, a

Remotely Variable Lead-Trans-

mission Gamma-Ray Dosimeter D. L. Gilliland
CF-55-4-122 Spectrometer Measurements of Fission -

Product Gamma Rays for the CFR E. P. Blizard 4-21-55

IX. Miscellaneous

ORNL-1864 Aircraft Nuclear Propulsion Project
 Quarterly Progress Report for Period
Ending March 10, 1955 A. W, Savolainen
(ed.) 4-13-55

-

"

229

 

 
 

T Y

(o

ni...,.

 

 

 

 

 

a s ey ey <o
 

ik iR A el b d i o i, bt S v - P R .

   

 

 

THE AIRCRAFT NUCLEAR PROPULSION PROJECT

 

 

 

GONFHDERHE
AT ‘
THE OAK RIDGE NATIONAL LABORATORY
JUNE 1, 1955

ANP PROJECT DIRECTOR W. H, JORDAN ’0
. CO-DIRECTOR S. J. CROMER RD
Do ASSOCIATE DIRECTOR & 1. STROUGH PHA
s . : . ASSISTANT DIRECTOR A, L MILLER RO

) D.HILYER, SEC.  RD

 

 

 

 

 

 

 

PRATT & WHITNEY AIRCRAFT REPORTS
R. k. STROUGH, PROJECT ENGINEER  EWA A. W, SAVOLAINEN ARE
4. GIANGREGORIO, ADM, ASST.  PWA ERATURE § .
0. A. BRADEN, SEC. ARE LITERATURE SEARCHES
H.C. GRAY, DESIGN ENGINEER ™ A L. DAVIS ARE

 

 

 

 

 

 

 

 

1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IR I

 

 

   

 

 

MRCRAFT REACTOR ENGINEERING DIVISION SUPPORTING RESEARCH
5. J. CROMER, DIRECTOR 70 AHP STEERING COMMITTEE W. H. JORDAN
A. J. MILLER
H. MCRATRIDGE, SEC, ARE W. H. JORDAN, CHAIRMAN
E. S. BETTIS
D. 5. BILLINGTON
E. P. BLIZARD
G. £. BOYD
S, J. CROMER
W. K, ERGEN
ASSISTANT T0 DIRECTOR AP, FRAAS
W. R. GRIMES
R. 5. CARLSMITH ARE C. E. LARSON
W, D. MANLY METALLURGY
A, J. MILLER
W T POPEENDIEK W. D. MANLY, STAFF ASSISTANT M
H. W, SAVAGE
E.D. SHIPLEY
R. 1. STROUGH
3. A, SWARTOUT
A. M. WEINBERG
EXPERIMENTAL ENGINEERING CHEWISTRY
] H. W, SAVAGE ARE W. R. GRIMES, STAFF ASSISTANT e
NOTE: THIS CHART SHOWS GNLY THE LINES OF TECHNICAL CCORDINATION OF THE ANP PROJECT. THE
VARIOUS INDIVIDUALS AND GROUPS OF PEQPLE LISTED IN THIS AND THE FOLLOWING CHARTS ARE EN-
GAGED EITHER WHOLLY OR PART TIME ON RESEARCH AND DESIGN WHICH IS COORDINATED FOR THE
BENEFIT OF THE ANP PROJECT IN THE MANNER INDICATED ON THE CHART. EACH GROUP, HOWEVER,
POWER PLANT ENGINEERING IS ALSO RESPONSIBLE TO ITS DIVISION DIRECTOR FOR THE DETAILED PROGRESS OF THE RESEARCH SHIELDING
A P, FRAAS ARE AND FOR ADMINISTRATIVE MATTERS. E. P, BLIZARD, STAFFE ASSISTANT AP
THE KEY TO THE ABBREVIATIONS USED IS GIVEN BELOW.
AC ANALYTICAL CHEMISTRY DIVISION - ORNL
AP APPLIED NUCLEAR PHYSICS DIVISION
RADIATION DAMAGE
ENGINEERING DESIGN ARE  AIRCRAFT REACTOR ENGINEERING DIVISION — ORNL. 5. BILLINGTON, STAFF ASSISTANT s
H. C. GRAY PHA BAC BOEING AIRPLANE COMPANY o3 8 :
8P BENDIX PRODUCTS, DIVISION OF BENDIX AVIATION CORPORATION
c CHEMISTRY DIVISION - ORNL
CT  CHEMICAL TECHNOLOGY DIVISION ~ ORNL
CY  CONSOLIDATED VULTEE AIRCRAFT CORPORATION
REACTOR CONSTRUCTION ENR  ELECTRONUCLEAR RESEARCH DIVISION HEAT TRANSFER AND PHYSICAL PROPERTIES
b 1 1] -
e 5 BETTIS - c INSTRUMENTATION AND CONTROLS DIVISION ~ ORNL 4. F. POPPENDIEK, STAFF ASSISTANT  REE
-3, M METALLURGY DIVISION — ORNL
MC  MATERIALS CHEMISTRY DIVISION - ORNL
ME  MERCK & COMPANY, INC. Lo e .
MH  MINNEAPOLIS-HONEYWELL REGULATOR COMPANY fa"m‘_ s;:{}%
P MATHEMATICS PANEL - ORNL . s
REACTOR PHYSICS NACA NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS REACTOR PHYSICS ;:ffi -@
A. M. PERRY eng PACKARD MOTOR CAR COMPANY W: K. ERGEN, STAFF ASSISTANT ARE 5
PRATT & WHITHEY AIRCRAFTY, DIVISION OF UAC WA
RESEARCH DIRECTOR'S DEPARTMENT — ORNL, :‘\;-W
REACTOR EXPERIMENTAL ENGINEERING DIVESION ~ ORNL
STABLE 1SOTOPE RESEARCH AND PRODUCTION DIVISION — ORNL ;
SOLID STATE DIVISION — ORNL
¥ UNITED STATES Aif FORCE
W. G. PIPER® K-25 *PART TIME F. R. BRUCE* cr

 

 

 
 

PRI I et e TR MR R Ry W W TR M T Y e on e

 

 

 

 
(¥

 

 

 

 

 

™

 

B

il i .

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. . i . . - - R e - e - ., LY e
'E
!
AIRCRAFT REACTOR ENGINEERING DIVISION |
i
\
RIDGE NATIONAI_ LABORATORY |
JUNE t, 1955 !
5. 3. CROMER, DIRECTOR RD }
H. MCFATRIDGE, SEC. ARE :
|
|
|
|
‘1
ASSISTANT TO DIRECTOR 1
R. S, CARLSMITH ARE
REACTOR PHYSICS REACTOR CONSTRUCTION POWER PLANT ENGINEERING EXPERIMENTAL ENGINEERING ENGINEERING DESIGN 7503 AREA CONSTRUCTION
A. M, PERRY ENR E.S, BETTIS ARE A. P, FRAAS ARE H. W. SAVAGE ARE H. C. GRAY PWA W. G. PIPER* K25
M. WILSON, SEC. ARE J. PARKER, SEC. ARE A, 5. THOMPSON PMC D. ALEXANDER, SEC. ARE 3. ZASLER ARE $. M, JANSCH,* SEC. K-25
P. HARMAN, SEC. ARE L. E. FERGUSON, SEC. ARE
S. J. FOSTER, CLERK ARE
M. C. OVERTON, CLERK ARE
H. W. BERTINI ARE FLUID MECHANICS R, CORDOVA ARE
W. J. FADER PWA ADMINISTRATION T, E. CRABTREE ARE
A. FORBES ARE G.D. w“géfl_” Qf,fi V. L. SCOTT ARE V. J. KELLEGHAN ARE
W, E. KINNEY ARE R. L. BREWSTER ARE s DESIGNERS F. R. MCQUILKIN ARE
M. TSAGARIS ARE A. M, SMITH ARE HYDRODYNAMICS AND THERMODYNAMICS A. A. ABBATIELLO ARE G. C. ROBINSON ARE
FILL AND DRAIN AND ENRICHMENT W. T. FURGERSON ARE ! A. H. ANDERSON ARe cR:' 2' \SVL%TING ::E
SYSTEM K M. E. LACKEY ARE 4 D. C. BORDEN PWA , F.
CONSULTANT G. SAMUELS ARE j W. G. co:BwE . ;\T'i
L. T. ANDERSON L. A MANN ARE W. J. STELZMAN ARE DEVELOPMENT TEST SERVICES 4. M. CORNWEL RE
HIGH-TEMPERATURE CRITICAL M, W, HINES, COMPUTER — ARE R. C. DANIELS Al
. 1. J, TUDOR, DRAFTSMAN  ARE E. R. DYTKQ PWA W. B. MCDONALD ARE J. Y. ESTABROOK ARE
EXPERIMENT D. P. HARRIS, SEC. ARE W. A, FRY PWA
W.C. TUNNELL ARE APPLIED MECHANICS AND PHYSICS PUMPS A. N, MONTGOMERY, CLERK  ARE B. C. GARRETT PWA
G. W, PEACH ARE M. E. LAVERNE ARE A. G. GRINDELL ARE W. 5. BARRIS PWA
L. E. ANDERSON PWA R. CURRY PWA TEST FACILITY ASSEMBLY R. E. HELMS ARE
CONTROLS F A FIELD USAF S, M. DECAMP ARE COORDINATION T. L. HUDSON ARE
E. R. MANN Ic 8. L. GREENSTREET ARE L. R. ENSTICE PWA E. STORTQ ARE A. D. JARRATT PWA
J. M. EASTMAN BP C.B. THOMPSON MH J. B. KERCHEVAL PWA J. L. CROWLEY ARE J. KERR ARE
R. G. AFFEL Ic C. H. WELL PWA J.W. KINGSLEY ARE W. H. KELLEY ARE W. €. LEUTLOFF PHA
S. C. SHUFORD PWA WL GALYON DRAFTSMAN ARE W. L. SNAPP PWA W. J. MUENZER ARE F. L. MAGLEY ARE
E. VINCENS Ic D. R, WARD ARE H. C. SANDERSON PWA C. MANTELL PWA
£. 5. WALKER Ic HEAT TRANSFER H. C. YOUNG PWA W. 1. NELSON PWA
R. D. SCHULTHEISS ARE CORROSION LOOPS J. A. GLSON PWA
T. J. BALLES PWA HEAT EXCHANGERS AND YALVES C. P. COUGHLEN ARE W. H, POND PWA
R. 1. GRAY PWA R. E. MACPHERSON ARE R. A. DREISBACH PWA H, R. ROCK PWA
H. C. HOPKINS PWA J. C. AMOS ARE P. G. SMITH ARE
M. M. YAROSH ARE -J. L. BRANDT PWA
M. 1. COOPER PyA ELECTRIC SERVICES DRAFTSHEN
SPECIAL PROBLEMS L. H. DEVLIN PWA E. M. LEES ARE W. C. GEORGE ARE
R. D. PEAK PWA D. L. CLARK ARE G. R. HICKS ARE
¥. g' gongflfiéTTE Cgf,_- J. 3. W. SIMON PWA W. F. FERGUSON ARE G. L. HOLLIS ARE
T wov PWA A. L. SOUTHERN ARE C. E. MURPHY ARE J. R, LARRABEE PWA
Wi . G M
H. WIARD e PROCUREMENT COORDINATION TEST OPERATIONS g i M:EESELSON ::E
W. F. BOUDREAU ARE J. J. MILICH PWA J. 0. NICHOLSON ARE
C. F. SALES ARE
CONSULTANTS SHOP AND PROCUREMENT TECHNICIAN GROUP |
A. H. FOX, UNION COLLEGE W. R. OSBORN ARE R. HELTON ARE '
W. LOWEN, UNION COLLEGE D.T.STOREY, CLERK  ARE L ARE
R. L. MAXWELL, UNIVERSITY OF 5. Ry ASHTON ARE G S, CHILTON ARE
TENNESSEE C. A. CLILPEPPER ARE 'M' COBURN ARE
G. F. WISLICENUS, PENNSYLVANIA STATE J. R, EVANS ARE j R CROLEY AEE |
COLLEGE D. L. GRAY ARE -
T LK ARE J, M, CUNNINGHAM ARE
. R. E. DIAL ARE
L MA A
§ G ETEESK,NS AQE J. R. DUCKWORTH ARE
E. B. PERRIN ARE W. H. DUCKWORTH ARE
W. K. R. FINNELL ARE
INSPECTION H, FOUST ARE
R. B. CLARKE ARE W, D, GHORMLEY ARE
H. HOOVER ARE T. L. GREGORY ARE
L w. LOVE PWA R. A. BAMRICK ARE
C. J. HENLEY ARE
J.W. TEAGUE ARE
J.R.LOVE ARE !
D. £, MCCARTY ARE l
IN-PILE LOOP TESTS G E. MILLS ARE ‘»
0. B. TRAUGER ARE B. H. MONTGOMERY ARE |
S. EUBANKS, SEC. ARE J. J. PARSONS ARE ‘
C. ¢ BOLTA PWA H. F. PENLAND ARE ,
L. P. CARPENTER ARE M. A REDDEN ARE
J. A. CONLIN ARE -5 D ARE
C. W, CUNNINGHAM ARE J. P SHUGART ARE
P. A. GNADT ARE A. G. TOHNS ARE
D. M. HAINES PWA B. C. WILLIAMS ARE
D. . MAGNUSON ARE C. A WALLACE ARE
€. J. GREEN ARE
E. M. GRIZZELL ARE
€. C. NANCE ARE
B. L. JOHNSON ARE CONSULTANTS i
F. J. SCHAFER ARE A F. AILEY UNIVERSITY OF TENNESSEE ¢
€. E. STEVENSON ARE W. R. CHAMBERS ;
W, K. STAIR UNIVERSITY OF TENNESSEE |
OFFICE SERVICES !
3P LANE ARE
F. M. LEWIS, CLERK ARE
R. E. THOMPSON ARE

 

 

 

 

 

 

rery

 

 

S

s v

 

T mm oy v

 
ey e e e e w el b B bt

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

|
|
!
AT :
THE OAK RIDGE NATIONAL LABORATORY
I
JUNE 1,1955 i
SUPPORTING RESEARCH |
W. H. JORDAN
A. ). MILLER
|
|
STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT 1 STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT
. K. ERGEN ARE E. P. BLIZARD AP W, R. GRIMES MC i W. D. MANLY M D. $. BILLINGTON S5 H. F. POPPENDIEK REE
i
i
REACTOR CALCUL ATIONS SHIELDING CHEMISTRY ANALYTICAL CHEMISTRY METALLURGY METALLOGRAPHY RADIATION DAMAGE HEAT TRANSFER AND PHYSICAL
R. A, CHARPIE* RO W. R. GRIMES MC M. T. KELLEY* AC W. 0. MANLY . M R. J. GRAY* M D. . BILLINGTON* 58 PROPERTIES RESEARCH
SHIELDING RESEARCH D. R. CUNEQ MC C. D. SUSANO* AC M. J, WHITHAN ) USAF M, J. FELDMAN 55 H. F. POPPENDIEK REE
D. E. CALDWELL, SEC. MC B. E. YOUNG,* SEC, AC G. M. ADAMSONY M G. W. KEILHOLTZ 55 T. K. CARLSMITH, SEC. REE
N, D. GIVEN* MP E. P, BLIZARD" AP 1. C. WHITE AC ¥. H. BRIDGES* ! M R. S. CROUSE M J. B. TRICE 58
W. C. SANGREN* WP R. RICKMAN, SEC AP E. M. ZARZECKI,* SEC. AC J. THOMAS, SEC. M M 2. C. WILSON 55
! T. M. KEGLEY
R. K. OSBORN AP L. 5. ABBOTT AP SPECIAL PROBLEMS ; E.L.LONG M HEAT TRANSFER
5. AUSLENDER PWA ! E. J. MANTHOS M
. F. NEWTON ! .0,
CONSUL TANT F. H, MURRAY AP R. F. NEWTO RD RESEARCH GENERAL CORROSION E. R. BOYD " MTR LIAISON oo EEE
J. A. NOHEL, GEORGIA TECH. R. B, MURRAY AP PHASE EQUILIBRIUM STUDIES A. S. MEYER, JR. AC E. E. HOF FMAN M B. F. DAY M H. V. KLAUS 58 Y W, HOFFMAN REE
H. E. STERN cv C. J. BARTON MC D. L. MANNING AC R. CARLANDER PWA R, L, FITZGERALD M HOT LAB FACILITY F. E. LYNCH REE
C.B. ZERBY AR L. M. BRATCHER MC B. L. MCDOWELL AC W. H. COOK M B. C. LESLIE M G. L. MULLER PWA
SHIELD DESIGN B. H. CLAMPITT MC J. M, PEELE AC C. F. LEITTEN, JR. M M. J. FELDMAN 55 | L. D. PALMER REE
1. B. DEE PWA R. E. MOORE MC W, J. ROSS AC J. E. POPE M C. ELLIS 55 $. J. PIAS PWA
Sy R. J. SHEIL MC L. R. TROTTER M G. E. KLEIN 55 J. L. WANTLAND REE
C. A, GOETZ PHA D. L. STOCKTON ME DEVELOPMENT W. B, PARSLEY 55 J. LONES REE
CRITICAL EXPERIMENTS J. SMOLEN PHA J. P. YOUNG AC DYHAMIC CORROSION R. N. RAMSEY 55 R. L. MILLER REE
LID TANK SHIELDING FACILITY FUEL PREPARATION RESEARCH G, GOLDBERG AC . M. ADAMSON M . E. §. SCHWARTZ 55 G. M. WINN REE
A. D, CALLIHAN® AP AND PHASE BEHAVIOR M. A, MARLER AC T. C. PRICE PWA CERAMIC RESEARCH L. C. TEMPLETON 5
M. L. RUEFF,* SEC. AP G.T. cmmgg Q;A F. F. BLANKENSHIP NC . A TABOADA M PHYSICAL PROPERTIES
W. J. tcCooL C. 4, BLOOD MC G. D, BRADY M J. M. WARDE* M s. 1. COHEN REE
J. M. MILLER AP F. P, BOODY MC W. F. VAUGHN AC R. M. EVANS M A. HOBBS,* SEC. M MECHANICAL FROPERTIES W, D. POWERS REE
J. 8. CRUDELE PWA D. K. TRUBEY AP R. E. CLEARY PWA C. R. WILLIAMS AC J. C. WILSON® 55 G. C. BLALOCK REE
J. F. ELLIS . AP H. C. WOODSUM PWA H. DAVIS PWA R. F. APPLE AC FABRICATION C. D. BAUMANN 39 R. M, BURNETT REE
L V. G. HARNESS AP E. BECKHAM AP F. L. DALEY MC D. E. CARPENTER AC J. H. COOBS M C. £, CURTIS* M W. E. BRUNDAGE 55 S. J. CLAIBORNE REE
JLLYNN AP H. JARVIS K H. A. FRIEDMAN M R. C. BRYANT AC M. R. D' AMORE PWA O W DONEY* M W, W, DAVIS 35 T. N. JONES REE
J. H. MARABLE AP J. R TAYLOR AP 5. LANGER MC L. R, HALL AC H. INOUYE M 5. b. FULKERSON® W A. 5. OLSON 53
B e e J. W, WAMPLER AP R. P. METCALF MC L. E. 1DOM AC R, W, JOHHSON M R. L. HAMNER" W J. C. ZUKAS $$
- R. F. W, MILES MC A, H. MATTHEWS AC o
E, V., SANDIN PWA BULK SHIELDING FACILITY R E. THOMA e C E. PRATHER A WELDING AND BRAZING M. P. HAYDON" M FLUID FUELS
D. SCOTT, JR. ARE F. C. MAIENSCHEIN AP A. D, WILSON AC P. PATRIARCA i M f," Fj‘ 4‘;’3&"; fl G. W. KEILHOLTZ 55
S. SNYDER PWA C. BOUNDS, SEC. AP CORROSION STUDIES C. . WILSON AC R. E. CLAUSING M I e, " W. E. BROWNING ss
R. M. SPENCER USAF T. V. BLOSSER AP F. KERTESZ MC R. L. HEESTAND PWA " A GRIFFIN® D. E, GUSS USAF
G. M. ESTABROOX AP H. J. BUTTRAM MC G. M, SLAUGHTER M H. L. HEMPHILL $5 FUEL REPROCESSING
J.D. FLYNN AP F. A, KNOX MC 0. E. CONNOR M M. F. OSBORNE S5 £, R. BRUCE* P
M. P. HAYDON? AP R. €. MEADOWS MC B. MCDOWELL M CONSULTANTS H. €. ROBERTSON 3 -
X. M, HENRY AP G. F. SCHENCK PWA C. E. SHUBERT M M. T. ROBINSON 55
E. B. JOHNSON AP J. M. DIDLAKE MC R. G. SHOOSTER M G. M. BUTLEg,R %RBTORg;lggm f?lPé:ANY C. C. WEBSTER 58 CHEMICAL DEVELOPMENT
M. T. CORY, HITE
$: f: JL%}:,'?ON ::,M PRODUCTION OF PURIFIED FUELS SPECTROGRAPHIC ANALYSIS b C. WILLiAMS M CORPORATION ¥- R WILLIS s D. E. FERGUSON® er
R.W. PEELLE AP G. J. NESSLE MC — 3. R. MCNALLY* sl PHYSICAL pHEMISTRY OF H. INSLEY FLUX MEASUREMENTS 6. .. CATHERS cT
W, ZOBEL AP J. P.BLAKELY MC 3. A. NORRIS* AND OTHERS sl LIQUID METALS T. N, MCVAY, UNIVERSITY OF ALABAMA J. B. TRICE 5 M. R. BENNETT cT
D. J. KIRBY AP C. R. CROFT MC G. P. SMITH | M T. 5. SHEVLIN, OHIO STATE UNIVERSITY J. KRAUSE PWA R. M. DUFF cT
J. SELLERS [ F. A. DOSS MC C. R. BOSTON M H, THURNAUER, MINNESOTA MINING AND
R. M. SIMMONS AP J. E. EORGAN MC J. V. CATHCART M MANUFACTURING COMPANY SOLID FUELS | UNIT OPERATIONS
S. SMIDDIE ic N. V. SMITH MC H. W. LEAVENWORTH PWA 0. SISMAN* $§ W. K. EISTER® T
G. G. STOUT ic J.TRUITT MC G. F. PETERSEN M R. M. CARROLL 3 1T, LONG e
H. WEAVER AP B. F. HITCH MC M. E. STEIDLLTZ M J. G. MORGAN 85 S H. STAINKER o1
W. JENNINGS MC R. W. ANDERSON M M, T. MORGAN 55 T
TOWER SHIELDING.FACILITY G. A, PALMER MC MASS SPECTROMETRY L, L, HALL M DESIGN
€. €. CLIFFORD AP B. C. THOMAS MC — c.r BALDOCK* 5l ; H. E. GOELLER* cr
.R, E . E.
E. McBEE, SEC. AP R. G. WILEY MC J.R. SITES sl b A Mmc:::;:f" PROPERTIES " METALLURGY CONSULTANTS R. P. MILFORD cr
1. L. HALL AP CHEMICAL EQUILIBRIA - _j,. H. DEVAN' M N. CABRERA, UNIVERSITY OF VIRGINIA PILOT PLANT
F. J. MUCKENTHALER AP Me . R. KENNEDY PWA N. J. GRANT, MASSACHUSETTS
W R AP L. G. OVERHOLSER - R : INSTITUTE OF TECHNOLOGY H. K, JACKSON* T
F. . SANDERS E. E. KETCHEN MC LR WEIR, JR. M H .
M. F. VALERINO NACA M. B. PARRISH MC W WOO!SS M J. L. GREGG, CORNELL UNIVERSITY W. H. CARR’ cT
J. VAN HOOMISSEN BAC N b REDMAN e "W, BOLING " W. D. JORDAN, UNIVERSITY OF ALABAMA W. H. LEWIS cT
F. N. WATSON AP B 1. STURM e )T EAST " E. F. NIPPES, RENSSELAER
E. D. CARROLL Ic L E ToPoL e ] I D, HUDON M — POLYTECHNIC INSTITUTE
I. D. CONNER AP C F WEAVER e V. G. LANE M W. F. SAVAGE, RENSSELAER
J, W. MONEY AP " B, MCNABD, JR. M POLYTECHNIC INSTITUTE
G. G. UNDERWOOD IC SUPPORTING STUDIES B.C. STOWI’ERS IR, M G. SISTARE, HANDY & HARMON
R. E. WRIGHT AP P. A. AGRON c K. THOMAS | M P. C. SHARRAH, UNIVERSITY OF
‘A o ARKANSAS
C. W, WA R M
4. B. SUTHERLAND ¢ t LE F. . TATMALL, BALDWIN-LIMA-HAMILTON
CONSULTANT NONDESTRUCTIVE TESTING E. C. WRIGHT, UNIVERSITY OF ALABAMA
H. A. BETHE, CORNEL L UNIVERSITY CONSULTANTS R. B. OLIVER | M
J, M, CARTER, CARTER LABORATORIES ¥ ’;‘?'EEE'; m
CONTRACTORS g' ?fisilé\[(" DUKE UNIVERSITY W, 5. MASON M
METAL HYDRIDES, INC, T. N. MCVAY, UNIVERSITY OF ALABAMA
NUCLEAR DEVELOPMENT ASSOCIATES,
INC. METALLURGY CONTRACTORS
CONTRACTORS BATTELLE MEMORIAL INSTITUTE
AMES LABORATORY BRUSH BERYLLIUM COMPANY
BATTELLE MEMORIAL INSTITUTE FERROTHERM
_CARTER LARORATORIES GLENN L. MARTIN COMPANY
MOUND LABORATORY ——  METAL HYDRIDES, INC,
A. R. NICHOLS, SAN DIEGO STATE NEW ENGLAND MATERIALS TESTING
COLLEGE LABORATORY
RENSSELAER POLYTECHNIC INSTITUTE
SUPERIOR TUBE COMPANY
UNIVERSITY OF TENNESSEE
*PART TIME
1
|
+ T e i Il

TN ERTOCER e CMOTERISRERT T T O oM T OTRETTER e R I e

 

 

   

 

 

o g