ORNL-1729.txt

v SYSTEMS LABRARIES

ARt mmu—.flrfalcnakc y
% A

. -:-av.%f-:flg 'fin"
__gfi\ Y J{?T "g_‘;-;g, "“& e

v

:
TR

 
 

ORNL-1729

This document censists of 155 pages.

Copy ?J; 255 copies. Series A,

Contract No. W-7405-eng-26

AIRCRAFT NUCLEAR PROPULSION PROJECT
QUARTERLY PROGRESS REPORT

For Period Ending June 10, 1954

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

DATE ISSUED

T g0 1954

OAK RIDGE NATIONAL LABORATORY
Operated by
CARBIDE AND CARBON CHEMICALS COMPANY
A Division of Union Carbide and Carbon Corporaotion
Post Qffice Box P
QOck Ridge, Tannessee

 

MARTIN Mantzrry ¢

 

WA T

3 4456 0360703 o

 

———

_
0P NO ;AL

. Hollaender

 
   

-,.;;:; . Larson
"S. Livingston
N, Lyon

FERRO-MMZOOER-PPMEPECNLIAZIONOAOTPNEMIOMONNRO

 

ORNL-1729

Progress

INTERNAL DISTRIBUTION

  

. B. McDoghld

   
  
  
 
 
   
   
 
    
 
 
 
     

. Adamson 43. W g
h 44. J. L. Meeg
45. A. J. Midter
46. K. Z.‘_E;gg‘s%rgan
47. E. J,Mur phy :
JP Murray (Y-12)

2 J. Nessle
. Patriarca :
F. Poppendiek
M. Reyling
W. Savage
W
D
S

    
  
   
 

. Bredig
. Bruce
. Callihan
. Cardwell
. Center
. Charpie
. Clewett
. Clifford
. Cotirell
. Cochran
. Cowen

romer

. Savolainen

. Shipley

isman

» Smith

. Smith (consultant)
. Snell

. Storrs

. Susano

. Swartout

DUOEPMIPMETCIPINLOEDOX

L. Culler . Taylor

B. Emlet (K-25) . frice

K. Ergen . Yan Artsdalen
P. Fraas . Vonderl.age
R. Grimes . Worde

E. Hoffman . Weinberg

. White

. Whitman

. Wigner (consultant)
. Williams

BMEO-PETIMEmM-O0P - OOM» I DI T

. Wilson
Winters

. D. Manly _ 95-97. Central Researé L ibrary

. A. Mann

 

 

Vii
98.

99.

100.

5 101,
%102,
103.
104-115.
116,
17,
118-122.
123.
124.
125.
126-128.
129.
130.
131-136.
137.
138.
139.
140.
141.
142.
143.
144,
145-149.
150.
151.
152.
153-155.
156-163.
164,
165.
166.
167-170.
171-172.
173-174.
175.
176.
177.
178.
179-182.
183.
184-185.
186.
187.
188-189.
150.
191.

 

EXTERNAL DISTRIBUTION

Air Force Engineering Office, Ogk Ridge

Air Force Plant Representative, Burbank

Air Force Plant Representative, Seattle

Air Force Plant Representative, Wood-Ridge

American Machine and Foundry Company

ANP Project Office, Fort Worth

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

Armed Forces Special Weapons Project, Washington (Gertrude Camp)
Atomic Energy Commission, Washington (Lt. Col. M. J. Nielsen)

Babcock and Wilcox Company

Battelle Memorial Institute

Bendix Aviation Corporation

Brookhaven National Laboratory

Buraau of Aernautics (Grant)

Bureau of Ships

Carbide and Carbon Chemicals Company (Y-12 Plont)
Chicago Patent Group

Chief of Naval Research

Commonwealth Edison Company

Convair, San Diego (C. H. Helms)

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

Detroit Edison Company

Division of Research and Medicine, AEC, ORO
duPont Company, Augusta

duPont Company, Wilmington

Duquesne Light Company

Foster Wheeler Corporation

General Electric Company, ANPD

General Electric: Compqny, APS, Richland

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

Hanford Operations Office

lowa State College

Knolls Atomic Power L.aboratory

Lockland Area Office

Los Alamos Scientific Laboratory

Moterlais Laboratory (WADC) ((. ol. P. L. Hill)
Mefcl!flrgtcal Project

Monséinto Chemical Company

Mouncl Laboratory

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

Newport News Shipbuilding and Dry DOck Company
New York QOperations Office

North American Aviation, Inc.

Nuclear Development Associates, Inc.

Pacific Northwest Power Group

 
Patent Branch, Washington
Phl”lps Petroleum Company (NRT

    
 

192.
193-19%,

  

     
  
 

&, M. Nelson)
202-211. o ~:»__:;_' w'smn (Fox Project)
212-213. B

214. :

215, Sylvanm E' Bt i foducts, Inc.

216. Tennessee V_ Authority (Dean)

217. USAF Hegdquarterl,

218. U.S. Nawal Rad:ologé%l Defense Laboratory
219.220. Umwé?é:ty of CaliforniciBadiation Laboratory, Berkeley
221-222. Up{‘varsuty of California R&fiwhon Lubomfory, Livermore
223, % ;
: # Westinghouse Electric Corporatich,
230-2 0. Wright Air Development Center (WCSKS, Col J. R. Hood, Jr)
241-255. Technical lnformahon Service, Oak Rldge

 
   

 
 
 
 

 
vi

 

Reports previously issued in this series are as follows:

ORNL-528
ORNL-629
ORNL-768
ORNL.-858
CRNL-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

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

This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL records
the technical progress of the research on the Circulating-Fuel Reactor and all other ANP research
at the Laboratory under its Contract W-7405-eng-26. The report is divided info three major parts:
. Reactor Theory and Design, Il. Materials Research, and Ill. Shielding Research.

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

The nucleus of the effort on circulating-fuel reactors is now centered upon the Aircraft Re-
actor Experiment — a high-temperature prototype of a circulating-fuel reactor for the propulsion
of aircraft. The equipment for this reactor experiment is now being assembled; the current status
of the experiment is summarized in Section 1 of Part 1. The supporting research on materials
and problems peculiar to the ARE — previously included in the subject sections — is now in-
cluded in this ARE section, where convenient. The few exceptions are referenced to the specific
section of the report where more detailed information may be found.

The ANP research, in addition to that for the Aircraft Reactor Experiment, falls info three
general categories: (1) studies of aircraft-size circulating-fuel reactors, (2) materials problems
associated with advanced reactor designs, ‘and (3) studies of shields for nuclear aircraft. These
phases of research are covered in Parts i, Il, and i, respectively, of this report.
FOREWORD .. ..
SUMMARY .....

CONTENTS

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

PART |, REACTOR THEORY AND DESIGN

1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT .. ... ... ... et e e

The Experimental
Pumps .....

Reactor System . . .. i v i it it i ittt it ettt aer i an s araae e

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

Heat exchangers . . .o it i it ittt it i i it it it s et i st e e e ’

Fluid circuits
Fuel-enrichment

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

BYSTEM & h v s i s s i e e h e s sie s e s s s e e s s e v e e e s e n e n

Loading facilities ... ... P
Fuel-sampling facility . ............. S
Fue[*UnlmdinngCillfy nnnnnnnnnnnnn :-n..onononlpuoonauu---au-.----.lc--:

Reactor control

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

Fuel System Mock-up Tests .. v i ittt i i i it it st st i st o st aa e o s e

Operation of the F

vel System . ......... e et e e s s e e e

Clecningnon.olonn.c-lconl.poululot.a..uno’ -------------------------

Pressure filling

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

Pressure and flow charaCteristics v v v v v v s o o s o s 0 s s 8 0 s o oo s n o nssseanncensenasens

Functional tests
Hot-gas test .
Pump Fabrication

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

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

AN T SIS v v ittt et e e e e e e e e e e e e s e ae e et

Reactor System Component Loop . .. ... .. e e e e e e h e e e e
Fuel Recovery and Reprocessing . ... .... i e s s e e e a et e e e e e

2. EXPERIMENTAL REACTOR ENGINEERING o v vt attiee e e ieneeeneanenns |

In-Pile Loop Component Development .. ... ...t ienserns e e -
Centrifugally sealed pump  « . . . . L it i it e e e e e e e e e e

Horizontal-shaft

SUMP PUMP v s v o v s v s s R Gt e e et e e e e s e

Vertical-shaft sump pump .. ..... M e e e e e e e e i e e e e e :
Hydraulic pump drives ..« v v v i i i it it i it i i et et e e e e e e h e s e e

Fluoride-to-wate

rheat exchanger . .. v i ittt it et e i e e P h e _
Forced-Circulation Corrosion Loops ... ... e et e et e e
Sodium-Beryllium-Inconel Mass-Transfer Test .. . ... . i i i e e e e ‘
Fluoride-to-Sodium Heat-Exchanger Test ... ............ e e e ‘
Gas-Furnace Heat-Source Development ... ... ..... .. e e e e e e |

Sodium Sampler . v v v vt i vt i e e e e e e ey ek e e e e T
Removal of Fluoride Mixtures from Equipment . . . . . i i i it i ittt it it i i e e i e

Bearing-Materials
Rotary-Shaft Seals

Tests v v e oo onnens e e e s e e m e e e s e e s

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

3., REFLECTOR-MODERATED REACTOR vttt it ittt et s e oot nnonassnennss R

A Comparison of Lithium- and Zirconium-Base Fluoride Fuels .. ... ....... ... .o oL,
Reactor Physics ... v i h i it v v s n v e w e s en e e e e s et aa e e

Reactor Calculations « v v v v v v vt e e v s v e e a e e e n e e e e e s e h e e e e :

 
PART 1. MATERIALS RESEARCH
4. CHEMISTRY OF MOLTEN MATERIALS .. i it i i i it i i et e es

Quenching Experiments with Fluoride Systems . . . . . v ittt ittt i i i ittt sce i aeaa
NaF-ZrF o e e s e
NaF-UF, e e e

Visual Observation of Melting Temperatures . . . v o i ittt it ittt i i it et e i e o nos

Differential Thermal Analysis of the NaF-ZrF, System .. ..... ...,

Filtration Analysis of Fluoride Systems . . .. v v i it it i i ittt i ittt s i ettt enn
NaF-ZrF, .o e e e
NaF-LiF-RbF-UF, . i i i e

Thermal Analysis of Fluoride Systems . . o v i it ittt ittt ittt i e b i o e n e e
e R
NaF-UF, o i e s
RbF-UF, o e e e e e
KE-LiF-UF, o i e i e as et
REF-LiIF-UF, i i s e
NaF-ZrF -UF -UF, i i e
NaF-ThF, . i i i i e
LiF-BeF -ThF,-UF, oo i i i e
L T R R

Thermal Analysis of Chloride and Mixed Chloride-Fluoride Systems . .. ... ... oot
UCL-UCT, o i i e s e e
UG -UF o e s
UG U, e i e
KLU, i i s s i i i s s e
RECLUCE, o i st e e
O O O T

X-Ray Diffraction Studies of the NaF-ZrF  System ........... ... .o i,

Chemical Reactions inMolten Salts . . .. o i i ittt ittt ittt it s ane s
Chemical equilibriain molten fluorides . . . v v v i v it i it il i i s i et es s et n o
Solubility of UF, in NaF-ZrF mixtures . ... ..o,
Solubility of UF, in NaF-KF-LiF mixtures .. ..oty
Solubility of UF, in NaF-RbF-LiF mixtures . ... oot
Chlorinationof UF, ..o vv it i e i s e
Preparation of UF3-QZrF4 ..............................................
Preparation of various fluorides . . v vt v i i it it il ittt ittt st e s e e

Fundamental Chemistry of Fused Salts . .. . 0 i ittt i e e it et v
EMF measurements o v v v it e it sttt o st ns i s st e e e
Physical chemisiry v i it ittt ittt ittt ettt ettt ettt

Chemical Effects of Fission Products . . v v v i vt it i i it i it it e et i s et e e e n e ones
Quantity of fission products v v v v vttt ot n e ottt v oot ettt e e e
Separation of solid phases . . . . i i i i it i i it e i ettt e e e s s
Effects on viscosity and heat capacity . . . . i it i it it i i i s e e e e
Effects on COMmOSIoN v v i i i it o it sttt s et s n st s st nossstnenonsossssnsnss

Purification and Properties of Alkali Hydroxides . .. ... ... 0.ttt
Purification of hydroxides ........... e e e r et a e et e
Reaction of sodium hydroxide with metals . . v i i ittt i e ittt e e e i e

Production of Purified Molten Fluorides ... 0 it it i ittt ittt it et et s e
Laboratory-scale production .« . i it it it i e e e et e
Experimental production . . i i it i it it i i i s e et e e e e
Large-scale production . ..
Production of enriched fuel for in-pile loop ..

« n o 9

5. CORROSION RESEARCH . .....

6.

¢ N e B8 4 0w

® B & ® B8 9 3 4 w a

a * a2 % % B B N A

® & 8 8 2 8 s o8 ¢ & & &

Effect of reduced phases in fluoride melt ... ..

Effect of fission products
Static and Seesaw Tests of Yarious Materials in Fluoride Mixtures and Liquid Mefols s e e e e

Dissimilar metals in NaF- ZrF UF
Molybdenum-coated Inconel in NaF LrFA -UF,
Stainless steel in lithium with lithium mtnde udded e e e e

Chromalloyed steels in liquid metals .
Cermets in NaF-ZrF ,-UF,

« n o & & 3 m = p

* & v o B o n

" s a % a0 00 4.8 0

Fluoride Corrosion of Inconel in Static and Seesaw Tests .. ...
Effect of chromium addition to fluoride meh .
Effect of temperature
Effect of surface area ...

* = & » @« @ e s
L ] o« » -
* s » 4@ o A W

® &4 o B & & 5 2

> w 8 P 4 2 B & 5 B 4 2 8 & » 4 8 4 4 O 4 & & e 3 a

¥ R ¢ 0 o 3 ® & @

s & 8 2 & & B 8 >

2 & 4 B & B.% B3 W & s & ¥ & ® & 0 4 2

Graphite in NaF-ZrF -UF ; and in sodium .. ...

NaF-ZrF
NaF-ZrF

4

L

° B 8 M 9 4 w

in special Inconel
in Hastelloy B .

in stainless steel
in Incone! with stainless steel or

2 ¢ e e w0

4 ® 8 8 & & s a 8 v .

5 m % B 8 A u

« 8 s .2 B 0 »
4 » & & * & = &

a 2 ¢ a =

Sodium in Inconel with beryllium inserts ... ...

Lithium in stainless steel
Fundamenta! Corrosion Research
Mass transfer in liquid lead

a 8 & »

¢ o

LI O L

4 8 s 4 w & B & s O 4 &

& 8 & B B & 0 & B B 3 B & 0w

Products of hydroxide-metal reactions . ......

Dehydration of sodium hydroxide

a a

¢ " o m e & e ¥ e

Color changes in fused hydroxides .. ........

METALLURGY

Stress-Rupture Tests of Inconel
High-Conductivity Metals for Radiator Fins . ... ...
Special Materials Research . ..

Hastelloys B and C

Nickel-molybdenum alloys

& @ & o B 8 s 8 @

" e A s A2 0 e v oa » @

* % 2 B @ R M A 2 4w 0.k s =

Stainless-steel-clad molybdenum and columbium

Columbium
Boron carbide

Welding and Brazing
Brazing alloy development
Beryllium test assembly

@ 6 & & D N s p

a & & @ o n

® #F & B 0 2 0 W B B & 8 & 3

* e & & 2 U & @

4 W 9 2 8 4 ® 6 5 8.8 B % a

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

7. HEAT TRANSFER AND PHYSICAL PROPERTIES

Heat capacity

Physical Properties Measurements

@ o

® A B S B s 8 8 % 4 I A e e ow

--------

2 w & » e & @ & »

Flueride Corrosion of Inconel in Thermal-Convection Loops . ..
Effects of UF, and mixtures of UF, and UF,
Effect of hydrogen fluoride . .......
Effect of temperature ... ...
Effect of exposure time
Thermal-Convection-Loop Tests of Various Materials
-UF,
-UF,
NoF-ZrF ,-UF
NoF-ZrF4_~UF4

in fluoride fuels

@ u 4 4 % W e Y &

e 3 8 & v & » 8

a 3 s o

------

LI L I I )

nickel inserts

* 8 % 3+ 3 & b0

* 4 » o & w & B A O

B e 8 8 a2 ¥ & e o

4 m 5 a2 5 x4 4w

-----

¢ e -

¢ 4 2 8 & 8 @ ®
« 5 & 5 & 3 & a'e
o » @ . w0
+ 0 % . 0]

------

* & 8 6 & 3 @ © o w & @ 8 3 + e - e & 6]
4 = a 3 o 3 % a @ * & v o2 4 4 . v e 6]
e A 2 e u - 2 s 8 » = . > & 3 4 s o 62
« v o & @ - » v =2 8 e a4 3 a 63
. - L ® 5 8 8 4 4 P 4 0 ® & & 3 o ¥ 63
L T T T T - s . . 63
4 ¥4 & 8 © 50 & 4 8 0 B A e o % 4 B ® B B g ; 64
4.3 8 4 4 & B ¢ o A & ® & e % 7 3 *+ e = w B 65
a % 4 & w 2 0 » & v ¢ 3 0 & % O & b o & 65
66
# P % 0 s 8 8 s @ s % s 8 D & a8 8 3 m & B 66
--------- - « oA 4 v 4 8 @ 66
» o a s » e o o 8 > . 5 66
LI L] s 4 s ? . s b & D B 66
a4 v @ * 2 e A e & m 9 8 o8 b @ v - . 68
s s & 2 a * ¢ & o8 2 e« ¥ 2 o * & o 70
. e "4 o % a3 B e . 2 a3 . - 72
. s 5 s & & 9 8 B . a o a . 72
o A+ 8 4 2 o ¢ @ » ® 4 5 o © & & > & o 74
2 ®w o ¥ a4 " s - - . . a @ ° 74
--------- * * 5 8 »# v oa .« a3 s 76
4 s s e B o oo . ° « e & - 77
4 o 2 % 8 . I ) » % p w8 e 77
+ & 3 b A a . s » o e 2 8 & 3 » 77
----- » a * 2 % p v & * s b . o ?7
a u v » # 0 4 % & 2 @ » L 77
FI a 7 3 & 8 & 3 ° & & v P e b e e = 79
b e " « - " oa e a2 3 s & 3 B » 79
------- . a2 s L A 79
e e e e oo 19
------- ° s » 2 % A e B B e @ 81
ooe 8 ® a2 @ A % 9 2 8 & w - a 84
---------------- » s = ° 85
= 4 % & B & D S A4 % A W B o & a & e a s @ 88
LI T I a v w o # - . v . . s 89
. a s ° a o ®» b © o a o 4 8 o @ « » 9]
CI ] @ * % 4 % & % ¥ F & o w » oa 8 B & 8o ow 9]
f et e RN 21
................ b e 93
® 0 % o P & B A 3 A & B & & & & * s . a : 93
o . + & 3 s 2 8w s ° 08 & ¢ o b & o 93
“ s 4 & & o @ 4 ® & 3w o0 a & o 8 o . 93
o s e * & & 0 * * 2 3 s » s 8 W l 94
B r 4 s A . * * @ e @ e s 8 94
......... . 97
e et e e s e e e cee . 99
e e e e C e e .. 99
--------------------- 2 ® 4 3 & o ¥ 4 . 99
F X
Density and viscosity o o i it i i it i e s e e e e 99
Thermal conductivity o v ittt i ittt it e it ittt ettt e e e e 100
Electrical conductivity oo it i i i it it i et i s i e ittt e e e 100
Vapor PreSSUFES v v v v v v v v v v ot m st st b s s e e ae e e e et e e e 101
Fused-Salt Heat Transfer .. o0 it ittt i i i ittt ettt ettt ennenas 101
Reactor Hydrodynamics .o v o v vt i i i i i e e e e e e 102
Heat-Removal Study of BSF Reactor . .o i it it i ittt it i it ittt ettt e nnanenns 102
Heat Transfer in Circulating-Fue! Reflector-Moderated Reactor . .. o v v ot e i o vt i h e v e 103
Heat Transfer in NaOH-Moderated Circulating Fuel Reactor . ... ... .. vttt i v et vt 103
8. RADIATION DAMAGE . ... it ittt ittt ittt it ettt et n et anennoneneennses 105
Radietion Stability of Fluoride Fuels .. . oo i it it et ittt i it e e 105
LITR Fluoride-Fuel Loop v v v e i ittt i i it i it i ittt e tan st e 107
In-Pile Stress Corrosion and Creep « v v vt v it e s i it te ot ettt naneesonsonenneess 107
Remote Metallography v v i i i i il it it it i et ittt e it et enter et nnenns 107
9. ANALYTICAL STUDIES OF REACTOR MATERIALS .. .. ... it ittt it i i e e 109
Analytical Chemistry of Reactor Materials . ..o . it ittt ittt i ie v e ee e 109
Oxidation states of chromium and uranium in Nquanbase fuels . .. i i e 109
Determination of oxygen in NaZrF -bose fuels . ... ... . oo, 110
Oxidation of trivalent uranium by hexavalent uranium .. . . v i ittt i i ittt it e e s e 111
Stability of trivalent uranium in hydrochloric acid solutiens . . . ... oo it i it e i i i 111
Removal of film from lnconel tubing .. .. .. i i it i i it i it e et ianan 1
Determination of sulfur innatural gas . . o oottt i i it i i i e i et 112
Petrographic Investigations of Fluoride Fuels .. .. . i iiii i, 112
Summary of Sercvie Anolyses ... o it i i i e e i e st e e e 112

10, LID TANK FACILITY o i i i ittt it s ettt a et a s aanensenenens 117
Slant Penetrations of Neutrons Through Water .. .. . . ittt it ie et i et enn 117
Secondary Gamma-Ray Study . o oo it it i i i i e et e e e e e 118

11, BULK SHIELDING FACILITY .. ittt ittt it st e st tnnns st anaesonsnns 121
Gamma-Ray Air-Scattering Caleulations .. . o it ittt it i i it s e e e 121
Thermal-Neutron-Flux Perturbation by Gold Foils inWater . . ... ... it inen. 128
Reactor Power Calibration Techniques . ... vttt ittt ittt nnenrannens 128
Leakage-Flux Changes Due to Control-Rod Settings . . . v v v v vt i ittt et innennnmnnsen, 135

12. TOWER SHIELDING FACILITY .. i ittt i ettt e ittt e e ananeen 136
Experimental Program . .. 0o it i i i it i it i ittt e et et 136
Operation of the Reactor ... ... vivn . P e i e e e s et ee ettt 136

PART IV, APPENDIX
13. LIST OF REPORTS ISSUED DURING THE QUARTER .+ .. it ii i ittt et e e, 141

 

xii
ANP PROJECT QUARTERLY PROGRESS REPORT

SUMMARY

PART I. REACTOR THEORY AND DESIGN

The main pump for the fuel system of the Aircraft
Reactor Experiment was installed, and it was
therefore possible to start the first operational
phase of the experiment — the water test of the
fuel system (Sec. 1). The objectives of the water
test are, in addition to c¢leaning of the system,
determining effectiveness of the pressure fill of
the system, checking tightness of the valves,
determining the flow characteristics of the system,
ascertaining the helium consumption rate, checking
ability to transfer liquid (from one fill tank to
another) while holding a fixed level in the pump
tank, determining operability of the fuel enrichment
system, and checking process instrumentation.
The main sodium pump has also been installed
and the stand-by fuel and sodium pumps will be
installed soon. The fuel-to-helium and sodium-to-
helium heat exchangers which had faulty welds
have been refabricated and reinstalled. Facilities
for loading ‘the sodium, fuel carrier, and fuel con-
centrate, for sampling the fuel, and for unloading
the fuel after the experiment have been or are
being constructed. Operation of a large-scale test
loop for circulating fluoride mixtures to test cor-
rosion and structural stability of ARE components
started June 5, 1954; the duration aim of the test
is 2000 hr. The loop contains an ARE-type pump,
a fuel-to-helium heat exchanger, and two reactor-
core hairpin tubes.

The experimental reacter engineering program
(Sec. 2) has included the development of com-
ponents for in-pile loops, the design of forced-
circulation ‘corrosion-testing loops, and the con-
struction of a unit for testing the mass-transfer
characteristics of a sodium-beryllium-Inconel
system. In addition a sodium-sampling device
was developed, and a method for removing fluoride
mixtures from equipment to be sclvaged was de-
vised, Further tests of bearing and shaft-seal
materials were also made. Development and hot
testing of the vertical-shaft, down-flow sump pump
for operation of the initial in-pile loop in the LITR
were completed, and a model of an air-driven
horizontal-shaft sump pump was fabricated.  In
additional development work on the centrifugally
sealed pump, it was found to be possible to de-
crease the size with no sacrifice in pump per-

formance. - A compact heat exchanger for  the

removal of fission heat from in-pile loops is being
fabricated. An lnconel corrosion-testing loop is
being designed for circulafing fluoride mixtures
that will provide high-velocity turbulent flow and
large temperature differentials. The tests of the
i-Mw, regenerative fluoride-to-sodium heat ex-
changer were terminated and the test unit is being
dismantled for examination. A 100-kw gas-fired
furnoce is being fabricated to determine its suita-
bility as a heat source for future heat-exchanger
tests. If tests indicate that a heat source of this
type will be satisfactory, a 1-Mw furnace will be
developed.

Components of the proposed &60-Mw Circulating-
Fuel Reoctor Experiment (CFRE) are now being
designed and constructed and are to be tested to
determine operational characteristics (Sec. 3). De-
tailed designs will be prepared from the data thus
obtained. A stress analysis of the reactor is being
prepared, and a series of charts has been con-
structed for use in determining temperature distri-
bution and thermal stress. A comparative analysis
of lithium- and zirconium-base fuels was made that
indicates ‘the superiority of the lithium-base fuels
for reflector-moderated reactors,
xenon-poisoning effect in the Reflector-Moderated
Reactor have emphasized the need for removing
the xenon during operation.

Estimates of the

An experiment is
being planned to determine whether adequate
purging of the xenon can be obtained. Calcu-
[ations of a set of 48 reactors are being made that
evaluate the effect of reactor dimensions or con-
centration of U?33 in the fuel, on total U235
investment, on peak-to-average power density in
the core, and on the fraction of thermal fissions
in the core. :

PART Il. MATERIALS RESEARCH

The intensive studies of the fluoride systems of
inferest as reacter fuels were continued with par-
ticular emphasis on systems in which the vranium-
bearing component is the less corrosive UF_ or a
mixture of UF, and UF, rather than UF, alone
(Sec. 4). It appears that attainable concentrations
of UFB in Zer—based fuels are insufficient to fuel
aircraft reactors; however, the use of UF, in ad-
dition finF4 in such mixtures appears to be
promising. The UF, is apparently sufficiently

" @ | o
ANP QUARTERLY PROGRESS REPORT

soluble in NaF-KF-LiF mixtures (which would re-
quire Li’7 for utilization) to make such fuels at-
tractive; the wuse of NaF-RbF-LiF mixtures as
solvents for UF, (without UF } has not yet been
shown to be possible. The elucidation of the
complex NaF-ZrF , system has virtually been com-
pleted, and exploratory phase-equilibrium studies
of the chloride-fluoride systems are under way.
An apparatus for the visual observation of liquidus
temperatures has been tested and found to be
satisfactory. The chemical effects of fission
products on aircraft reactor fuels are being ex-
plored.

The static, seesaw, and thermal-convection loop
facilities were used extensively to further test
the corrosion resistance of various materials in
fluoride mixtures and in liquid metals (Sec. 5).
It has been demonstrated in both seesaw and
thermal-convection-loop tests that Inconel is not
attacked by ZrF -based fuels containing UF,
instead of the customary UF . Since UF, is not
sufficiently soluble in NaF-ZrF | and several other
fluoride mixtures of interest, tests are under way
for determining the amount of UF, required to
eliminate corrosive attack on lnconel by a fuel
which contains UF, and UF .

The effects of temperature and exposure time
have been investigated further and additional con-
firmation of the relationship of these variables
to mass transfer in fluoride-Incone! systems was
obtained. The amount of mass transfer increases
with an increase in temperature or an increase in
temperature difference between the hot and cold
legs of a thermal-convection loop. The previously
reported rapid initial attack and slower secondary
attack after 250 hr of exposure of Inconel to NaF-
ZrF ,-UF , were again demonstrated in a loop oper-
ated for 5000 hr. The depth of attack increases
about 4 mils for each 1000 hr of exposure.

Reductions in depth of attack by UF4-bearing
fluoride mixtures, in comparison with the attack
on standard Inconel, were found in loops con-
structed of Hastelloy B and of a special Inconel
with a portion of the chromium replaced by mo-
fybdenum. In both seesaw and thermal-convection-
loop tests it was shown that, when Inconel and
type 316 stainless steel are exposed in the same
system to o fluoride mixture, the steel is inferior
to Inconel in its resistivity to attack. Two type
316 stainless steel thermal-convection loops filled
with high-purity lithium operated for 1000 hr with

only small amounts of mass transfer and no indi-
cation of plugging, in contrast to those operated
previously, which plugged in 200 to 300 hr. The
increased life was probably due to the higher
purity of the lithium and, especially, to the de-
crease in the lithium nitride content. In the study
of corrosion and mass-transfer characteristics of
materials in contact with liquid lead, it was found
that quartz thermal-convection loops which con-
tained types 410 and 446 stainless steel speci-
mens had much longer life prior to plugging than
similar loops which contained pure iron and chro-
mium or one of the 300-series stainless steels.
Studies are continuing on the identification of
compounds produced by hydroxide-metal reactions.

The metallurgical research effort has been de-
voted to studies of the mechanical properties of
Inconel in contact with fluoride mixtures, to in-
vestigations of materials suitable for high-thermal-
conductivity fins, to searches for container ma-
terials other than Inconel for fluoride mixtures,
and to the development of fabricational techniques
(Sec. 6). Tests have shown that Inconel exposed
to fluoride mixtures has o much longer rupture
life when stressed under uniaxial conditions than
when stressed under the multiaxial conditions that
will prevail in reactor applications. Therefore
tube-burst or other multiaxial-stress tests are to
be emphasized in further studies.

In previous investigations of materials suitable
for high-thermal-conductivity fins for radiators, it
was found that copper fins clad with types 310,
446, or 430 stainless steel or with Inconel were
quite satisfactory in the unstressed condition if
a suvitable diffusion barrier was provided; however,
in oxidation tests under stress it was found that
high stresses greatly increased the oxidation of
the fin material. Experimental work was started
on the preparation of high-boron-content material
for use as shielding between the moderator and
the heat exchanger and between the heat exchanger
and the pressure shell of the Reflector-Moderated
Reactor,

High-temperature oxidation tests of several
brazing alloys have shown that the majority of
the nickel-chromium-base alloys are suitable for
service in an oxidizing atmosphere at 1500°F and
that several of the alloys retain this resistance
at 1700°F. A new semiautomatic heliarc-welding
process for the production of tube-to-header joints
is described and its use in the construction of a
prototype sodium-to-air radiator is illustrated. An
assembly has been fabricated for use in de-
termining the effect of thermal stresses and
thermal cycling on beryllium metal and for studying
the compatibility of sodium and beryllium.

The physical properties of several fluoride mix-
tures and other materials of interest to aircraft
reactor technology were determined, and the heat-
transfer characteristics of reocctor fluids were
studied (Sec. 7). The enthalpies and heat ca-
pacities of the ARE fuel (NaF-ZrF -UF,, 53.5-
40.0-6.5 mole %) and of K CrF were determmed
Density and viscosity meosurements were made
for molten RbF-LiF (57-43 mole %), and thermal
conductivities were measured for molten RbF-LiF
and solid NaF-KF-LiF (11.5-42.0-46.5 mole %).
Electrical conductivity measurements were ob-
tained on molten NaOH over the temperature range
of 625 to 1490°F. A Lucite model of the circu-
lating-fuel Reflector-Moderated Reactor was fabri-
cated and is to be used to study the hydrodynamic
structure in that system. The results of a mathe-
matical analysis of convection are presented for
the case of forced flow between parallel plates
of Hluids with volume-heat sources; this analysis
is useful in estimating the temperature structure
in the flow annuli of refiector-moderated reactor
cores. A study has been initiated to investigate
the heat-transfer and fluid-flow characteristics of
a NaOH-moderated circulating-fuel reactor system.

The radiation domage program included ad-
ditional irradictions in the MTR of fluoride mix-
tures in Inconel capsules, construction of in-pile
circulating-fuel loops, and development of creep-
testing equipment for use in the MTR (Sec. 8).
The Inconel capsules now being irradiated in the
MTR contain UF,- or UF  -bearing fluoride mix-
tures so that additional comparative data can be
obtained on the effect of UF; in decreasing the
corrosiveness of fluoride mixtures. Temperatures
of the capsules are being carefully controlled so
that the Inconel-fuel interface temperature will
remain approximately 1500°F throughout each test.
A re-examination of a group of Inconel capsules
from the earlier irradiations showed that previously
reported excessive grain growth and deep pene-
tration of the Inconel had occurred in only a few
capsules and in those capsules the Inconel-fuel
interface hdd been heated to temperatures muc:h

higher than the desired ISDOOF

' PERIOD ENDING JUNE 10,.1954

The analytical studies of reactor materials in-
cluded the problems of separating UF from U'i‘:4
in NquFs—b‘ase fuels and fuel solvents, the formu-
lation of a method for determining oxygen in these
mixtures, stability tests of trivalent uranium in
hydrochloric: acid solutions, and petrographic ex-
aminations of Zermbased fuels (Sec. 9). In the
studies of the separation of UF, from UF4,
methods were investigated for the conversion of
the fluorides to chloride salts and the simul-
extraction of the chlorides
Petrographic examination of

taneous info non-

aqueous solutions.
NaF-ZrF -UF, fuels reduced with metallic zir-
conium, w:th mefqllac vranium, or with hydrogen
has shown that the predominant phase in the re-
duction complexes obtained is alwagys a solid
solution of U4 and U3 ¥ in Nog Zr F

8 471°
PART Hl. SHIELDING RESEARCH

The Lid Tank Facility has been used primarily
for studying special attenuation problems which
arise in aircraft shield design (Sec. 10). The
crew-shield plastic sides attenuate nevtrons which
arrive at slant incidence, and therefore experi-
ments are being carried out to measure the effecy
of slant incidence on the attenuation. At 60 deg
to the normal, the short-circuiting effect becomes
quite important. The gamma-ray attenuvations in
lead and bismuth have been compared as a function
of the neutron flux at the metal in a metal-water
shield. A first study of the data shows little
difference between the two materials, but more
work is to be done to clarify this point.

The work at the Bulk Shielding Fqcrhfy has
included gomma-ray air-scattering calculations, the
determination of the thermal-neutron-flux pertur-
bation by gold foils in water, a study of reactor
power calibration techniques, and a determination
of the effect of control-rod settings on leakage
flux (Sec. 11). The calculation of air scattering
for the divided shield with a lead shadow shield
was carried out some time ago for gamma rays by
using the spectral data obtained on the reactor
shield mock-up at the Bulk Shielding Facility,
The data have now been extended to include at-
tenuation by the crew shield so that a direct
comparison is possible with earlier Shielding
Board calculations. For scattered radiation there
is good agreement, but the leakage around the
shadow shield was badly underestimated in the
ANP QUARTERLY PROGRESS REPORT

early work., This will require redesign of the
shield configuration but will not introduce an ex-
cessive weight penalty. The correction for flux
depression by detector foils is known for indium
foils, but since much work is done with gold foils,
the determination of flux depression in water has
been extended to this element. In connection with
intercalibration of the several shielding reactors
(Bulk Shielding Facility and Tower Shielding Fa-
cility reactors at ORNL and a reactor at Convair),
it has been demonstrated that bare-foil activation
gives a reliable measure of reactor power density.
The effect of sofety and control-rod positions on

fast-neutron leakage from the reactor has been
detetmined experimentally for use in correlating
data taken on two otherwise similar reactors.
Fortunately, the effect appears to be small, as is
the effect of small changes in fuel loading.

The Tower Shielding Facility reactor was made
critical for the first time at the site on March 12
(Sec. 12), Since that date much time has been
spent in determining the operational character-
istics of the equipment, and the reactor now oper-
ates regularly in the air at powers of up to 4 kw.
It will soon be operated at up to 100 kw, the
design maximum power level,
Part |

REACTOR THEORY AND DESIGN
1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT

E. S. Bettis

J. L. Meem

ANP Division

The first operational phase of the Aircraft Re-
actor Experiment is now under way. With the
installation of the main fuel pump, it was possible
to start the water test of the fuel system. The
water was charged to a fill tank and then forced
by gas pressure into the system. Subsequent
circulation of the water by the fuel pump carried
the gas from pockets in the system, which will
not pressure fill, to the pump tank where the water
degassed. The system probably became gas free,
that is, full of liquid, but since there is no posi-
tive indication of fullness, vacuum filling of the
system is being considered. Data on the pressure
and flow characteristics of the system are being
obtained.

The main sodium pump has also been installed
and the stand-by fuel and sodium pumps will be
installed soon. The fuel-to-helium and sodium-to-
helium heat exchangers which had faulty welds
have been refabricated and reinstalled. Minor
modifications were made during refabrication that
are expected to improve the performance of these
heat exchangers. The thermal-barrier doors, which
gre lowered to permit preheating of the heat ex-
changers to above the melting point of the fuel
or the sodium, were modified to eliminate thermal
buckling. Operation of the helium supply was
checked, and the arrangements were made for
assuring an adequate supply of helium for oper-
ation. The original fuel-injection system, which
was dependent on the operation of valves, was
modified to permit injection of the fuel concentrate
hy a more reliable gas-pressure equalization tech-
nique. Facilities for loading the sodium, fuel
carrier, and fuel concentrate, for sampling the
fuel, and for unloading the fuel after the experi-
ment have been or are being constructed. Plans
hove been made for a leak check of the fuel
system at operafting temperature. Helium spiked
with krypton will be introduced into the fuel
system and the helium in the annulus around the
fuel piping will be checked for krypton.

A large-scale test loop for circulating fluoride
mixtures is being operated to test corrosion and
structural stability of ARE components. The loop
contains an ARE-type pump, a fuel-to-helium heat

exchanger rebuilt from one originally purchased for
the ARE, and two hairpin tubes purchased as
spares for the ARE reactor core. The system is
operating isothermally at about 1375°F with a flow
rate of 20 gpm. The duration aim of the test,
which started June 5, 1954, is 2000 hr.

THE EXPERIMENTAL REACTOR SYSTEM

W. B, Cottrell J. Y. Estabrook
J. K. Leslie
ANP Division
G. J. Nessle J. E. Eorgan

Materials Chemistry Division

G. A, Cristy E. Wischhusen

Engineering and Mechanical Division

Pumps.

The main fuel pump and the main sodium pump
have been installed in their respective systems.
The stand-by sodium pump, which is on hand, and
the stand-by fuel pump will be installed in the
immediate future. All these pumps have been
satisfactorily hot checked with sodium at 1300°F
for 100 br.. '

Operation of the main fuel pump with water re-
vealed that the maximum oil flow in the lubricating
system was too low (~ 2 gpm) at high pump speeds
(~ 1800 rpm). Consequently, the auxiliary lubri-
cating pump, now a single-phase Chempump, will
be paralleled with a three-phase Chempump, and
the single-phase pump will be heid in reserve.
Tests with the three-phase Chempump show thot
the oil flow is ample (>3 gpm) at all speeds of
the primary pump. :

Heat Exchangers

Disassembly, refabrication, and reinstallation of
the fuel-to-helium heat exchangers and the sodium-
to-helium heat exchangers which had faulty welds
have been completed. The welding was com-
pleted by the two qualified ARE welders dhead
of schedule. There were two significant design
changes in the fuel heat exchangers as refabri-
cated. FEach fuel heat exchanger now has three
parallel passes, rather than two, in series with
ANP QUARTERLY PROGRESS REPORY

three similar paralle! passes. Also, a small by-
pass line was eliminated to simplify fabrication
of the fuel heat exchangers. Without the bypass
fines, however, the heat exchangers are not com-
pletely drainable.

Two insulated and heated barrier doors are pro-
vided for each heat exchanger to permit preheating
of the heat exchanger to above the melting point
of the fuel or the sodium before these liquids are
loaded into their respective systems. These
doors, which are lowered within the closed duct
to block the flow of helivm through the heat ex-
changers, were tested at operating temperature
(1300°F) before the heat exchangers were removed
for refabrication. The doors stuck at temperatures
above 900°F, and therefore they have been modi-
fied to eliminate thermal buckling. These doors
have not yet been retested at operating tempera-
ture because of the water tests now in progress.

Fluid Circuits

The major fluid circuits (fuel and sodium) have
been virtually completed for some time. The re-
welding of some pipe lines necessitated by the
removal and reinstallation of the heat exchangers
has been completed. Except for the bypass around
the reactor in the sodium system, all fuel and
sodium piping is completely welded, instrumented
(thermocouples located every 3 ft), heated, and
insulated. The heafer electrical connections are
essentially complete, but a substantial number of
thermocouple connections remain to be made.

An operational check of the helium supply system
disclosed o few leaks which have been repaired.
The helium supply problem has been resolved.
Three railroad tank cars have been allocated for
ARE use so that a helium tank car should be on
hand ot all times. Arrangements have been ef-
fected to have two helium trailer trucks (capacity
20,200 scf) on hand at all times to transport the
helium from the tank cars to the ARE building.
The bank of 12 helium cylinders (capacity 2640
scf) held in reserve at the experiment will be
supplemented by a bank of 15 larger cylinders
(capacity 16,000 scf), which are now being fabri-
cated.

Fuel-Enrichment System

The original fuel-enrichment system provided
for the injection of the fuel concentrate into the
system below the liquid-gas interface in the pump

tank. Since such an arrongement is inherently
dependent upon the operation of the system
valves, freeze valves were inserted in the ¥%-in.
transfer lines to back up the bellows valves in
the system, which are of questionable reliability
at the temperatures involved (1300 to 1400°F).
These freeze valves were merely sections of
tubing from which the heat could be removed so
that the liquid fuel inside the tubing would freeze.
However, in tests of similor freeze sections the
3/8-in. tubing frequently ruptured upon thawing,
and this injection technique could not be con-
sidered sufficiently reliable.

It was therefore decided to effect transfer of
the fuel concentrate into the fuel system by a
gas-pressure equalization technique. In order to
employ this technique, only two modifications to
the existing system were necessary. First, the
transfer line had to be installed so that the fuel
concentrate would be injected in the pump above
the liquid-gas interface, and second, gas equalizer
lines, both with gas valves, had to be provided
from the storage tank to the transfer tank and
from the transfer tank to the pump. The concen-
trate will now be injected through the pump flange
into the pump tank. Since the pump flange is
maintained at 600°F, it was necessary to develop
a special resistance-heated annular fitting which
could attain a temperature of 1400°F in the center
fuel passage.

The transfer tank, originally fabricated of type
316 stainless steel, has been refabricated of In-
cone! to conform to the rest of the system., The
fuel-enrichment system will be tested with water
as soon as the transfer tank is installed and
connected.

Loading Facilities

Facilities for loading the sodium, fuel carrier,
and fuel concentrate into their respective con-
tainers have been, or are being, fabricated. Each
of these facilities will effect transfer of the liquid
from a portable container into the tank provided
for it at the ARE. The transfer will be effected
by gos pressure in each case.

The sodium, which will assay less than 0.02

wt % oxygen, will be received in four 55-gal
drums. The fuel carrier, NaZer, will be received
in 13 cylinders, each containing approximately

250 Ib of the fluoride. The fuel concentrate,
Na,UF,, will be received in 15 cylinders, each
weighing approximately 75 1b. All these materials
are on hand. In all cases the fluid to be trans-
ferred will be maintained under helium pressure.
The transfer temperatures are 600, 1200, and
1300°F for the sodium, carrier, and concentrate,
respectively. ' :

.FueI-Scmpling Facility

It is anficipated that from 5 to 10 samples of
fuel will be drawn from the ARE between the time
the fuel system is filled with fuel carrier and the
time the reactor is operated ot an appreciable
power level. For the chemical analyses to be
performed on these samples, only 10 to 20 g of
material are required; however, to ensure that the
sample is actually representative of the flowing
fuel, at least 300 to 400 g of material must be
withdrawn.  Equipment for accomplishing this
sampling operation has been designed, fabricated,
and successfully tested. '

The sampling apparatus consists, as shown in
Fig. 1.1, of a nickel pot containing a machined
graphite liner and a flanged top through which

PERIOD ENDING JUNE 10, 1954

pass the fluoride entrance tube, three electrical
contacts, the gas manifold connection, and a lever
for moving the grophite-lined sample cup under
the flucride entrance tube. This apparatus will
be connected to the system by a heated 3{g-in.
Inconel line containing a bellows valve and will
be placed in the reactor pit at a level 4 to '8 ft
above the highest fuel level in the ARE system.
With the sampling cup out of the fuel delivery
stream, a vacuum will be drawn on the nickel pot;
when sufficient fuel has been drawn into the
graphite liner to contact the lowest probe, the
sample cup will be swung under the delivery
When the sample cup has overflowed
sufficient fuel to contact the second probe, an
atmosphere of helium will be admitted to stop the
flow of fuel. As soon as the apparatus is suf-
ficiently cool to handle, it will be disconnected
and replaced by an identical apparatus. It is
anticipated thot sampling can be accomplished
without closing the valve in the sample line and
without resorting to the use of freeze valves in
the delivery lines. '

stream.

 

Fig. 1.1. Fuel-Sompiing Apparatus.
ANP QUARTERLY PROGRESS REPORT

Four complete tests of the apparatus have been
made with four different values for pressure above
the fuel and inside the sampling apparatus. These
tests, in which the sampling apparatus was not
heated, were satisfactory; it is anticipated that,
unforeseen complications in the ARE,
the fuel! circuit should proceed

barring
sampling of
smoothly.

Fuel-Unloading Facility

The radioactive fuel mixture will be discharged
from the hot-fuel dump tank through a l-in. pipe
to aluminum cans for removal to fuel recovery
and reprocessing equipment. A mock-up of the
unloading equipment was constructed and tested
with the fuel carrier NaZrF . The feasibility of
unloading the ARE with equipment of this type

has been satisfactorily demonstrated.,
Reactor Control

The control actuator assembly has been mounted
over the reactor, and the shim and confrol rods
have been installed obove the reactor. Accurate
measurements have been made from the bottom of
the reactor to assure proper location of the rods.
The fission-chamber circuits have been checked,
and the chambers are now being installed. In
order to prevent a chamber from catching where
the sleeve changes from 3 in. to 2]/2 in. in di-
ameter, a tapered transition section will be pro-
vided. Except for the final check on the ion
chambers, the nuclear instrumentation has been
completely checked. The actual installation of
the ion chambers and shielding plugs will be
accomplished just before the fuel carrier is loaded
into the system.

FUEL SYSTEM MOCK-UP TESTS

H. W. Savage A. G. Grindell
W. G. Cobb W. R. Huntley
ANP Division

The complete filling of the ARE fuel system
would be most easily accomplished with the
system under vacuum, but the problems involved
in obtaining and holding vacuum are formidable,
Therefore, a mock-up was constructed and oper-
ated to test filling and degassing characteristics
under a helium pressure greater than atmospheric.
The mock-up was also used to test the drainability
of the circuit.

The test unit was an essentially full-scale
mock-up of the ARE fuel system without the heat

10

exchangers. Included were a reactor core, a pump,
a fill tank, piping, and valves with which to dupli-
cate ARE opressure drops. The fill tank, core,
and pump were set ot levels corresponding to the
respective levels in the ARE building. The
working fluid used was tetrabromoethane becouse
its room temperature density and viscosity (sp
gr = 2.95 un =93 cp at 77°F) are similar to those
of the fluoride fuel at operating temperature. A
helium blanket at 20 to 30 psig was maintained
in all tests,

The primary objective of the tests was to quanti-
tatively determine the flow rate required to com-
pletely fill the reactor core and sweep all gas
bubbles from it. A secondary objective was a
determination of the degassing and self-priming
characteristics of the pump.

In all the tests, flow rates of 60 to 64 gpm were
required for filling all six parallel circuits in the
core and sweeping out all gas bubbles. Since
the tubes in the mock-up were 1.00 in. in inside
diameter and the ARE core tubes are 1.115 in.
in inside diameter, it is estimated that flows of
85 gpm may be required for filling the ARE. In
the mock-up it could be ascertained by visual
observation that all gas had, been removed from
the core, but visual observation will not be pos-
sible in the ARE.

The pump degassed well in all tests. In no
case was there sufficient surface turbulence to
endanger the probes or gas lines by splashing (the
liquid surface was visible through glass portholes
in the pump casing). In each filling operation,
the pump lost its prime several times but regained
it each time without adjustments of the controls,

In the drainability tests, gravity draining unas-
sisted by gas pressure resulted in approximately
25% removal of liquid from the reactor core.
Helium-pressure-assisted drainoge (15 psi differ-
ential between pump and dump-tank pressures) re-
sulted in @ maximum of approximately 60% removal.

OPERATION OF THE FUEL SYSTEM

R. G. Affel W. B, Cottrell
G. D. Whitmon
ANP Division

The installation of the main fuel pump es-
sentially completed the fuel system, and the water
test of the system is now under way. While this
represents attainment of the first operational phase
described in ‘the ““ARE Operating Procedures,””’
the immediate objectives of the water test are
(1) cleaning of the system, (2) determining ef-
fectiveness of the pressure fill of the system,
(3) checking tightness of the valves, (4) de-
termining the flow characteristics of the system,
(5) ascertaining the helium consumption rate, (6)
checking ability to transfer liquid (from one fill
tank to another) while holding a fixed level in
the pump tank, (7) determining operability of the
fuel-enrichment system, and (8) checking process
instrumentation.

Cleaning :

In order to remove the scale of calcium meta-
silicate now deposited throughout the system as
a result of previous tests with water and Conklene
(o detergent — sodium metasilicate), the system
will be washed with o 1% solution of the tetra-
sodium salt of ethylenediaminetetroacetic acid in
water. According to the analytical results (cf.
Sec. 9, "Analytical Studies of Reactor Materials’’),
this solution will dissolve the calcium deposits
and leave no residue when it, in turn, is flushed
out of the system. |t was recommended that the
solution be circulated for 4 hr in the system at
170°F. Furthermore, because of the significant
holdup volume of the system (about 1 cu ft of each
6 cu f1), the rinsing step should be repeated six
or seven ftimes to ensure complete removal of the
washing agent. To date, the initial cleaning and
rinses hove been effected, and the water
shows the expected decreasing concentration of
impurities. = The water for the second rinse is
being used in the operational tests described
above., '

two

Pressure Filling

The water was first charged to the fill tank and
then forced into the system by gas pressure. Al-
though the system (reactor and heat exchanger)
contained gas pockets which would not pressure
fill, subsequent circulation of the water by the
fuel pump carried the gos pockets around to the
pump tank where the water degassed. Although
it is probable that the system thus became gas
free (full of liquid), there does not appear to be

e —ma— e

 

IW. B. Cottrell, ARE QOperating Procedures, Parts [,
1, and 1H, ORNL C¥F-54-2-68 (Feb. 11, 1954).

PERIOD ENDING JUNE 10, 1954

any positive indication available, The gas re-
maining in the system is of little consequence in
water operation, but it would be
operation with fuel.  Unless operational pa-
rameters permit a more positive defermination of
the extent’' of filling,

vacuum fill the system.

intolerable in

it may be necessary to

Vacuum filling of the
system has thus far been avoided because oxygen
could conceivably enter from the vent system and
contaminate the hot fuel. ‘

Pressure and Flow Characteristics

Many data have been taken on the pressure and
flow characteristics of the fuel system. It is
apparent that the flow through the bypass (in
parallel with the heat exchangers) is less than
that through the heat exchangers. This is un-
doubtedly caused by the pressure drop across the
bypass throttling valve being greater than an-
ticipated and that across the hear exchanger being
less. The pump speed and system pressure-drop
data do not check with the system flow data. |t
is most probable that the flowmeter (a Rotameter
which fransmits a signal from a coil within which
moves a tapered iron core attached to the Ro-
tameter bob) is not calibrated correctly. This
possibility is being checked. The Rotameter has
shown a tendency to drift, and it has not been

- possible therefore to moke occurate measurements

of the system flow chamcterfisfics.

Functional Tests

In the operation of the system to date, the dump-
line valves appear to be tight, although there is
some indication that the valves to individual tanks
leak slightly under pressure. Such slight leaks
are tolerable, although not desirable, and the leak
tightness of the valves may improve upon repeated
operation.

The helium consumption rate during the test
with water in the fuel system has been about
3 cfm. It is therefore probable that the maximum
helium consumption at any time during the ex-
periment will be about 10 ctm.

It was found to be possible to transfer water
from one fill tank to another through the system
and to maintain o fixed liquid level in the pump.
This operation could be of significance should
it be desirable to replace the initial batch of
carrier and/or to replace the hot fuel at the con-
clusion of the test,’

11
ANF QUARTERLY PROGRESS REPORT

Hot-Gas Test

Although it would be desirable to leak-check
both the fuel and the sodium systems at operating
temperature, there is no completely satisfactory
method for doing this except with fuel and sodium
because the systems are not completely drainable.
However, the inadequacy of u cold check in com-
parison with a hot check is so apporent that a gas
check of the fuel system at the operating tempera-
ture will be effected. Basically this test will
consist of heating the system to 1200°F, filling
the fuel system with one gas under pressure,
tilling the fuel-system annulus with another gas
at lower pressure, and spectrographically ana-
lyzing the gas in the annulus for the presence of
the gas from the fuel system. Although in prin-
ciple any two gases would suffice for this test,
it would be impossible, for example, to detect a
small leak of helium into nitrogen because of the
background of acir in the annulus system. Ac-
cordingly, helivm spiked with krypton will be used
in the fuel system, and the helium in the annulus
will be examined for krypton.

PUMP FABRICATION AND TESTS

H. W. Savage W. G. Cobb
A. G, Grindell W. R, Huntley
R. E. MocPhersen
ANP Division

P. Patriarca G. M. Slaughter
Metallurgy Division

All six ARE-pump rotary elements have been
assembled and were found to meet hot-test re-
quirements in tests made by using the two im-
pellers which have been accepted. Four of the
six rotary elements have passed all acceptance
tests.2 The other two rotary assemblies were
rejected because of unsatisfactory upper seals
which permitted excessive gas leokage. Re-
lapping of the seals resulted in no improvement
because the lapped surfaces were scored in sub-
sequent testing. It was thought that the lapping
abrasives (1500-grit diamond dust, and others)
were not being completely removed from the soft
nose of the upper seal. As o check on this possi-
bility, the adequacy of lapping with water-soluble

 

24, W, Savage et al., ANP Quor. Prog. Rep. Mar. 10,
1954, ORNIL-1692, p 10.

12

scouring compounds (for example, Bon Ami) is
being evaluated.

One of the accepted elements, with a cast im-
peller {to be exchanged for a fabricated impeller
before start-up of ARE), has been installed in the
fuel system, and one was installed in the sodium
system. One accepted element has been assigned
to K-25 for use in o heat-exchanger test. One
unaccepted element is installed in the cold shake-
down test unit for seal tests., The other unac-
cepted zlement is being used in the hot shakedown
test unit for hot-testing fabricated impellers. Two
fabricated impellers have passed all acceptance
tests,® and one has been furnished to K-25 for
the heat-exchanger test,

Shorting of the probes used for liquid-leve! indi-
cation in the sodium pump has shown the clear-
ance between the riser wall and probe to be
inadequate. The original design called for a
3/8-in. schedule-40 pipe riser which allowed bridg-
ing of sodium condensate between the inner pipe
wall and the 3/32-in.—dia probe., The size of the
riser was increased to 1/2 in, and no shorting of
the probes occurred during subsequent festing.
All the sodium pumps have been modified to
utilize ]/:’L,-in. risers.

The operation of an ARE-type d-c pump drive
motor in a dry helium atmosphere continued un-
eventfully during the quarter. Approximately 3500
hr of operating time have been accumulated at a
test temperature of 130 to 140°F. This test will
be terminated at 4000 hr,

The detail and assembly drawings of the pumps,
made in the fall of 1953, are being revised to
incorporate the modifications in the pump cooling
and lubricating systems and other modifications.

The fabrication of five [nconel impellers for
ARE pumps by machining and welding has been
completed. The rough-machined parts are shown
in Fig. 1.2. These parts, except the vanes, were
subjected to a stress-relief annecaling heat treat-
ment so that precise dimensional tolerances could
be maintained during subsequent machining. The
stress-relief thermal cycle consisted of heating
at 125°C/hr to a temperature of 1000°C, main-
taining this temperature for l/2 hr, and cooling at
125°C/hr to room temperature.

The drive-shaft hub was then tack welded to
a carbon steel strong-back to prevent warping
during welding, and the vanes were heliare welded

 

3ibid., p 11,
PERIOD ENDING JUNE 10, 1954

into place, as shown in Fig. 1.3. The stress- sition, and the joints between the hub and the
relief heat treatment was then repeated. After  vanes were heliarc welded as far as they were
the vanes had been machine finished, the fluid- accessible, as shown in Fig. 1.4, The stress-
entrance hub was heliarc plug-welded into po- relief heat treatment was again repeated. ' The

%, D ;w.
i
o
il

.

 

 

AN AN D
Yoy

  

Fig. 1.3. Impeller After Heliarc Welding of ‘
Yanes to Drive-Shaft Hub. Impeller is tack welded Fig. 1.4. Impeller with Fluid-Entrance Hub

to carbon steel strong-back., Plug-Welded into Position,

13

 
ANP QUARTERLY PROGRESS REPORT

welded impeller was then removed from the strong-
back and machine finished. The completed as-
sembly is shown in Fig. 1.5,

UNCLASSIFIED
Y107 9a

 

Fig. 1.5. Completed Impeller.
REACTOR SYSTEM COMPONENT LOOP

H. W. Savage A. G. Grindell
W. G. Cobb W. R. Huntley
ANP Division

The services of the K-25 Technical Division
have been secured for constructing a large-scale
system for circulating fluoride mixtures and testing
reactor-system components. The major components
of the system are the pump (mentioned above), a
fuel-to-helium heat exchanger rebuilt from one
originally purchased for the ARE, and two hairpin
tubes purchased as spares for the ARE reactor
core (Fig. 1.6). The system is operating iso-
thermally at 1375 + 25°F, with o flow rate of
20 gpm. The duration aim of the test, which
started June 5, 1954, is 2000 hr. When the oper-
otion is terminated, the heat exchanger and hairpin
tubes will be examined and evaluated in terms of
corrosion, structural stability, etc.

FUEL RECOYERY AND REPROCESSING

D. E. Ferguson G. |, Cathers
Chemical Technology Division

The chemical process currently being con-
sidered* for recovery and decontamination of 235
from the ARE fuel mixture NaF-ZrF ,-UF , consists
of dissolution of the fuel in an aqueous aluminum
nitrate—nitric acid solution, solvent extraction
with 5% tributyl phosphate in a kerosene diluent,

 

4p, E. Ferguson, G. I. Cathers, and O. K. Tallent,
A!‘]J.g] Quar. Prog. Rep. June 10, 1953, ORNL-1556,
[ .

14

and isolation on an ion-exchange resin column.
In one solvent-extraction c¢ycle a gross beta-
decontamination factor of greater than 10* and
vranium recovery of greater than 99.9% will be
obtained.

Additional work has now been completed on
dilute-TBP extraction of uranium solutions having
high nitrate-salting strength, In batch counter-
current extraction experiments with six TBP con-
centrations in the range 1 to 30 vol %, with feed
nitric acid concentrations of 0.5 and 3 M, ru-
thenium and total-rare-earth beta-activity decon-
tamination factors improved with decreasing TBP
concentration in the solvent (Fig. 1.7). The
anomalous zirconium ond niobium beta decon-
taminations obtained affected somewhat the gross-
beta decontamination factor determinations in the
low TBP concentration range. However, the best
gross beta decontamination was obtoined at a
TBP concentration of 7.5% or less regardless of
acidity. There was some evidence of o maximum
in gross beta decontamination at 5 to 7.5 vol %
TBP with a feed acidity of 0.5 M. Maximums were
also observed in the zirconium and niobium decon-
tamination factors at a low acid concentration.

Three extraction and scrub stages were used,
with stoichiometrically neutral aluminum nitrate
solution as the scrubbing agent. The salting
strength was controlled by varying the aluminum
nitrate content to obtain a uranium distribution
coefficient of about 4 at the feed plote and 0.5
at the last scrub stage. Figure 1.8 is a plot of
the distribution coefficients of the various ac-
tivities at the feed plate. These data could be
correlated, in general, with the conclusions drawn
from Fig. 1.7.

Testing of the data by log-log plots (Figs. 1.9
and 1.10) showed that the ruthenium and total-
rare-earth distribution coefficients have close to
second-order dependence on TBP concentration at
constaont uranium distribution, The equilibrium
constant for this case is represented by the
equation

D'C"(Ru Bor TRE f)

(TBP conc)”®

 

where n, the dependence, is equal to 2. A least-
squares analysis showed the average dependence
for all data in the case of ruthenium activity to
be 2.10 and in the cose of total rare earths to be

1.95.
PERIOD ENDING JUNE 10, 1954

UNCLASSIFIED
ORNL-LR-DWG 1688

 

 

PUMP

 

 

 

Lo

 

PUMP TANK

 

 

 

 

   
 

O PIPE

 

 

= sosmafin-

 

 

I

=

!
~

 

 

 

HEAT EXCHANGER

 

 

D)

 

 

iy

 

 

 

 

 

    
 

VENTURI

W HAIRPIN TUBES

 

H
!
Il FiLL AND DRAIN
l1 TANK
B

 

 

 

Fié. 1.6. Fuel#o-Helium Heat-Exchanger Test Loap.

15
ORNL-LR-DWG 1683

ANP QUARTERLY PROGRESS REPORT

 

ORNL.- L'-DWG 1620

       

 

 

)

 

 

 

 

 

 

 

%
|
0]

3

CONCENTRATION OF TBP IN AMSCQ 125

 

 

cs ,

- [
Lozr B 1
10

0.5 M NITRIC ACID
—— 3.0 ¥ NITRIC ACID

 

  

 

 

 

 

 

 

 

 

 

CONCENTRATION OF TBP IN AMSCO 125-90W (vol %
Fig. 1.7. Decontamination vs TBP Concentration in Solvent Extraction.

B =

 

 

 

 

=+ GROSS

S

 

 

 

 

 

 

 

 

 

-—

801074 NOILYNIWVYLINOD3Q

%)

-3QW {vol

Fig. 1.8. Activity Distribution Coefficients at Feed Plate vs TBP Concentration in Solvent Extraction,

16
PERIOD ENDING JUNE 10, 1954

   
   

 

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 . ORNL-LE-DWG 1692
10" oot I 1T
- 2 g ool AR I -
= o F e
L i@ L (o 05 M NITRIC ACID
-3 e, 30 M NITRIC ACID
% = Sz .
z 2 88
S P & 110
g 0 ¥ a0 e T R AT T
o & H [ £E= 5 CootnInh
. LT Tl A Bo 5 e T e e
&g : . LT 8
g < ; - /] - T 05 M NITRIC ACID | =
oo - <L L b LT I,,,J o< 2
G | ; | ]T ‘ i i ‘ o
{ b ,,.fi,..__._._.'_i.l,fif_,. . i ,.,,,,_,, ____!._}__ ! ___“,!, : q - i -
1075 2 5 %2 5 072 5 0?2 s o 1070 2 s 0% 2 5 107 2 5 wF g 5 107
DISTRIBUTION COEFFICIENT OF RUTHENIUM ACTIVITY DISTRIBUTION COEFFICIENT OF TOTAL RARE EARTH ACTIVITY
Fig. 1.9. Dependence of Ruthenium Activity on Fig. 1.10. Dependence of Total-Rare-Earth Ac-

TBP Concentration. & tivity on TBP Concentrotion,

17
ANP QUARTERLY PROGRESS REPORT

2. EXPERIMENTAL REACTOR ENGINEERING

H. W, Savage
ANP Division

Development of components for in-pile radiation-
damage test loops has continued. Major emphasis
was placed on completing development and hot
testing of the vertical-shaft, down-flow sump pump
for operation of the initial in-pile loop in the LITR
and on the design and fabrication of the first hot-
test model of an air-driven horizontal-shaft sump
pump. Further development work has been done
on the centrifugally sealed pump, and it has been
possible to decrease the size with no sacrifice
in pump performance. Tests of hydraulic motors
for driving in-pile pumps are being conducted. A
compact heat exchanger has been designed for
removal of fission heat from in-pile loops. Fabri-
cation of the first exchanger is in progress.

Forced-circulation corrosion-testing loops are
being designed for obtaining information on the
corrosion of Inconel in high-velocity turbulent fluo-
ride mixtures with large temperature differentials
in the system. A test unit for ascertaining the
mass-transfer characteristics of a sodium-be-
ryllium-Incons| system was constructed and is
now operating with turbulently flowing sodium and
a high temperature differential,

Tests of the 1-Mw, regenerative fluoride-to-
sodium heat exchonger were terminated and the
test unit is being dismantled for examination. A
100-kw gas-fired furnace is being fabricated for
testing to determine its svitability as a heat
source for future heat-exchanger tests. If tests
indicate that a heat source of this type will be
satisfactory, a 1-Mw furnace will be developed.

A sampling device for taking an unrestricted
number of samples of molten sodium from high-
temperature systems was completed and several
samples were successfully taken. Also, a satis-
factory solvent for removing fluoride mixtures from
tnconel was developed, tested, and put into
general use.

Several  bearing-material were
tested in molten fluorides under strictly controlled
conditions. {t was found possible to screen out
some combinations, while others showed sufficient
promise for further testing.

combinations

 

'W. B. McDonald et al.,, ANP Quar. Prog. Rep. Mar,
10, 1954, ORNL-1692, p 15.

18

IN-PILE LLOOP COMPONENT DEVELOPMENT

W. B. McDonald
ANP Division

Centrifugally Sealed Pump

J. A. Conlin
ANP Division

The deaeration and filling problems encountered
with the previously described! centrifugally sealed
pump have been essentially solved, und a series
of modifications have been made for increasing
pump performance and decreasing the over-all pump
dimensions. Some success in reducing pump di-
mensions has been achieved, but no change in
head and flow characteristics has been obtained.

A satisfactory method of deaeration was es-
tablished by bleeding off 5 to 10% of the total
pump flow from the pump discharge and returning
it to the centrifugal-seal cavity. The centrifugal-
seal cavity octs as a centrifuge for removing the
gas from the fluid and replacing the bleed-off with
clear fluid from the seal cavity., Deaeration at a
low developed head is, as far as can be de-
termined, complete. However, as the develeped
head is incregsed by increasing pump speed, de-
aeration becomes less satisfactory; in fact, some
aeration occurs slowly when the pump speed is
increased ofter complete deaeration at low speed.
At high speed the gas pressure in the seal cavity
is above the pump suction pressure by about 65%
of the developed head. After an investigation it
was concluded that gas (in this case, air) is
dissolved in the fluid (water) under pressure in
the seal cavity ond then comes out of solution
in the form of small bubbles at the low pressure
(pump suction) portion of the loop. However, the
mechanism of aeration of the water is not defi-
nitely known. It is possible that the turbulent
surface of the liquid in the seal cavity af high
pump speed may increase the rate of solution of
the oir in the water or even entrap air in the form
of bubbles which then pass into the system through
the seal. The precise degree of aeration hos not
been measured, but it is not very great, and, of
course, the extent of aeration which would occur
with fluoride mixtures under a helium gas is not
known.

The introduction of the degeration bleed into
the seal cavity caused the seal to leak when fluid
was added to the system. This seal leakage was
corrected by the addition of a deflector to keep
the bypass fluid away from the junction between
the inner tube in the seal cavity and the pump
impeller.

In its present state, the pump may be started
empty in a horizontal position and filled by simply
adding water to the system, with care being taken
not to overfill the system. A series of tests is
now under way for developing filling and operating
techniques that will be satisfactory for operation
of the pump with fluoride mixtures.

in an attempt to improve pump performance, an
impeller that had been modified by drilling the
flow ports tangentiolly 1o the core was tested.
However, the resulting performance was somewhut
worse instead of better.

A pump casing with an unconventional dischorge
passage was tested in an attempt to reduce the
over-all diameter of this or any similar pump. The
modified casing (Fig. 2.1) has a narrow annular
passage around the outside diameter of the im-
peller and a wvolute-like passage formed in an
axial direction from the annulus. The volute turns,
at the water deflector (cut-water), into a passage
which discharges in an axial direction parallel
to the pump inlet at a maximum distance from the
axis of the pump no greater than the outside di-
ameter of the annulus around the impeller.

Performance of this casing was slightly befter
than that of the casing used previously which had
a conventional annular coliector ring and a tan-
gential discharge nozzle. However, the results
were inconsistent because of the geration con-
ditions, and it is believed that under identical
conditions an equally well-designed conventional
casing would show slightly better performance than
that of the modified casing, The principal ad-
vantage in the axial-discharge design is that the
discharge nozzle does not extend radially beyond
the pump casing. However, this advantage must
be balanced by the greater difficulty of fabricating
such a casing. The velocity-to-head conversion
could probably be improved by modifying the cut-
wafer to extend over the impeller and thereby
obtain more benefit from the high fluid velocity
in the annulus.

PERIOD ENDING JUNE 10, 1954

UNCLASSIFIED
ORNL-LR-DWG 1693
DISCHARGE PASSAGE —

CLUTWATER —
\(flf

CASING -

 

 

IMPELLER

Fig. 2.1, Modified Axial-Discharge Pump Casing.

Horizontal-Shaft Sump Pump

D. F. Salmon
ANP Division

A horizontal-shaft air-driven sump pump for hot
testing with a fluoride mixture is being assembled
(Fig. 2.2). A plastic model of the pump was tested
with 1-1,2-2 tetrabromoethane (sp gr ~ 3, u ~ 9 cp).
Priming was accomplished by periodic sudden ap-
plications of pressure in the sump to force slugs
of liquid into the pump inlet. A longer time was
required to degas this liquid than water. For
faster degassing, part or all of the flow was
directed through the sump instead of directly into
the pump inlet.

Vertical-Shaft Sump Pump

D. R. Ward
ANF Division

Two vertical-shaft sump pumps' were tested with
sodium at 1500°F and delivered to the Solid State
Division for operation in in-pile loop tests. Two
identical pumps are being retained for more testing

19
0c

AIR—DRIVEN MOTOR -—\

DISCHARGE -
\ fe———— ———— § 1/q. in,. —————

\ ‘ ‘

 

 

 

UNCLASSIFIED
ORNL-LR-DWG 1694

 

 

 

 

 

 

 

 

 

 

 

 

INLET

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L, A
COOLANT OQUT

FACE SEAL

LIQuUiD LEVEL——/ SUCTION
HEAT INSULATORS

Fig. 2.2. Horizontal-Shaft Air-Driven Sump Pump for an In-Pile Loop.

 

i30dIY SSTUO0Ud ATHILEVAD dNV
and development aimed ot simplitying fabrication
and increasing pump capacity in anticipation of
more severe requirements, both in experimental
test loops and in other in-pile loop tests,

Hydraulic Pum§ Drives

J. A. Conlin
ANP Division

Hydraulic motors are being considered aos pos-
sible in-pile pump drives because they have high
starting torque and high power output from a very
small unit, and there might be a possibility of
using the hydraulic oil both as a power source and
as a coolant. Two commercially available motors
were ordered - a gear type and an axial-piston
type. Since neither of these was immediately
available, an axial-piston motor similar in design
but of much lorger capacity than the one ordered
was put on a no-load duration test to determine
its operating characteristics. The normal maximum
operating speed of this motor is 3600 rpm; how-
ever, the test was run at the opproximate in-pile
pump speed of 5000 rpm.

The motor completed a 500-hr test with no indi-
cation of trouble except for a constant shaft oil-
seal leak of 0.47 cc/hr. The seal leakage was
aggravated by two things. First, during initial
tests the shaft, for reasons unknown, became ex-
cessively hot. Second, it was necessary to by-
pass about 0.5 gpm of oil through the motor casing
and bearings to prevent excessive motor tempera-
tures at the high speed. This coused o 0.7- to
1.0-psig buck pressure against the seal, which
was a rubber-lip type that was not intended for
use under pressure. An operating oil temperature
of between 100 and 120°F was maintained during
the test.

Fluoride-to-Water Heat Exchanger

D. F. Salmon
ANP Division

A heat-extraction unit is needed for in-pile test
loops which is small enough to be inserted in the
reactor hole and does not require a complicated
external circuit of auxilicries, since shielding for
external apparatus is cumbersome and expensive.
A concentric-tube double-wall heat exchanger has
been designed that has a stagnant annulus of
fluoride mixture between circulating streams of

PERIOD ENDING JUNE 10, 1954

molten fluoride mixture on one side and cooling
water on the other. The stagnant mixture will be
solid at the water wall and molten at the circu-
lating-fluoride-mixture wall, and there will be a
temperature drop of approximately 1000°F radially
across the annulus. The circulating ceoling water
will either be dumped or cooled outside the re-
actor, : '

The exchanger is designed around an Inconel-
sheathed electrical tubular heater for preheating
the unit prior to starting circulation of the fluoride
mixture stream. The hairpin design to permit
differential thermal expansion is shown in Fig.
2.3. With the fluoride mixture NaF-KF-LiF [11.5-
42.0-46.5 mole %, K (solid) = 3.5 Btu/hr.sq ft
(°F/ft) and K (liquid) = 2.6 Btu/hr-sq ft (°F/ft)]
in the annulus, approximately 12 kw/ft can be
removed from a 1500°F stream of fluoride mixture
i\faF-ZrF"A-UF"4 (50-46-4 mole %) and adsorb?d by

the water stream,

FORCED-CIRCULATION CORROSION LOO PS

W. C. Tunnell

ANP Division
W, K. Stair, Consultant
J. F. Bailey, Consultant
University of Tennessee

Previous attempts to circulate fluoride mixtures
at simultaneously high velocities and large temper-
ature differentials were only partially successful.?
Therefore 'a small loop is being built that is de-
signed to have a temperature differential of about
135°F, with a maximum temperature of 1500°F,
and a Reynolds number of about 4500. Elecirical
resistance heating of the 0.300-in.-OD, 0.025-in.-
wall container tube will be used, but the power
consumption will be larger than in the previous
apparatus. This loop will be used to confirm
engineering design estimates of performance, but
it may not yield significant metallurgical infor-
mation, : :

The Metallurgy Division has provided specifi-
cations for six additional loops.
to have maximum Reynolds numbers of 10,000 and
temperature differentials of 100, 200, and 300°F,
and the other three loops are to have temperature
differentials of 200°F and Reynolds numbers of
1,000, 3,000, and 15,000. Heat economizers will

Three loops are

 

zD. F. Salmon, ANP Quar, Prog. Rep., Mar. 10, .7954,
ORNL-1692, p 19, :

21
ANP QUARTERLY PROGRESS REPORT

~INCONEL PLATE
- 56 -in.~LONG GE CALROD 48423

5 _ j

% HEATER [INCONEL SHEATH) / ;INCONEL RING

\ !
4 / / ~INCONEL. ADAPTER (CLOSED)

    
   
 
 

/ —Y=in. 0D x 0.020-in-WALL INCONEL TUBE
/ / s~ Vg=in OD x0.035-in.- WALL INCONEL TUBE

UNCLASSIFIED
ORNL-LR-CWG 1895

INCONEL CAP —,

INCONEL ADAPTER \
(OPEN) —

INCONEL RETURN HEADF.R7

INCONEL RING /
INCONEL PLATE 7 /
/ INCOMEL
/ END-PLATE -

   
   

    

 
 

 

 

 

 

 

3-in. 0D x 0.035-in.-WALL

 

 

 

 

x12-in-LONG INCONEL
TURE —

  

 

 

Y4-in 0D x 0.035-in-WALL /
INCONEL TUBE - /

 

INCONEL SHELL

\
— STAINLESS STEEL NOZZLE

Y g=in DIA x Y-in, LONG SPACER (INCONEL ROD)

 

 

 

 

 

  

Y4-in. OD x 0.035-in-WALL INCONEL TUBE ~/

1 2 3

| . l =
— 3

INCHES

Fig. 2.3. Concentric-Tube Double-Wal! Heat Exchanger for Use Inside a Reactor Hole,

be used, in at least the first three units, to con-
serve electric power, space, and fabrication costs,

The unit with o 300°F AT and a Reynolds
number of 10,000 is designed to dissipate 80 kw
of heat to air and therefore to have an 80-kw re-
heating capacity; resistance heating will be used.
The unit will also include a vertical-shaft sump
pump? for circulating the molten fluoride mixture

at 5000 tb/hr.

SODIUM-BERYLLIUM-INCONEL MASS-
TRANSFER TEST

L. A, Mann
ANP Division

A test unit has been constructed for investi-
gating mass transfer of beryllium to Inconel in a
flowing sodium system under conditions which
approach those proposed for the Reflector-Moder-
ated Reactor. A series of tests will be run to
determine whether it will be necessary to protect
the beryllium reflector-moderator from the sodium
coolant during operation of the reactor.

The test unit is designed to operate with a
sodium flow of 3 gpm through a venturi-shaped

 

3W. B. McDonald et al., ANP Quar. Prog. Rep. Mar.
10, 1954, ORNL.-1692, p 15,

22

beryllium insert with a minor diameter of about
0.25 in. The sodium temperature will range from
1200°F maximum (at the beryllium) to less than
900°F minimum. The Reynolds number will be
about 100,000 at the beryllium minor diometer and
about 50,000 at the coldest section of the loop.

FLUORIDE-TO-SODIUM HEAT-EXCHANGER
TEST

A. P. Fraas R. W, Bussard
R. E. MacPherson
ANP Division

As was previously reported,? test operations on
the 1-Mw regenerative fluoride-to-sodium heat ex-
changer were interrupted by failure of the sodium
heater coil. The coil was reploced and an attempt
was made to continue operation of the equipment
to obtain endurance information. However, a leak
developed from the sodium side of the heat ex-
changer into the fluoride side, and the unit has
been shut down for dismantling. The heat ex-
changer will be delivered to the Metallurgy Di-
vision for examination,

4R. E. MacPhersen, H. J. Stumpf, and J. G. Gallagher,
ANP Quar. Prog. Rep. Mar. 10, 1954, ORNL-1692, o 18.

 
GAS-FURNACE HEAT-SOURCE DEVELOPMENT

A. P. Fraas R. W. Bussard
R. E. MacPherson
ANP Division

Past experience and future plans for the testing
of prototypes of the Reflector-Moderated Reactor
intermediate heat exchanger have indicated a need
for high-capacity, compact heat sources. As a
step in this direction, a loop is currently being
assembled to check the operating characteristics
of a 100-kw gas-fired furnace. This unit, if
operable, will provide a basic design capable of
being enlarged to accommodate the large heat
loads which will be demanded in the near future.
The furnace, os shown in Fig. 2.4, is a drastic
departure from conventional furnace design. It
employs natural gas and pressurized air feed, the
air being preheated to 1000°F before entry into

 

PERIOD ENDING JUNE 10, 1954

the combustion zone, by a counfercurrent heat ex-
changer, with flue goses. :The flame temperature
is expected to approach 4000°F. No ceramic
lining is utilized in the combustion chamber; in-
stead, the design provides for cooling of all ex-
posed metal surfaces by the normal sodium flow
through the furnace. The combustion chamber is
a cylindrical opening 6 in. in length ond 5 in. in
diameter. The sides of the chamber are formed
by sodium coils and the ends by the gas and air
inlet arrangements. After combustion, the flue
gas passes over the coils to give up its heat to
sodium and is then channeled into a spiral dis-
charge duct where it flows countercurrent to feed
air to provide an economizer arrangement.

If operability is proved in this test, it is planned
to construct a 1-Mw furnace employing the same
basic design. This furnace will serve as the heat
source for tests of the second prototype of the

UNCLASSIFIED
ORNL~LK--DWG 1696

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

] - ) )
) . s Z(‘,__.‘[_k__ ~ o
7 : - FILAME  CONTROL
PILOT LIGHT ..~ \ ' / MOUNTING
MOUNTING —" 1 _ o R
i <% - GAS INLET TO
& ‘ _ - COMBUSTION CHAMBER
EXIT SOGIWM SRR 00 R - TR AR INLET TO
HEADER -1 0~ OA A O [ PSS COMBUSTION CHAMBER
| siclelalofiolelelols e |
N Py = -
s O = -1 i T COMBUSTION CHAMBER
JBE8E BEEHE |
O J (0
SGOIUM  TUBING - I \
L <
~— COMBUSTION AIR
- PLENUM
COMBUSTION AIR e ] M
ENTRANCE e s
‘:-—4-".; ;
FLUE GAS EXIT ="

Fig. 2.4. 100-kw Gas-Fired Furnace.

23
ANP QUARTERLY PROGRESS REPORT

Reflector-Moderated Reactor intermediate heot ex-
changer.

SODIUM SAMPLER

L. A, Mann
ANP Division

A unit has been designed, constructed, and test-
operated for repetitive sampling of molten sodium
at temperatures of up to 1500°F. A device is
included in the unit for separating the sodium from
its oxide by flotation in mercury and for furnishing
the sodium and oxide in separate flasks for ti-
tration., The basic method reported by Pepkowitz
and Judd® for analyzing Na,O in Na is used, and
the sampler is o simplified adaptation of a similar

device developed at KAPL.

REMOYAL OF FLUORIDE MIXTURES FROM
EQUIPMENT

L. A. Mann
ANP Division

A solvent for NaF-ZrF , and NaF-ZrF [-UF  mix-
tures has been developed, tested, and put into
general use. The solution has the composition (by
weight):  15% HNO,, 5% H,BO,, 4% F’b(ND3)2,
76% H,0. It is used at a temperature of 180 to
200°F, with the solution either flowing or being
stirred. A hood is required because of the NO,
fumes evolved. An approximately ],é
NaF—Zer-UF on Inconel was completely removed
(visual inspection) in 2 hr, with an lnconel cor-
rosion rate of approximately 1 mg/sq in. hr. In
further tests, NaF-ZrF, was completely removed
in 10 min except from very small openings which
took longer to clean. This cleaning method is,
of course, not recommended for thin-walled tubing

-in. coating of

or for equipment to be subsequently tested or
examined for corrosion,

BEARING-MATERIALS TESTS

W. C. Tunne!l
ANP Division

Tests have continued, with the equipment previ-
ously described,® in attempts to find combinations
of materials usable for journal bearings in high-

 

SL. P, Pepkowitz and W. C. Judd, Determination of
Sodium Monoxide in Sodium, KAPL-P-452 (Ocr. 1952);
L. P. Pepkowitz, W. C. Judd, and R. Downer, Determi-
nation of Sodium Monoxide in Sodium, An Addendum,
KAPL-972 (Aug. 5, 1953).

24

The early tests
showed metal build-up between the test speci-
It is believed that the build-up material
was the product of corrosion of the stainless steel
containers. Recent tests have been conducted in
Inconel containers, and although there is still
some evidence of metal build-up, the specimens
remain sufficiently clean during the test peried
so that wear patterns can be obtained. Figures 2.5
and 2.6 are photographic records of some of the
test results; Fig. 2.5 shows the wear that resulted
in operating a tungsten carbide pin against a
tungsten carbide plate, and Fig. 2.6 shows the
comparatively little wear that results when a tung-
sten carbide pin was tested against a titanium
carbide plate. Additional records that are made
of each test include the surface finish, flatness,
and hardness values of each specimen, before and
after each run, and a Faxfilm replica of the surface
condition of the pin and of a typical section of the

temperature fluoride mixtures.

mens.

plate. (Faxfilm is a surface evaluation process
furnished by the Brush Development Company. A
clear solvent is applied to the surface to be re-
produced. A clear plastic film is then pressed
against the surface and solvent, and aofter o few
seconds for drying, the duplicate surface is lifted
off with the plastic film. This film is then
mounted on a slide and can be projected.) The
fluoride mixture is analyzed after each run for
corrosion products to determine whether constitu-
ents of the specimens have been introduced into
the fluoride mixture. No corrosion products have
yet been found in the fluoride mixtures analyzed.
The materials tested and the test results are
shown in Table 2.1, which lists all the combi-
nations tested to date. Of the materials tested,
Cr,C, appears to be a poor possibility for use
as a bearing material because it shows excessive
wear, Hot-pressed beryllium oxide and Al O, have
been eliminated as possible bearing materials
because of excessive corrosion and/or erosion,
Tungsten carbide and Kennametal titanium carbide
appear to be possible bearing materials; both
materials showed very little wear and very little
corrosion in the tests, These materials also
showed good wear resistance when tested in
sodium, as reported by Vail.” Boron carbide also
indicated promise in wear and corrosion resistance,

SW. C. Tunnell ef al., ANP Quor. Prog. Rep. Mar. 10,
1954, ORNL.-1692, p 22,

’D. B. Vail, Compatibility of Materials in Liquid
Metals, KAPL-589 (Aug. 18, 1951),
Fig. 2.5. Tungsten Corbide Pin Operated Against
a Tungsten Carbide Plate in Contact with a Fluo-
ride Mixture in an Inconel Container at 1200°F for

2 hr.

PERIOD ENDING JUNE 10, 1954

 

Fig. 2.6. Tungsten Carbide Pin Operated Against
a Titanium Corbide Plate in Contact with a Fluo-
ride Mixtfire in an Inconel Container at 1200°F for

2 hr.

  

25
ANP QUARTERLY PROGRESS REPORT

TABLE 2.1. RESULTS OF TESTS OF BEARING-MATERIALS COMBINATIONS EXPOSED TO

 

 

 

 

 

 

 

 

 

PLATE MATERIAL
2
=
a
o 2
@ 3
PIN MATERIAL - = RESULTS
" < @
o = o s
o 5 > — o = O
% : o | o gl % Q% :
5 % W - = “ o R < g" 5
XL o = < < (s1] O - b= G xI
Hot-pressed BeC 29 27 X Sliminated
BeO-Ni X X
wC X 21 30 25 23 24 31 28
A|203 X X Eliminated
AIZOB-MQO X X
BAC X X
Cr3C2 X 22 X 33 18
19
TiC-Ni X 20 X 26
TiC-Co X X
Graphitar X X
_High-density graphite X X

 

 

 

 

 

 

 

 

 

 

 

 

 

 

*The numbers are test serial numbers.

not yet tested or results not yet available.

ROTARY-SHAFT SEALS

¥W. C. Tunnel! P. G. Smith
ANP Division

Tests are under way to determine the feasibility
of sealing fluoride mixtures in an in-pile pump
with a packed seal on a horizontal shaft. One
packing that was tested consisted of o mixture
of 95 wt % Asbury Graphite and 5 wt % MoS,, re-
tained by bronze wool at each end of a stuffing
box. The stuffing box was 3!/2 in, long with a
]/4-in. annulus., For this test, the apparatus was
operated for a total of 382 hr, and the test was
then terminated because a heater burned out as
a result of leakage of the fluoride mixture. Leak-

26

The notation X indicates unsatisfactory combinations.

Blank spaees mean

age from this seal (test No. 33)® was at the rate
of 0.7 cc/day throughout the test. Postrun ex-
amination revealed only slight shaft wear at re-
gions where the packing retainers were located
and in the glond region. The pot showed no sign
of packing leakage from the hot end of the seal,

In another test of the same packing material
(test No. 35), an inert blanket was installed on
the cold end of the seal to protect the section of
the shaft under the gland from wear by oxidation
products, A pillow-block beuring was installed on

8w. C. Tunnell et al., ANP Quar. Prog. Rep. Mar. 10,
1954, ORNL-1692, p 26.
the 1/z-in. shaft to improve alignment, The type
316 stainless steel shaft was coated with ¥ -in.-
thick Colmonoy ground to a 20-gin. finish, The
test apparatus was operated for 603 hr with NaF-
ZrF -UF, at a temperature of 1250 to 1300°F and
with a pressure of 2 psi on the seal. The leckage
rate from the seal was about 0.5 cc/day and ap-
peared to be mostly packing material. The power
requirement of the seal varied from 100 to 200
watts, Termination was caused by shaft seizure.
Postrun examination revealed shaft wear of about
the same magnitude as in test No. 33. A section
of the shaft surface at the hot end appeared to
have peeled off or to have corroded slightly.
Analysis of a sample from the pot showed that

- PERIOD ENDING JUNE 10, 1954

an unsatisfactorily large amount of the packing
material had ledked inte the pot. There was
evidence of corrosion on the hot end of the gland,
which indicates a need for an inert-gas bianket
between the gland and the stuffing box, os well
as between the gland and the shaft, Another test
is planned for the near future with an improved
inert-gas blanket and improved gland alignment.

Seal test No. 31 was the same as seal test No.
28,% except that Asbury Graphite was substituted
for the National Carbon Graphite in the 10 wt %
BaF ,~90 wt % graphite packing material, The
results obtained were very similar to those for test
No. 28 in that excessive leakage occurred. The
duration of the test was 165 hr,

27
ANP QUARTERLY PROGRESS REPORTY

3. REFLECTOR-MODERATED REACTOR

A. P. Fraas
ANP Division

Components of the proposed 60-Mw Circulating-
Fuel Reactor Experiment (CFRE) are now being
designed ond constructed ond are to be tested to
determine operational characteristics. Plans are
being made for tests of full-scale fuel pumps and
of elements of full-scale heat exchangers. Also,
thermal-stress cycling tests for both Inconel and
beryllium are to be performed. A fairly complete
detailed stress analysis of the reactor is well
under way. A series of charts was prepared to
facilitate the determination of the temperature dis-
tribution and thermal stresses in media having
uniformly distributed volume-heat sources.! Also,
a comparison of lithium- and zirconium-base fluo-
ride fuels for aircraft reactors was made that
indicates the definite advantages to be gained
through the use of the lithium-base fuel.

Cstimates of the xenon-poisoning effect in the
Reflector-Moderated Reactor have confirmed the
great incentive for removing the xenon during
operation,  Therefore an experiment is being
planned to determine whether the xenon can be
adequately purged. Studies of reactor dynamics
indicate that the design of the Reflector-Moderated
Reactor precludes the possibility of the coupling
of mechanical oscillations with nuclear oscil-

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

lations to create antidamped oscillations that
would destroy the reactor.

Calculations of a set of 48 related reactors are
weil under way. The effects of reactor dimensions
on concentrotion of U233 in the fuel, on total
U235 jnvestment, on outside peak-to-average power
density in the core, ond on the per cent thermal
fissions in the core ore presented.

A COMPARISON OF LITHIUM- AND ZIRCONIUM-
BASE FLUORIDE FUELS

A key decision in the design of a circulating-
fuel reactor is the choice of the fluoride fual to
be employed. Of the fuels that cuirently seem
most promising, the sodium-zirconium-uranium fluo-
ride fuel has the poorest physical properties but
would be the easiest and cheapest to prepars,
while the sodium-lithium-potassium-uranium fluo-
ride fuel has the best physical properties but
would require Li’ — an expensive material. The
properties of greatest interest for these two fuels
for vranium concentrations of 25 Ib/cu it are given
in Table 3.1.

In a comparison of the two fuels it is evident
that the lithium-base mixture has both a lower
melting point and a much lower vapor pressure.
The lower vapor pressure is particularly advan-
tageous in that it should eliminate the many
troubles experienced with sublimation of Zrl:4

TABLE 3.1, PHYSICAL PROPERTIES OF TWO FLUDRIDE FUELS CONTAINING

 

PHYSICAL PROPERTIES

25 b OF U23% PER CUBIC FEET OF FLUORIDE MIXTURE AT 1472°F

 

Melting point, °F

Vapor pressure at 1500°F, mm Hg

Prandtl number, CP,U/K |
Viscosity,* cp

Thermal conductivity,* Btu/hr-sq ft (OF /ft)
Specitic heat,* Btu/lb

Density,* Ib/cu ft

 

28

chF-ZrFA-UF4 Nc:F-LiF-KF-UF4
(50-44.5-5.5 mole %} (11-45-41-3 mole %)
....... e s
970 850
8 <1
3.71 1.24
6.5 2.5
1.25 1.95
0.295 0.40
L 203 132

 
and the formation of high-melting-point deposits
on the roofs of expansion tanks, in pressure gage
lines, etc. The higher viscosity of the zirconium-
base fuel gives a lower Reynolds number in the
reactor core and hence a more severe boundary-
layer heating problem; the temperature differential
between the peak temperature in the boundary
layer and the mean fuel temperature is about 50%
greater than that for the lithium-base fuel.?2 Thus
the zirconium-base fuel, with .its higher viscosity
and its much higher vapor pressure, would be
more likely to give trouble with boiling in local
hot spots ond hence, conceivably, violent fluctu-
otions in power. '

Recent multigroup calculations show that for a
representative core geometry the critical mass for
a lithium-bose fuel will be about 16% greater thon
that for a zirconium-base fuel because of thermal-
neutron absorption in the potassium, The reactor
under consideration had an 18-in.-dia core with
a 12-in.-thick reflector and a 4-in,-thick fuel an-
nulus, and the beryllium in the island and reflector
was canned in 0.010-in.-thick Inconel. The core
shells were Y -in.-thick Inconel.

A further consideration in favor of the alkali-
metal fluoride fuels is based on radiation damage;
the alkali-metal fluorides are very stable com-
pounds, and their recombination rate con be ex-
pected to be larger than that for ZrF .

One of the most important design criteria is the
effect of the physical properties of the fuel on the
heat exchanger. While many design compromises
are possible, the most important of these can be
deduced from Fig. 3.1, which ‘was prepared for a
series of fuel-to-NaK heat exchangers. The pres-
sure-stress limitations are usually controlling, ond
if the same allowable stresses are to be used,
the pressure drop must be held constant. The
designs represenfed by Fig. 3.1 were prepared
by determining the number of tubes required for
each tube length to give a 50-psi pressure drop
through the tubes in the NaK circuit. The tube
spacing was then adjusted to give a 40-psi drop
in the fuel passages. As shown in Fig. 3.1, the
number of tubes in the heat exchanger must be
increased to provide more flow passage area as
the tube length is increased. -

The curves for the temperature differential be-

 

2y, E. Poppendiek and L. D. Palmer, Forced Con-
vection Heat Transfer in Pipes with Volume Heat

Scurces Within the Fluids, ORNL-1395 (Dec. 2, 1952).

PERIOD ENDING JUNE 10, 1954

tween the fuel and NaK are instructive. If the
temperature difference for the zirconium-base fuel
is to be the same as that for the lithium-base fuel,
the tube length must be increased, for example,
from 6.0 to 7.6 ft if the temperature difference is
kept ot 91°F. This would require a 15% increase
in the number of tubes. The Reynolds number
would drop from 4850 to 2100, that is, from a value
well within the turbulent range to one likely to
give laminar flow and hence even poorer heat
transfer than that estimated. [f the same tube
spacing were used and a higher pressure drop
accepted, the comparison would be less unfavor-
able in most respects. However, the pumping
horsepower would need to be increased by 50%
with the zirconium-base fuel, and many stresses,
such as those in the NaK outlet header sheets,
would be more than doubled..

The effects of heat-exchanger volume on beth
shield weight and the activation of the NaK in
the secondary circuit are vitally important. -The
weight of the reactor, heat exchanger, and reactor
shield assembly, together with the radiation dose
from the NaK in the secondary circuit at a point
50 ft from the engines, is given in Fig. 3.2 es a
function of the temperature loss in the fuel-to-NaK
heat exchanger. The data are for a 200-Mw reactor
with a shield designed to give 10 r/hr ot 50 ft
from the reactor at full power. It can be seen
that for a given design temperaiure loss the
lithium-base fuel is markedly superior to the zir-
conium-base fuel in that it gives one-half the
radiation dose and a 6000-lb saving in shield
assembly weight.

The limitations on high-temperafure heat-ex-
changer design imposed by pressure stresses merit
further explanation. The many closely spaced
holes in the header sheets generolly make the
header sheets much less strong than the thin tube
walls. In the spherical-shell heat exchanger for
the Reflector-Moderated Reactor, the NaK must
enter the heat exchanger at about 100 psi to take
care of the roughly 50-psi pressure loss in the
heat exchanger, the 40-psi loss in the radidators,
and the 10-psi loss in the connecting lines. Most
of the estimated 50-psi pressure drop in the fuel
system occurs in the heat exchanger. On this
basis the pressure difference across the header
sheet at the hot end is less than 10 psi, while
that at the cold end is about 100 psi. Fortunately,
the strength of Inconel is about five times greater

29
ANP QUARTERLY PROGRESS REPORT

ORNL-LR-DWG 1603

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4000
wn
el
m
, : ‘ e
S : L e S EE - — 3500 +
: W
O
| | 2
350 |——-r{- - - ‘ - : 3000
g —
4
T 2 300
L
QoW
b
o o
< 2 250
= @ =
© 2 200 14 £
g u
w5 5
Z X 12 2
[n et
Ld
10 [T
=
<1
T
Ly
-~
—
o
o T
—~ {20
la
a
Ll
O {10
=
1
i
Ll
& 100
o
w
« 90
=
}._
<{
&
a. 80
=
wh
]—.
¥ 70
Qo
<
o
K 60
ol
w
z
50 5000 @
m
>
4000 32
w
0
e PETTP ‘ e a —_— . he—— 3000 2
Zr-BA =
‘ SE FUEL TUBE SPAGING = 0.020in E_.
————— S -Zr-BASE Fug( = ot — 2000 -
0.027i
. i l ! 'n- 2
| ‘ i 1000
6 7 8 g 10

TUBE LENGTH {ft)

Fig. 3.1. Effects of Tube Length on Design Variables for a 60-Mw Fuel-to-NaK Heat Exchanger for
50.psi Pressure Drop in the NaK Circuit.

30
REAGTOR-HEAT EXCHANGER-SHIELD ASSEMBLY WEIGHT {tb}

Fig. 3.2. Weight of Reactor, Heat Exchanger, and Reactor
the NaK in the Secondary Circuit at a Point 50 ft from the
Loss in the Fuel-to-NaK Heat Exchanger.

88,000

85,000

80,000

75,000

70,000

20

ORNL-LR-DWG 1604

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FUEL-TO-NaK TEMPERATURE DIFFERENCE (°F)

0.80

0.60

0.20

PERIOD ENDING JUNE 10, 1954

RADIATION DOSE FROM NaK AT 50ft FROM ENGINES (r/hr)

-Shield Assembly and Radiation Dose from
Engines as a Function of the Temperature

31
ANP QUARTERLY PROGRESS REPORT

at the cold end than at the hot end. It has been
shown? that the optimum tube diameter is close
to % . in. and that the minimum thickness of liga-
ment between holes in the header sheet should
be at feast 0.10 in. for satisfactory welding. |f
an essentially cylindrical header sheet, such as
that shown in Fig. 3.3, is used to minimize the
stresses, the ligagents between the tubes will
be in tension. |f allowances are made for both
the cross-sectional area lost in the holes and the
increases in the stresses around the holes that
arise from geometric effects, the maximum tensile
stress in the ligaments will be about five times
that for a simple cylindrical shell of the some wall
thickness with ne perforations. The 1000-hr rup-
ture strength for Inconel at 1200°F is 12,000 psi,
and therefore the allowable stress should not
exceed 6000 psi to give some factor of safety. |f
the header-sheet thickness were arbitrarily chosen
as 0.25 in., its rodius of curvature would be 3 in.
(The allowable stress from simple loop tension
considerations is 60 times the pressure differential
for the unperforated cylinder, or 12 times that for
the perforated cylinder. Thus the ratio of header-
sheet thickness to radius of curvature becomes
1:12.)  While other plausible designs can be
evolved on a similar basis, that illustrated in
Fig. 3.3 appears to be about the best devised to
date.

 

3A. P. Froas and M. E. LaVerne, Heat Exchanger
Design Charts, ORNL-1330 (Dec. 7, 1952).

UNCLASSIFIED
ORNL-LR-0OWG 1605

 3e-in,-CD TUBES, 0.020~in.-THICK WAL
/
/

 

—~ NaK INLET TUBE

e —

T F

 

K*‘* HEADER SHEET

Fig. 3.3. Section Through Header Region of o
Tube Bundle for a 60-Mw Heat Exchanger.

32

REACTOR FPHYSICS

W. K. Ergen
ANP Division

Estimates of the xenon poisoning? of the 60-Mw
Reflector-Moderated Reactor (power density 1.35
kw/cc) indicate that unless the xenon is rather
rapidly purged from the fue!l during operation a
reactivity loss of the order of a few per cent will
occur. Also, the maximum xenon concentration
after shutdown would give a reactivity loss of
about 10% or more. A control mechanism capable
of compensating for such reactivity losses would
be rather complicated, and the added uranium con-
centration would be highly undesirable. Therefore,
an experiment is now being planned to determine
whether the xenon can be adequately purged.

The elaborate coding of the multigroup calcu-
lations, mentioned in the previous report,” has
temporarily been set aside. Instead, o three-
group three-region code is being prepared which
does not exceed the present capabilities of the
ORACLE. This code is adequate for a variety
of problems, and the use of the ORACLE is far
more convenient than the use of on out-of-town
machine.

The possibility of the coupling of mechanical
oscillations with nuclear-power oscillations in the
Reflector-Moderated Reoctor was studied, since
such coupling is known® to be capable, in prin-
ciple, of creating antidamped oscillations. In
particular, the possibility of the reactor oscillating
mechanically like a Helmholtz resonator was in-
vestigated. The antidamped oscillations cannot
occur if the frequency of the mechanical oscil-
lations is very much higher than the frequency
of the nuclear-power oscillations, as will be the

case in the 60-Mw CFRE.

REACTOR CALCULATIONS
M. E. LaVerne
ANP Division
C. S. Burinette
United States Air Force

Core radii of 20, 30, 40, and 60 cm, fuel thick-
nesses of 5, 10, 15, and 20 cm, and extrapolated

 

4J. L. Meem, The Xenon Froblem in the ART, ORNL
CF-54-5-1 {May 3, 1954).

SR. R. Coveyou and R. B. Bate, ANP Quar. Prog.
Rep. Mar. 10, 1954, ORNL-1692, p 40.

6P. R. Kasten, Linearized Stobility Criterio for HRT
Type Reactors, ORNL CF-54-4-183 (Apr. 21, 1954).
reflector thicknesses of 20, 30, and 40 cm were
selected for calculations of the set of related
reactors referred to previously.” Poison (sodium
and Inconel) distributions in the reflector and
island were obtained from a previous study.? The

 

’M. E. LaVerne and C. S, Burtnette, ANP Quar. Prog.
Rep. Mar. 10, 1954, ORNL-1692, p 39, :

8R. W. Bussard et al., Moderator Cooling System for
the Reflector:Moderated Reactor, ORNL-1517 (Jun, 22,
1954).

 

REFLECTOR THICKNESS = 20 CM

A
o>

S

o

£

=

<"
7

/

UEIT CONCENTRATION (MOLE %)

o

/-

£ %/
rne
o

UEE pONCENTRATION (MOL

PERIOD ENDING JUNE 10, 1954

fluoride fuel NaF~ZrF4-UF4_ (53.5-40.0-6.5 mole %)
was used for the entire set of calculations. From
the set of 48 reactors determined by the specified
independent parameters, it was found possible to
select 30 reactors that would give results that
could be used to cross-plot and interpolate the
properties of the 18 omitted reactors to an ac-
curacy sufficient for this survey. The results
obtained are summarized in Figs. 3.4, 3.5, 3.6,
and 3.7.

—

ORNL-LR~DWG 1£77

7
/

 

 

REFLECTOR THICKNESS =40 CM

Fig. 3.4. Effect of Reactor Dimensions on Concentration of U235 in Fuel,

33
ANP QUARTERLY PROGRESS REPORT

In view of stringent limitations on permissible
uranium concentrations in the fuel, it is clear
from Fig. 3.4 that there exists a considerable
advantage in increasing reflector thickness from
20 to 30 em but exceedingly little advantage in
a further increase to 40 em. The remaining figures
have, therefore, been restricted to a 30-cm re-
flector thickness.

EXTRAPOLATED REFLECTOR
THICKNESS =30 CM

TOTAL UESS INVESTMENT LB)

 

EXTERNAL FUEL VOLUME = 0O FT?

S INVESTMENT (LB)

I
>

7O074L L€

 

EXTERNAL FUEL VOLUME = 4 FT?>

 

The effect on total uranium investment of the
reactor dimensions and the external fuel volume,
such as would exist in a heat exchanger external
to the core, is shown in Fig. 3.5, The surface
shown for zero external volume is, of course,
simply that for critical mass., At a core radius
of 30 c¢m the critical mass is virtually independent
of fuel thickness. In the surfaces for both the

ORNL-LR~-DWG 1176

 

 

 

 

EXTERNAL FUFL VOLUME = 8 FT?2

Fig. 3.5. Effect of Reactor Dimensions and External Fuel Yolume on Total U233 Investment.

34
ORNL-LR-DwG 156

EXTRAPOLATED REFLECTOR

THICKNESS = 30 tm ’

    
  

FEAE -TO-AVERAGE POWER DENSITY RATID
AT
-

 

 

/‘?%} P = l L .
g PO - i) e
o E ., 1 ) - o Lo
% - <5
"o \\\\ - - 11/// Q_t;";D
~ L & e
Mo, 2| R
<.‘Lg‘=9 ~ L/
]
C‘,,p,/ Sy

Fig. 3.6. Effect of Reactor Dimensions on Out.
side Peak-to-Average Power.Density Ratio in Core
of Reflector-Moderated Reactor.

ORNL*&! - !WG 1197

EXTRAPDLATED REFLECTOR
THICKNESS = 30 oM

0 8g
g
& -~
=l
=1 4
%3
By
3
&
& -

0

Lk
/%{6

Fig. 3.7. Effect of Reactor Dimensions on Per
Cent Thermal Fissions in Core of Reflector-
Moderated Reactor,

Z- and 4-cu ft external volumes, simple visual
examination indicates the probable existence of
an optimum set of proportions. Whether any closed
contours {as required for an absolute minimum)
actually exist has not yet been determined. The
reversal of scales for the 8-cu ft surface was
necessitated by the existence of an extreme peak
at what was the front corner.

PERIOD ENDING JUNE 10, 1954

The effect of reactor dimensions on outside
peak-to-average power-density ratio is shown in
Fig. 3.6. The effects here are not large; an in-
crease in core radius from 20 to 60 cm at a con-
stant fuel thickness gives a decrease in the ratio
of about 20%, while an increase in the fuel thick-
ness from 5 to 20 cm at o constant core radius
results in an increase in the ratio of about 33%.

The per. cent thermal fissions as a function of
reactor dimensions is shown in Fig. 3.7. The
least-thermal reactor (28%) has the thickest fuel
layer and the smallest core, while the most-thermal
reactor (45%) has the thinnest fuel layer and the
largest core.

When the 60-Mw design reactor’ was modified
by deleting the Inconel tubes and their surrounding
stagnant sodium and fitting the resulting voids
with beryllium, the critical mass decreased by
29%, that is, from an original value of 40.7 b to
29.0 Ib. This percentage decrease was duplicated
by NDA? in a calculation on a 300-Mw reflector-
moderated reactor for which the same method of
poison reduction was employed.

The Curtiss-Wright Corporation, under contract
to USAF, is performing an extensive series of
multigroup, multiregion calculations on the IBM-
701 for the ORNL-ANP Project. Specifications
for these calculations are being prepared by the
ORNL-ANP General Design and Reactor Physics
Groups in cooperation with Curtiss-Wright.

Critical mass limits computed by Curtiss-Wright
for the two-region, Teflon-uranium foil core, be-
ryllium-reflected critical experiment as affected
by foil thickness and space-point interval are
listed in Table 3.2. These results agree quite
well with the UNIVAC calculation previously re-
ported.’

 

qM. R. Adolph, R. C. Ross, and M. S. Silberstein,
Circulating-Fuel Reactor Studies, Part [, NDA 10-122
(Apr. 4, 1954).
TABLE 3.2. CRITICAL MASS LIMITS FOR
RHOMBICUBOCTAHEDRON CRITICAL

 

 

 

EXPERIMENTS
FOIL THICKNESS Ar CRITICAL MASS
(mils) (em) (Ib of U233
o+ 1.828 14
10 0.457 18

 

 

 

*Homogeneous mixture.

35
Part i

MATERIALS RESEARCH
4. CHEMISTRY OF MOLTEN MATERIALS

W. R. Grimes
Materials Chemistry Division

The experimental studies of the chemistry of
molten materials have been. devoted almost ex-
clusively to fluoride systems of interest as fuels.
Considerable attention has been paid, during the
quarter, to systems in which the vraniferous ma-
terial consists wholly or in part of UF ..

Studies of phase equilibrium by the method of
thermal analysis have been continued. This tech-
niqgue has been applied to exploratory studies in
mixed chloride-fluoride systems, Quenching tech-
niques have been significantly improved and have
been used, with petrographic and x-ray examination
of the melts so produced, to virtually finish the
elucidation of the complex NaF-ZrF, system.
Differential thermal analysis and filtration tech-
niques have been used in several phase-equilibrium
experiments, and an apparatus for visual obser-
vation of liquidus temperatures has been tested
and found to be satisfactory. '

The solubility of UF, in several molten-sait
systems has been examined experimentally. |t
appears that attainable concentrations of UF, in
Z¢F ,-based fuels are insufficient to fuel aircraft
reactors; however, use of UF, and UF, in such
mixtures appears to be promising. The UF, com-
pound is apparently sufficiently soluble in NaF-
LiF-KF mixtures (which would require Li’7 for
vtilization) to make such fuels atiractive; the use
of NaF-RbF-LiF mixtures as solvents for UF,
{without UF4) has not yet been shown to be pos-
sible,

The large-scale (250-lb-capacity} equipment for
the production of fluoride mixtures has been re-
activated and is now in three-shift five-day oper-
ation for furnishing the ZrF -based materials re-
quired for large-scale testing at ORNL and other
installations, '

QUENCHING EXPERIMENTS WITH
FLUORIDE SYSTEMS
R. E. Thoma ‘R. E. Moore
M. S, Grim C. J. Barton
Materials Chemistry Division
G. D. White H. Insley, Consultant
Metallurgy Division
The major part of the quenching work with com-
positions in the NaF-ZrF, system has been com-

pleted and a revised diagram for the system is
presented. Work on the Na F-UF , system continued,
but further work will be required before a revised
diagram can be completed. It is anticipated that
attention will shift from the binary system to the
NaF-ZtF UF , system and possibly the NaF-ZrF -

UF, system in the near future,

NaF-Z:F,

The methods and apparatus for performing quench-
ing experiments were described previously. ' During
this quarter, all the guenching experiments were
made by heating the sample up to, but not above,
the quenching temperature, ,

The incongruent melting point of the compound
Na,Zr F ., was established as 538°C by quenching
mixtures with 57 and 60 mole % ZrF ;. The liquidus
temperatures of mixtures with 57 and 60 mole %
ZrF, were found to be 546 + 5°C and 609 t 5°,
respectively. The 57 mole % ZrF, composition
(almost 100% NayZr,F,,) was converted to ZrF,
and liquid at 538°C. From the shape of the liquidus
curve and from x-ray diffraction studies of large
samples of slowly cooled mixtures, it appears that
the ZrF, primary-phase field extends to about 56
mole % erf:‘l in the NaF-Zer system.

Quenching and glass-devitrification experiments
indicate that the liquidus-solidus temperature of
the 47 mole % ZrF, composition is about 522°C,
and the liquidus-solidus temperature of the 50 mole
% ZrFA composition is about 511°C. Recently,
very careful thermal-analysis comparison of the 47
and 50 mole % ZrF4 mixtures showed that the 47
mole % mixture melts and freezes sharply at 520°C
and the 50 mole % mixture freezes sharply at
510°C. These results mean that the single phase
which is always obtained on cooling 47 mole %
ZrF4 compositions is a congruently melting com-
pound (NagZr F,,), while the 50 mole % ZrF,
composition is near a euftectic of Na Zr,F . and
Na,Zr F .

A phase referred to as R-3 has appeared? in
association with glass in compositions ranging

 

1C. J. Barton et al., ANP Quar. Prog. Rep. Dec. 10,
1953, ORNL-1649, p 54.

2R, E. Thoma et al., ANP Quar. Prog. Rep. Mar, 10,
1954, ORNL-1692, p 52-54.

39
ANP QUARTERLY PROGRESS REPORT

from 45 to 57 mole % ZrF,. This phase forms only
when large quench samples are used, that is, under
poor quenching conditions. Experiments with very
small samples at various temperatures have failed
to produce R-3. It was concluded from these ex-
periments that R-3 does not have a stable existence
in the binary system but is an easily crystallized
metastable compound (probably NaZrF;) that is
produced by fairly rapid cooling below the equi-
librium liquidus temperature,

Petrographic and x-ray diffraction examination
of slowly cooled preparations led to the postulation
of the existence of o weakly birefringent compound
Na,Zr,F,, and a series of solid solutions between
it and Na,ZrgF .. Petrographic examinations of
samples of 43 and 45 mole % ZrF , compositions
heated to 483°C and held for 20 hr, before quench-
ing, definitely pointed to an immiscibility gap in
this series between about 42 and 44 mole % ZrF .

Petrographic examinations of a large number of
quenched samples with between 39 and 45 mole %
ZrF, led to the conclusion that Na,Zr F | is not
stable at the liquidus temperature. In 47 to 42
mole % ZrF, compositions, the solid solution
NagZrgF ., is the primary phase. The liquidus
temperatures . of compositions with 47, 45, 43, and
42 mole % ZrF, are about 522, 520, 517, and
S08°C, respectively. In 39 and 40 mole % ZrF

compositions, the primary phase is Na,ZrF . The
liquidus temperature for the 39 mole % Zri, com-
position is above 513°C, ond for the 40 mole %
ZrF, composition it is near 510°C, For the 41
mole % composition the liquidus temperature is
about 502°C, and this composition is near a eutectic
between Na,ZrF, and the NayZr F,, solid so-
lution.  Below about 488°C, Na,Zr,F,, appeors
as fine-grained fibers. Slowly cooled compositions
with about 40 mole % ZrF, usually undergo a
reaction just after complete solidification that
produces a fluff of very fine-grained material,
According to x-ray diffraction results, this material
is a mixture of Na,Zr,F,, and NaZr F .. It is
likely that the reocfron is the solldmsfofe conversion
at about 488°C of the eutectic into these com-
pounds, The formation of Na,Zr,F,, in the solid
state explains why it has olways Leen found as
fine grains,

A number of different forms of Na,ZrF have
been observed, and quenching experiments have
been conducted to establish the transition tempera-
tures., This work is not yet complete, but the
approximate temperatures are known, The phases
which have been found and the optical properties
and the approximate temperatures at which these
phases are stable are given in Table 4.1. The
temperature of transition from phase 5 to phase 4

 

 

 

 

TABLE 4.1, CRYSTAL MODIFICATIONS OF Na2ZrF6
APPROXIMATE TEMPERATURE
PHASE OPTICAL PROPERTIES °C)
-1 | sotropic 545 to 620
2 Uniaxially positive, O = 1.406 540 1o 545
E = 1.408
3 Uniaxially positive, O = 1.376 505 to 540
E = 1.386
4 Biaxially positive, y = 1.412 Below 460
a = 1.408
5 Biaxially negative, y = 1.419
a = 1.412
6 Biaxiolly negative, y = 1.429
a = 1.420

 

 

 

40
is not known, but it is below 460°C.

Phases 3 and 4 have been found in quenches of
mixtures with 37 ond 40 mole % ZrF, at the ap-
propriate temperatures. The refractive indices
were the same as for the 33 mole % ZrF, compo-
sition, and thus there appears to be ||f1e or no
solid-solution range for these phases,

The phase diagram presented in Fig. 4.1 repre-
sents the data discussed here and in previous
reports. 13 Thermal data obtained from cooling
curves, together with visual observation of the
Na,ZrF, compound, were the chief sources of
IIqUIdUS temperatures in the 0 to 30 mole % ZrF
region. The Na,.ZrF_, compound was observed to
melt rather sharply between 840 and 850°C to a
clear liquid that permitted observation of turbidity
on cooling to a temperature slightly above 850°C,
The liquidus temperatures in the 60 to 100 mole %
ZrF , region were obtained from thermal data,

NaF-UF,

Eight samples covering the range in the NaF-UF ,
system from 18 to 67 mole % UF4, chosen to fill
insofar as possible the gaps in the knowledge of
the system,? were mixed and hydrofluorinated to
minimize oxide contamination. These slowly cooled
specimens were examined by petrographic and x-ray
techniques. The results of this examination, some
of which were published previously, are shown in

Table 4.2,

 

BR. E. Moore, C. J, Barton, and T. N. McVay, ANP
Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609, p 59.

ORNL - LR -DWG 204

1000

 

s
W
i
par
=
o
x
W50

=

I}
*_

NoZiFe (MZTASTABLE)
400 [ : bl
'duEZrFS -
Naz o fyy ‘Nag ZrgFay
o 2 20 30 40 50 60 70 80 S0 4 100
Zriy (mate %)

Fig. 4.1. Phase Diagram of the NaF.Z:F,

System,

PERIOD ENDING JUNE 10, 1954

TABLE 4.2, PHASES PRESENT IN SLOWLY COOLED
NaF-UF , MIXTURES

 

PHASES IDENTIFIED BY
PETROGRAPHIC AND
X-RAY DIFFRACTION AMALYSES

COMPOSITION
{mole % UFA)

 

18 B Na,UF o, NajUF,, NaF

22 ‘8 NuZUF NaF

25 '83'N°2UF6" N03UF7, NaoF (trace)

28 [‘3 Na,, UF (major}, Na,UF

30 f)) NaZUFfi (major), Na UF

31.5 [‘3 Na,UF

33.3 B NcQUF

35 B -NQZUFfi (major), NaUF ¢

37 [) a,UF o (major), NaUF ¢

38.5 fi NuzUF {major), NnUF5

40 3 --Nm2 g NaUFg

43 [33-N02UF6, NaUF

46 fi?’-quUFé, NuUF:5 (major)

48 NaUF ., UF, {trace)

50 NaUF  (major}, fia-Nc:ZUFé, UF .
U0, (trace)

53 NaUF (major), UF

67 NaUF UF

5!

 

Small portions of these specimens have been
used with conventional techniques in quenching
experiments, From 18 to 30 quenches have been
made on portions of all specimens except the
Na,UF, composition, for which a total of about
80 attempts has been made. By quenching speci-
mens from temperatures as high as ]022“(:; to as
fow as 399°C, it has been established that four
crystalline forms of the Na ,UF, composition exist,
These crystalline modifications are a (cubic), B,
(hexagonal), B, (hexagonal), and y (orthorhombic}.
In no case could the Na,UF, be quenched com-
pletely to glass; the a and y forms occurred in
quenches from 637°C or higher. The y form appears
to be most prevalent in the more rapid quenches.
The B, form has been observed as the major phase
in large, slowly cooled melts and in most quenches
from temperatures below 637°C. The 3, form has
been prepared in a practically pure state in the

41
ANP QUARTERLY PROGRESS REPORT

temperature range of 612 to 589°C by using samples
somewhat larger than those customarily used in
quenching experiments.

In general, the quenching data indicate that
NalUF is the most uranium-rich compound in the
NaF-UF , system. It is anticipated that a definitive
diagram of this binary system will be prepared
during the next quarter,

VISUAL OBSERVATION OF MELTING
TEMPERATURES

R. J. Sheil C. J. Barton

Materials Chemistry Division

An apparatus has been constructed that permits
visual observation of fused fluoride and chloride
mixtures at high temperatures under an inert atmos-
phere, It consists of a 2‘/2-in.-OD metal pipe
threaded at the ends and mounted vertically. Glass
disks gasketed with Teflon are fitted to the tube
ends, Large nuts are used to compress the gaskets
which form virtually gosstight connections. Water-
cooled copper coils are used to cool the furnace
ends and to protect the Teflon gaskets. The inert
gas, usually argon, is maintained at slightly above
atmospheric pressure; a slow flow of gas enters
near the bottom of the apparatus and leaves near
the top.

Samples are placed in 20-ml nickel crucibles
supported on a tripod near the center of the furnace.
A bare chromel-alumel thermocouple junction is
immersed in the fluoride sample. The thermocouple
was checked at the sodium chloride melting point
after several exposures to molten fluorides; it was
found that the slight corrosion had not affected
the emf output.

Visual observations have been confined thus far
to a few compositions in the NaF-ZrF, and the
NaF-UF, systems. However, the usefulness of the
equipment for confirmation of liquidus and solidus
temperatures has been demonstrated, and its use
in phase studies will undoubtedly increase.

DIFFERENTIAL THERMAL ANALYSIS OF THE
NaoF-ZrF, SYSTEM

R. A. Bolomey

Materials Chemistry Division

A critical study of the differential thermal-analy-
sis data collected for the NaF-ZrF
that they are consistent with petrographic and x-
ray analysis data obtained from quenching and

system shows

42

devitrification experiments, A comprehensive re-
port on the technique for differential thermal analy-
sis is being prepared, The report will describe the
theory of differential thermal analysis in a quali-
tative manner, the preparation of the samples, the
experimental technique, the interpretation of the
data, and the precision and accuracy of the method.
Suggestions for further improvements of the method
will be included,

FILTRATION ANALYSIS OF FLUORIDE SYSTEMS

R. J. Sheil C. J. Barton
R. E. Thoma
Materials Chemistry Division
G. D. White
Metallurgy Division

The filtration apporatus was diverted from phase
studies to solubility determinations during this
quarter because of the urgent need for data on the
solubility of UF; in alkali fluorides and in NaF-
ZrF, mixtures. A few experiments were made,
however, to demonstrate the solubility of UF, in
the NaF-RbF-LiF eutectic and to define the pri-

mary phase in some NaF-ZrF4 mixtures.

NaF-ZrF“

One filtration was carried out in the NaF.ZrF,
system to supplement information obtained by other
techniques, A mixture containing 63 mole % ZrF
was filtered at 552°C, and 79.5 wt % of the material
filtered. The filtrate contained 55 mole % ZrF,
and was predominantly Na,Zr,F . with a small
amount of NayZr,F,,. The residue analyzed 78.5
mole % ZrF, and was predominantly ZrF , with
some Nc:32r4F1 and NagZr F .. These results,
which demonstrate that ZrF | is the primary phase
in the 63 mole % er:4 composition, lend support
to the revised diagram for the NaF-ZrF, system
(Fig. 4.1).

NaF-LiF-RbF-UF

The solubility of UF, in the NaF-LiF-RbF eu-
tectic at 500°C was determined by filtering a
mixture of the following calculated composition
(in mole %) 5.8 NaF-40.3 LiF-49.9 RbF-4.0 UF,.
A cooling curve obtained prior to the filtration
showed o questionable thermal effect ot 542°C and
a definite break at 428°C, that is, slightly below
the melting point of the ternary alkali fluoride eu-
tectic (435°C). About 90% of the charge material
possed through the filter at 500°C. All the uranium
was apparently in the form of Rb JUF, complex.
Petrographic examination showed fhat the filtrote
contained about 10 to 20 wt % of this uranium
compound and that the residue contained approxi-
mately 50 wt %, Chemical analyses of the sepa-

rated phases gave the following compositions (in
mole %):

In Filtrate Ih Residue
LifF 48.5 . 30.3
NaF 6.9 9.4
RbF 41.7 50.2
UFA 2.8 (10.2 wt %) 10.1

The chemical analyses appear to confirm the
petrographic finding that a RbF.-UF, complex is
the primary phase that separates from the NaF-
LiF-RbF-UF, mixture. This result was not un-
expected becuuse phase studies of the alkali
fluoride—UF, systems have shown that Rb,UF,
has the h|ghesf melting point and is the most
stable of the UF, complexes that can be formed
by the components present in the NaF-LiF- RbF-
UF, system.

THERMAL ANALYSIS OF FLUORIDE SYSTEMS

L. M. Bratcher A. B, Wilkerson
Materials Chemistry Division
G. D. White T. N. McVYay, Consultant
Metallurgy Division -

Preliminary data on the NaF-ZrF ,-UF, and NaF-
ZrF sUF (-UF, systems disclose that conventional
thermal-analysis techniques do not give a reliable
indication of liquidus temperatures in these po-
tentially useful systems. The apparatus developed
for thermal analysis of chloride systems, which
consists of a flanged stainless steel container
which holds the samples in nickel crucibles, has
been found useful for thermal onalysis of these
UF ;-bearing systems. An investigation of phase
behqvmr in the NaF-CrF, system has been initiated
because of the importance of CrF. in corrosion of
Inconel and stainless steels by molten fluorides.
Work on the NaF-ThF  system has been concluded,
although the results obtained are not completely
satisfactory; should this system become important
other techniques will be requ;red for establishing
the phase behavior, -

PERIOD ENDING JUNE 10, 1954

LiF-UF,

Efforts to obtain data on mixtures in the LiF-UF,
system were made rather early in the phase-study
program, but the results were unsatisfactory be-
cause of the presence of UF, and UQ, in the fused
mixtures. Relatively pure LiF-UF, mixtures can
be produced by adding 100% excess of uranium
metal to a previously fused LiF-UF, mixture, The
few data obtained for the relatively pure mixtures
indicate that there is a eutectic containing about
25 mole % of UF,
770°C.,

that melts at approximately
F’efrogruphlc examination of the fused
mixtures showed only LiF and UF,. It is though’r
that the LiF-UF, mixtures probably compr:se a
simple eutectic sysfern

NaF-UF,

Data previously reported?+® indicated the exist-

ence of one eutectic and one compound in the
NaF-UF, system, When a mixture containing 33
mole % UF, and 67 mole % NaF was prepared by
fusing NaF with previously prepared UF,, thermal
effects were noted at 750, 695, 650, and 595°C,
These data agree quite well with the earlier data,
but the petrographic examination of the mixture
showed the presence of a green, presumably oxi-
dized, phase in addition to the lavender NaF-UF,
complex. Further study of this system is planned.

RbF-UF,

Very few compositions prepared in the RbF-UF,
system were free of UD, and UF, complexes, The
only compound which has been observed (either
Rb,UF, or RbyUF ) is described as a brown-
orange |so’rmp|c phcse with a refractive index of
1.44, The 33.3 mole % UF, composition showed
a thermal effect at 860°C. The method of preparing
alkali fluoride~UF ; mixtures by adding an excess
of uranium metal to UF, mixtures is not suitable
for the RbF-UF, system because the reduction of
RbF by uromum gives volatile metallic rubldlum.

KF-LiF-UF,

A KF-LiF-UF, (48-48-4 mole %) mixture was
prepared by reducing a previously fused KF-LiF-
UF mixture with 100% excess uranium metal.
Thermal effects were noted on the coeling curve

 

4\’. S. Colemun and W C. Whitley, ANP Quar. Prog.
Rep. Sept. 10, 1952, QRNL-1375, p 79.

5W C. Whitley, V. S. Coleman, and C, J. Barton, ANP
Qudr. Prog. Rep. Dec. 10, 1952, ‘ORNL- 1439, p 109.

43
ANP QUARTERLY PROGRESS REPORT

at 558, 540, and 487°C. It is probable, although
not yet established, that the thermal effect at
558°C represents the true liquidus temperature for
this composition. The fused melt contained about
15% of an orange, low-birefringent phase of re-
fractive index 1.43 and colorless isetropic ma-
terials with the refractive indices of KF and LiF,
These two alkali fluorides apparently do not form
a complex comparable to the LiF-RbF complex.

RbF-LiF-UF,
A Rbl'"_-!..if:~Ui'"“3 (56-40-4 mole %) mixture was

prepared by mixing uranium metal and previously
prepared UF, with the fused alkali fluorides. No
thermal effect was observed above 475°C, the
melting point of the RbF.LiF eutectic. Since the
filtration experiments reported above indicate that
the liquidus temperature for a similar composition
is above 550°C, it is concluded that the first
separation of a UF j-containing phase from ternary
mixtures in this region cannot be readily detected
by conventional thermal-analysis techniques. Pet-
rographic examination revealed that the melt con-
tained a colorless birefringent phase, presumably
a LiF-RbF complex, and a brownish isotropic
phase that was probably a RbF-UF, complex.

NaF.ZrF UF,-UF,

A series of samples with various amounts of the
UF, was prepared in sealed nickel capsules in
mixture that contained initially NaF-ZrF -UF,
(53.5-40-6.5 mole %) with sodium metal added to
reduce the UF; to UF,. When the capsules had
been heated to maximum temperatures of about
P00°C they were inverted so that the contents
would mix, Duplicate cooling curves were ob-

toined with each composition by reheating and
remixing the contents. The data obtained are
presented in Table 4.3. In the data, the dilution
of the melt by the NaF produced in the reaction is
disregarded,

The dota seem to indicate a minimum melting
point when a little more than one-half the uranium
is reduced. Petrographic examination of the fused
melts showed a regular grodation of properties in
the series. Two phases were observed in the un-
reduced mixture, the Naq(Zr,U) F“

8 solid solution
ond a small amount of colorless Nog.ZrF

o LrF . Four
phases were observed in the reduced mixtures:
UF,, o yellow isotropic material, a yellow-green
solid solution, ond Nu2ZrF6. With increasing
degrees of reduction, the solid solution became
increasingly  yellow, oand the isotropic yellow
phase, fres UF,, ond yellow-green Na, ZrF  ap-
peared in increosing amounts, This crystalline
form of Na,ZrF ., usually referred to as “normal”
Na,ZrF ,, does not, according to present knowledge,
dissolve tetravalent urgnium,

NaF-ThF,

Thermal analysis of the NaF-ThF, system has
given data which are difficult to interpret. Petro-
graphic examination of the fused melts has failed
to clarify the situation; Na,ThF . was the only
crystalline phase thot could be definitely identified.
The further study that would be needed for clarifi-
cation does not appear to be justified at this time,
and therefore the available dota are presented in
the form of a tentative diagram (Fig. 4.2)., Petro-
graphic examination of some melts made recently
indicated that the 50-50 composition is close to a
pure compound and suggested that another crystal-

TABLE 4.3. THERMAL EFFECTS OBRSERVED WiTH NaF-ZrF“-UF(UFa COMPOSITIONS

 

 

 

 

 

COMPOSITION THERMAL EFFECTS (°C)
(mole % UF,, calculated) First Cooling Curve Second Cooling Corve
0 565, 530 (halt), 490 557, 528, 492
1.5 555, 535, 510 576,525
3.0 510, 500 507, 500 (halt)
4.5 507, 495 509, 495
5.5 540, 519 (halt) 550, 520
525, 490

 

44

6.5 525, 487

 

b A0
DRN LR -DWG 1531 7

1000 b

Cle s ‘R } e

00 | - |\ e
wol e N e s LT
_,il\ffifi{%fi .

20 20 A0 80 &0 7 82 21 Thr,
Thi, [ mole %)

 

TEMPERATURE (°C)

 

 

 

 

Fig. 4.2. Tentotive Phase Diagram of the NaF-
ThF, System,

line phase observed in the 25 and 30 mole % ThF
mixtures might be Na,ThFg. Zachariasen® re-
ported the existence of this compound. These
examinations also showed that more oxide was
present in compositions formed by adding ThF, to
the 50-50 mixture than in those formed by adding
NaF. Therefore it seems probable that oxide and/
or water in the ThF  is the major source of oxygen
in the fused NaF-ThF , composition.

Lif-BeF,-ThF -UF,

Preliminary examination has been made of the
LiF-BeF -ThF ,«UF , system which has been con-
sidered as a p055|b|e fuel for a breeder-power
reactor,’ Cooling curves obtained with two mix-
tures of LiF- BeF2-T|1F UF, (48.6-48. 6-2.7-0.1
mole %) showed thermal effects at 515, 392, and
325°C., The thermal effect at 515°C was slight but
reproducible and probably represents the first
separation of a ThF -containing phase. This sepa-
ration should be confirmed by other techniques,
such as filtration or quenching, if this composition
continues to be of interest. ‘It is possible that a
composition with a higher LiF and a lower BeF,
conterit might have a lower liquidus temperature,
although presumably such a mixture would be less
efficient as a moderator.

 

Sw. H. Zachariosen, J. Am. Chem. Soc. 70, 2147
(1948).

S, Peterson, Reacfor Chemistry, ORNL CF-54-4-198
{May 12, 1954).

PERIOD ENDING JUNE 10, 1954

Na F-CrFa

Mixtures containing 5 to 65% Crf_ in the NaF-
CrF, system have been exomined, The thermal
dum do not serve to define the system uniquely.
Petrographic examination indicates the existence
of a compound which seems to be NaCrF,. Two
other crystalline materials have been observed but
neither has the optical properties of the Na,CrF,
compound previously described. & Thermal effects
are observed below the solidus temperature ot 25
mole % CrF,; it is possible that several crystal-
line modtfucahons of Na,CrF . exist.

THERMAL ANALYSIS OF CHLORIDE AND
MIXED CHLORIDE-FLUORIDE SYSTEMS

A. B, Wilkerson C. J. Barton
Materials Chemistry Division
T. N. McVay, Consultant
Metallurgy Division

Studies of chloride systems have been continued
at a somewhat diminished rate during the paost
quarter, The systems studied have included a few
mixed chloride-fluoride compositions and some
binary and ternary systems containing UCI .

UcCl,-Uci,

Conflicting statements about the UCI,-UCI, sys-
tem are found in the literature. It was stm‘ed by
Kraus? that the two compounds are only slightly
soluble in each other in the liquid state, and he
reported a liquidus break for only one composition.
The group ot Ames, lowa, !0 stated that there is a
eutectic in the system at 88 mole % UCI , with a
melting point 14°C lower than that of the UCI .
Five compositions in this system were prepared
in sealed capsules, and thermal effects were noted
on the cooling curves obtained after the mixtures
had been heated to about 850°C and the capsules
had been inverted to obtain mixing of the contents,
Samples of UCI and UCI were also heated in
sealed capsules. In uddn‘lon, one UC! -UCI, (80-
20 mole %) mixture was prepared in a mckel con-
tainer at atmospheric pressure (helium), but it was
not heated to such a high temperature as were the

 

8[... M. Bratcher and C. J. Barton, ANP Quar.
Rep Dec. 10, 1952, ORNL-1439, p 116.

%c. A. Kraus, Phase Diagrams of Some Complex 3alts
of Uranium with Halides of the Afkali and Alkaline
Earth Metals, M-251 (July 1, 1943).

mJ. J. Katz and E. Rabinewitch, The Chemistry of
Uranium, p 488, McGraw-Hill, New York, 1951,

Prog.

45
ANP QUARTERLY PROGRESS REPORT

sealed capsules because of the high vapor pres-
sure of UC|4. The data are given in Table 4.4.

The thermal data give little indication of a eu-
tectic but are not easily interpreted because of the
anomalous thermal effects noted with the twe
components in the sealed capsules. In open con-
tainers under an inert atmosphere, only one break
was noted on the UCl, cooling curve, at 840°C,
and one with UCI,, at 565°C. Both UCl; and UCt,
were noted petrogrophically in the fused ucl,
probably because of reduction of UCI, by the con-
tainer, but this does not seem to account for all
the thermal effects noted, Both UCl, and ucl,
were found in all the mixtures, except for the 12%
UCl, compesition, in which no UCI, was visible
under the microscope. It is possible that oxidation
of the UCI; may have occurred before the petro-
graphic examination could be completed,

UCI,-UF,

Only one composition was prepared in the uct -
UF, system, a 50-50 mixture corresponding to the
compound UCL,F, reported in the literature. 1 It
gave a single thermal effect at 520°C with an indi-

 

”J. W. Gates, Jr., G, H. Clewett, and H. A. Young,
Preparation of a New Mixed Halide of Tuballoy, CD-454
(duly 20, 1944); J. W. Gates, Jr., et al., A Mixed Tub-
alloy Tetrahalide, CD-460 (Aug. 26, 1944),

TABLE 4.4, THERMAL EFFECTS ON COOLING
CURVYES FOR MIXTURES OF UCI3 AND U(.‘.I4

THEORETICAL
COMPOSITION

 

THERMAL EFFECTS

 

(mole % UCI3) Cc)
100 835, 768, 520
80 797, 535, 488
60 798,(9) 540,(%) 405
40 762, 542, 497
20 622,%) 545, 531, 497
20() 670, 555
12 720, 548, 500
0 602, 540,(%) 515, 498

 

(u)lndicqfes that undercooling cccurred.
(b)Questionable thermal effect. '

(C)Prepared at atmospheric pressure.

46

catiori of undercooling, Petrographic examination
revealed three crystalline phases: ucl,, Ur,, and
a bright greenish-blue phase with refractive indices
near 1.747. The latter was the principal phase and
is believed to be the UCI,F, campound. Other
workers who have attempted to produce the com-
pound by the fusion method'? have also failed to
obtain a pure product, A probable melting point
of 460°C, based upon their work, was reported’3
for the compound.

UCI,-UF,

Five compositions in the UCI,-UF, system have
been prepared in sealed capsules. The thermal
data are too meager to permit definite conclusions
to be drawn, and petrographic examination of the
fused melts was rather difficult, The 50-50 mole
% mixture showed only one thermal effect, at 615°C,
but it is not known whether the mixture was com-
pletely liquid above this temperature, The fused
mixture was predominantly a single phase with
optical properties different from those of the two
components; this suggests the possibility of a
compound such as U,CI,F,. Such a compound
would be a novelty, since, insofar as is known
here, no mixed chloride-fluoride compound of tri-
valent uranium has been reported.

KCI-UCi,
The KCI-UCI, system, studied by Kraus,’ has

received considerable attention in this laboratory
because thermal data obtained here 46 for a part
of the system were at variance with the earlier
work, The data, particularly the liquidus breaks
in the 30 to 60 mole % UCI, region, have not been
very reproducible. The reaction of UCH, with the
nickel containers is responsible apparently for
some of the difficulty. A crystalline growth, identi-
fied as NiCl,, appears in the cooler part of the
opparatus above the fused melt; apparently it is
more volatile than the UCi, from these mixtures,

 

12J. C. Warf et al.,, Mixed Uranium Halides, CC.1785
(Sept. 10, 1944); N. W, Gregory, Compounds of Tuballoy
Containing Two or More Different Halogen Atoms, R.L.
4,6,905 (Jan. 8, 1945).

}3J. J. Katz and E. Rabinowitch, op. cit.,, p 542.

4R, J. Sheil and C. J, Barton, ANP Quar. Prog. Rep.
Mar, 10, 1953, ORNL-1515, p 109,

15R. J. Sheil and C. J. Barton, ANP Quar. Prog. Rap.
June 10, 1953, ORNL-1556, ¢ 42,

I6R. J. Sheil, S. A. Boyer, and C. J. Barton, ANP
Quar. Prog. Rep. Sept. 10, 1953, ORNL.-1609, p 58.
Petrographic examination of the fused melts indi.
cated the probable existence of three compounds
in this system, but only K,UCI, has been prepared
in sufficient purity to be readily identified. The
likely compositions for the other two compounds
are KUCI, and KU,Cl,, but since they were not
produced free of the other crystalline phases in
the system, it seems probable that they melt
incongruently. The diagram shown in Fig. 4.3,
which is based upon both petrographic and thermal
studies of the system, is believed to give a
reasonably accurate representation of the phase
relationships. No further work on the system is
planned at the present time.

RbCI-UCI,,

The RbCI-UCI, system, for which preliminary
data have been reported,!” has been studied con-
siderably less than the KCI-UCI, system. Cooling
curves with mixtures in the RbCI-UCI system con-
taining 5 to 65 mole % UCI, yielded very scattered
and confusing data, Petrographic examination of
several fused mixtures did not aid materially in the
interpretation of the thermal data because none of
the crystalline phases observed in the fused melts
had been prepared in sufficient purity for definite
identification. However, the available dato seem
to indicate that melting points ond phase relation-
ships in this system are not markedly different from
those shown for the KCI-UCl, system in Fig. 4.3.

 

”C. J. Barton and §. A, Boyer, ANP Quaor, Prog. Rep.
Dec. 10, 1953, ORNL.-1649, p 52.

OfiNL-! R-DWG 1532

   

 

 

TEMPERATURE (°C)

 

 

 

 

 

 

 

 

LGl {rmole %)

Fig. 4.3. Tentative Phase iDimgmm of KCI-UCH,
System. |

PERIOD ENDING JUNE 10, 1954

KCi-CuCl.UCI,

A eutectic in the KCI-CuCl system that melts
at 150°C was reported. A mixture corresponding
to the reported eutectic composition (65 mole %
CuCl) was prepared by using metallic copper to
reduce the CuCl, in the CuCl. The resulting mix-
ture, which showed a single thermal effect at
140°C, was colorless but contained some opaque
inclusions in the crystalline material. No thermal
effects were noted on the cooling curves when 2.5
or 5 mole % UCl; was added. Petrographic ex-
amination of the fused mixture was unsatisfactory,
however, since it disclosed only a mass of very
fine crystals with opaque material in the aggre-
gates, It is not certain, therefore, that the UCI,
dissolved in the fused KCI-CuCl mixture.

XRAY DIFFRACTION STUDIES OF THE
NaF-ZrF, SYSTEM -

P. A, Agron M. A. Bredig
Chemistry Division

A high-temperature x-ray diffractometer is being
utilized to further understanding of the phase
tronsformations that occur in the NaF-ZrF , system.
The equipment, which was originally developed '8
for the temperature range of 1500 to 2200°C, was
modified to cover the 400 to 900°C range. In order
to avoid decomposition of binary complexes by
partial vaporization, the solid salt is enclosed in
a 15-cc beryllium container, The walls were ma-
chined to a 10-mil thickness to allow free passage
of the x-ray beams. This flanged container has
been successfully gasketed to a stainless steel
mounting with a combination of gold and copper
rings. (In an initial aftempt to use a gold gasket
in contact with the beryllium, there was rapid
deleterious alloying at about 500°C,) The sample,
about 0.4 g of solid material, is held in a flat,
nickel holder. This arrangement was successfully
tested on the CsCl transition. A more detailed
description of this equipment, which is useful only
for the examination of solids, will be given ot a
future date.

In the region of 33 to 40 mole % ZrF , several
polymorphic forms of Na,ZrF, are indicated. At
the 60% NaF-40% ZrF, composition the question
of whether a solid solution in the cubic form ()}

 

IBM. A. Bredig, Chem. Div. Quar, Prog. Rep. Mar. 31,
1951, ORNL.-1053, p 114, :

47
ANP QUARTERLY PROGRESS REPORT

of NQZZrF{) {as in the analogous Na, UF , sys-
tem'?) or a new compound exists, or both, has yet
to be settled. Nine separate samples from this
composition range were heated successively over
temperature intervals from 450 to 575°C, A period
of 1 hr at a fixed temperature appeared to be suf-
ficient for equilibrating the solid phases.

An analysis of the x.ray data obtained is pre-
sented in Table 4.5, The choice of temperatures
was influenced by the results of differential thermal
andlyses and the quenching experiments described
above, The data indicate that wide ranges of solid
solutions exist in the regions examined. Diffuse
scattering would have increased the background in
the x-ray diffraction patterns if liquid was present
along with the solid. It appears that one or two
samples approximating 37 mole % ZrF, may have
reached the liquidus region at 350°C, the highest
temperature explored.  Further experiments will
be carried out before a phase diagram covering this
region is drawn.

The phase B-1 was observed for the first time
in a high-temperature x-ray pattern and appeared
to be easily preserved even on slow cooling, as
indicated for the compositions of 33,3, 33.9, and
35.0 mole % ZrF, (Table 4.5). The diffraction
pattern of the (X) phase at elevated temperatures
appears to be better developed than the pattern
observed at room temperature with a quenched
sample, ‘T-12"", The agreement of the (Z) phase
pattern with the pattern of a quenched ‘‘T.5"
sample does not appear to be so good.

The x-ray pottern at 495°C for the 40 mole %
ZrF, composition indicates the presence of B-]
and “Z5", The diffraction lines for the (40) phase
are coincident with those of *Z5"" and therefore
cannot be positively identified if the (40) phase is
present with an appreciable amount of **Z5"". Alse,
the presence of three phases would indicate that
equilibrium had not been established. At room
temperature the pattern was that for the (40) phase,
some low-temperature polymorphs of “Z6’’, and a
small amount of the **Z5"" phase. The petrographic
examinationZ2? of this mixture at room temperature
showed the presence of B-1 and the {(40) phase,
but the (40) phase here gave a slightly higher index
of refraction than that observed on the original
sample. The above observations would indicate

 

195, J. Katz and E. Rabinowitch, op. cit., p 377-382.

2OISampIe examined by H. Insley, Consultant, Metal-
lurgy Division.

48

that the (40) phase is not stable in the region of
495°C,

Some structural relationships among the com-
pounds in the NaF-ZrF  binary system have been
found to be similar to those in the NaF-UF / sys-
tem, The stucture of the compound NaUF . has
been given by Zachariasen'? as rhombohedral with
6 molecules per unit cell. The space group is R3
(ng‘)' The positions of the cations are given in
general positions, that is, H{x y z) (y z x) (z x y),
with the following parameter values:

x y z

3 } 9
U A3 43 13

6 2 5
Na 13 A3 43

There is a pseudocubic cell which contains 3.7
metal atoms and 9.2 fluorine atoms, and its struc-
ture closely resembles that of fluorite. The phase
“75""21 is isomorphous with NaUF .. The relative
intensities in the x-ray patterns of the uranivm and
the zirconium compounds are similar,

It has been noted?? that in kilogram preparations
of the NaZrF, and of compositions containing 47
mole % ZrF, the latter nonstoichiometric compo-
sition froze into a single glass-like phase, whereas
the former invariably gave mixtures of the phases
““Z5'" and "'E2"’,

For a discussion of the structure, a comparison
of the directly measured density with the density
derived from the x-ray data appeared necessary,
Therefore the densities of 5 samples of the glass-
like phase were measured, 23 and an average value
of 4,11 + 0.01 g/cc was found, Sampling from
different portions of the melt gave no apparent
difference in density value, It had also been
noticed that no change in cell size of the ‘‘Z5"
structure was apparent in the range from 43 to 52
mole % ZrF4; however, some variations in intensity
were observed. On the basis of 6 molecules per
unit cell the calculated x-ray density is 4.04 g/cc,
In order to resolve the discrepancy between the
values 4.04 and 4.11, a structure model of ‘‘Z5"
This model showed that
it was possible to add another sodium cation at

was built and examined.

 

21, tzsm phase refers to the extensive solid
solufion range for the NaZrF5 molecule.

22G. W. Watson, Materials Chemistry Division,

235. l. Cohen, Reactor Experimental Engineering Di-
vision,
6y

TABLE 4.5. HIGH-TEMPERATURE PHASES OF THE NaF-Zrf, BINARY SYSTEM

 

 

 

TEMPERATURE AND PHASES FROM HIGH-TEMPERATURE X-RAY STUDY

 

 

 

COMPGOSITION . — , -
(mole % ZfF4) Order of Heating Stages
a b c d o f
33,3(® 485°C; B-11®) 25°C: B-1 475°C: B-1 503°C: (x)¢<) 575°C; sample
major major expanded out of
sample holder
33.9 460 to 470°C; 490°C;: B-1 25°C: B-1 530 to 535°C: 550 to 555°C: 520°C: (v)
8.1 major, plus major major (X)(c) (Y){d)
unidentified
35.0 (1) 480°C: B-] 25°C: B-1 480°C: B.1 575°C: (z)y®) | a75°%C: 540°C: (¥)
major major major plus o (F unidentified
35,0 (11) 485°C: B.1 500 to 503°C: 505°C: (X) 540 to 550°C: 494°C; B.1
plus **Z5'"(?) B-1 and {X)} (X) plus & plus unidentified
36.8 460°C: B-2(9) 495°C: B-1 507°C: () 523°C: 550°C:
plus (40)“') plus unidentified plus unidentified unidentified unidentified
37.9 ~450°C; Be2 ~500°C: Bl 517°C: (x) 520°C: (X) 25°C; B-2, B.3,("
ptus (40) plus ““Z5"(?) plus unidentified and unidentified
40.0 460°C: —{40) 495°C; B-1,"Z5". | 25°C: (40) major,

 

plus unidentified

 

and {40)(7)

 

“2610’ Qnd llzsll

 

 

 

 

(a)Scmp'led fram composition C, a large batch preparation of qusz5° Remaining samples were taken from the hydrofivorinated '* X-mas tree’’ pre-

parations.,

(B

B} .phase shown petrographically to be bioxially pesitive, R.l. ¥ 1.41, Examined by G. D. White, Ceromics Department.

(C}(X) phase related to *'T-12"" phase quenched in this temperature region, Quenches of these compositions reported by R, E. Mcore, Materials

Chemistry Division,

(d)(y

) phases not identified.

{e (Z) phase may be related to "' T-5"" phase; quenched phase identified petrograghically,

(f)

Q. phase corresponds with cubic Na,ZrF .

(Q)B-Z phase may be the polymorphic form of NQZZrFé just below B.l,

LE W pattern of this (40) phase is similar to that obtained at room tempearature with a 40 mole % .Z.r'i:;‘ s.arnple..

(i)B

2 and B+3 are two biaxial opticelly negative forms of *'Z8"" normeally found at room temperature,

rso6l '_'Ol INAF ONIAN3 QOl3d
ANP QUARTERLY PROGRESS REPORT

the (0 0 0) position and that it seemed feasible to
also place an additional fluorine anion in the
F> (00 I/2) positions of the pseudocubic cell with-
out causing a noticeable change in the cell size.
The density calculated for this model was 4.16
9/cc at 53.8% NafF -46.2% ZrF ;. Maximum thermal
stability of stoichiometrically odd compositions of
solid solutions was noticed in other cases,24+25
On further additions of NaF, the cell size appears
to remain essentially constant, The structural
model may explain this, if it is assumed that, below
46.2 mole % ZrF,, substitutions of No™ for Zr*"
and vacancies in the anion positions occur and
counterbalance each other with respect to cell
size. (The introeduction of the larger sodium cation
may be offset by the removal of fluoride ions.) The
**Z5'" structure is ordered, and as more Naf is
substituted into the unit cell the ordered arrange-
ment is disrupted.

At 60% NaF —~40% ZrF , on the NaF-rich side of a
two-phase region the cations are in a disordered
arrangement which gives a unit cell similar to what
was refeired to above as the pseudocubic cell in
the “‘Z5"" lattice. Here that unit cell contains,
statistically speaking, 2.4 molecules of NaF and
1.6 molecules of ZrF4. The petrographic obser-
vations2® of this phase at room temperature have
indicated some slight birefringence. Quenching
experiments and high-temperature x-ray diffraction
studies give some indication that perhaps both
phases exist near this composition, in different
temperature regions.

An attempt was made to evaluate the effect of
changes in composition in the range between 44.2
and 30 mole % ZrF, upon the relative x-ray dit-
fraction intensities. Factors were calculated for
the ““Z5" structures for both the 6 NaF-6 ZrF
(in molecules per unit cell) and the 7 NaF ~6 ZrF4
compositions by considering only the scattering by
the cations. However, the omission of the contri-
bution by the numerous anions may account for the
lack of agreesment between observed diffraction
intensities and those calculated from structure
factors. A more favorable situation is envisaged
for a computation of the fluorite structure with the
inclusion of both cation and anion scattering.

 

25\1. A. Bredig, J. Phys. Chem. 46, 747 (1942).

26\.. G. Overholser, J. D. Redman, and C. F. Weaver,
ANP Quar. Prog. Rep. Mar. 10, 1954, ORNL -1692, p 56.

50

CHEMICAL REACTIONS IN MOLTEN SALTS

F. . Blankenship L.. G. Overholser
W. R, Grimes
Materials Chemistry Division

Chemical Equilibrio in Molten Fluorides
J. D. Redman C. F. Weaver

Materials Chemistry Division

Yalues for equilibrium constants ot 600 and at
800°C for the reduction of UF, in molten NaZrF
by Cr° and Fe® were presented in the previous
report.?® No comparable study has been completed
during the past quarter; however, a number of
experiments have been made to answer gquestions
suggested by the earlier work.

The solubility of FeF, in NaF-ZrF -UF, (50-
46-4 mole %) is 0.7 and 7.5 wt % at 600 and 800°C,
respectively. These values, which agree well with
those found previously for NaZrF ., show that small
quantities of UF, have o negligible effect on the
solubility of FeF .

Attempts to measure the solubility of Fef, in
NaZrF . at 800°C in nickel equipment demonstrated
that the reaction

2FeF, + Ni —> Nif, + 2FeF,

proceeded nearly to completion. Virtually all the
iron appeared in the filtrate as FeF,, along with
about 60% of the stoichiometric quantity of NiF,.
Examination of the solid residue indicated clearly
that the theoretical quantity of NiF, was formed
but that solubility of this material at 800°C fimited
the quantity which appeared in the filtered liquid.
The solubility of Nif, in NaZrF 5 at 800°C appears
to be about 2.5 wt %.

Experiments at 800°C have shown that the re-
action

QFer + Fe®—> 3FeF2

proceeds essentially to completion in a short time;
When

no frivalent iron was found in the filtrate.
the reaction

2Fe|:3 + 3Cr° o> 3CrF2 + 2Fe°

was attempted, no FeF, and only 200 ppm (1%) of
the original quantity of soluble iron which was
added were detected in the filtrate. The Cr**
concentration which was found approximated the
theoretical concentration.

Unsuccessful attempts have been made to measure
the solubility of CrF, in NaZrF . and the NaF.
ZrF (-UF , mixture described above. The molten
salts were difficult to filter; a solid phase which
clogged the filter or, as is somewhat less likely,
a large increase in viscosity of the melt may have
been responsible. The analyses of the few filtrates
obtained were quite inconsistent. |t does appear,
however, that virtually all the Cr3% is reduced to
Cr** by reaction with the nickel apparatus.
Some data: on the reactions

Fe® 4+ 2CrF3—>2CrF2 + FeF2
and ' '
Cro + 2CrF, —> 3GiF,

in molten NaZrF, at 600 and 800°C are shown in
Table 4.6, The results obtained at 800°C indicate
that in each case quantitative reduction to Cr*?
occurs, It is not clear from these data what occurs
at 600°C,

Preliminary aottempts have been made to study
the equilibrium

Fe + 2UF4*‘.,.:'3 QU'F3 + F@F2

in molten NaZrF, by adding UF, and FeF, rather
than Fe and UF | to the melt. Equilibrium concen-
trations of the soluble constituents agree quali-
tatively with those previously presented. iore
detailed study of this system will be conducted in
the near future. ‘

Preliminary attempts to determine the equilibrium

 

PERIOD ENDING JUNE 10, 1954

turnings are used to reduce UF, in molten NaZrFs
at 800°C have been made. The equilibrium concen-
trations are considerably fower than those found
when pure chromium is used; the activity of chro-
mium in Inconel is apparently considerably lower
than unity,

Solubility of UF, in NaF.ZrF , Mixtures
G. M. Watson C. M. Blood -

Materials Chemistry Division

The technique used in solubility experiments
was, briefly, the following: About 3 kg of an
appropriate NaF-ZrF -UF , mixture was prepared
and purified by hydrogenation and hydrofluorination
in the usual manner. The UF, was reduced to UF,
by addition, under an inert atmosphere, of an excess
of Zr® or U° to the melt at 800°C, followed by
continuous sparging with purified hydrogen, usually
overnight, to permit equilibration of the mixture.
Samples of the mixture were then taken through
filter sticks with the filtration medium consisting
of sintered nickel. Equilibration times of at least
1 hr
permitted each time the temperature was changed.

The results obtained by filtration at seven tem-
peratures in the range 600 to 900°C of o mixture
containing 50 mole % NaF, 46.8 mole % ZrF,, and

with continuous hydrogen sparging were

 

 

 

 

47
concentrations of CrF, and UF, when Inconel 3.2 mole % UF, are shown in Fig. 4.4. Since the
TABLE 4.6. REACTION OF Ee® OR Cr° WITH CrF, IN MOLTEN NquFs(a)

PRODUCTS OBSERVED IN FILTRATE

REDUCINjG(fi‘)GENT TEMPERATURE Fo (wr %) T 7 N[
USED (°C) A
Fe++ Total Cr++ Total {ppm)

Fo 800 0.38 0.37 1.50 1.02(¢) 10
800 0.41 0.38 1.45 1.01(e) 10

600 0.07 0.18 0.79 0.53 335

600 0.08 0.08 0.69 0.33 80

Cr 800 1,53(¢) 1.510€) 10
800 1.55(¢) 1.47(€) 10

600 0.96 0.74(d) 90

600 0.75 ~ 0.55( 10

 

 

 

 

 

 

 

(b)

c .
)Theorehccl value.

(a)'l wt % of Cr3T added as CrF3 to each sample.
5wt % of reducing metal added to sample.

d)Low values may be due to low solubility of CrFy at 600°C.

51
ANP QUARTERLY PROGRESS REPORT

ORNL-LR=-0OWG 1202

TEMPERATURE (°G)
700°C 600°C
I I I

 

 

 

 

 

 

URANIUM IN SOLUTION {wt %}
o
I
]

 

 

 

 

FID N

8.0 8.5 2.0 9.5 10.0 1C.5 1.0 15

4
O ek

Fig. 4.4, Solubility of UF, in NoF-ZcF .

mixture, as prepared, contained 7 wt % uranium, it
is obvious from the analyses that all the uranium
which waos added was soluble above about 815°C.
A composite sample of the residue which remained
after removal of about 10% of the batch as samples
contained 6.8 wt % uranium by analysis; it is
apparent that the pretreatment and the considerable
period at high temperatures affected the sample
only slightly if ot all, The accuracy of the data
obtained is not sufficient to establish that the
curvature of the line below 800°C in Fig. 4.4 is
real. Heats of solution as calculated from the
slope of this curve range from 8300 cal/mole at
the higher temperatures to 5600 cal/mole at the
lower temperatures,

Previous observations indicated that the
solubility of UF, increased with increasing ZrF,
concentration in the concentration range near 50
mole % ZrF . Filtration at 600°C of five different
preparations by the general technique described
above has quantitatively confirmed these obser-
vations. The dato obtained are shown in Fig. 4.5.

An increase in ZrF | concentration from 45 to 57
mole % apparently at least doubles the solubility
of UF,. Since the UF, dissolved by the 57 mole %
ZrF | mixture agreed within experimental error with
the amount added, the solubility of UF, in this
mixture at 600°C must be equal to or greater than
5.9 wt % uranium,

27,28

 

27F, F. Blankenship et al,, ANP Quar, Prog. Rep.
Dec. 10, 1952, ORNL-1439, p 119.

28J. D. Redman et al,, ANP Quar, Prog, Rep, Mar, 10,
1953, ORNL-1515, ¢ 110,

52

ORNL-LR-DWG 11398

 

o
o

o
o}

 

&
O

 

 

 

URANIUM {wt 7%}
3 |
o

g
o

 

 

 

 

 

 

 

o

 

 

 

 

 

 

40

Zrf, (mole %)

Fig. 4.5. Solubility of UF  at 600°C in NaF.

ZrF4.

The increase in solubility of UF, with increasing
ZrF, concentration implies that UF, is not com-
plexed strongly by free fluoride ion but that dis-
sociation of UF, is aided by the presence of free
ZrF . Whether disproportionation of UF, into UF,
and U° is likely to become apprecicble at a lower
temperature in such melts is, as yet, unknown.

Some auxiliary experiments performed in smaller
scale apparatus by other members of the ANP
Chemistry Group (R. J. Sheil and C. J. Barton)
have shown that a mixture prepared to contain 52.7
mole % NaF, 40.9 mole % ZrF, and 6.4 mole %
UF, when filtered at 600°C yields a filtrate con-
taining only 1.4 wt % vuranium, The residue from
this filtration comprised about 20% of the total
sample and was predominantly UF. with some
yellow crystals that were shown to be mainly
NagZr F ., with opaque material which was almost
certainly Zr°® or ZrH,, These data, which are in
general agreement with those cited above, serve to
demonstrate that UF, and not a complex uranifer-
ous fluoride is the primary phase in the concens
tration range near 50 mole % ZrF .

Solubility of UF; in NaF-KF.-LiF Mixtures

G. M. Watson C. M. Blood
Materials Chemistry Division

Two experiments were performed to determine
the solubility of UF; in the NaF-KF-LiF eutectic.
In both experiments, the UF, was obtained by
reduction in situ of the dissolved Ul , with uranium
metal, The technique used in obtaining the filtrates
for solubility determinations was otherwise identi-
cal to that used in the corresponding experiments
with NaF-ZrF | mixtures, as described above.

When a two-fold excess of metallic uranium was
used, potassium metal was distilled from the re-
actor at temperatures of 800°C and above. When
tilter sticks were introduced into the melt, the
filter media immediately became clogged. Filtrates
weighing only 1 to 15 g could be obtained in
contrast to - the 50.g filtrates obtained in normal
operations. At the end of the experiment, a metal-
lic alloy was observed on the reactor walls and as
a crust on the solidified melt,

The formation of potassium is explained when
the equilibrium constant of the reaction

KF {solution) + %U (crystal).
= K (vapor} + %UFS (solution)

is calculated from published values of free energies
of formation. At 800°C the equilibrium partial
pressure of potassium, with the assumption of ideal
solutions, may be estimated to be from 15 to 20 mm
Hg.

The constituents of the metallic alloy were
identified as nickel and uranium. The alloy forma-
tion may be rationalized from consideration of the
U-Ni phase dicgrum,29 which shows that eutectic
melting of nickel and uranium can occur at about
738°C, The eutectic phases are UNi and U Ni.

In order to avoid or minimize side effects in the
second experiment, only the stoichiometric quantity
of uranium metal was added with 10% excess of
zirconium metal, The mixture was heated to only
750°C, Much mere satisfactory filtrates were ob-
tained (an average of about 45 g above 600°C), and
no metallic alloy was observed. The concentrations
of uranium in the filtrates and in residues of beath
experiments are summarized in Table 4.7.

Petrographic examination of composite samples
of the residue from each experiment faoiled to show
the presence:of tetravalent uranium but detected
(in quantity) a bright-orange phase believed to be
a KF-UF, compound. The tabulated results appear
to show essentially constant uranium concentration
at the temperatures studied, This effect would be
shown if all the uranium were in the liquid phase
throughout the temperature range. Unfortunately,
the results are not consistent between experimenfs
and the filtrate (or residue) analyses do not agree

29The Reactor Handbook, Vol. 3, p 495, Technical
Information Service, AEC, 1953, .

 

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

PERIOD ENDING JUNE 10, 1954

TABLE 4.7, SOLUBILITY OF UF, IN
NaF-KF-LiF EUTECTIC

 

 

 

 

TEMPERATURE | SOLUBILITY OF UF, (wt % U)

S R-153* R-154%*
800 15.4
750 17.3 16.0
700 17.3 16.1.
650 17.3 163
600 17.2 16.2
550 164
208 15,9

Residue 20.1 | 17.8 -

 

 

*22.0 wt % uranium added.
**18.6 wt % uranium added.

within experimental error with the amount of uranium
added. [t is obvious, however, that the solubility
of UF, in the NaF.KF-LiF eutectic is higher than
15 wt % uranium,

Solubility of UF, in NaF-RbF-LiF Mixtures

R. J. Sheil - C. J. Barton
Materials Chemistry Division

The solubility of UF, in the NaF-RbF-LiF eu-
tectic has been determined at three temperatures
by equilibration of a melt to which sufficient UF,
and excess uranium metal had been added to
produce 4 mole % UF,, filtration of a specimen at
the desired temperaoture, and chemical analysis of
the resulting solution. Because of the high RbF
content of these mixtures, they are extremely
hygroscopic; great care in handling the samples is
necessary fo obtain reliable values for UF, solu-
bility.

The data obtained are shown in Table 4.8. The
values obtained show that 597°C is probably above
the liquidus temperature for this composition; ¢ wt
% of uranium is probably not the maximum that can
be dissolved at this temperature. Petrographic
examination of the filtrates indicated that no UF,
or complex fluorides containing tetravalent uranium
were present, The samples contdined no free UF;
no identification of the complex U3* fluoride has
yet been made. '

53
ANP QUARTERLY PROGRESS REPORTY

TABLE 4.8, SOLUBILITY OF UF5 IN NoF-RbF-LiF EUTECTIC

 

 

 

 

 

 

TEMPERATURE SAMPLE OF UF, ~—
©C) FIL TERED (%) (wt % U) LiF RbF NaF UF3
497 67 4,4 43.8 46.8 8.2 1.2
552 90 6.8 39.6 47.6 10.7 2.1
597 99 2.0 41.1 46,5 9.6 2.84

 

 

 

 

 

 

+ + +
*Calculated from analyses for Li , Na', Rb, and total U.

Chlorination of UF3
G. M, Watson C. M. Blood

Materials Chemistry Division

Analytical determinations of the UF; dissolved,
with or without added UF ,, in fluoride melts gave
values considerably lower than those anticipated
from petrographic examination or from the stoichi-
ometry of the preparation reactions, The descrip-
tion3® of the preparation and the properties of
UF,Cl suggested that chlorination of the UF, to
this compound ond subsequent determinations of
Cl~ in the product might be successfully used,

Attempts to chlorinate 1- to 2-g portions of a
200-mesh UF;-UF, mixture contained in ceramic
boats in a glass-tube reactor by exposure to dry
Cl, at 350°C for various intervals of from 2‘/2 to
20 hr yielded a green product of uniform appear-
ance. Petrographic examination showed the product
to be markedly birefringent and uniaxially (or
possibly biaxially) negative, with refractive indices
of 1.56 to 1.539. X-ray diffraction revealed no free
UF, but did reveal an unknown diffraction pattern
unrelated either to UF, or UF,. Chemical analyses
of samples chlorinated 21{2 to 41/2 hr indicated 5.7%
total chlorine; since this is roughly one-half the
expected value for chlorine, it appears that the
chlorination was incomplete or that some hydrolysis
and oxidation of the sample had occurred,

Attempts will be made to improve the contact
between the powder and the Cl, by passage of the
gos through the solid contained on a sintered-glass
filter, |If these attempts are successful, studies
with UF ;-bearing fuel mixtures will be attempted.

 

30) . Warf and F. Edwards in Chemical Research —
Generol Chemistry Report for Period of March 10 to
April 10, 1944, CC-1496 (May 11, 1944); J. C. Warf et
al., Mixed Uranium Halides, CC-1785 (Sept. 10, 1944),

3]\/. S. Coleman, C. J. Barton, and T. N. McVay, ANP
Quar. Prog. Rep. Dec. 10, 1952, ORNL.-1439, p 117,

54

Preparation of UF.2Z/F
F. P. Boody H, A, Friedman

Materials Chemistry Division

The compound UF,.2ZrF, was identified®! in
1952 as one of the materials obtained on cooling
ZrF ,-bearing melts containing large quontities of
UF,. Since large quantities of the compound had
not been prepared, the preparation of a 3,5-kg batch
was attempted to make available a quantity suf-
ficient for measurement of vapor pressure and other
physical properties. The method used was a modi-
fication of the standard hydrofluorination procedure
which involved reduction of a UF  -bearing mixture
of proper composition with zirconium metal chips
at 850°C, The product appeared homogeneous and
was a dark reddish-brown in color, Only 655 g was
filtered, since unavoidable cold spots in the trans-
fer line permitted freezing of this relatively high-
melting-point material (740°C). The compound ma-
terial, unlike virtvally all other melts observed
here, appeared to expand slightly on freezing.

Petrographic examination of the product showed
it to be approximately 99% homogensous, red-
orange, uniaxially positive UF;.2ZrF, and about
1% free UF,. X-ray diffraction analysis confirmed
the predominant phase to be UF;.2ZrF,, with
small amounts of free UF,. On warming in air, the
color of the product changes from orange-red to
oliveegreen and the refractive index increases,
Structurally the crystals do not appear to be altered
by the color change, and no evidence of oxidation
other than the color is visible.

Preporation of Various Fluorides

B. J. Sturm E. E, Ketchen
.. G. Overholser

Materials Chemistry Division

Several of the simple structural-metal fuorides,
as well as complexes of these materials with the
alkali-metal fluorides, were prepared for corrosion
experiments and some physical and chemical
property testing., In an effort to prepare the struc-
tural-metal fluorides in a high state of purity,
attempts were made to refine crude NiF, and CrF,
by treatment with BrF,. In a modification of the
apparatus originally assembled by the ANP Ana-
lytical Group, the crude material was treated with
Brf , ot 350 to 400°C for 1to 3 hr, and the product
was outgassed by thorough pumping at about 350°C.
Analyses of 10 samples indicated that considerable
conversion of oxides and/or oxyfluorides was
achieved by this procedure; however, the purity of
the product was still not of the order desired.
Large quantities (1.5 to 2 kg) of K3CrF6 and of
K,FeF, were prepared by heating appropriate
qucn’rmes of 2CtF,.7H,0 and FeF,.3H,0 with
KHF,. Several bctches of FeF,, FeF,, FeCl,,
NiF,, NiCl,, CrF,, and CrF, were syn'rhesued by

mefhods prewously described.

FUNDAMENTAL CHEMISTRY OF FUSED SALTS

EMF Medasurements

| L. E. Topol
Materials Chemistry Division

Measurements of decomposition potentials of
KCi and of a variety of chlorides in solution in
molten KCI at 850°C were attempted, Cathodes of
platinum or nickel and anodes of carbon, nickel,
zirconium, platinum, electrolytic chromium, Inconel,
and stoinless steel have been employed. Morganite
(Alzos) crucibles were used to contain the melt
in each case,

The solutions were prepared by heating 2 moles
of KC! with 1 mole of the desired anhydrous chlo-
ride to above the melting point of the mixture in a
sealed, evacuated quartz tube. More dilute so-
lutions were prepared by the addition of KCI to the
materials so obtained. The data obtained are
tabulated in Table 4.9, If the value of 0.90 volt
for electrolysis of KCl with uranium anodes is
combined with that of 3.15 volts for electrolysis
of this material with inert anodes, E for the reaction

x
U + 5 Cl, —> UCl_
appears to be 2.25 volts. This value agrees well
with the estimated E° (2.26 volts) for UCl,. The
value (1.27 volis) obtained by electrolysis of KCli

'PERIOD ENDING JUNE 10, 1954

with an anode of electrolytic chromium is less than
that previously obtained (1.5 volts) with an anode
of Cr° plated on gold wire.?2

The potentials observed for electrolysis of KCI
with Inconel electrodes suggest that Cr° is dis-
solved first and that, subsequently, Ni®is attacked;
no dissolution of Fe® was detected. However, when
the anode is a 300-series stainless steel, it appears
that first iron,and then chromium,is dissolved.

Three possible reactions of CrCl,, 4

(n 3CCl, - Cr? 4+ 2GCH,
= (.84 volt,
(2) CrCI2 -3 Cr° + CI2 ,
E° = 1.36 volts,
(3) CrCI + Ni® -—-}NICIZ + Cr

= 0, 54 volt,

may qualitatively explain the experimen’ral obser-
vations on electrolysis of solutions of CrCl,. It
appears that with a nicke! anode, reaction 3 occurs
at both CrCl, concentrations tested. With the inert
carbon anode, reaction 1 is controlfing at high
concentrations of CrCl,, while reaction 2 is im-
portant in the dilute solutions.

The experimental electrolysis of ZrCl, solutions
with nickel anodes checks well with E° values for
reactions producing ZrCl,, ZrCl,, and Zr° with
NiF,, as described prewously. 32” The values ob-
ramed with the graphite anodes may well be due to
the reactions

(4) L, — I 4 2C,

(5) Z0, + C—Zr" + CO, ,

for which the best estimates of E°are 1.9 and 1.24
volts, respectively. Attack on some of the graphite
anodes indicates that reaction 5 occurs.

The electrochemical behavior of UCI
be similar to that of CrCl

appears to
bt is likely that the

reactions
(6) 4uCi, — 3UCl, + U° ,
: E° = 1.08 volts,
(7) - 2UCH, —> 2U° + 3CH, |
E° = 2.26 volts,
(8)  2UCI, + Ni —=3NiCl, + 2U° ,

E° = 1.44 volts,

 

32, E. Topol, ANP Quar. Prog. Rep. Mar. 10, 1954,
ORNL-1692, p 63. _ |

55
ANP QUARTERLY PROGRESS REPORT

TABLE 4.9, OBSERVED DECOMPOSITION POTENTIALS FOR REDUCIBLE CHLORIDES IN

 

 

 

 

MOLTEN KCI AT 850°C
SOLUTE COEEES‘;TLTGEON ELLECTRODES BLANKETING | POTENTIAL OBSERVED
ADDED Cathode Anode ATMOSPHERE (volt)
{mole %)
None Pt U He 0.89
Pt U H2 0.90
Nij Zr He 1.03
Ni Cr* H2 1.27
Ni Incone! H, 1.85, 2.03
Ni Inconel He 1.85, 2.03
Ni Type 316 H, 1.55, 1.85
stainless steel
Ni Type 347 H2 1.43, 1.85
stainiess steel
CrCI2 33.3 Ni C H2 0.88 to 0,92
33,3 Ni Ni H:2 0.53 1o 0.57
2.0 Ni C He 1.30 to 1.43
2,0 Ni Ni He 0.65 to 0.69
ZirCI4 33.3 Ni C H2 1.67 to 1.80
33.3 Ni Ni H2 0.50 to 0,57, 0.80, 1,01
25.0 Ni C H2 1.33, 1.80 to 1.90
25,0 Ni Ni H2 0.63, 0.94 to 1.06
UCI3 100 Pt Ni H, 0.80 to 0.84, 1.01 to 1.10
100 Pt Ni He 0.80 to 0.87, 0.98
33.3 Pt C H, 0.78 to 0,85
33.3 Pt Ni H, 0.80 to 0.82
33.3 Pt Ni He 0.83 to 0.85
33,3 Ni C I'12 0.68 te 0.72
33.3 Ni Ni H2 0.78 to 0.82
2.0 Pt Ni H2 1.25 10 1.35
2.0 Pt Pt He 2.10
2.0 Pt C He 1.15, (2.08), 2.48 to 2,56
1.0 Pt Ni He 1.36

 

 

 

 

 

 

*Electrolytically prepared.

are responsible for most of the effects observed.
The ogreement between volues under He and H
suggests that UCI, is only slightly affected by H,
in the dilute solutions with relatively poor contact
efficiency used, When nickel cathodes were em-
ployed, an alloy of uranium and nickel resulted.
The general constancy of the potential obtained
with both pure UCL, and K,UCI, seems to indicate
some solubility of UCI, in ucl,.

56

A zirconia crucible (fabricated by Titanium Alloys
Division of National Lead) was heated at 750°C
for 4 hr in a hydrogen atmosphers with a charge of
NofF-KF-LiF-UF,. The vessel proved to be some-
what porous; the only evidence of attack observed
was a slight reaction to produce UO, at the cru-
cible melt interface., It is hoped that vessels of
this material may be useful for electrochemical
studies of fluoride melts,
Physical Chemistry®?

E. R. VYan Artsdalen
Chemistry D:wsnon

The density and electrical conductivity of the
molten system KCl-NaCl have been determined
over the entire composition field from temperatures
a little above the melting point of the individual
mixtures to about 1000°C. Density can be ex-
pressed as a linear function of temperature, and the
specific conductance can be expressed as a quad-
ratic function of temperature. A plot of the molar
volume vs the composition shows a slight positive
deviation from linearity, with a maximum deviation
of about 0.5% at 50 mole % KCl. Equivalent con-
ductance, A, may be expressed in the form

1000 1000
C + A —-:I-:———-+ 2302.6B IOg -—-*%‘"“* ’

log A =
where T is‘in degrees Kelvin, C is a constant of
integration, and A and B are the intercept and the
slope, respectively, of the straight line which
results from the graphical differentiation of log A
vs (1000/T). The possibility that this form of
equation is general in nature and applicable to
many molten salt systems is being investigated.
A plot of equivalent conductance vs mole per cent
KCI shows that the system is far from additive but
that it resembles the molten KCI-LiCl system re-
ported previously.®4 There is a very shallow
minimum in equivalent conductance for the KCl-
NaCl system at about 85 mole % KCl, The expla-
nation for this is apparently the same as the one
advanced in the case of the molten system KCI-
LiCL

An emf cell with transference has been designed
for use with fused salts. With such a cell it is
possible to make the potential measurements re-
quired for obtaining important
cross-checks on transference numbers and activity
coefficients of ions in melts, An initial study with
the molten system NaNO;-AgNO, indicated that
useful results can be cbtmned by measuremenfs of
cells with transference. However, the attainment
of good reproducibility poses a difficult problem.

Radiotracer techniques are being used for making
measurements of the self-diffusion coefficient of

 

33For details of this work see Chem. Div. Semiann.
Prog., Rep. Dec., 20, 1953, ORNL.1674, and Chem. Div,
Semiann, Prog. Rep. June 20, 1954 {to be published).

Mg R, Van Artsdalen, ANP: Quor. Prog. Rep. Dec,
10, 1953, ORNL 1649, p 58. _

information - and

PERIOD ENDING JUNE 10, 1954

This work has

necessitated the development of new techniques
for handling molten materials as well as for pre-

individual ions in molten salts.

cision control of high-temperature furnaces over
periods of several days. Measurements of the self-
diffusion of sodium ions in molten sodium nitrate
indicate that the diffusion coefficient is of the
order of 1 or 2 x 10~° sq cm sec™ ! at temperatures
a little above the melting point. This work is
being pursued actively and is expected to yield
important results applicable to transport processes.

A high-precision, adiabatic, heat-capacity calo-
rimeter has been built for operation in the range
up to about 1000°C. AH preliminary tests have
been completed and the calibration should be com-
pleted soon. It is anticipated that heats and free
energies of formation can be determined with good
accuracy for a number of important reactor ma-
terials. In addlhon\ fundamental information should
be obtained concerning the solid state as well as
the molten state of salts.

CHEMICAL EFFECTS OF FISSION PRODUCTS

R. F. Newton
ANP Division

Other than the absorption of neutrons, the pos-
sible undesirable effects of fission products in a
reactor are the decrease of heat transfer by sepa-
ration of o solid phase on heot exchange surfaces
and by change of viscosity or heat capacity of the
molten mixture, and the acceleration of corrosion.
Any such effects will be dependent on the QUanti-
ties of fission products which accumulate.

Quantity of Fission Products

The fission products were calculated on the
assumption that about 0.1 mole of uranium under-
goes fission per kilogram of melt at a uniform rate
during 1000 hr. The quantities given in Table 4,10
are for the end of the 1000-hr period. The yields
for the various products were taken from the smooth
curve of Siegel.3® The quantities are as large as
or farger than any which occur at earlier times and
not greatly different than they would be if the
fissions were spread over a greater time period.

Separation of Solid Phases

Of the elements listed, molybdenum, ruthenium,
rhodium, ‘and palladium are expected to separate

 

35). M. Siegel, J. Am. Chem. Soc. 68, 2411 (1946).

57
ANP QUARTERLY PROGRESS REPORTY

TABLE 4.10. FISSION-PRODUCT YIiELD AT END OF 1000 hr

 

 

 

PRODUCT YIELD* PRODUCT YIELD PRODUCT YIELD
Se 0.4 Ch 7.0 | 1.2
Br 0.25 Me 18.6 Xe 18.4
Kr 3.5 Te 5.1 Cs 16.9
Rb 4.0 Ru 14.0 Ba 6.7
Sr 12.2 Rh 1.0 La 6.5
Y 5.3 Pd 1.3 Ce 26.0
Zr 24.8 Te 1.4 TRE** 53.3

 

 

 

 

 

 

*In atoms per 100 fissions occurring uniformly over 1000 hr.

**Total rare eavths.

as the metals. Molybdenum and palladium alloy
with nicke!, and it is presumed that ruthenium and
rhodium do likewise. If they form a solid solution
with nicke! or nickel alloy and slowly diffuse into
the heat-exchanger wall, they will probably do no
harm; if they deposit on the surface and remain
there, they may reduce heat transfer becouse of
roughening of the surface.

The possible separation of either simple or com-
plex solid fluorides, because of their poor heat-
transfer properties, would be much more serious
than the deposition of metal. Rubidium may be
dismissed immediately, since it is o desirable
constituent of the fuel; chemically, cesium is also
desirable.  Zirconium likewise may already be
present in the fuel, in which case the zirconium
formed by fission will change the relative amount
of zirconium in the fuel to a negligible extent, If
zirconium intentional constifuent, it
appears impossible that it could precipitate from a
fluoride-based fuel not containing oxygen.
fluoride and sodium fluoride form a eutectic at
820°C that contains 35 mole % BaF,. The extrapo-
lated solubility of the eutectic is 14 mole % at
600°C, Corresponding data are not available for
SrF,, but an extrapolation from the eutectic of
CaF,-NaF gives a solubility of CaF, at 600°C of
12.3 mole %. Since the properties of strontium
compounds are generally between those of calcium
and barium, it is safe to assume a solubility of at
least 10 mole % for the strontium compounds, thot
is, nearly 100 times the amount produced.

is not an

Barium

The only remaining substances worthy of con-
sideration are the fluorides of yttrium and the rare
earths. These substances are so similar chemi-

58

that they undoubtedly form a continuous
seties of solid solutions and can practically be
treated as a single compound. From the phase
diagrams of alkali-metal fluorides with LaF, ond
with YF3, the lowest solubility (YF3 in KF) ex-
trapolated to 600°C is 8 mole %. Since the total of
rare earths plus yttrium is of the order of 0.5 mole
%, even here, there seems to be an ample factor of

cally

safety regarding the possibility of precipitation.

Effects on Yiscosity and Heat Capacity

The replacement of one wuranium atom by two
atoms of fission products would be expected to
increase the heat capacity per gram, but the density
would be decreased at the same time, with very
little change in the heat capacity per cubic centi-
meter. Similarly some fission products would in-
crease the viscosity while others would decrease
it. In ony case, it is safe to predict that the
changes will be no greater than about 1%,

Effects on Corrosion

It is apparent that the total number of equivalents
of cationic fission products exceeds that of anionic
fission products plus the fluoride made available
by fission of the uranium (assuming the uranium as
UF ) only if molybdenum is included as one of the
cationic elements. However, it the relative affinity
of molybdenum and nickel for fluorine is considered
(Glassner3® gives free energies of formation of
~56 and ~63 kg-cal/gram-atom of fluorine for MoF
and NiF,, respectively), it is evident that molyb-

 

36A. Glassner, A Survey of the Free Energies of For-
mation of the Fluorides, Chlorides, and Oxides of the
Elements to 2500°K, ANL-5107 (Oct. 22, 1953).
denum will appear as the element along with
In such cases,
the fission of uranium occurring as UF

ruthenium, rhedium, and palladium.
. 4 will' be
accompanied by the solution of a corresponding
amount of nickel, chromium, or other wall con-
stituent. The imbalance can be corrected by using
a fuel in which one-half of the uranium which
undergoes fission is UF,. Since it is anticipated
that about 20% of the uranium. present will undergo
fission, the change required to prevent solution of
nickel because of this cation-anion imbalance is
to have 10% or more of the uranium present as UF,,
From preliminary data available on the solubility
of UF,, it appears that even the zirconium-base
fuels should be capable of dissolving this amount,
while the alkali-metal-fluoride fuels possess a
large margin of safety in this regqrcfl

PURIFICATION AND PROPERTIES OF
ALKALI HYDROXIDES

.. G. Overholser F. Kertesz
Materials Chemistry Division

Purification of Hydroxides
E. E. Ketchen L. G. Overholser

Materials Chemistry Division

Eight batches of NaOH were purified by filtering
a 50-wt % aqueous solution through a fine sintered-
glass filter to remove the Na,CO, and then dehy-
drating at 400°C under vacuum. The product, as in
the past, contained less than 0.1 wt % of H,0 and
of Na,CO,. Four batches of commercial LiOH
were dehydrated at 200°C under vacuvum to yield
material containing less than 0.1 wt % of H,0 ‘and
0.2 wt % of Li,CO;. Approximately 5 Ib of potas-
siuvm, after bemg dwnded into l/-I‘o portions, was
reacted with water, and the resulflng solutions of
KOH were dehydrated at 400°C under vacuum, The
dehydrated KOH contained approximately 0.1 wt %
of Na and less than 0.1 wit % of K,CO, and of
H,0.

Reaction of Sodium Hydroxide with Metals

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

The hydrogen equilibrium pressures developed in
the sodium hydroxide~nickel reaction were further
investigated with the use of the previously de-
scribed quartz-jacketed nickel reaction tube which

PERIOD ENDING JUNE 10, 1954

was connected to a manometer. [t was planned that
hydrogen would be added to the jacket to reduce
the time necessary for establishment of the pressure
equilibrium, but in practice this arrangement proved
unsatisfactory, The hydrogen added to the jacket
intfroduced an imbalance in the system and thus
increased - the pressure inside the reaction: tube.
It was also recognized that errors were introduced
in the previous measurements by hydrogen being
removed when the system was evacuated during
heating up to 600°C, because the sodium hydroxide
was thus subjected fo heat and to lowering of the
pressure, - The current measurements are being
made by evacuating the system at near the melting
point of the hydroxide. The redetermined values,
which seem to approximate steady-state conditions,
were found to be higher than those reported previ-
ously. The redetermined values are given in Table
4,11, These data yield a straight line when log

P,, s plotted against 1/T as indicated in Fig. 4.6.

In addition to these measurements of the hydrogen
pressures developed during the reaction, an attempt
was made to study the mechanism involved by
determining the amount of metal dissolved. In the
first series of experiments, nickel copsules con-
taining hydrogen were sealed in quartz envelopes;
in the second series, the capsules were unjacketed
and the evolved gas was allowed to escape, The
capsules were heated for periods which ranged from
2 to 256 hr.,

It was found that the nickel concenfrqhon in the
hydroxide in the capsules with the quartz jackets
reached o plateau rather early and remained con-
stant during the prolonged heating periods. It may
be assumed that the nickel concentration reached
an equilibrium value in the jacketed system. The
nickel concentration continued to increase with
time in the unjacketed capsules, '

TABLE 4.11. HYDROGEN PRESSURE OVER
NaOH IN NICKEL

 

TEMPERATURE

 

PRESSURE
(°C) {mm Hg)
600 46
700 77
800 | 126
900 | 160
1000 205

 

59
ANP QUARTERLY PROGRESS REPORT

Similar series of experiments with sodium hy-
droxide in copper capsules showed similar be.
havior. After 256 hr of exposure at 600°C, the
copper concentration in the jocketed capsules was
about 700 ppm, while it was as high as 8000 ppm
in the wnjocketed capsules. For the jacketed
capsules the copper concentration rose to about
800 ppm at 700°C and 1100 ppm at 800°C., The
unjacketed capsules that were heated at 800°C in
a helium-filled dry box had relatively low copper
concentrations, but the walls were covered with
copper crystals., This is the first time that mass
transfer of copper has been observed under these
conditions, A larger percentage of the nickel
impurity that originally existed in the copper
capsule was found in the transferred metal than
in the wall material of the capsule.

PRODUCTION OF PURIFIED MOLTEN
FLUORIDES

F. F. Blankenship G. J. Nessle

Materials Chemistry Division

Laboratory-Scale Production

C. M. Blood G. M. Watson
F. P. Boody H. A. Friedman

Materials Chemistry Division

A total of 28 batches of various molten fluorides
was processed during the quarter; these materials
were used for phase studies, for corrosion testing,
for chemical separations research, and for studies
in kinetics of high-temperature chemical reactions,
The largest usage of these purified materials was
in the studies of UF, solubility,

Interest has recently developed in the alkali
tluoride eutectics as UF, solvents, Since there
had previously been some difficulty in preparing
the NaF-KF-LiF eutectic in a high state of purity
by the hydrogenation-hydrofluorination technique
without excessive corrosion of the apparatus, a
study of the behavior of these materials during
purification was attempted.

Two batches of the NaF-KF-LiF eutectic have
been prepared successfully in apparatus which was
assembled with special care to avoid leaks, The
solid fluorides were admitted to the apparatus and
dried at 350 to 380°C under flowing helium before
elevation of the temperature, The fluorides were
mainfained under hydrogen during the heating to
800°C and HF was admitted only after this temper-
oture was reached, One batch was treated for 100

60

UNCLASSIFIED
ORNL-LR-DWG 1333

TEMPERATURE (°C)

 

 

 

 

 

 

 

 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

1000 900 800 700 600
300 |-
200 |-
100 | |
2 I
T I
£
£ L
£ | AH=2.3R(2000)]
7 29200 cal |
oy
Ut — SR -
o
.
1
20 - -
f
T }
10 ‘l . ! ! I _____________
06 07 08 0% 10 11 1.2 1.3
‘/r(oK)Xios
Fig. 4.6. Hydrogen Pressure over Sodium Hy-

droxide in Quartz-Jacketed Nickel Capsules.

min with HF and stripped with 500 liters of H,;
the other batch was treated with HF for 70 min and
stripped with 600 liters of H,. The HF concen-
tration in the exit hydrogen from the first batch
was 2.5 x 104 mole/liter and from the second,
1.8 x 104 mole/liter.

The color of the product wes dead white in each
case, and the product appeared to be quite homo-
geneous, Chemical analyses of the products have
not been completed. Although there were deposits
of metal around the thermocouple well, on the
reactor walls, and as a partial plug in the gas-inlet
tube that were shown spectrographically to consist
of NaF, KF, and LiF, along with an alloy of nickel
and iron, attack on the container seemed to compare
faverably with that observed in processing NaF-
ZrF , mixtures. Therefore it appears that the alkali-
metal-fluoride melts can be prepared in a high
state of purity by existing techniques if the pro-~
cedure described above is followed.

Experimental Production

J. E. Eorgan C. R, Croft
J. Truitt J. P. Blakely
Materials Chemistry Division

A total of 162.7 kg of various fluoride compo-
sitions was prepared in the experimental production
facility and dispensed to various requesters in the
ANP program, The pilot-scale production facility
which will somewhat more than double the present
pilot-scale capacity is nearly completed and should
be available for operations before June 1. The
experimental production facility will then be
adapted for safe operation with BeF .

Large-Scale Production

F. A. Doss R. Reid
R. G. Wiley
ANP Division
M. S, Freed
Pratt & Whitney Aircraft Division
F. L. Daley E. F. Joseph
E. L. Youngbiood J. P. Blakely
Materials Chemistry Division

To meet the markedly increased demand for
processed fluorides, the 250-ib production facility
was placed on three-shift operation, and a total of
1358 kg of fluorides was processed. Most of this
material was transferred to containers which ranged
in size from 2 to 25 kg. Proft & Whitney Aircraft
Division received opproximately 450 kg of this
material, and, except for about 20 kg that was
delivered to Battelle Memorial Institute, the re-
mainder was dispensed to requesters in the ANP
program at ORNL.

The increasing demands for purified fluondes
necessitated, early in this calendar year, an in-
creased output of ZrF4 from the production plant
operated by Y-12 personnel. With the available
equipment, this production increose could be ac-
complished only by decreasing the hydrofluorination
time of ZrCl in the converfer and by increasing
the temperature of sublimation of the crude ZrF .
The quality of the ZrF , was, accordingly, reduced;
the iron content of the ZrF , increased from a mean
of 250 to about 1000 ppm, while the nickel content
rose from about 50 to about 200 ppm.

Although it is possible to prepare fluoride mix-
tures to rigorous specifications frem this ZrF“,
use of this material has reduced the efficiency of
the production facility to a level considerably
below that obtained during 1953 when the ARE fuel
solvent was being processed. - For example, the
gas-inlet tubes frequently become plugged by

PERIOD ENDING JUNE 10, 1954

crystals of nickel and iron during the hydrogen-
flushing operation. Since the amount of iren to be
stripped (by reduction of FeF, with hydrogen) is
much larger than that previously considered normal,
the time required for the hydrogen stripping has
increased from cbout 30 to nearly 100 hr, and
hydrogen requirements have risen from 10,000
liters to nearly 25,000 liters per 250-Ib batch of

fluoride.

Since the anticipated demand for fluorides will
probably preclude a slower production rate of ZrF,
and since improved purity at the present production
rate would require expansion of the facility,
aftempts are being mode to obtain pure Na,ZrF
from commercial sources. While ZrF4 must be
added to this material to reach the fuel compo-
sitions required, the amounts of ZrF , needed would
be well within the capacity of the Y-12 Plant for
ZrF , of acceptable purity.

It is anticipated that pure Zer (or its equivalent)
will be available within the next six to -eight
weeks. Should no commercial concern be able to
meet purity specificotions (ot a competitive price),
it may prove more economical in the long run to
use hafnium-free ZrF, for all fuel production with-
out regard to intended use. This rather startling
conclusion (based on price of production of about
3500 Ib of hafnium-free material} is dve, in part, to
the better initial purity of the hafnium-free ZrCl,
and, in part, to its higher bulk density which
permits larger loadings of converters and sublimers
and, accordingly, larger throughput of material in
existing ZrF , production equipment,

Production of Enriched Fuel for
{n-Pile Loop

J. E. Eorgan
Materials Chemistry Division

To provide the fuel for an in-pile forced-convec-
tion loop (cf., Sec. 8, “Radiation Damage’’}, two
batches of about 6 kg each of o mixture of 62.5
mole % NaF, 12.5 mole % ZrF,, 25 mole % UF
were prepared from product-level UF, by Y-12
personnel with some technical assistance from the
ANP Chemistry Group. Actual production of the
material  was accomplished during the interval

March 15 to March 29, 1954,

61
ANP QUARTERLY PROGRESS REPORTY

5. CORROSION RESEARCH

W. D. Manly
Metallurgy Division

W, R, Grimes

. Kertesz

Materials Chemistry Division
H. W. Savage
ANP Division

The static and the seesaw corrosion testing fa-
cilities were used for further studies of Inconel,
stainless steel, various cermets, ond graphite
exposed to fluoride mixiures or liquid metals.
Chromium metal additions to fluoride mixtures with
and without uranium contained in Inconel capsules
have been shown to give some protection to the
Inconel. Experiments conducted at various temper-
atures confirmed previous indications that the
chromium content of a fluoride mixture contained
in Inconel increases to a maximum and subsequently
decreases with increasing temperature. The es-
sentially linear relationship between chromium
concentration of the fluoride mixture and the surface
area of the exposed Inconel indicates that equi-
Jibrium concentrations are not attained when Inconel
specimens are exposed to fluoride mixtures in
graphite crucibles., Fluoride mixtures containing
UF, were tested in Inconel capsules in the seesaw
apparatus and, as in thermal-convection-loop tests,
no evidence of attack on the Inconel wos found.
Experiments are under way to determine the effects
of fission products on the corrosion of Inconel by
fluoride mixtures.

Tests have shown that when Inconel and type
316 stainless steel are exposed in the same system
to fluoride mixtures, the steel is inferior to Inconel
in its resistivity to attack even when the steel is
in the cold zone., Static tests confirmed the re-
ported increase in corrosiveness of lithium metal
when lithium nitride is added. Impregnation of the
surfaces of steels with chromium did not improve
the corrosion resistance of specimens tested in
sodium and several liquid-metal alloys. Extensive
tests of various cermets that are being considered
for applications as bearing materials in contact
with fluoride mixtures are under woy. The cor-
rosion resistance of cobalt-base bonded specimens
has been found to be superior to that of nickel-
Special tar-impregnated
graphite crucibles showed better corrosion resist-
ance in a fluoride mixture than in sodium. Reactor-

baose bonded specimens.

62

grade {C-18) graphite exposed to sedium in a type
304 stainless steel container was unaltered, where-
as o similarly exposed, tar-impregnated graphite
was badly cracked and spalled.

Thermaleconvection loops have been used for
additional investigotions of corrosive attack of
fluoride mixtures and liquid metals on Inconel, a
special Inconel-type alloy, stainless steel, Hastel-
loy B, and Inconel with stainless steel, nickel, or
beryllium inserts. It has been shown that it is
possible to greatly reduce the attack in Inconel
loops when UF, is used with zirconium fluoride-
base mixtures instead of the customary UF,, One
difficulty with these mixtures is the formation of
layers on the hot-leg surface which may be uronium-
Inconel alloys. Fluorides with mixtures of UF,
and UF , are being investigated, Mass fransfer of
chromium was found to oceur in thermal-convection
foops with hot-leg temperatures as low as 1200°F,
The amount of mass transfer increases with an
increase either in operating temperature or in
temperature difference between the hot and cold
legs.  When graphed, the values for depth of
attack obtained in a loop operated for 5000 hr fell
on a straight line with the values obtained in
shorter times. The previously presented data
which showed @ rapid initial ottack with a slower
secondary attack after 250 hr were confirmed. The
increase in depth of attack is about 4 mils for each
1000 hr of exposure. Reductions in depth of attack
were obtained with loops of Hastelloy B and a
Incone! with a portion of the chromium
In almost every loop

special
replaced by molybdenum,
constructed with combinations of Inconel and type
316 stainless steel, evidence of increased attack
on the steel was observed. A reduction was
usually noted in the attack on the Inconel but mass
transfer wos increased. Dissimilar-metal mass
transfer was found with beryllium metal inserts in
Inconel loops in which sodium was circulated at
1500°F. More mass transfer was found with a hot-
leg insert, but it was present even with a cold-leg
insert, Two lithium-filled type 316 stainless steel
thermal«convection loops operated for 1000 he with
hot-leg temperatures of 1445 and 1500°F, respec-
tively. Only small omounts of mass transfer oc-
curred, and ot no time was there any indication of
plugging. These results are in contrast to results
obtained with previous loops of this type which
would only operate in the neighborhood of 200 to
300 hr prior to plugging. The increased life was
probably due to the higher purity of the lithium
used and, especially, to a decrease in the lithiom
nitride content. '

In the study of the corrosion and mass-transfer
characteristics of materials in contact with liquid
lead, it was found that quartz thermal-convection
loops containing types 410 and 446 stainless steel
specimens had a much longer life prior to plugging
than similar loops containing pure iron and chro-
mium or one of the 300-series stainless steels.
Two possible explanations of these findings are
being investigated., A tenacious oxide film might
act as o diffusion barrier between the container
material and the moltern lead, or it may be that «
thin film of sigma phase forms and acts as a dif-
fusion barrier,

Studies are continuing on the identification of
the compounds produced by hydroxide.metal re-
actions. Current research is concerned with the
reaction of NaOM and nickel when hydrogen is
allowed to escape from the system. In the past,
color changes were observed when NaOH was
heated to above 500°C, Present studies indicate
that this is not due to a change in concentration
of metallic iohs from the container but that it must
be due to the existence in thermal equilibrium of
some species other than the expected sodium ions
and hydroxy! ions, Additives such as oxygen,
hydrogen, and water vapor have been studied, The
interactions of these materials on hydroxides do
not seem to be greatly altered by changing the
cation from sodium to cesium, but they do seem
to be associated with the anion.

FLUORIDE CORROSION OF INCONEL IN
STATIC AND SEESAW TESTS

H. J. Buttram R. E. Meadows
J. M. Didlake F. Kertesz
Materials Chemistry Division

Effect of Chromium Addition to Fluoride Mel?

An investigation was made of the effect of
chromium additions to the fluoride mixtures NakF-

-

PERIOD ENDING JUNE 10, 1954

ZrF“-UF‘4 :(53.5-40.0-6.5 mele %) and NaF-ZrF‘
(53-47 mole %) on the tluoride corrosion of inconel.
In a series of seesaw tests the chromium was
added as coarse powder and as slugs of three
sizes with surface-area ratios of 1:2:4 to determine
the effect of varying the surface orea of the chro-
mium, Inconel capsules which contained the fluo-
ride mixtures with and without the chromium ad-
ditions were tested. The capsules without chromium
showed light attack to a depth of about 1 mil. A
thin metallic loyer was formed on the walls of a
capsule tested with chromium powder added to the
uranium-bearing mixture; no layer was found on a
capsule containing chromium powder and the non-
vranium-bearing mixture.

The weight losses of the solid chromium speci-
mens which were added to the capsules were found
to be independent of the size of the specimen used,
but there was a marked difference between the
weight losses caused by the two fluoride mixtures.
The weight losses of the solid chromium speci-
mens averaged 0.62 g in the uranium-bearing mix-
ture and 0.11 g in the NaF-ZrF ; mixture. It was
obvious that the difference between the weight
losses in the two types of mixtures, 0.51 g in this
test, did not correspond to the amount of reduction
according to the equation

UF, + C° —>CrF, + 2UF, .

A simple calculation shows that the weight lost
by the chromium specimen would hove been suf-
ficient to reduce all the UF , contained in the 30-g
charge. However, the chromium content gained by
the fluoride mixture (about 2000 ppm) was much too
small for the reduction of the UF ;. Also, an x-ray
diffraction study of the solidified mixture revealed
only the normal pattern of Na Zr(U)gF,, and none
of the extraneous lines which would be expected if
there were a high concentration of UF,, Petro-
graphic examination showed just a tinge of the
vellow, reduced phase. The protective action
exerted by the chromium was undoubtedly due to
the higher activity of the pure metal in comparison
with that of the chromium in the Inconel. Since the
chromium content of the melt after the test was
much lower than the amount lost by the specimen,
the chromium must have been deposited and alloyed
with the wall, especially at the cold end of the
capsule. '

Effect of Temperature

The effect of temperature on corrosion of Inconel
in fluoride ‘mixtures was examined previously to

63
ANP QUARTERLY PROGRESS REPORT

establish the relationship between reaction temper-
in the melt.
These studies suggested that the chromium concen-
tration of the melt increased to a maximum and
subsequently decreased with increasing temper-
ature,

ature and chromium concentration

To check the validity of these rather surprising
tentative conclusions, itwo series of tests were
conducted, In each test both NaZrF. and NaF-
ZrF“-UFz4 (50-46-4 mole %) were used. Both series
of tests were made under static, isothermal cone
ditions for 100 hr; the test temperatures ranged
from 600 to 1000°C. In one series of experiments
the fluoride mixtures were contained in graphite
capsules to which Inconel specimens were added.
The second series of tests was made with the
fluoride mixtures contained in Inconel capsules.

Metallographic examination of the Inconel speci-
mens from the graphite capsules showed no attack
after exposure at 600°C, while after exposure af
700°C, light to moderate subsurface-void formation
to a depth of about 2 mils was observed. The
attack was more marked at 800°C, and some of the
voids were associated with carbide precipitation
on the alloy. The depth of attack increased ot
?00°C, whereas at 1000°C a slight reduction in
void formation was observed. Metallographic ob-
servation of the walls of the inconel capsules used
in the second series of tests showed considerable
attack at the lower temperatures; at the higher
temperctures the subsurface-void formation was
light and scattered, but some areas of very deep
intergranular penetration were observed. It is
possible that this general decrease in number of
subsurface voids at the higher temperatures is due
to a much higher rate of diffusion of chromium
under these circumstances and a consequent slight
depletion in chromium concentration over a large
region rather than to void formation.

The results of chemical analyses of the fluoride
mixtures for chromium paralleled to some extent
the metallographicy observations, The chromium
content of the fluoride mixtures in the inconel
capsules certainly did not increase in a regular
fashion with temperature, as might be expected.
While there appear to be slight tendencies toward
maximums in the curves of chromium concentration
vs temperature, it is likely that the differences
observed are within the experimental errors of
sampling and specimen analyses. For the samples
tested in graphite capsules, however, the chromium

64

content of the fluoride mixture increased with
temperature up to 900°C, and a straight line was
obtained when the logarithm of chromium concen-
tration was plotied against the reciprocal of
absolute temperature. The chromium values for the
mixtures tested at 1000°C appear to be somewhat
lower than the values for those tested at 900°C,

The appearance of the maximum values for the
tests at 900°C, if it is real, is difficult to explain.
The increase in chromium concentration with in-
creasing temperature suggests that equilibrium is
attained quite slowly at the lower temperatures and
"that some stage in the corrosion mechanism has a
considerable activation energy. The difference
between the behavior of specimens in graphite
capsules and those in Inconel capsules may well
be due to the formation of a carbide film which
reduces the rate at which chromium is available to
react with the melt., No effects attributable to the
presence of uranium in the fluoride mixture could
be observed in these tests.

Effect of Surface Area

In static, isothermal tests in which fluoride mix-
tures were contained in graphite capsules, two
different sizes of Inconel specimens were used in
identical quantities of fluoride mixture to establish
the influence of ratio of corroding surface to fluo-
ride volume on the quantity of chromium dissolved.
Tests were conducted at temperatures from 600
to 1000°C at 100-deg intervals; the test time was
100 hr in each case,

The surface areas of the two sizes of Inconel
specimen were in the ratio 1.88:1. Except for the
tests at 600°C, in which the chromium concentra-
tions were very similar for the two specimen sizes,
the larger specimen gave higher chromium concen«
trations in the fluoride mixture. The ratio of chro-
mium concentration in the fluoride mixtures ranged
from o minimum value of 1.5 to a maximum of 2. 1.

The essentially linear relationship between chro-

mium surface area exposed
cannot, at the chromium concentrations observed,

concenfration and

be ascribed to prior surface oxidation of the speci-
mens. It seems likely that equilibrium concentra-
tions are not aftained, possibly because of the
formation of a carbide film on the specimens, and
that the concentrations, which therefore are meas-
urements of reaction rate, are contralled by area
of surface exposed.
Effect of Reduced Phases in Fluoride Melt

The fuel containing trivalent uranium which gave
low corrosion: values in thermal-convection-loop
tests has been tested in seesaw experiments, The
UF, was obtained by adding metallic uranium in
the melt of NaF-ZrF,-UF, (53.5-40.0-6.5 mole %).
After 100 hr of exposure in the seesaw apparatus,
the unreduced control batch of the NaF-ZrF -UF,
mixture showed intermittent, moderate, subsurface-
void formation to a depth of 1 to 1.5 mils at the hot
end of the capsule, whereas none of the batches
containing UF, showed any meosurable attack.
Chemical analyses confirmed the metallographic
observations; the chromium concentration of the
unreduced control batch was higher than that of
the mixtures containing reduced uranium com-
pounds. |

Effect of Fission Products

Preliminary ' experiments were run to determine
the effect of fission products on the corrosive
properties of NaF-ZrF ,«UF, on Inconel. The total
amount of fission products per megawatt day of

PERIOD ENDING JUNE 10, 1954

operation was taken as about: 1.1 g or about 6.6 g
for four days of operation at 1.5 Mw,

Most of the fission products would be present in
amounts less than 10% of the uranium undergoing
fission; so the maximum concentration of the
product in highest abundance should be not more
than about 1 ppm because the products will be
diluted in about 1500 Ib of the fuel. Such low
concentrations are not expected to cause difficulty
from the corrosion standpoint. Since it would be
difficult to handle such small quantities in the
regular corrosion tests which require only a total
of 30 g of fluoride mixture, the concentrations of
the simuloted fission products were scaled up
one-thousandfold. '

In order to mock up a fuel containing fission-
product elements, a number of arbitrary substi-
tutions had to be made because some of the desired
elements are not availoble, Zirconium was not
included as o fission-product element because the
mixture already contained a large amount of that
element, The additions thot were made to the

NaF-Zer—UFd mixture are shown in Table 5.1,

TABLE 5.1. MATERIALS ADDED TO SIMULATE FiSSION-PRODUCT ELEMENTS IN A
' NaF-ZrF («UF, MIXTURE

 

 

QUANTITY ADDED

 

T

FISSION PRODUCT

 

 

 

 

 

 

 

 

 

MATERIAL (meiq/kg) SIMULATED
Added Before Hydrofluorination of Mixture
BaF, 110 Ba
CsF 40 Cs
CeF, 425 Rare earths
LaF, 110 Rare earths
RbF 76 Rb
YF 122 Y
Added After Hfdrofluorination of M:ixfure*
MoBr. 147 Br
Te 70 Te
RuF4 157 Pt group
Nal 83 t

 

 

 

*These materials were added after hydrofiucrination because they would have decomposed or volatilized during

hydrofluorination.

65
ANP QUARTERLY PROGRESS REPORT

Metallographic study of the Inconel-capsule walls
after the 100-hr seesaw test showed heavy to very
heavy subsurface-void formation 2 to 5 mils deep.
In some cases the voids were so concentrated near
the grain boundary that the attack appeared at first
to be of an intergranular nature, Closer inspection
revealed that the generally observed subsurface-
void attack was present, ond the voids showed a
predominant tendency to form near or at the grain
boundaries. Chemical analyses of the fluoride
mixtures showed very high chromium concentrations
that ranged from 0.23 to 0.25 wt % ond thus con.
firmed the severe attock observed metallographi-
cally.

Tests are under way to determine the effects of
the simulated fission-product materials individually
in as small amounts as possible. Preliminary re-
sults show that after the usual 100-hr seesaw test,
yttrium fluoride caused the greatest attack, with
the moximum depth of subsurface-void formation
being 3 mils, while molybdenum bromide coused
the next greatest. Chemical results confirmed, in
general, the metallographic observations. Tests
are in progress to establish the exact amount of
chromium removed from the Incone! by the various
quantities of simulated fission products.

STATIC AND SEESAW TESTS OF VARIOUS
MATERIALS IN FLUORIDE MIXTURES AND
LIQUID METALS

E. E. Hoffman J. E. Pope
W. H. Cook L. R. Trotter
Metallurgy Division

Dissimilar Metals in NuF-ZrF4wUF4

A series of tests has been conducted to determine
the effect of having dissimilar metals (type 316
stainless steel and Inconel) in contact with the
fluoride mixture NaF-ZrF -UF, (50-46-4 mole %)
in a dynamic system with a temperature differential,
Each tube was one-half type 316 stainless steel
and one-half Inconel and the joins were made by
heliarc welding. The hot zone in the seesaw
apparatus used for these tests was at 1500°F and
the cold zone was at 1256°F. Both the type 316
stainless steel and the Inconel sections were
tested in the hot zone for 100 and 300 hr, The
results of these tests are presented in Table 5.2,
The hot and cold zones and the welds joining the
two materials were examined after each fest. It
may be concluded from the results of these tests

66

that type 316 stainless steel is inferior to Inconel
in resistivity to attack by molten fluorides even
when the steel occupies the cold zone. The
heaviest attack found was to a depth of 6 mils in
the type 316 stainless steel tube adjacent to a
weld; the Incone!l tube adjacent to the opposite
side of the weld was attacked to a depth of 1 mil
{(Fig. 5.1).

Molybdenum-Coated Inconel in NaF-ZrF ,-UF

Inconel specimens which were coated with

molybdenum by the Crawford Fitting Company were
tested in static lithium at 1500°F for 100 hr,
Lithium was chosen as the comrodent in order to
determine whether the molybdenum coating was
continuous, because molybdenum is resistant to
attack by molten lithium at this temperature and
Inconel is heavily attacked. Examination of the
as-received sample revealed a ]/2-» to 3/,’4-mil molyb-
denum-rich surface layer. The molybdenum-rich
phase penetrated the grain boundaries to o depth
of 'lTé mils beneath the surface layer. Examination
of the tested specimen showed very heavy inter-
granular attack to a depth of 35 mils; therefore the
molybdenum surface layer was either not corrosion
resistant or not continuous.

Stainless Steel in Lithiom with Lithium
Nitride Added

Static tests were made to confirm the reported
increase in corrosiveness of lithium metal when
lithium nitride is added. Type 316 stainless steel
tubes were exposed at 1600°F for 100 hr to lithium
metal with 0.1, 0.25, and 1% of lithium nitride
added. The test temperature of 1600°F was chosen
because lithium nitride has a melting point of ap-
proximately 1553°F., In all the tests the 35-mil
walls of the type 316 stainless steel containers
were completely penetroted intergranularly (Fig,
5.2). In standard tests under similar conditions,
the maximum intergranular attack of type 316 stain-
less steel by lithium (without lithium nitride) is
S mils or less.

Chromalloyed Steels in Liquid Metals

Seesaw tests of chromalloyed stainless steels
in ligquid sodium and in several liquid alloys were
run to determine whether the chromium-rich layer
improved

the corrosion resistance of the base

material, The surfaces of somples of type 304

stainless stee!l and type 1035 carbon steel were
PERIOD ENDING JUNE 10, 1954

TABLE 5.2, RESULTS OF SEESAW TESTS OF DISSiMILAR METALS WELDED TOGETHER AND
EXPOSED TO NaF»Z.r-F4-'UF4 (50+46.4 mole %)

 

 

- EXPOSURE .
MATERIAL TEMPEORATURE TIME DEPTH Oi ATT CK

(FF) (he) {mils)}
Inconel (hot zone) 1500 100 0.3
Type 316 stainless steel (cold zone) 1256 2
Inconel (hot zone) 1500 300 0.5
Type 316 stainless steel (cold:zone) “1256 1
Type 316 stainless steel (hot zone) 1500 100 2
Inconel (cold zone) 1254 G
Type 316 stainiess steel (hot zone) 1500 300 ; 2.5
incone! (cold zone) ' 1256 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 5.1. Attack Adjacent to Weld Joining Type 316 Siuinl@ss Steel and Inconel After Exposure to
NnF-ZrF4-U‘F4 (50-46-4 mole %) in Seesaw Apparatus for 100 hr. During the test the temperature in the
weld zone was approximately 1382°F. 500X. fi

impregnated with chromium by the process which
uses a reaction between a gaseous chromium com-
pound and iron as a means of exchanging iron
atoms for chromium atoms at the surface of the
The parts to be treated were packed in o
compound containing
energizer, and an inert material,

steel.

powdered chromium, an

Both the type 304 stainless steel and the type
1035 carbon steel after being chromalloyed were
tested for 100 hr with hot- and cold-zone tempera-
tures of 1500 and 1220°F in sodium, 38% Pb-62%
Sn, 82% Pb-18% Cd, 32% Cd-68% Sn, 60% Bi-40%
Cd, and 45% Pb-55% Bi. In all cases the chro-

miumerich' layer was found to have cracked, and

67
ANP QUARTERLY PROGRESS REPORT

 

 

| UNCLASSIFIED 1
Y- 10604

,
e

Fd

vt
-,

 

Q.01

 

tNCH

002 |

 

Q03

 

 

 

Fig. 5.2. Intergranular Penetration of Type 316 Stainless Steel Exposed ot 1600°F for 100 hr to Static
‘Lithjum with 0.1% Lithium Nitride Added. Top edge exposed to liquid. 100X,

therefore the corrosion resistance of the base
material was unimproved. Figure 5.3 shows photo-
micrographs of a chromalloyed type 304 stainless
steel specimen before and after testing in 45%

Pb-55% Bi.
Cermets in NGF-Z.!F4»UF4

Various cermets that are being considered as
bearing materials were exposed to NaF-ZrF ,-UF ,
(50-46-4 mole %) in seesaw apparatus. The speci-
mens were restricted to the hot zone of Inconel
tubes and were exposed to the fluoride mixture for
100 hr. During the tests the hot zones of the tubes
were at 1500°F, and the cold zones were at ap-
proximately 1292°F, The results of the tests are
presented in Table 5.3. Of the various cermets of
titanium carbide with nickel binder, one specimen
with 20% of binder and another with 30% of binder
had the best resistance to attack. However, tests
of duplicate specimens showed slightly heavier
attack. This discrepancy may be due to variations
in the specimens. Kennaometal D4675 (tungsten
carbide plus 2.5% cobalt} showed very good resist-
ance, and additional tests are planned to confirm
this result. Since Corboloy X3505, which is the

same as Carboloy 608 except that it has no nickel

68

binder, was not attacked, it appears that the nickel
binder in Carboloy 608 (Cr,C, plus 2% WC and
15% Ni) was attacked preferentially {(Fig. 5.4),

A second group of cermet samples submitted
by Kennameta! lnc., and the Sintercast Corp. of
America, was exposed under the some conditions,
The exact compositions of one of the Kennametal
samples and all the Sintercast samples are yet to
be obtained from these firms. All these specimens
had fair to good resistance to the fluoride mixture,
as may be seen in Table 5.4, In general, the cor-
rosion resistance of the cobalt-base bonded speci-
mens was equal to or slightly superior to that of
the nickel-base bonded specimens, In all cases
the attack appedred o be along the titanium car-
bide particle surfoces that separate the particles
from each other and from the binder, The corrosion
along the surfaces of the titanium carbide particles
might have been the attack of a fabrication reaction
product the titanium carbide and the
binder. !t has been reported that a finely dispersed
reaction product is usually found in the metallic
areas of cermets containing cobalt or nickel.'

between

 

IC. C. McBride, H. M. Greenhouse, and T. 5. Shevlin,
J. Am, Ceram,. Soc. 35, 29-30 (1952),
PERIOD ENDING JUNE 10, 1954

 

 

 

 

 

 

 

 

 

 

Fig. 5.3. Chromalloyed Type 304 Stainless Steel Before and After Exposure to 45% Pb-55% Bi in
Seesaw Apparatus. (a) As received. Specimen nickel plated prior o examination to prevent rounding of
edge during polishing. (b) After exposure. 500X.

 

 

 

 

 

 

     

 

 

 

ST A

Fig. 5.4. Carboloy 608 (Cr,C, plus 2% WC and 15% Ni) Before and After Exposure to NaF-ZrF ,-UF
(50-46-4 mole %) for 100 hr at 1500°F in Seesaw Apparatus. (o) As received. (b) Affer exposure. 250X.

69
ANP QUARTERLY PROGRESS REPORT

TABLE 5.3. RESULTS OF SEESAW TESTS OF CERMETS EXPOSED TO
NuF-Zer-U Fa (50-46-4 mole %) AT 1500°F FOR 100 hr

 

 

MATERIAL COMPOSITION
Kentanium K-150-A TiC + 10% Ni
Kentanium K-151-A TiC + 20% Ni
Kentanium K-151-A TiC + 20% Ni
Kentanium K-152-B TiC + 30% Ni
Kentanium K-152-B TiC + 30% Ni

 

Attacked to a depth of 4 mils
A zone of slightly affected material to a depth of 1 or 2 mils

Macroexamination revealed many blisters on the surface of

the sample; attack extended to a depth of 6 mils

Attacked in only a few areas to a maximum depth of 2 mils;

much less attack than on either K<150-A or K-151-A

Attacked somewhat erratically to a depth of 4 to 7 mils; did

not resist attack so well as the other sample of K-152-B

Kentanium K-162-B TiC + 25% Ni

and 5% Mo

Kennametal D4675 WC +2.53% Co

Firth Sterling 27 TiC +7% Cra(:2

and 50% Ni

Carbolay 608 Cr3C2 +2% WC
and 15% Ni

Carboloy X3505 CryC, +2% WC

{no nickel binder)
Metamic L T-]
(not heat treated)

77% Cr--23% A|203

Metamic LT-1
(heat treated)

77% Cr~23% Al203

 

 

Macroexamination revealed a surface which had the
appearcnce of a plated specimen with portions of the

plating removed; attacked to a depth of 9 mils
Subsurface-void attack to a depth of 1 mil

Attacked to a depth of 2 to 5 mils

Attacked to a depth of 4 mils

No attack detected; both as-received and tested specimens
very brittle and similar in appearance

Completely penectrated

Completely penetrated

 

These limited tests indicate that corrosion resist-
ance is low for the specimens with the greatest
binder content, whether nickel or cobait.

Titanium boride, zirconium boride, and molybde-
num boride samples were also exposed to NafF-
ZrF ,-UF , (33.5-40.0-6.5 mole %) at 1500°F for 100
hr in seesaw apparatus, The specimens were
restricted to the hot zone of Inconél tubes. The
results of the tests are presented in Table 5.5,
The evidence of attack on the zirconium boride
specimen may be seen in Fig. 5.5.

Graphite in NaF-ZrF UF, and in Sodium

Special tar-impregnated graphite crucibles were
exposed to NaF-ZrF4-UFd (50-46-4 mole %) and
to sodium at 1500°F for 100 hr in static tests. As
can be seen in Fig. 5.6, the sodium completely
penetrated the walls of the crucible and caused
them to crack and crumble. The crucible containing

70

the fluoride mixture had a weight loss of 0.012%,
and the surfoce which was in contact with the
fluoride had an etched appearance, The fluoride
mixture did not penetrate the crucible walls,

Static tests were also run fo compare the cor-
rosion resistance in molten sodium at 1500°F for
100 hr of two types of graphite in type 304 stain-
less steel containers. The types of graphite tested
were C-18 (reactor-grade) and a special tar-impreg-
nated graphite that was impregnated ond fired 16
times during its preparation. The purpose of the
repeated impregnation and firing was to produce o
high density and a tough skin that could possibly
help to reduce penetration by various liquids., The
C-18 graphite was practically vnaltered by the test
(Fig. 5.7a). The surface of the specimen had an
etched appearance and the sharp corners had been
rounded slightly. The tar-impregnated cylinder was
cracked badly and had spalled, as may be seen

in Fig, 5.7b.
PERIOD ENDING JUNE 10, 1954

TABLE 5.4. RESULTS OF SEESAW TESTS OF TITANIUM CARBIDE CERMETS EXPOSED TO
' MaF-ZrF («UF , (50-46-4 mole %) AT 1500°F FOR 100 hr

 

MATERIAL

COMPOSITION

METALLOGRAPHIC NOTES

 

Kentanium K-138-A

Kentanium K-153-B

Kentanium K-161-B
Sintercast No. 1
Sintercast No, 4
Sintercast No. 3

Sintercast No, 8

Sintercast Ne. 10

 

65% TiC—20%

Co-15% CbTiTaC,

60% TiC~40% Ni

TiC + Ni and Mo

TiC + Co

TiC + Co

TiC + Ni

TiC + Ni

TiC + Ni

 

Attack appeared to be on binder to a depth of 0.5 to 1 mil

Attacked to a uniform depth of 3 mils; attock was along the TiC
particle surfaces that separate the TiC particles from each’

other and from the binder

No attack; tested surface slightly more irregular thon as-

received surfoce

Attacked to a depth of 0.5 to 2 mils; many voids throughout

specimen probably formed during fabrication

Affucked to a depth of 1 mil; some TiC particles had a second

phase within them

No attack; binder appeared to be firmly attached to the TiC

particles throughout the specimen

Attacked to a depth of 2 mils; small section on corner of speci-
men not uniform; appeared to be binder without TiC particles

and did not appear to be attacked

Attacked to a depih of 5 mils along the TiC particle surfaces;

aftack just enough to separate particles from the binder

 

 

 

 

GO0

INCH

 

0002

G003

 

 

 

 

Fig. 5.5. Zirconium Boride Before and After Exposure to NaF.ZrF -UF {53.5-40,0-6.5 mole %) for
100 hr of 1506°F in Seesaw Appuratuso {@) As received. (b) After exposure. 1000X. :

71
ANP QUARTERLY PROGRESS REPORT

TABLE 5.5, RESULTS OF SEESAW TESTS OF VARIOUS BORIDES EXPOSED TO
NaF«ZeF «UF , (53.5-40.0-6.5 mole %) AT 1500°F FOR 100 hr

 

 

 

DEPTH OF ATTACK

 

 

SAMPLE (mils) METALLOGRAPHIC NOTES
TiB:2 2 As-received sample had what appeared to be a pressing
crack; tested sample had a cracked corner
ZrB, 2 Irregular shaped dark and light grains; as-received
sample cracked and porous
MO2B 16 Corrosion was definite and macrescopically visible in

the polished metallographic sample; sample porous

and very brittle

 

 

 

 

UNCLASSIFIED

Y-11097 ¥-12009

  

Fig. 5.7. C.18 Graphite and Tar-lmpregnated
Graphite After Exposure to Static Sodium in Type
304 Stainless Steel Containers at 1500°F for 100
hr. (a) C-18 graphite. (b) Tar-impregnated graphite.

FLUORIDE CORROSION OF INCONEL IN
THERMAL-CONVECTION LOOPS

G. M. Adamson
Metallurgy Division

 

Fig. 5.6. Special Tar-lmpregnated Graphite Effects of UF; and Mixtures of UF; and UF,

Crucibles After Static Tests in Seodium ond in in Fluoride Fuels

NuF.Z:F -UF, (50-46-4 mole %) at 1500°F for An extensive experimental program is under way
1C0 hr. ?a) Tested in sodium. (b) Tested in flueo- to determine the effect on corrosion of Inconel
ride mixture, thermal-convection loops by circulating fluoride

72
fuels containing UF; and mixtures of UF; and UF .
it has been determined that fluoride fuels con-
taining uranium as UF3 are less corrosive fo
Inconel than UF ;-bearing fuels. However, UF is
not sufficiently soluble in the fluoride mixtures
presently being considered for use in the Reflector-
Moderated Reactor and mixtures of UF, and UF,
are therefore being investigated, An effort is
being made to determine the amount of UF; required
to ‘eliminate corrosive attack by a fluoride mixture
containing UF ; and UF ;. A few experiments have
been completed with zirconium-base fluoride mix-
tures ond tests with alkali-metal fluoride mixtures
are planned. :
The fluoride mixture NaF-ZrF ,-UF, was circu-
lated for 500 hr in a standard Inconel thermal-
convection loop with a hot-leg temperature of
1500°F (AT = 195°F). The uranium concentration
was low, being only 2,8 wt %.: Examination of the
loop after 500 hr of operation showed no subsurface
voids or intergranuvlar type of attack. The hoi-leg

PERIOD ENDING JUNE 10, 1954

surface was partially covered by a thin, unidentified
layer, but no change in the uranium analysis was
found.

A mixture with a higher, but still low, uranivm
concentration was circulated in a second loop.
The uranium concentration was 5.8 wt % and to
obtain this high a concentration, a mixture of UF,
and UF ; was required, Estimates and cnalyses of
the relative amounts of the two compounds have
varied from 100% UF, to 100% UF,. While a
definite figure cannot be obtained, a reasonable
value for the UF,
and three-fourths of the total vranium. This loop
also showed no aottack after 500 hr of operation
with o hot-leg temperature of 1500°F. Layers were
present in both the hot and the cold legs, but the
layer in the hot leg was thicker than the layer in
the cold feg (Fig. 5.8). . The average uranium
concentration of the fluoride mixture after circu-
lation was 5.3 wt %, with none of the values as
high as the value reported for the mixture before

seems to lie between one-half

 

Fig. 5.8. Layer on Surfoce of Hot Leg of an Inconel Thermal-Convection L00p After Circulation for

500 hr ot 1500°F of a NaF-ZrF , Mixture Containing UF and UF .

1400X.

73
ANF QUARTERLY PROGRESS REPORT

circulation. It is possible that the deposited ma-
terial is a uranium-Inconel alloy formed by the
dissociation of the UF, from the fluoride mixture.
This is the main question that must be answered
before a decision can be made as to whether these
fluoride mixtures will be useful as reactor fuels
circulating in Inconel,

An intermediate uranium mixture of 2.7 wt % UF,
and 8.6 wt % UF, or o total uranium concentration
of 8.7 wt % was circulated in another loop. These
proportions were obtained by calculations, but the
analytically determined uranium concentration of
8.63 wt % makes the calculated value appear to be
reasonable. After 500 hr of circulation, a moderate
concentration of subsurface voids with a maximum
penetration of 4 mils was found. No layer was
found in either the hot or the cold leg. The final
analytical values have not been received.

Effect of Hydrogen Fluoride

Small additions of hydrogen fluoride gas were
made to a batch of NaF-ZrfF,-UF (50-46-4 mole %).
Portions of this batch were then circulated for
varicus times at 1500°F in Inconel thermal-con-
vection loops in order to determine the effect of
hydrogen fluoride on the corrosiveness of the fuel
mixture, The data obtained are summarized in
Table 5.4.

The doubling of the depth of penetration in loop
415 in comparison with that in loop 421 confirms
the adverse effect of small amounts of hydrogen
fluoride on the initial rate of attack. In comparison

with the maximum penetration of 17 mils found in
loop 417 after 2000 hr of operation, an estimate
based on results obtained previously indicates that
the maximum penetration in a loop operated for
2000 hr with the as-received fluoride mixture would
be 12 mils. The effect of the hydrogen fluoride
addition on maximum penetration oppears to be
about the same after 2000 hr of operation as after
500 hr of operation,

EHect of Temperature

The effects of temperature drop and of maximum
hot-leg temperature on mass transfer in Inconel
thermal-convection loops are being studied. The
use of thermal-convection loops in the study of the
effects of temperature gradients is not entirely
satisfactory because the flow in a loop is a function
of the density of the fluid ond therefore the flow
rate changes with changes in the temperature drop.
The thermal-convection-loop resulis are of some
value, however, because the flow rates are so low
that the changes in flow rate ore not so important
as the direct changes in temperature drop.

Two series of loops in which the cold-leg temper-
ature was varied while the hot-leg temperature was
held at 1500°F were operated with NaF-ZrF -UF ,
(50.46-4 mole %) as the circulated fluid, The
variation in the cold-leg temperature was accom-
plished by adding insulation or by blowing air on
the cold leg. The data from these tests are pre-
sented in Table 5.7,

of attack is caused by impurities and therefore

The first, more rapid stage

TABLE 5.6, EFFECT OF HYDROGEN FLUORIDE ADDITIONS ON CORROSIVENESS OF
NaF-ZrF -UF, IN INCONEL THERMAL-CONVECTION LOGPS

 

 

 

 

 

CORIGINAL FLUORIDE- 7
LOOP HYDROGEN FLLUORIDE MIXTURE IMPURITY |CIRCULATION MAXIMUM
NO CONCENTRATION {ppm) TIME PENETRATION HOT-LEG ATTACK
" | (moles/liter of purging gas) (hr} {mils)
Ni Fe

421 1.2 x 105+ <10 110 500 5 intergranular, moderate
to heavy

415 3.4 x 108 300 10 Intergranular, moderate
to heavy

416 6 x 108 <10 345 1000 12 Heavy intergranular

417 6 x 10° <10 195 2000 17 Intergranular, moderate
to heavy with large
voids

 

 

 

 

 

 

 

*Normal concentration of hydrogen fluaride in as-received NqF-Zer-UFd mixture,

74
would not be expected to be greatly influenced by
the temperature drop. The differences found appear
to take place in the mass-transfer portion of the
attack which usually occurs after the first 250 to
500 hr of operation. Although no definite con-
clusions should be drawn from such meager data,
the rate of mass transfer seems to vary quite
markedly with changes in temperature drop.

A series of loops filled from one batch of NoF-
ZrF ,-UF , was operated for 2000 hr in an effort to
determine the effect on rate of mass transfer of

PERIOD ENDING JUNE 10, 1954

data obtained are presented in Table 5.8. The
increases .in depth of attack and incidence of cold-
leg deposits indicate that the rate of mass transfer
increases with increasing hot-leg temperatures,
L.oops operated previously for 500 hr with maxi-
mum hot-leg temperatures of 1200°F showed maxi-
mum penetrations of 3 to 5 mils, and in a loop
operated for 500 hr with a maximum hot-leg temper-
ature of 1250%F the maximum penetration was 3
mils. The previously operated loops were not filled
from the batch of fluoride mixture used for filling

variations in maximum hot-leg temperatures,

The

the loops listed in Table 5.8, but the small vari-

TABLE 5.7, EFFECT OF TEMPERATURE GRADIENT ON MASS TRANSFER IN INCONEL
: THERMAL-CONYECTION LOOPS CIRCULATING NaF-ZrF ,-UF,

 

 

 

L ooP MAXIMUM OPERATING MAXIMUM , .
NO Ar TIME PENETRATION HOT-LEG ATTACK
: (°F) (hr) {mils)
447 65 1000 5 Intergranular, light to mederate
443 195* 1000 10 General and intergranular, moderate to
heavy; the deepest penetrations were
scattered
446 220 1000 12 Infergranular, moderate to heavy
422 130 500 5 Intergranular, moderate to heavy
423 140 2000 6 Intergranular, moderate to heavy
442 195* 500 10 Intergranular, moderate to heavy

 

 

 

 

 

*Standard temperature drop.

TABLE 5.8. EFFECT OF VARIATIONS IN MAXIMUM HOT-LEG TEMPERATURE OF
THERMAL-CONVECTION LOOPS OPERATED FOR 2000 hr

 

 

 

 

HOT-LEG MAXIMUM :
ngp TEMPERATURE | PENETRATION HOT-LEG CONDITION COLD-LEG CONDITION
’ (“F) (mils)
391* 1200 7 "Very heavy general attack with No deposit
small veoids :
449 1250 6 Intergranular, moderate to No deposit
heavy
392 1400 8 ‘Heavy general attack to 4 mils Intermitfent
" with oceasional deeper metallic deposit
intergranular penetratiens
393 1600 13 Antergronular, moderate to Thin but continuous
heavy with large voids metallic deposit

 

 

 

 

 

*This run wos terminated after 1313 hr when circulation was stopped for the second time by a power failure.

75
ANP QUARTERLY PROGRESS REPORT

ations encountered in comparing different batches
are not sufficient to invalidate the comparison
between the previously operated loops and loops
391 and 449. The difference between the maximum
penetration in loop 391 and the maximum penetration
in the loops operated previously indicates that
mass transfer takes place even at a hot-leg temper-
ature as low as 1200°F,

Effect of Exposure Time

A maximum penetration of 27 mils was found in
loop 344 which circulated NaF-ZrF -UF , (50-46-4
mole %) for 5000 hr at 1500°F, The fluoride mixture
in this loop was from the same batch as that used
for a series of tests reperted previously. When
graphed, the values for the depth of attack in 5000
hr fell surprisingly close to the straight line for
the second stage of attack (after 250 hr), as shown
in Fig. 6.6 of the previous report,? The average
increase in depth of attack is about 4 mils for
every 1000 hr of operation ofter the first 250 hr.
A layer about !{‘-in. thick of dendritic chromium
metal crystals was found in the trap at the bottom
of the cold leg of the loop operated for 5000 hr
(Fig. 5.9).

Another series of experiments is under way to
confirm the time curve and to provide samples from
which a reaction rate may be obtained by deter-
mining the volume of the subsurface voids. The
data obtained for the time study (tabulated in
Table 5.9) confirm the data obtained previously,?

including the change of slope of the curve af
around 250 hr, However, the depth of attack found
in the loop that operated for 1000 hr is out of line
both with the other data obtained in this series of
experiments and with the data from previous experi-
ments.,

 

2G, M. Adamson, ANP Quar. Prog. Rep. Sept. 10,
1953, ORNL-1609, p 79, esp. Fig. 6.6.

 

 

Fig. 5.9. Dendritic Chromium-Meta! Crystals in
Trap at Bottom of Cold Leg of an Inconel Therma!.
Convection Loop That Circulated NoF-ZrF -UF,
(50-46-4 mole %) for 500 hr at o Hot-Leg Tempera-
ture of 1500°F. 250X. Reduced 19.5%.

TABLE 5.9. EFFECT OF OPERATING TIME ON DEPTH OF ATTACK BY NaF-ZrFA-UF4 CIRCULATING IN
INCONEL THERMAL-CONVECTION LOOPS AT A HOT-LEG TEMPERATURE OF 1500°F

 

 

 

 

 

 

LoOP OPERATING MAXIMUM
NO. TIME PENETRATION HOT-LEG ATTACK
(hr) {mils)
431 10 2 Light with intermittent moderate areas
432 50 3 Moderate
433 100 4.5 Light to moderate
434 250 8 Moderate to heavy, primarily intergranular
435 500 8 Heavy, primarily intergranular
436 1000 14 Heavy, intergranular
445 1500 13 J Heavy, intergranular

 

76
THERMAL-CONVECTION-LOOP TESTS OF
VARIOUS MATERIALS

i‘lm:l"'-iz.fF“-Ul‘:4 in Special inconel

G. M. Adamson
Metallurgy Division

A special series of alloys with slight changes
from the usual Inconel analys:s were vacuum cast
af ORNL ond drawn into /-m. tubing with a 0.045-
in, woll by the Superior Tube Co. The first lot of
tubing was made into a loop which circulated NaF-
ZrF4-UF4 (50-46-4 mole %) for 500 hr at a hot.leg
temperature of 1500°F, A portion of the chromium
normally found in Inconel had been replaced in
this special alloy by molybdenum. An analysis of
the special alloy showed 6.0% Cr, 10.0% Fe, 74.4%
Ni, and 9.8% Mo. Examination of the tubing after
500 hr showed light attack with a maximum pene.
tration of 3.5 mils, Standard Inconel loops of
commercial ,‘/f‘,-in. tubing developed moderate to
heavy attack 4 to 5 mils deep under the same
conditions.® Although the attack on the special
alloy was somewhat less than the attack on com-
mercial Inconel, the special alloy was not as cor-
rosion resistant as had been expected, Additional
loops are to be operated for Jonger periods to study
the effects of mass transfer.

NaF-ZrF -UF, in Hdsfelloy B

G. M. Adamson
Metallurgy Division

A loop of lé-in. Hastelloy B tubing was fabri.
cated by seam welding and then solution annealed
for 1 hr at 2150°F and aged 30 hr at 1950°F, The
fluoride mixture NaF-ZrF ,-UF, (50-46-4 mole %)
was then circulated in this loop for 500 hr at a
hot-leg temperature of 1500°F, Examination of the
tubing after 500 hr showed that the inner surface
of the hot leg was rough and had light subsurface
voids to a depth of 2 mils. The weld bead at the
seam showed much less aftack than the tubing
(Fig. 5.10). The cold leg showed similar attack
except that there were fewer subsurface voids.

NuFfsz“-UF“ in Stainless Steel

G. M. Adamson
Metallurgy Division

A second type 430 stainless steel loop? was
operated with. NaF-ZrF UF, (50-46-4 mole %) for
500 hr at a hot-leg temperature of 1500°F. The

PERIOD ENDING JUNE 10, 1954

hot leg developed very light, widely scaitered
subsurface voids to o maximum depth of 7 mils.
A phase change of the stainless steel to what
appeared to be martensite was noted in the vicinity
of the voids. [n an area up to 10 mils deep from
the surfoce, the carbides had apparently gone back
into solution and grain growth had occurred: (Fig.
5.11). Particles of metal had deposited on the
surface of the cold leg, and metallic crystals were
found in the odhering fluorides, While the attack
was less than that found in Inconel loops, more
metallic crystals were visible, The attack was
also slightly greater than that found in the first
type 430 stainless steel loop.*

NGF-ZIF4-UF4 in Inconel with Stainless Steel
or Nickel Inserts

G. M. Adamson
Metallurgy Division

Mass transfer of dissimilar metals was investi-
goted in a series of loops with various combinations
of Inconel and type 316 stainless steel, The fluo-
ride mixture NqF-Zer-UF‘t. (50-46-4 mole %) was
circulated in the loops for 500 hr ot a hof-ieg
temperature of 1500°F,

Inconel loop 315 had a 6-m. insert of type 316
stainless steel in the upper portion of the hot leg.?
The lnconel above the insert showed very light,
widely scattered attack to a depth of 3 mils.: The
surface of the type 316 stainless steel insert was
very rough and there was attack to o depth of 12
mils, There was a 0.3-mil-thick deposit on 'rhe
Inconel cold leg.

In inconel loop 430, the lower half of the hot leg
was replaced with type 316 stainiess steel. The
Inconel showed moderate attack to o depth of 8
mils, At the welds the Inconel was rough but
showed no other evidence of ottack. The stainless
steel had severe intergronular voids to a depth of
4 mils. There was no deposit and no attack in the
cold leg.

Loop 429 was constructed with a hot leg of type
316 stainless steel and o cold leg of Inconel. The
surface of the hot leg was very rough and showed
severe intergranular penetrations to a depth of 4

 

35, M. Adamson, ANP Quar. Prog. Rep. Mar. 10, 1954,
ORNL-1692, p 71.

4G. M. Adamson, ANP Quar. Prog. Rep. Dec. 10, 1953,
ORNL-1649, p 74. .

56, M. Adamson, ANP Quar. Prog. Rep. Sept. 10,
1953, ORNL-1609, p 77.

77
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 5.10, Hot-Leg Surface and Weld Bead of o Hastelloy B Loop After Circulation of NaF.-ZeF -UF,
(50-46-4 mole %) for 500 hr ot o Hot-Leg Temperature of 1500°F. 250X,

mils. At the welds the stainless steel was heavily
attacked to a depth of 3 mils. At the lower weld
the Incone! was unattacked, but at the upper weld
there were many subsurface voids to a depth of 3
mils in the Inconel. A scattered metallic deposit
was found on the Inconel cold leg.

In loop 428, the cold leg was type 316 stainless
steel and the hot leg was Inconel. The Inconel hot
leg showed heavy intergranular attack to a depth
of 10 mils.
rough and there was a light concentration of sub-
surface voids to a depth of 0.5 mil. The stainless
steel at the welds showed severe but intermittent
attack to a depth of 2 mils. The surface of the
stainless steel cold leg was very rough with heavy
intergranular aftack to a depth of 2.5 mils. An all-
Inconel loop (435) filled from the same batch of
fluorides showed heavy hot-leg attack to a depth
of 8 mils.

At the welds the Inconel surfoce was

78

 

 

 

pat Wl s | e
e T T a :
o % » i L
. s
- _
// o
. - ’ 4
’ 'j//‘\»"‘“'- “ * " =
a 5 . S P e
i A t ; *
1 < - = 4 .
) ) .. : . S ~
“ . - -
Lo - . " P a
: - . ¢ N B - T -
e ks b ¢ ..
: s S g s oL
LSS e R A g
- s [ " ’ ™~ . P - - Cooh—ed
- a, 1o N T e T s -

Fig. 5.11. Hot-Leg Surface of Type 430 Stain-
less Steel Loop After Circulation of NaF.ZrF -
UF4 (50-46-4 mole %) for 500 hr at o Hoi-Leg
Temperature of 1500°F. 250X. Reduced 35.5%.
in every loop tested, there was much deeper
attack of the type 316 stainless steel insert;espe-~
cially when it was in contact with Inconel in the
hotter portion of the loop, than there would have
been in an all-steel loop. The effect of the stain-
less steel on the Inconel is not so definite, but in
most cases there seemed to be a reduction in the
attack on the:Inconel, Since the data showed rather
large variations in wall thickness, the thicknesses
are being rechecked, but it seems that considerable
reductions in wall thickness occurred in all the
type 316 stainless steel tubing.

A nickel insert was also placed in the cold leg
of an Inconel loop. With this combination neither
material seemed to be affected. This experiment
confirms a previous experiment® in which a nicke!
hot-leg insert in an inconel loop was also shown
to have practically no effect,

Sodium in Inconel with Beryllium Inserts

G. M. Adamson
Metallurgy Division

Two Incone! loops were constructed with cylin-
drical hotspressed beryllium inserts about 6 in.
long. Both loops circulated sodium for 500 hr at
a hot-leg temperature of 1500°F. In one loop the
insert was in the upper portion of the hot leg. The
outer surface of the insert, which was separated
from the Inconel by a 0,020-in, annulus filled with
slow-moving sodium, showed subsurface-void at-
tack to o maximum depth of 14 mils. The inner
surface of the insert did not have the void type of
attack, but it was rough, and some even removal
had probably occurred.. The Inconel adjacent to
the beryllium was rough, and there was a metallic-
appearing deposit on the surface. A black scale
was present in all sections of this loop, and
metallic crystals were found both in the annulus
and in the cold leg. These crystals were predomi-
nantly nickel and beryllium. The black scale
produced a very complicated diffraction pattern
which could not be identified.

In the other loop the beryllium insert was in the
fower cold leg. No attack was found in the Inconel
hot leg of this loop. A thin metallic deposit was
visible on the surface of the Inconel below the
insert but not on the surface above it. The insert
showed some attack on both faces as surface

 

6(5. M. Adamson, ANP Quar. Prog. Rep., Mar. 10, 7954,
ORNL-1692, p 73.

PERIOD ENDING JUNE 10, 1954

roughness, intergranular penetrations, ond voids
to a maximum depth of 2 mils. Again, some even
removal had probably occurred. No metallic crys-
tals were found in this loop, but the same black
scale was found in all sections. Three additional
loops are now operating with similar hot-leg in-
serts, These loops are cnrculahng sodium at 900,

1100, and 1300°F.

Lithium in Stainless Steel

E. E. Hoffman W. H. Cock
J. E. Pope L. R. Trotter
Metallurgy Division

Two lithium-filled type 316 stainless steel ther-
mal-convection loops completed 1000 hr of oper-
ation. One loop operated with a hot-leg temper~
ature of 1500°F and a cold-leg temperature of
1220°F. The other loop operated with a hot-leg
temperature of 1445°F and a cold-leg temperature
of 1337°F. At no time during the tests did the
loops give any indication of plugging. There was
very little difference in the appearance of the hot
leg and the coid leg in either loop. There was no
macroscopic evidence of mass-transferred crystals
except at the bath-level line below the fill pipe
where there was a fine ring of crystals in both
loops. Metallographic examination of these Ioops
is not yet complete. '

FUNDAMENTAL CORROSION RESEARCH

G. P. Smith
Metallurgy Division

Mass Transfer in Liquid Lead

J. V. Cathcart
Metallurgy Division

The investigation of corrosion and mass transfer
in liquid lead has revealed that small quortz
thermal-convection loops containing types 410 and
446 stoinless steel test specimens require from
two to three times longer to plug than similar loops
confaining pure iron and chromium or one of the
300-series stainless steels.” An effort has been
made to find an explanation for this phenomenon.
As in previous experiments, all tests were con-

 

7ANP Quar. Prog. Rep. Sept. 10, 1953, DRNL.1609,
p 80; ANP Quar. Prog. Rep. Mar. 10, 1953, ORNL-1515,
p 128; ANP Quar. Prog. Rep. June 10, 7953, ORNL-
1556, p 64.

79
ANP QUARTERLY PROGRESS REPORT

ducted in small quartz thermal-convection loops.®

Two possible explanations for the increased
resistance to mass transfer of the 400-series stain-
less steels in liquid lead have been considered.
First, the influence of the very thin oxide films
on the specimens was investigated. It was thought
possible that the oxide formed on the 400-series
stainless steels possessed a structure which made
it especially resistant to ottack by liquid lead.
In such an event a thin oxide film would present a
barrier to the solution of material from the speci-
men walls and thus reduce the rate of mass trans-
fer. It has been demonstrated’ that relatively
thick oxide films (approximately 1000 1&) greatly
reduced mass transfer in loops containing types
304 and 347 stainless steel specimens.

An attempt was olso made to remove completely
the oxide film from type 446 stainless steel speci-
mens mounted in a loop. A small quartz tube was
sealed into the loop ot the bottom of the bend in
the cold leg and, before the loop was loaded, the
specimens were heated to 950°C while hydrogen,
which had been dried over magnesium perchlorate,
was allowed to flow through the loop. After 1 hr
of this treatment the side arm was sealed off with
a hand torch, and the loop was immediately loaded
with liquid lead. The loop plugged ofter 768 hr of
operation with hot and cold-leg temperatures of
805 and 525°C, The long time required for the
plugging of this loop was taken as an indication
that the presence of an oxide film on the specimens
was not the correct explanation for the relatively
high resistance to mass transfer in liquid lead
exhibited by type 446 stainless steel. Of course,
it is possible that the oxide on the test specimens
was only partially reduced or that, even if it were
completely reduced, encugh oxide remained in the
lead to reoxidize the surface of the test specimens.
In such an event, the results of this test would be
inconclusive; however, it is believed that the
corrosion characteristics of type 446 stainless
steel cannot be explained on the basis of the
influence of thin oxide films on the surface of the
test specimens,

The second opproach made to the problem in-
volved the assumption that a thin layer of sigma
phase formed near the surface of the types 446 and

 

8ANP Quar. Prog. Rep. Dec. 10, 1952, ORNL-1439,
p 148.

9ANP Quar, Prog. Rep. Dec. 10, 1953, ORNL.-1649,
p74.

80

410 stainless steel hot-leg specimens and served
as a barrier to solution of the specimens. The
presence of such a layer has not yet been revealed
either by metallographic exomination of the test
specimens or by an x-ray study of filings taken
from the surface of a type 446 stainless steel hot-
leg specimen, If this explanation is correct, hows-
ever, only a very thin layer of sigma phase would
be required to produce the increased resistance to
mass transfer and it may not be surprising that
such o layer has not been found.

In order to test the effect of sigma phase on
resistance o mass transfer in iron-chromiumenickel
alloys, a special alloy having a composition of
16% Ni—~37% Cr—47% Fe was prepared. This alloy
was chosen because its composition corresponds
to that in the sigma-plus-austenite region of the
nickel-chromium-iron phase diagram. Two sets of
specimens were prepared; one set was annealed
for 80 hr at 810°C to facilitate the precipitation of
the sigma phase; the other set was annealed at
1200°C and water-quenched to leave only ferrite
and austenite phases in the alioy.

Two loops containing sigma-phase (sigma and
austenite} specimens of this alloy were operated.
The first loop plugged ofter 456 hr of operation
with hot- and cold-leg temperatures of 820 and
510°C. The second loop plugged after only 306 hr
with hote and cold-leg temperatures ot 805 and
550°C. Metallographic examination of the second
loop has not yet been completed, but a transverse
section of the attacked region on the hot-leg speci-~
men from the first loop is shown in Fig. 5.12. As
may be seen, the attack was very irregular; thin
filaments of the original metal extend into o lead
Close examination revealed that each
filament consisted of a core of sigma-phase ma-
terial surrounded by a thin film of matrix metal,
The sigma-phase regions of the test specimens
appeared to have suffered much less attack than
the austenitic regions.

matrix,

The nonsigma-phase (ferrite and austenite) speci-
mens of the alloy were also tested in two loops.
The first loop plugged after 287 hr of operation
with hot- and cold-leg temperatures of 810 and
520°C. The second loop plugged after 101 hr with
hot- and cold-leg temperatures of 810 and 510°C,
A transverse section of the hot-leg specimen from
the first loop is shown in Fig., 5,13,

The lack of repreducibility in the plugging times
for both the sigma- and nonsigma-phase loops can
 

 

PERIOD ENDING JUNE 10, 1954

 

 

0,004

 

 

 

 

 

 

Fig. 5.12. Transverse Section of a Special Sigma-Phase Aflpy (16% Ni-37% Cr-47% Fe) Specimen
from the Hot Leg of a Quartz Thermal-Convection Loop Which Circulated Liquid Lead. 750X,

probably be accounted for in terms of differences
in the degree of transformation which occurred
during the heat treatments of the specimens., The
hot-leg temperature during the operation of the
loops was very close to the sigma-to-ferrite trans-
formation temperature; thus it is not unlikely that
an appreciable amount of sigma phase was either
formed or destroyed during the operation of the
loops and therefore produced the variation in
plugging times which was observed.

The results obtained with the 16% Ni-37%
Cr-47% Fe alloy indicate a definite increase in
resistance to mass transfer for sigma-phase alloys
in comparison with the nonsigma-phase alloys.
However, no definite conclusions can be drawn as
to the role of the sigma phase in the resistance to
mass transfer exhibited by the 400-series stainless
steels until a sigma-phase layer is actually ob-
served in the test specimens from a loop containing
one of these steels, |

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

Products of Hydroxide-Metal Reactions

H. L. Yakel, Jr. G. P. Smith
‘ Metallurgy Division

Studies are being made to identify the compounds
produced by hydroxide-metal reaction and to deter-
mine their properties. Previous studies were con-
cerned with the action of lithium and sodium
hydroxides on nickel in the presence of oxidizing
agents.'® Current research is concerned with the
reaction which occurs between sodium hydroxide
and nickel when hydrogen is allowed to escape
from the system, _ fi

Miller and Williams'' have shown the existence

 

19 D. Dyer, B. S. Borie, Jr., and G. P. Smith, Alkali-
Metal Nickel Oxides Containing Trivalent Nickel, ORNL-
1667 (Feb. 26, 1954).

”Reported by D. D. Williams and C. T. Ewing in
Sixth Progress Report on Thermal and Related Physical
Properties of Molten Materials, Naval Research L.ubora-
tory Problem No. 32C11.06 (1953) and preceding reports
in this series.

81
ANP QUARTERLY PROGRESS REPORT

 

Fig. 5,13. Tronsverse Section of a Special Nonsigma-Phase Alloy (16%
from the Hot Leg of a Quartz Thermal-Convection Loop Which Circulated Liquid Lead. 750X.

of an equilibrium of the type Ni + NaOH = H, + un-
known products. They reported that when this
equilibrium was displaced by evacuating at 900
to 1000°C until hydrogen evolution ceased, an
“amorphous’’ product was obtained which had a
gross composition corresponding to 1 mole of
sodium oxide and 1 mole of nickelous oxide.
Woltersdorf!? found that mixtures of sodium oxide
and nickelous oxide react at temperatures as low
as 250°C to produce various crystalline sodium
nickelate(ll) compounds. He was able to establish
that sodium orthonickelate(l!) hos the empirical
formula Na,NiO,. Kruh'3 found that a mixture of
NoOH and nickel heated to 700°C under flowing

argon gave x-ray powder patterns which did not

 

126, Woltersdorf, Z. anorg. Chem. 252, 126 (1943).
13R’eporfed by R. F. Kruh, Report for the Period April

 

1, 1953 through June 30, 1953, p 5, University of

Arkansas, Institute of Science and Technology, Fayette-
ville.

82

 

 

 

 

 

UNCLASSIFIED &
Y-12420

0.0G4

0.002

I

0

Z

0.003

0004

10,005

 

 

 

Ni—37% Cr.-47% Fe) Specimen

correspond to any patterns in the literature.
Kertesz and Knox'# induced a reaction between
sodium hydroxide and nickel by slowly removing
a relatively small amount of hiydrogen gas. X-ray
powder patterns of the resulting material did not
show any identifiable lines.

During the past few months the reaction which
takes place when hydrogen is rapidly evacuated
from the NaOH-Ni system has been studied. The
primary product of this reaction is a crystalline
sodium divalent nickelate of an as yet undetermined
formula.

In these studies the reaction chamber was a
1.27-cm-00D, 44-cm-long nickel tube (low-carbon
grade), The tube was welded closed at the bottom
end ond fastened at the top end to the rest of the
all-glass apparatus by means of a standard, ground
ball joint. The nickel tube could be evacuated

 

g, Kertesz, private communication.
through either of two circuits by turning a stopcock.
One circuit was relatively short and allowed fast
pumping speeds. The other circuit contained a
ltquid-nitrogen cold trap. The pressure of the
system was measured by means of a Pirani gage
and an Octoil-S manometer similar to that described
by Biondi.'®>  This combination of instruments
provided pressure measurements from 1 to 1000 p
with a precision of 1/5 u at the low-pressure extreme
and 20 p at the high-pressure extreme. The vacuum
was produced by a Kinney CVM 3534 mechanical
pump. Dry carbon-dioxide-free argon could be
admitted to the apparatus when desired. The
nicke! reaction tube was heated by a tube furnace.
Temperatures were measured with chromei-alumel
thermocouples which were wired to the nickel tube.
The method used for opening the reaction tubes at
the end of an experiment did not make it practical
to weld the thermocouples to the tube. Hence, the
highest temperatures reported (950°C) may have
been in error by as much as +20°C,

The nickel reaction tube was loaded in air with
about 10 g of cp-grade sodium hydroxide and then
quickly attached to the apparatus. Dehydration of

 

150, A. Biondi, Rev. Sci. Instr. 24, 989 (1953).

PERIOD ENDING JUNE 10, 1954

the hydroxide was accomplished by slowly heating
it to about 400°C under a vacuum. The circuit
containing the cold trap was used to remove the
water evolved, The course of the dehydration was
followed by observing pressure changes, and some
observations were made that are described in the
following section. | f

After the dehydration step the temperature was
increased to a predetermined value while the gases
evolved were pumped off through the short circuit.
After reaction had proceeded as far as desired, the
furnace was rapidly cooled, Dry carbon-dioxide-
free argon was then admitted to the apparatus, and
the reaction tube and its contents were removed to
a dry box for examination.

Reactions were carried out over the temperature
range of 770 to 950°C, and. the reaction products
were examined., At 770°C the pressure (assumed
to be due to hydrogen) rose to a maximum of 200
and slowly decreased to 70 p over a period of about
7 hr. The course of another reaction is shown in
Fig. 5.14, It had been intended that this reaction
would be carried out at 950°C, However, as may
be seen from Fig.5.14, the pressure passed through
its maximum before 950°C was reached. The shaded
area connects the extremes of pressure flucty-
ations, | |

UNCLASSIFIED
ORNL-LR—~DWG 1346

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000 T - 1000
T T ITT
1 : .
’ TEMPERATURE
i
!

800 e e — ] 80O
’:,"‘ P
£ &
5 600 b e e e et 4 800
£ L

&
£ £
Z 3
a )

qa,
D400 o X e e e 400 =
x : =
.
PRESSURE
200 b dopln MTm"_m— ----------------- el 200
0 L 0
D 10 20 30 40 50 60 70 g0 a0 100
TIME (min)

Fig. 5.14. Gas Evolution in the Reaction of Nickel with Sodium Hydroxide in a Continuously Evacu-

ated System,

R AT e R A LR S 8 4 8y S AR e R R LS 4S8 e At e

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

83
ANP QUARTERLY PROGRESS REPORT

In Fig. 5.14, an area A taken under the curve
for any time interval Atz is proportional to n, the
number of moles of gas evelved during the time
interval Af, The pumping speed s is defined to be
s =dV/dt, where dV is the increment of gas velume
passing through a given cross section of tubing
in the time interval 4z, By applying the ideal gas
law in its usuval form it is seen that

sP

dn = —— dt
RT

If s is constant over the pressure range under con-

sideration,
s Ly sA
n o= P dt = |
RT RT

and therefore
n ~ A

For the pressure range represented in Fig. 5,14,
the pumping speed varies by about 20% for the
Kinney CVM 3534 pump.

Application of this analysis to Fig. 5.14 shows
that the reaction NeOH + Ni = nickelate + H,
cannot be said to cease at any definite time under
the experimental conditions. The effect which the
accumulation of the reaction product has on the
diffusion of sodium hydroxide to the nickel wall
is undoubtedly the reason for this.

The nickel tubes containing the reaction products
were opened in a helium-filled dry box., Microscopic
examination showed that the reaction products
consisted of fibers and a powder, which were
subsequently separated.

The fibers, when very thin, were transparent
green, They showed no recognizable crystal faces
and were frequently bent or portially split along
the fiber axis. They dissclved readily in T N
hydrochloric acid, had an oxidizing power of less
than 0.88 meq/g (probably due to cobalt and man-
ganese impurity), and reacted with moisture and
carbon dioxide in the air to form a crust of sedium
carbonate monohydrate. The chemical data sug-
gest that the compound is a sedium nickelate(ll).

A single fiber was sealed in a glass capillary in
the dry box immediately after the nickel reaction
tube was open. This fiber, which was found to
be a single crystal, was used for preliminary
structure determination work. Rotation, Weissen-
berg, and precession photographs of the single-
crystal fiber show orthorhombic symmetry (Laue

84

symmetry mmm). Unit cell dimensions were meas-
vred from all three types of photograph, and the

suitably averaged results are:
4 - 828,A 100114,
b = 10.140 A + 0,008 A ,

o

c = 281, A + 0.008A ,
V - 236.67 x 107 cc .

The extinctions observed in the single-crystal
diffraction pottern are consistent with the three
space groups Amam, Amae2, and Amo2,, Piezo-
electric or pyroelectric experiments might distin-
guish between the centrosymmetric space group
Amam ond the other noncentrosymmetric possibili-

+

It

ties, but these experiments have not been performed
because of the lack of experimental equipment,
A statistical averaging of the observed intensities
may also permit a distinction to be made between
the possible space groups.

If the compound is assumed to consist only of
Na*, Ni2* and 0?2, chemical analysis for sodium
and nickel of alcohol-washed products gives a
composition which cannot exist. If, on the other
hand, it is assumed that nickelous hydroxide is
present, the chemical arnalysis is quantitively
satisfied by a mixture of Na,NiO, with 40 wt %
Ni(OH),. Such o mixture would be plausible. It is
not unlikely that sodium orthonickelate(ll) is the
original product of the reaction, that it reacts with
atmospheric water to form nickelous hydroxide and
sodium hydroxide, and that the sodium hydroxide
is removed by the alcohol washing.

Dehydration of Sodium Hydroxide

G. P. Smith
Metallurgy Division

It is well known that sodium hydroxide can be
dehydrated by heating to 400°C under o vaccum,
However, the details of this process have not
previously been reported. In studies of the re-
action of sodium hydroxide with nickel, as de-
scribed above, it was possible to measure the
evolution of water during the course of dehydration.
It was found that virtually all the woter evolution
occurred near the melting points of sodium hy-
droxide monchydrate and sodium hydroxide. This
result is not surprising in view of the enhanced
diffusion rates in the liquid as compared with the
solid. However, the quantitative extent of this
effect is greater than might have been predicted.
Figure 5.15 is a plot of pressure and temperature
vs time during one of four similar dehydrations.
As was shown in the preceding section, an area
under the pressure-time curve for a specified time
interval is proportional to the amount of gas
evolved during that time interval, Hence, most of
the water was-evolved within the relatively narrow
limits of two peaks or bands, the first of which
corresponds to the melting of sodium hydroxide
monchydrate and the second of which corresponds
to the melting of sodium hydroxide., The shaded
area on the second peak connects the upper and
lower limits of rapid pressure fluctuations which
were probably caused by the bursting of water-
vapor bubbles,

Color Changes in Fusea Hydroxides

C. R. Boston M. E. Steidlitz
Metallurgy Division

Color changes observed in fused hydroxides
indicate the existence of species in thermal equi-
librium other than the expected sodium ions and
hydroxyl ions. Several years ago it was noted in
this laboratory that colorless fused sodium hy-
droxide in a nickel container under an atmosphere
of hydrogen turned green when heated to above
500°C and again became colorless when cooled to
below 500°C, This color change was at first
thought to be due to a changing concentration of

PERIOD ENDING JUNE 10, 1954

nickel ions with changing temperature. Subse-
quently it was found that the same color change
took place ‘when the hydroxide was contained in
silver. These observations indicated that the
color was not due to the presence of heavy-metal
ions. This qualitative study of color changes has
been greatly extended.

The experimental method used consisted of
placing the hydroxide in a crucible which was con-
tained in a quartz tube which could be evaocuated
with a mechanical pump and to which various
atmospheres could be admitted. The crucible of
hydroxide was heated by a small pot-type furnace
in which the hydroxide could be observed. The
gases were either dried by passing them through a
liquid-nitrogen cold trap or they were saturated
with water vapor by bubbling them through water.
In most of the experiments the crucibles used were
pure sintered aluminum oxide (Morganite) to ensure
the absence of heavy-metal ions. Morganite is
very slowly and unifoermly attacked by fused hy-
droxides. However, of the varicous ceramic cru-
cibles available, Morganite was by far the most
resistant to corrosion. In a few tests, nickel
crucibles were used. The lithium, sodium, and
potassium. hydroxides used were cp grade. The
rubidium and cesium hydroxides were supplied by
L. G. Overholser of the ORNL Materials Chemistry

Division.

UNGLASSIE(ED
ORNL LR —DWE 1347

 

400

300 -

PRESSURE microns)

 

 

 

 

 

 

  
 

 

 

 

 

 

 

( I 400
! ! . s
st TEMPERATURE - e T T o) 300)
\S//M
s —_
‘ 4/ / } é}
- A ‘ =
4
200 b e e e e e X e e g 200 &
- 5 G
; ‘ a
r | 2
//, -
L[] T R et mmaaaanan e ____,,A___,,_,,__[___________,___,_“____ 100
PRESSURE '
!
e Q
50 60 70 80 90

 

TIME {(min)

Fig. 5.15. Water Evolution During the Dehydration of Sodium Hydroxide ot a Slow Heating Rate,

85
ANP QUARTERLY PROGRESS REPORT

A series of experiments was conducted in which
the hydroxides were heated to approximately 600°C
under specified atmospheres. At 6009C the initial
atmosphere was evacuated and various atmospheres
were successively admitted and evacuated. Finally
the hydroxide was cooled to room temperature,
During each step the color changes were noted.

The colors of the hydroxides (contained in
Mcrganite) which developed on standing in specified
atmospheres at 600°C are shown in Table 3.10.
For each combination of fused hydroxide and
atmosphere which was studied, the hydroxide was
colorless below approximately 500°C and distinctly
colored at 600°C. The colorless-colored transition
was reversible with changing temperature. When
the system was evocuated or filled with a specified
gas, the color changed sometimes with startling
abruptness after a brief induction period,

Lithivm hydroxide did not lend itself well to
testing because of its tendency to decompose with
the evolution of water. Sodium hydroxide, after
being heated for a long period in either air or
vacuum, changed in color from yellow to a deep
red-orange.  This color, however, disappeared
promptly on cooling or when the atmosphere was
changed,

Several experiments were conducted with sodium
hydroxide contained in nicke! crucibles, The
colors found were the same as those observed with
a Morganite container except when an atmosphere
of air wos used. In this instance, reaction with
the nickel quickly turned the hydroxide black.

Experiments were performed in which various
sodium compounds were added to sodium hydroxide
and the mixtures were heated under a vacuum fto
a temperature of about 400°C, These additives

TABLE 5.10, COLORS OF HYDROXIDES IN VARIOUS ATMOSPHERES AT 600°C

 

were of commercial purity. Metallic sodium and
sodium hydride additions gave, at first, a blue
color which, in time, turmed to green. Sodium
peroxide additions imparted an orange color. With
sodium oxide additions the hydroxide was yellow.
It should be noted that these colors were formed
ot a temperature at which the pure hydroxide would
have been colorless. All the colors were retained
vpon cooling the hydroxide to below its melting
point.

The results of these experiments are still too
incomplete and too qualitative to permit specific
conclusions as to the species giving rise to the
various colors, Nevertheless, a number of tentative
speculations can be made. First, various simple
atmospheres such as oxygen, hydrogen, and water
vapor can interact with the fused alkali-metc!
hydroxide to produce different chemical species
which can be characterized, at least to a limited
extent, by their color, These interactions do not
seem to be greatly altered by progressively chang-
ing the cation from sodium to cesium, Hence, they
seem to be associated with the anion. The extent
of interaction is rapidly established and thereafter
seems to change only very slowly with time at
constfant temperature, The extent of interaction
seems fo increase markedly with incregsing temper.
ature.

It is significant that these hydroxide-gas inter-
actions take place in the absence of free metal, It
is well khown that the atmosphere plays a dominant
role in controlling the mass transfer of nickel, It
is not entirely unlikely, then, that the chemical
reactions that cause mass transfer do not take
place directly between nicke! and the hydroxyl ion,
as is generally believed, but rather between nickel

 

 

 

 

 

 

HYDROXIDE
Air Vacuum
LiOH Yellow
NaOH Yellow, Yellow,
orange orange
KOH Yellow Yellow
RbOH Yellow Yeliow
CsOH Yellow Yellow

 

 

 

 

 

 

86

 

 

H H Water Vapor
2 € in H2 or He
b -
Green Yellow Colorless
Green
Graen
Green
and a chemical species whose concentration is
dependent on an interaction between the hydroxyl
ion and the atmosphere.

Preliminary measurements have been made of the
electrical conductivity of fused sodium hydroxide
contained in silver exposed to the air at temper-
atures up to 550°C. The results are in fair agree-

PERIOD ENDING JUNE 10, 1954

ment with those reported by Aradt and Ploetz'®
who give data up to 450°C, It is interesting to
note that the conductivity was found to rise rapidly
over the temperature range for which a color chonge
is first noted.

 

161, Arndt and G. Ploetz, Z. physik. Chem. 110, 237
(1924).

87
ANP QUARTERLY PROGRESS REPORT

6. METALLURGY

W. D. Manly
Metalturgy Division

The studies of the mechanical properties of
[nconel in contact with the fused fluorides have
shown that material stressed under uniaxial con-
ditions has a much longer rupture life than material
tested under multiaxial conditions. Since in ony
ANP type of reactor the material will be under
multiaxial stresses, the testing of alloys for this
service should be performed by using tube-burst
tests. The uniaxial test conditions are much
easier to control and will give a much better in-
sight into the effect of environments, but for ob-
taining actual design data, the tube-burst tests
or other multiaxial-stress tests ore preferred. It
has been found that the load-carrying abilities of
Inconel, as well as the corrosion, are also quite
dependent on the ratio of surface area to volume
of fluoride mixture. As the rotio of surface area
of container material to volume of fluoride mixture
decreases, the rupture life for a given siress is
also decreased.

Investigations of materials svitable for high-
conductivity fins have been primarily concerned
with the protection of copper from oxidation and
with the prevention of diffusion between the
cladding material and the copper. This work has
shown that copper clad with types 310, 446, or
430 stainless steel or with Inconel is quite satis-
tactory in the unstressed condition if a suitable
diffusion barrier is provided. In oxidation tests
of the clod material under the stress, however, it
was found that high stresses greatly increase the
oxidation of the material. Additional tests will
be made at lower stresses and for {onger times.

In the search for container materials for fluid
fuels, Hastelloys B and C were found to have very
poor extrusion characteristics. A comparison with
the results of extrusion experiments with nickel-
molybdenum special alloys indicates that the diffi-
culty in extruding Hastelloy B was probably caused
by the impurities in the material and the poor
melting practice used in its preparation, rather
than poor lubrication during extrusion. The work-
hardening characteristics of columbium were de-
termined,

Sheets of high-boron-content material will be

88

required for shielding between the moderator and
the heat exchanger and between the heat exchanger
and the pressure shell of the Reflector-Moderated
Reactor. Powder compacts of boron carbide
blended with copper and with silver powders were
hot pressed, and boron densities which approached
the density required were obtained. It is thought
that the required density can be obtained by warm-
pressing compacts of boron carbide and copper,

High-temperature oxidation-resistance tests of
several brazing alloys at 1500 and 1700°F for 200
and 500 hr have shown that the majority of the
nicke! and nickel-chromium-base alloys are suit-
able for service in an oxidizing atmosphere at
1500°F and that severa! of the alloys retain this
resistance at 1700°F,

A new semiautomatic heliarc-welding process for
production of tube-to-header joints is described
and its use in the construction of a prototype
sodium-to-air radiator is illustrated. The fabri-
cation of a radiator with high-conductivity fins
requires a brazing alloy that has good strength
at 1600°F and flowability at approximately 1900°F.
Coast Metals alloy No, 52, which will svitably
wet stainless steels ond flow in dry-hydrogen
atmospheres, has been found to be satisfactory
for this use. In addition to having a suitable
alloy, the furnace temperature must be very care-
fully controlled so that there will be a minimum
of thermal variation over the assembly. To test
the brazing alloys and te check the controls of
the furnace, a prototype radiator with high-con-
ductivity fins was constructed. The successful
fabrication of this prototype radiator indicates
that such complicated configurations can be con-
structed,

An assembly consisting of a beryllium plate
canned in an Inconel assembly in which the be-
ryllium will be heated by electrical resistance and
cooled by sodium passing through holes in the
beryllium plate was fabricated. This unit will be
used for determining the effect of thermal stresses
and thermal cycling on beryllium metal and for
studying the compatibility of sodium and beryllium,
STRESS-RUPTURE TESTS OF INCONEL

R. B. Oliver D. A. Douglas
J. W. Woods
Metallurgy Division

The previous report! presented results for a
series of stress-rupture tests of tubular Inconel
specimens under multiaxial stress in contact with
NGF-ZfF“*UF; (50-46-4 mole %). Some additional
dota which compare fime to rupture for multiaxially
stressed tubular specimens with that for uniaxially

 

1R, B. Oliver, D. A. Douglas, and J. W. Woods, ANP
Quar. Prog. Rep. Mar. 10, 1954, ORNL-1692, p 84.

PERIOD ENDING JUNE 10, 1954

stressed flat specimens in the same fluoride mix-
ture are presented in Fig. 6.1. The previous re-
sults were for specimens tested with the fluoride
mixture inside the tube and purified orgon gas
outside. Metallographic examination of the tubes
after rupture did not show the void formation which
characterizes the type of corrosive attack found
on the straight tension tests. The surface ap-
peared uneven, but there was no evidence of inter-
granulor attack., With the fluoride mixture inside
the tube, there is a ratio of about ten units of
surface area to one of volume, It was thought
that the change in the corrosion rate could be

CRNL-LA-DWG 1587

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

104 ,
- j ‘ 1 ( { ’ J J
R | TS A S S O T Y R A ,
- ot L1 — e
5 em m e rm i e b kb e saaed o aiiqgaaen L e barmee mmmmmnnf e — e
""""""""" T \ T
2 . :
w . o = \ . - -
o 0.010-in.~WALL TUBING WITH — ‘/r--0.062—|n.—THICK SHEET IN TENSION
’5.‘:’ osr —  FLUORIDE MIXTURE INSIDE -y : \fi— 1 IN FLUORIDE MIXTURE
7 : - \
o b I e e L e
i
0.010 -in.- WALL TUBING WITH
FLLUORIDE MIXTURE OUTSIDE -
1 :
! l
10° A L
10 2 5 10% 2 5 10° 2 5 10%

TIME TO RUPTURE (hr)

Fig. 6.1. Stress vs Time to Rupture for inconel Sheet and Inconel Tubing in NaF-ZrF -UF, {50-46-4

mole %) ot 1500°F.

89
ANP QUARTERLY PROGRESS REPORT

explained on the basis that the surface-to-volume
ratio was too high to sustain the corrosive attack.
With this point in mind, tests were run with the
fluoride mixture outside the tube and argon inside,
with a resulting ratio of one unit of surface area

UNCLASSIFIED
Y-12495 -

 

Fig. 6.2.

the Fluoride Mixture Was Inside the Tube and the Resulting Surface-to-Yolume Ratio Was 10 to 1.

etched. (b) Etched. 250X.

   
  

. ' -

UNCLAS&?F!ED
¥-12513

Fig. 6.3.

 

Inconel Tubing After Exposure to NaF.ZrF -UF

to eight units of volume. The types of corrosive
attack on these specimens are illustrated in Figs.
6.2 and 6.3. The attack shown in Fig. 6.3, which
is similar to that found in uniaxial tests, differs
radically from the attack found when the fluoride

    
  
   
 

-

TUNCLASSIFIED
Y-12530

Incone! Tubing After Exposure to NaF-ZyF -UF , ot 1500°F in Stress-Rupture Test in Which

(@) Un-

’ fi*‘ '&g“f&xngt‘f l“&xi‘flf o

o
SigTh ,;? AP UNCLASSIFIED +;
: M.‘ Ao Y174 .’3

  

e

. . N } e e
fiv'“ *h t.‘l?j}w % “‘h\ N“x’-‘*\'t ‘ w‘fi ’fi(:éw*}

at 1500°F in Stress-Rupture Test in Which

the Fluoride Mixture Was Outside the Tube and the Resulting Surfoce-to-Volume Ratio Was 1 to 8. (a)

Unetched. (b) Etched. 250X.

90
mixture was inside the tube (Fig. 6.2). It is
therefore concluded that the surface-to-volume
ratio is as importont in stress-rupture tests os it
is in corrosion tests. It appears that under these
test conditions a certain minimum volume of fluo-
ride mixture in relation to the surface area is
required for corrosion to proceed. Since stress-
rupture tests in fluoride mixtures are being run
at a number of testing laboratories, some of which
will be in support of the efforts of ORNL, it seems
imperative that the best surface-to-volume ratio
for reproducibility be determined. Otherwise, a
comparison of the data from the various labe-
ratories may lead to confusion or erroneous con-
clusions. ’

Another variable which should be considered in
analyzing tube-burst data is the wall thickness of
the specimen being tested. The data presented
in Fig. 6.1 are for tubular specimens with 0.010-
in.-thick walls. Naturally, corrosive attack will
have more effect on this type of specimen than
on a thick-walled tube. Specimens with wall thick-
nesses of (0.020, 0.040, and 0.060 in. are being
tested so that the effect of this variable can be
studied, [t is expected that all the tubular speci-
mens will fail in shorter times than will corre-
sponding specimens tested in. straight tension,
One reason for this prediction is that in the tube-
burst test there is a tangential-to-axial stress ratio
of 2 to 1. It is generally recognized that this
stress distribution will result in the minimum
ductility for any given material. No strain meas-
urements are made during the tube-burst test; how-
ever, after-teést measurements do not show meas-
urable deformation that would be evidence of the
brittle behavior of Inconel in this type of test.
The lowered ductility of the multiaxially stressed
specimen is believed to be the prime reason for
its early failure. :

Since the available data from tube-burst tests
indicate that multiaxial stresses reduce rupture
life in comparison with rupture life under the
‘uniaxial stresses used in the more conventional
straight-tension tests, it appears that more em-
phasis should be placed on the tube-burst tests.
Information on the behavior of Inconel tubing under
multiaxial stress will be particularly pertinent to
ANP-type reactor design. '

PERIOD ENDING JUNE 10, 1954

HIGH-CONDUCTIVITY METALS FOR
RADIATOR FINS

E.S. Bomar ‘H. lnouye
J. H. Coobs R. W. Johnson
Metallurgy Division
Investigations of fin material have been primarily
concerned with the protection of copper from oxi-
dation and with the prevention of diffusion between
copper and the cladding material. The materials
that were found to be satisfactory as cladding were
types 310, 446, and 430 stainless steel and In-
conel, provided a suitable diffusion barrier was

placed between the cladding ond the copper. Since

the purpose of the investigation was to develop
a high-thermal-conductivity fin, a large ratio of
copper to cladding was obviously desirable. With
the roll-cladding techniques employed, the minimum
clad that could be applied successfully was about
2 mils thick, although clads as thin as 1 mil have
been made successfully by other methods. The
minimum cladding thickness is apparently de-
pendent only on ability to fabricate the composite.
Under the conditions that the fin material will be
used it must be assumed that there will be
stresses. | herefore tensile-strength tests of the
composites under oxidizing conditions are being
made. _ -

A few exploratory tests were made on 8-mil-thick
type 310 stainless-steel-clad copper composites;
the cladding was 2 mils thick on each side of the
copper. The composites made by the General
Plate Compony were tested in the as-received
condition. The results of a few tests are pre-
sented in Table 6.1. Additional tests at different
stresses and temperatures are contemplated.. No
definite conclusions can be made at this time, but
it is evident that low stresses caused failures
that were not evident in tests under no-load con-
ditions.

SPECIAL MATERIALS RESEARCH

E. S. Bomar H. Inouye
J. H. Coobs R. W. Johnson
Metallurgy Division

Hastelloys B and C

Seven rod extrusions of as-cast Hastelloys B
and C were made to determine the optimum temper-
atures and the general fabrication characteristics

91
ANP QUARTERLY PROGRESS REPORT

of these alloys. The extrusion data are presented
in Table 6.2. These extrusions were successful
in that they demonstrated that the material could
be extruded. However, in nearly every instonce
the extruded rod cracked at the leading edge, and
in one instance the extrusion shattered. At the
present stage of investigation it is thought that
the cracking of the leading edge is due to inter-
mittent welding of the alloy to the die (‘‘rattle-
snaking'’) because of insufficient lubrication.
Fractures of a different kind occurred near the

center of the extruded rod that probably resulted
from faults in the billets — the melts were made
in air. Shattering is believed to be caused by
low-melting phases in the grain boundaries.

These extrusion difficulties may be solved by
(1) making the melts in vacuum and thus mini-
mizing atmosphere contamination and probably the
low-melting constituents, since purification might
also be accomplished, (2) changing the die shape
to cause more working in the leading edge, (3)
canning the billet in such a manner that the alloy

TABLE 6.1. TENSILE STRENGTH TESTS OF TYPE 310 STAINLESS-STEEL-CLAD COPPER AT 1500°F IN AIR

 

 

 

 

 

 

 

 

 

 

 

 

 

STRESS TEST DURATION ELONG;\TION REMARKS
{psi) (ki) (% in 2 /2 in.)
500 500 5 Surface blistery but otherwise sound
1000 500 20 Siringers of oxide on surface; a few
nodules of copper oxide
1500 500 47.5 Numerous copper oxide nodules on
surface; arsas oxidized throughout;
considered to be a failure
1800 336 50 Sample ruptured
2000 96 39 Sample ruptured
TABLE 6.2. EXTRUSION DATA ON HASTELLOYS B AND C
Die: CHW* preheated to 572°F, 45-deg cone
Lubrication: glass wool in container
Ram sleeve: 3]/6 in., preheated to 692°F
Billet: 3% in. long, 3 in. OD, heated in Houghton's No. 1550 salt bath
EXTRUSION HEATING EXTRUDED RAM PRESSURE ({tons)
BILLET TEMPERATURE TIME SIZE
°F) (he) (in.) Start Run
Hastelloy B 2300 % 1Y 548 400
2300 1y, 1 550 420
2100 17% 1% 560 390
2100 20 17, 615 450
Hastelloy C 2100 18 1‘/2 530 390
2100 20} 1, 565 460
2100 4 1 570 450

 

 

 

 

 

 

*Manufactured by Latrobe Electric Steel Company.

92
is placed where lubrication is known to be ade-
quate, that is, by using a dummy in front of the
billet. ' :

The short billets (3]/2 in.) were extruded at high
velocities, and in many instonces the tools were
broken when the ram struck the die. This diffi-
culty in stopping the ram movement at a precise
moment was prevented by using a brass dummy
block behind the hillet so that the ram could be
allowed to move until the press stopped.

Nickel-Molybdenum Alloys

Two vacusm melts of nominal composition 24%
Mo~76% Ni and 20% Mo-80% Ni have been made.
Attempts to extrude the 24% Mo-76% Ni alloy ot
2350°PF at reductions of 13 to 1 and 9 10 1 failed
because of tool breakage. In the attempt at ex-
trusion at a ratio of 9 to 1, the die insert fractured.
A solid die was made, and attempts to extrude
the 20% Mo—-80% Ni alloy with the new die failed
because of mandrel breakage. A straight mandrel
and a 25-deg cone die were used. The extruded
material that was obtained before the mechanical
failures occurred had good surfaces and ne evi-
dence of ‘‘rattlesnoking.”’

Stainless-Steel-Clad Molybdenum and Columbium

Molybdenum tubing (0.620-in.-OD, 0.098-in.-wall)
was clad on the outside with 0.065-in.-thick type
321 stainless steel. The composite tube was
successfully bent 105 deg on a 3-in. radius at
temperatures around 400°F and made into the
shape of a conventional thermal-convection loop.
Attempts to weld the molybdenum were unsuc-
cessful because of porosity and extreme brittle-
ness. Another compliceting problem was that, ot
the temperatures required to fuse the molybdenum,
the cladding melted ond alloyed in the weld.
These difficulties were circumvented by assem-
bling the loop with threaded molybdenum joints
and welding the stainless steel cladding. Two
such loops were made. Type 310 stainless-steel-
clad columbium tubing has been bent into the
shope of a conventional thermal-convection loop,
but the joint has not yet been welded. The
bending was accomplished at room temperature by
using Cerrobend inside the tube.

Columbium

Fansteel columbium sheet 0.250 in, thick was
reduced to 0.193 in, in thickness and annealed in

PERIOD ENDING JUNE 10, 1954

vacuum to an ASTM grain size of 6 to 7. The
work-hardening characteristics were determined by
making Tukon hardness measurements at various
stages of reduction up to 89.6% in thickness. The
results of these tests are given in Table 6.3.
The data show that hardening in cold-worked co-
lumbium does not increase rapidly, and therefore
the metal may be severely deformed without danger
of fracture. A similar investigation? showed some-
what higher hordness values for the same re-
duction., '

TABLE 6.3. WORK-HARDENING CHARACTERISTICS
OF COLUMBIUM SHEET AT VARIOUS STAGES
OF REDUCTION OF THICKNESS

 

 

 

THICKNESS REDUCTION iN HARDMESS
Gn) THICKNESS VPN
(%) '
0.193 Annealed 72.4
0.170 1.9 98.2
0.152 21.2 100.2
0.135 30.0 108.3
0.115 40.3 111.1
0.095 50.2 115.5
0.077 60.1 123
0.058 70.0 127
0.040 79.3 . 134
0.025 87.0 144
0.020 89.6 150

 

 

 

 

Boron Carbide

Thin high-boron-content shielding is required for
the heat exchanger of the Reflector-Moderated Re-
actor. This shield will be in two ]/s-in.—thick
layers, one inside and one outside the heat ex-
changer, and it should have as high a boron
content as practical. A concentration of 1.5 g/cc
of boron, equivalent to B,C compacted to 80%
density, would be sufficient.

The temperature of the shield, which will be
about 1500°F, severely limits the choice of
bonding materials for use in consolidating the

 

2w. o. Alexander, Observations on the Fabricating
Properties of Columbium Sheet, BR-627 {July 12, 1945}.

93
ANP QUARTERLY PROGRESS REPORT

boron carbide. lron and nickel both react with
B,C to form eutectics and brittle intermetallic
phases, but they have been used to bond B,C by
hot-pressing or high-temperature sintering. Copper
and silver are more attractive since they do not
react with boron carbide and would therefore retain
their strength and conductivity after being formed
into a composite. However, silver is somewhat
objectionable because of its relatively high cross
section and associated gamma activity,

Hirsch® has reported fabricating strong compacts
of boron bonded with 15 vol % of copper or of
silver with horon densities as high as 1.52 g/cc
by warm-pressing at 500°C with 50-tsi pressure.
In the hope of producing satisfactory compacts by
duplicating this work with high-grade B,C, compo-
sitions containing 15 vol % of copper or 20 vol %
of silver were prepared. Fine, annealed electro-
lytic copper and precipitated silver powders were
used with blends of various particle-size fractions
of B,C. These mixtures were then pressed at
350°C with 40-tsi pressure, and the density was
determined from the measured volume.

The maximum boron density obtained from these
compositions thus far is about 1,25 g/cc for high-
grade B4C bonded with 20 vol % of silver. Den-
sities attainable by using sized fractions of metal-
lurgical-grade B,C are 1.1 g/cc of boron with
20 vol % of copper binder and about 1.2 g/cc with
20 vol % of silver binder. By using compositions
of high-grade B,C (>80% boron) with copper and
by pressing at 500°C, it should be possible to
obtain compacts with boron densities at least as
high as 1.3 g/cc. This density is somewhat lower
than desired but may still be considered. The
attainment of higher densities will require one of
three alternatives:

1. the use of metallic boron bonded with copper
by warm-pressing, as reported by Hirsch,

2. the high-temperoture hot-pressing of B, C, with
or without iron or other reactive bonding agent,

3. the high-temperature sintering, at 1900 to
2000°C, of carefully prepared B,C-Fe compo-
sitions, for which boron densities as high as
1.45 g/cc have been reported.?

A wmore complete investigation of the warm-
pressing technique is to be made. For this in-

 

SH. Hirsch, J. Met. Ceram., TID-67, p 7 (May 1949).

YKnolls Atomic Power Laboratory, Report of Metal-
furgy Section for December 1952 and Janvary and Feb-
ruary 1953, KAPL-894.

94

vestigation, 15 to 20 vol % of copper in BAC will
be pressed at 500°C with 40- to 50-tsi pressure.
A new die of high-alloy steel will be required for
satisfactory service at this temperature.

In addition to the problem of fabricating a satis-
factory shield, the possible effects of boron dif-
fusion must be studied. Since the shield will be
in contact with Incone! at operating temperatures
of 1500°F, extensive diffusion of boron may occur.
Samples of Incone! coated with Al,O, are being
obtained for compatibility tests with boron-con-
taining compacts,

WELDING AND BRAZING
C. E. Shubert

P. Patriarca
G. M. Slaughter K. W. Reber
B. McDowell J. M, Cisar

Metallurgy Division

Brazing Alloy Development

The extent and rate of high-temperature oxidation
of brazing alloys during service are among the
important factors which will determine the useful
life of ANP radiators. Inconel T joints brazed
with 15 different high-temperature brazing alloys
were subjected to static air at 1500 and 1700°F
for periods of 200 and 500 hr. The results of these
tests, as determined by metallographic examination
(100X magnification) in the as-polished condition,
are presented in Table 6.4. It appears that the
majority of the nickel and nickel-chromium-base
alloys are suitable for service in an oxidizing
atmosphere at 1500°F, aond several are suitable
at 1700°F. However, since G-E alloy No. 81 has
shown considerable promise for tube-to-fin joints
and as a backup for tube-to-header joints, its
unexpected behavior ot 1700°F is being investi-
gated further. |t was encouraging to find that the
Coast Metals alloy No. 52 was very oxidation
resistant, since its favorable flow point makes
this alloy useful for joining high-conductivity fin
materials,

A prototype sodium-to-gir radiator utilizing In-
conel-clad-copper high-conductivity fins was fabri-
cated by a combination heliarc-welding and dry-
hydrogen-brazing technique. The assembly, shown
in Fig. 6.4, contains 180 heliarc-welded tube-to-
header joints and approximately 4300 brazed tube-
joints. As an added precaution against
undesirable stress concentrations and leaks during
service, the tube-to-header joints were back brazed.

to-fin
§6

TABLE 6.4. OXIDATION RESISTANCE OF DRY-HYDROGEN-BRAZED T JOINTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(O)Very slight, tess than 1 mil of penetration.

Stight, 1 to 2 mils of penetration.

Moderate, 2 to 5 mils of penetration,

Severe, greater than 5 mils of peretration.

Complete, fillet completely destroyed,

)y oid formation caused by internal oxidation at braze alloy fillet~Inconrel interface.

 

 

OXIDATION IN STATIC AIR(Q}
COMPOSITION BRAZING S >
BRAZING ALLOY (wt %) TEMPERATURE At 1500°F A+ 17007 F
e o
F) For 200 hr For 500 hr For 200 hr | For 500 hr
Commercial Alloys
i ! f r
Nicrobraz 7O Mi-14 Cr—6 Fe-58-45i-1C 2150 Slight Slight { Stight j Stight
L ow-melting-point Microbraz | 80 Ni—6 Fe--5Cr-55i-3 B-1C 1920 Slight Stight | Stight Stight
Coast Metals alloy No. 52 89 Ni-5S5i—-4 B-2 Fe 1840 Very slight Slight Slight Slight
~Mond Nickel Co. alloy 64 Ag~33 Pd-3 Mn 2150 Severe internal | Severe internal | Complete
oxidation oxidation , ¢
= s
Copper 100 Cu 2050 Complete | E
Experimentui Nickel-Bose Alloys
GE alloy No. 81 66 Ni-19 Cr-10Si-4 Fe-1l Mn 2150 Very stight Slight Severs®! Severe(b)
Ni-Ge 75 Ni~25 Ge 2150 Stight Slight ! Moderate Moderate
Ni-Ge-Cr 65 Ni-25 Ge=10 Cr . 2140 Yery slight Stight Modercfe{b) Severe
Electroless Ni-P 88 Ni-12 P ] 1740 Very slight Slight
80 Ni=10 P=10 Cr 1830 Very slight Slight
68 Ni-32 5n 2150 Slight Moderate Severe Complete
40 Ni=60 Mn 1920 Complete |
Precious-Metal-Base Alloys
Au-Ni 82 Au~18 Ni 1830 Very stight R!ight Moderate Moderate
Au-Cuy 80 Au—20 Cu 1740 | Moderate Complete
Pd.al 92 Pd-8 Al 2020 Very slight Very slight Very slight Sligh-t

 

#S6L ‘Ol ANNT ONIAON3 QONRI3d
ANP QUARTERLY PROGRESS REPORT

 

Fig. 6.4. Sodium-to-Air Radiator.

All tube-to-header joints were heliarc welded by
a semidutomatic process. In this process, a com-
mercially ovailable heliarc torch is rotated around
the tube periphery by a variable-speed motor-driven
offset-cam mechanism which can be odjusted to
any desired diameter of swing. A photograph of
this equipment is shown in Fig. 6.5.

Preliminary experiments were conducted on
typical tube-to-header samples to investigate the
effects of arc current and welding speed upon weld
size, hole constriction, and weld penetration. It
was shown in these tests that welds comparable
to those produced manually by a skilled operator®
could be made semigutomatically.
satisfactory welds having a penetration of at least
one tube-wall thickness could be made ot a weld
time of 7 sec over a welding current range from
40 to 50 amp. The hole consiriction resulting from
welds made over this current range did not exceed

S%.

For example,

 

3p, Patriarca, G. M. Slaughter, and J. M., Cisar, ANP
Quar, Prog. Rep. Dec. 10, 1953, ORNL-1649, p 86,
Fig. 7.3

96

UNCLASSIFIED
Y.12494

 

 

 

Fig, 6.5. Semicutomatic Tube-to-Header Welding
Machine.

As would be expected, welds made at other
welding speeds required corresponding changes
in the welding variables. Further experiments are
now being conducted to study more thoroughly the
relationships between the joint parameters and
the welding variables. Since it is expected that
the maximum weld penetration will be dependent
to a significant extent on the radius of electrode
swing, experiments will be conducted to determine
this effect.

The fabrication of a radiator containing high-
conductivity fins requires close control of the
furnace temperatures during brazing fo assure a
minimum of thermal variations over the assembly,
Experiments were conducted which indicated that
the large Globar pit furnace available in the
Welding Laboratory would permit a temperature
control of $5°F over the assembly at 1870°F and
hence would be satisfactory for this fobrication,

Coast Metals alloy No. 52 (Table 6.4) was se-
lected as the brazing material because of its
superior oxidation resistance and its relatively low
flow point of 1840°F.

A preheat temperature of 1520°F was utilized
in this dry-hydrogen-brazing operation to permit an
equalization of temperatures over the assembly
before brazing began. The prehecot portion of the
therma! cycle aided fillet formation by minimizing
selective flowing of the brazing alloy to the fins.
Boron diffusion from this alloy was not a problem
during preheating because very little sintering
occurs at this temperature, It also appears that
the fin distortion on the assembly was kept to a
minimum as a result of the preheating and the
careful temperature control during brazing.

Good flowability of the alloy on both the tube-
to-fin and tube-to-header joints was obtained
during brazing for 30 min at 1870°F. Since the
copper exposed by the fin-punching operation can
be protected adequately by the brazing alloy, it
is believed that the oxidation resistance of the
tybe-to-fin joints should be satisfactory.

Beryllium Test Assembly

An Inconel assembly for determining the effects

PERIOD ENDING JUNE 10, 1954

of thermal stresses and thermal cycling on be-

- ryllium was fabricated. This unit, which will be

used to evaluate the temperature distributions in
the beryllium, as well as the compatibility of
sodium and beryllium, consists of 46 tube-to-
header joints and extensive manifold welding. The
beryllium block is enclosed in the center section
of the assembly shown in Fig. 6.6. All joints were
manually heliare welded by%quol-ified operators and
were leak-tight to helium.

Since heating will be achieved by the electric
resistance heating of sodium, it was necessary
to fabricate two lominated-copper bus bars and
to join these to the Inconel assembly. A silver-
copper eutectic brazing alloy was used to join
the copper laminations into a solid electrical con-
nection. However, since this silver brazing alloy
will not wet inconel satisfactorily in the available
dry-hydrogen atmosphere at the brazing temperature
of 1510°F, it was necessary to copper-braze nickel
sheets to the Inconel cans before subsequent
joining of the bus bars to the test assembly with
the silver-copper eutectic alloy.

97
98

ANPF QUARTERLY PROGRESS REPORT

 
    

Fig. 6.6. Beryllium Test Assembly.

UNCLASSIFIED
Y-12546

o T b
PERIOD ENDING JUNE 10, 19-54

7. HEAT TRANSFER AND PHYSICAL PROPERTIES
| H. F. Poppendiek |

Reactor Experimental Engineering Division

The enthalpies and heat capacities of the ARE
fuel NoF-ZrF -UF, (53.5-40.0-6.5 mole %) were
detfermined; the heat capacity in the solid state
over the temperature range 260 to 490°C was
found to be 0.19 cal/g-°C, and.the heat capacity in
the liquid state over the temperature range 590 to
920°C was found to be 0.23 cal/g-°C. Enthalpy
and heat-capacity measurements for K;CrF  in the
solid state were also obtgined. Density ond
viscosity measurements were made for molten
RbF-LiF (57-43 mole %); the viscosity varied
from aobout 8 centipoises {cp) ot 550°C to about 3
cp at 750°C. The thermal conductivity of molten
RbF-LiF (5743 mole %) was determined to be
about 1.2 Btu/hr-sq ft (°F/ft}. The thermal con-
ductivity of solid MNaF-KF-LiF (11.5-42.0-46.5
mole %) was found to be about 3 Btu/hresq ft
(°F/ft). Electrical-conductivity measurements of
molten NaOH were obtained over the temperature
range 625 to 1490°F.

Additional forced-convection heat-transfer meas-
urements of molten NaF-KF-LiF were made at
higher Reynolds numbers than -had previously been
obtained. Some thermal-conductivity information
on K,CGrF, o wall-deposit material, was obtained.
A Lucn‘e model of the circulating-fuel reflector-
moderated reactor has been fabricated ond will be
used to study the hydrodynamic structure in that
system. The results of a mathematical analysis
of convection dare presented for the case of forced
flow between parallel plates which are ducting
fluids with volume heat sources; this analysis is
useful in estimating the temperature structure in
the flow annuli of reflector-moderated reactor
cores. A study has been initiated to investigate
the heat-transfer and fluid-flow characteristics' of
a NaDH-moderated circulating-fuel reactor system,

PHYSICAL PROPERTIES MEASUREMENTS:

Heat Capacity
W. D. Powers - G. C, Blalock

Reactor Experimental Engineering Division

The following enthalpy and heat-capacity meas-
urements were made with Bunsen ice calorimeters:

NaF-ZrF -UF ; (53.5-40.0-6.5 mole %)
Solid (260 to 490°C)
Hy — Hpoe = —4.1+ 0,197
C, = 0.19 +0.02

Liquid (590 to 920°C) ;
Hyp ~ Hypoe = 34 + 0.23(5)T

C, = 0.23(5) + 0.03

‘K CrF

Sohd (50 to 830°C)

Hi = Ho. = —4 + 0.228)T
| -

p 0.22(8) * 0.006

NaF-KF-LiF (11.5-42.0-46.5 mole %)
Solid (60 to 455°C)

H -«.H\ = «-26+027T+098x10“"12

T
C, = 0.27 + 1.96 x 1077

In these expressions H is the enthalpy in cal/g,
Cp is the heat capacity in cal/g-°C, and T is the
temperature in °C. The enthalpy of K CrF, (an
insoluble deposit material) is shown in th 7.1,
The enthalpy and heat-capacity data for NaF-KF-LiF
(11.5-42.0-46.5 mole %) in the solid state corre-
spond to o much greater temperature range than
was previously reported. ‘A copper calorimeter
has been constructed that will be used to supple-
ment the Bunsen ice calorimeters now in use. The
heat content of the sample will be measured by the
temperature rise of the copper housing.

Density and Viscosity
S. |, Cohen T. N. Jones

Reactor Experimental Engineering Division

Density and viscosity measurements were made
on two fluoride mixtures. The viscosities of
RbF-LiF (57-43 mole %) were determined  with
modified Brookfield and capillary viscometers,
ond the values varied from about 8.0 ¢p at 550°C
to about 3 cp at 750°C, The density measurements
of RbF-LiF in the molten state are represented by
the relation

p = 3.39 — 0.00085T ,

99
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIED
ORNL-LR—-DWG 41414

-

 

200 - R

   

 

ENTRALPY {cal/q)

 

 

400 o 00 800 1000
TEMPERATURE (°C)

Fig. 7.1. Enthalpy-Temperature Relationship for
K,GF .

where p is in g/cc and the temperature range is
500°C < T < 700°C.

The density and viscosity values! for NaF-KF-LiF
(11.5-42.0-46.5 mole %) that were determined in the
early stages of the ANP Project were checked,
and the recent determinations were in substantial
agreement with the early ones. The viscosities
for the mixture vary from about 8 cp at 500°C to
about 3 cp at 750°C, and the molten densities are
represented by the relation

2.53 ~ 0.00073T ,
600°C < T < 900°C .

A report? published recently lists all the density
measurements that have been made on molten
fluoride mixtures for the ANP Project and includes
a correlation of the data that can be used to pre-
dict densities of molten fluoride mixtures of known
composition.

p"'_.‘

Thermal Conductivity

W. D. Powers S. J. Claiborne
R. M. Burnett

Reactor Experimental Engineering Division

The Deem-type apparatus has been used to meas-
ure the thermal conductivity of RbF-LiF (57-43 mole
%). The conductivity was determined to be 1.2

 

VANP Physical Properties Group, Physical Property
Charis for Some Reoactor Fuels, Coslanis, and Miscel-
A.'lanec;us Materials, 3d ed., ORNL CF-53-3-261 (Mor. 20,

953).

100

Btu/hr-sq ft (°F/ft) at an average sample temperc-
ture of about 1050°F.

A flat-plate method for determining the thermal
conductivities of solid salts has been developed
in which the salt is cast in the form of a slab,
The heat that is passed through the sample is
measured by the rise in temperature of the water
that circulates through the cooling plate, The
temperature drop through the sample is also meas-
vred. The conductivity may be calculated casily
from this information and the physical dimensions
of the slab. The preliminary thermal conductivity
value obtained for NaF-KF-LiF (11.5-42.0-46.5
mole %) by using this method was about 3 Btu/hr-sq
ft (°F/f1).

Additional thermal- conductivity measurements
have been made on solids by using the transient
cooling or heating technique. An improved method
of casting nonporous spheres has been developed
which invelves cocling the mold from the bottom at
a very slow rate. Improved methads of cooling and
heating the spheres have also been developed.
The preliminary thermal-conductivity value obtained
by using this method on NaF-KF-LiF (11.5-42.0-46.5
mole %) in the solid state was about 2.7 Btu/hr-sq
ft (°F/&).

Electrical Conductivity
N. D. Greene

Reactor Experimental Engineering Division

A preliminary determination of the elecirical con-
ductivity of molten NaOH at temperatures of up to
1500°F has been made with the platinum conduc-
tivity cell, This determination has extended the
temperature range of the electrical-conductivity
measurements of NaOH to approximately 450°F
higher than the maximum temperoture reported in
the literature. The experimental conductivity data
are shown in Fig. 7.2, together with the published
measurements? for low temperatures,

Checks on the platinum conductivity cell have
been made by studying three molten salts for which
conductivity values have been published. The
salts investigated, together with the deviation
between ORNL and literature values and the tem-

25, 1, Cohen and T. N, Jones, A Summary of Density
Measurements on Molten Flucride Mixturas and a Cor-
relation Useful for Predicting Densities of Fluoride
Mixh.):res of Known Composition, ORNL-1702 (May 14,
1954).

3!nfernafiona! Critical Tables, vol 6, p 149,
HNCLASGIFISD
CRNL-LR -DWG 1415

 

 

 

 

 

7.00 oo
6.00
1
E
5 5.00
E
~ 4.00
D
l:'.
= 3.00
[ ‘
g
g 2.00 P/ e bl ..__J....---
W e ORNL EXPERIMENTAL VALUES ‘
&4 LITERATURE VALUES
100 —_ .]._.._..{ - ...}.........r_........\....._,. - [ 4 e —
O e ,A,._i. ......... ‘. ............... .J ................... l...,.._.,
500 800 1000 ' 1200 1400 1600
TEMPERATURE {°F) :
Fig. 7.2. Electrical Conductivity of Moliten
NaOH.

perature ranges, are listed in the following:

‘Mean Deviation Temperature Range

| (%) (°F)
NaOH +2.7 625 to B35
LiMC\3 +4.4 534 to 862
KNO, -3.3 875 to 900

The measurements are difficult to make becouse
of polarization within the cell, and therefore some
alternate method of measurement has been sought.
It is felt that a hemispherical graphite crucible in
which it would.be possible to position leads so
that both current and potential measurements would
be obtained simultanecusly might be useful. Since
the potential measurement would be independent of
the degree of polarization at the current-carrying
electrode ond an alternating current would be used,
it is estimated that good accuracy would be ob-
tained. An. attempt to verify the tesults of the
measurements made with the platinum conductivity
cell will be made with this alternate method during
the coming guorter, The conductivities at various
pertinent temperatures of some of the salts which
are of interest to the ANP Project will also be
determined.

Vapor Pressures

R. E. Moore C. J. Barton
Materials Chemistry Division
In an attempt to resolve some anomalies, ob-
served in another laboratory,? in vapor pressures of
some NaF-ZrF , mixtures, the method and apparatus

PERIOD ENDING JUNE 10, 1954

of Rodebush and Dixon® have been applied to a
NaF-ZrF, mixture contoining 75 mole % ZrF,.
Preliminary values obtained for thissmaterial are
37, 54, 66, 74, 83, 115, and 152 mm Hg at 769,
786, 796, 805, 816, 828, and 844°C, respectively.
A straight line through these values on a log P vs
1/T plot intersects the plot of vapor pressure of
ZrF, ot 762°C. Since the primary phase at 75
mole % ZrF, in this system is ZrF,, the inter-
section should correspond to the ligquidus tempera-
ture for the mixture, which thermal-analysis data
indicate to be 755°C, The difference between the
two tempemtures is only slightly iarger 'rhcm the
expected error of either method.

It was discovered that there was a consistent
temperature error in the vapor pressure data given
in o previous report® for the 50 mole % NaF~50
mole % ZrF, mixture. Therefore the correct pres-
sures ore Jower than those given before.  The
corrected equation for this composition is

7425
log P H = e e
og P (mm Hg) =

oK)

+- 7.889 .

FUSED-SALT HEAT TRANSFER

H. W, Hoffman
Reactor Experimental Engineering Division

J. Lones

Previous NaF-KF-LiF heot-tronsfer experiments
made in systems in which no tube-wall deposits
were formed were limited to rather low Reynolds
numbers because of the temperature limitations
inherent in the use of nickel tubes. It had been
observed that molten fluorides could be contained
in type 316 stainless steel for short periods of
time without serious corrosion results or the forma-
tion of the insoluble fluoride deposits, such as
K 3CrF , that are found in Inconel systems. There-
fore in order to get data at higher Reynolds numbers
it was decided to conduct a heat-transfer experi-
ment of short duration in type 316 stainiess steel
tubes, The results of the experiment are shown in
Fig. 7.3. As was to be expected when there was
no wall deposit, the data fell on the curve for
normal turbulent-flow for ordinary fluids, that is,
the curve expressed by j = 0.023 Re=02, The

 

44, R. Nelson and R. W. Dayton, Progress Report for
January 1954, BMI-902 {(Feb. 1954).

5W. H. Rodebush and A. L. Dixon, Phys. Rev. 26,
B51 (1925).

SR, E. Traber, Jr., R. E. Moore, and C. J. Barton,
ANP Quar, Prog. Rep. Dec. 10, 1953, ORNL-1649, p 99.

101
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 1416

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.040 - _ ] - . - .
0.009 —— e — - l e L_ }— -
0.008 |- — — i — : | .
0.007 L | | a NICKEL TUBE
| ‘ # TYPE 316 STAINLESS STEEL TUBE
0.006 |——o s — ] T
\
z 0.005 [—— L T - e : -
g J 3 Re~ O ‘
= 0.004 ;i - .- | e . JR ‘1*%‘
) .
b * l T Lol
| : 1 |
™~ 0.003 \ ‘
& |
o i - INCONEL TUBE (WALL DEPQOSITS}
Q
O 0.002 - Afe 1 ){.& - F - L - —
o . J ’
’ T
/ iy
1 /,; 1 !
J | : 1 ‘ ‘ 1 1
. ,‘ J . i ‘
o004 L— —— ,W‘,J_‘,_,,J ,,,,, . [ J | e [ [
103 2 5 10% 2 5 103

REYNOLDS NUMBER

Fig. 7.3. Colburn j-Function vs Reynolds Number for Molten MaF.KF-Lif,

previously obtained nickel-tube and Inconel-tube
data are also shown in Fig. 7.3.

Some preliminary determinations show the thermal
conductivity of K,CrF ( to be about 0.13 Btu/hr.sq
ft (°F/ft) for the temperature range 100 to 200°F.
The salt was hydrostatically molded into the shape
of a cylinder, approximately ]/2 in. in diameter and
4 in, in length. The density of the molded salt was
96% of the theoretical value, The cylinder was
heated to a uniform temperature and then suddenly
exposed to a cool air stream where it was transiently
cooled. The thermal conductivity was calculated
from the time-temperoture measurements obtained
in this experiment,

REACTOR HYDRODYNAMICS

J. O. Bradfute L. D. Palmer
G. M. Winn

Reactor Experimental Engineering Division

A Lucite model of the reflector-moderated reactor
core has been fabricated, The fluid velocity dis-
tribution in the voriable-cross-section flow annulus
is to be determined by photographing tiny particles
which will be dispersed throughout the flowing

 

7L. D. Palmer and G. M. Winn, A Feasibility Study of
Flow Visualization Using a Phosphorescent Particle

Method, ORNL CF-54-4-205 (Apr. 30, 1954).

102

water, An improved method of increasing the light
output of the Strobolume (the flashing fight source)
has been devised. The timing unit for the Strobolume
has nearly been completed,

An investigation was made of a technique’ for
visual observation of the nature of fluid flow in
tronsparent duct systems. It was found that the
features which determined the average velocity
profile, as well as the velocity fluctuations, under
turbulent flow conditions could be observed, A
beam of ultraviolet light was flashed through o
region of the flowing fluid (containing suspended
phosphorescent particles) for which the velocity
profile was to be observed. The beam of phos-
phorescent particles generated in the flowing fluvid
by the wltraviolet light ‘was slowly distorted by
the flow into a curve which exactly described the
fluid velocity profile. It is believed that this
technique can be used to visualize flew character-
istics in reactor cores,

HEAT-REMOVAL STUDY OF BSF REACTOR
D. C. Homilton F. E. Lynch
Reactor Experimental Engineering Division

A study has been initiated for determining the
heat-removal rate that can be provided by free con-
vection in a bulk-shielding-facility type of reactor.
The problem divides naturally inte two coses de-
pending on whether or not nucleate boiling is per-
mitted. When boiling is not permitted, the limiting
fuel-plate temperature will be equal to the satura-
tion temperature. If nucleate boiling is permitted,
the exit temperature of the water may not exceed
the saturation temperature.

The analysis is somewhat similar to on cmoinis
performed on a thermal convection harp.? There
are two distinct regions in the flow circuit of the
reactor, the reactor region and the jet or stack
region. The reactor region presents a problem in
flow and heat transfer between parallel plates
with a known but not uniform wall-heat-flux dis-
tribution. At lower reactor powers, the flow is
superposed - free-and-forced viscous convection,
At higher powers, the Reynolds modulus may
approach the lower limit for isothermal turbulent
flow; in either case, the heat-transfer and friction
data are not available in the literature. The manner
of mixing of the jet with the surrounding pool will
determine the contribution of the jet to the flow
through the reactor. An accurate consideration of
the jet contribution will be difficult.

HEAT TRANSFER IN CIRCULATING-FUEL
REFLECTOR-MCDERATED REACTOR

H. F. Poppendiek L. D. Palmer
G. M. Winn
Reactor Experimental Engineering Division -

A heat-transfer analysis? was made for the case

of forced convection between parallel plates of
infinite extent, which are ducting fluids that contain
uniform volume-heat sources, with heat transferred
uniformly fo or from the fluids through the parallel
plates. The laminar and turbulent flow data in
terms of dimensionless wall-fluid temperature dif-
ferences are plotied in Fig. 7.4 as a function of
Reynolds moduli for the case of no wall-heat
transfer. 1f it is desired to obtain the solution for
the composite case, that is, a flow system with
wall-heat transfer as well as volume-heat sources,

 

8p. C. Hamilton, F. E. Lynch, and L. D. Palmer, The
Nature of the Flow of Ordinary Fluids in a Therma!
Convection Harp, ORNL-14624 (Feb 23, 1954).

H. F. Poppendiek and L. D. Palmer, Forced Con-
vection Heat Transfer Between Parallel Plates and in
Annuli with Volume Heot Sources Within the Ffurds,
ORNL-1701 (May 11, 1954).

PERIOD ENDING JUNE 10, 1954

the data given in Fig. 7.4 must be superposed on
the known data for the case of wall-heat tronsfer
with no volume-heat sources. It will be noted that
the data shown in Fig. 7.4 pertain to liquid metals
as well as to ordinary fluids and may also be used
to estimate heat transfer in annulus systems such
as the reflector-moderated reactor for which inner-
to-outer radius ratios do not differ significantly
from unity.

A series of experiments was conducted recently
to determine the influence of electrical-current
flow on the hydrodynamic characteristics of a
liquid flowing in o tube under both laminar and
turbulent flow conditions. This problem arises
in convection experiments in which the volume
heat sources are generated electrically. No in-
fluence of the electrical-current flow on the hydro-
dynamic structure has yet been detected.

HEAT TRANSFER IN NaOH-MODERATED
CIRCULATING-FUEL REACTOR

J. O. Bradfute M. F. Poppendiek

Reactor Experimental Engineering Division

A study has been initiated to determine the
feasibility, of @ NaOH-moderated circulating-fuel
reactor in terms of heat tronsfer and fluid flow,
In such a reactor, the molten sodium hydroxide
would be the moderator as well as the coolant
necessary to reduce the high wall temperatures in
the fuel region. Because of the corrosion difficulty
that arises when NaOH is used with the typical
structural materials of reactor systems that operate
at temperatures above 1000°F, it is further neces-
sary to cool the fuel-tube wall sufficiently (by
forced circulation of the NaOH) to yield a NaOH-
wall interface temperature of below 1000°F. There-
fore it is necessary to know whether this cooling
can be accomplished with NaOH with reasonable
pumping powers and pressure drops. Some pre-
fiminary heat- and momentum-transfer analyses have
been made for the case in which fuel circulates
through tubes that are force-cooled by molten
NaOH which circulates in a direction parallel to
that of the fuel; the number of tubes was varied in
the analyses. The preliminary results suggest
that the pumping powers and pressure drops are
not prohibitively high. Further calculations on
multitube, as well as annulus and channel, systems
are being made.

103
ANP QUARTERLY PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 180A

 

10° 2 5 10® 2 5 104 2 5 105 2 5 1o}

Fig. 7.4. Dimensionless Differences Between Wall ond Mixed-Mean Fluid Temperatures os Functions
of Reynolds and Prandt] Moduli for o Paralle! Plate System (Walls Insulated).

104
PERIOD ENDING JUNE 10, 1954

8. RADIATION DAMAGE

J, B, Trice
Solid State Division
A, J. Miller
ANP Division

The program of irradiating fused-salt fuels in the
MTR continued with mixtures containing UF_- and
UFA-»bearing fuels. The fuel-lnconel interface
temperature is being maintained ot more nearly
1500°F than in previous tests with UF, fuels, A
re~examination was made of a group of Inconel
capsules from the earlier runs along with a study
of the temperature control system used and the
thermal histories. It was concluded that in mest
capsules the interface temperature was in the
region of 100°F higher than the desired 1500°F and
that 2 to 4 mils of intergranular corrosion occurred.
Excessive grain growth and deep penetration of the
[nconel occurred only in a few capsules and those
capsules had been heated to much higher tempera-
fures. '

The first in-pile circulating fuel loop was in-
stalled in a horizontal beam hole of the LLITR but
difficulties developed before start-up of the reactor
which caused the loop to be removed for inspection.
Development continued on smaller loops for iin-
sertion in LITR and MTR beryllium A-pieces.

Creep-test equipment was revised to permit creep
testing in the MTR, and development of in-pile
stress-corrosion equipment continved, Metallo-
graphic studies were made for Pratt & Whitney
Aircratt Division on sandwich-type UO,~type 347
stainless steel fuel elements irradiated in the MTR,

RADIATION STABILITY OF FLUORIDE FUELS

G. W, Keilholtz M. T. Robinson

J. G. Morgan W. R. Willis

H. E. Rebertson W. E. Browning

C. C. Webster M. F. Osbome
Solid State Division

As previously reported,! some of the Inconel
capsules containing fluoride fuel which were
irradiated in the MTR were found to have developed
large inconel grains and deep intergranular cracks.
Since such corrosion effects indicated excessive
heating, a re-examination of the thermal conditions

 

]G. W. Keilholtz et al., ANP Quar. Prog. Rep. Mar.
10, 1954, ORNL-1692, p 102,

that had been imposed on the capsules during the
irradiations was undertaken. Temperature control
had been based on data from several thermocouples
welded to the outside surface of the capsule and
was aimed at maintaining this outside temperature
at a level that would provide a 1500°F fuel-Inconel
interface temperature, ' |

In order to evaluate the accuracy of the control
method, a series of tests was made to determine
the relationship between the temperature of the
outside surface of the capsule as indicated by a
thermocouple to the actual wall temperature in an
equivalent portion of the capsule that was unper-
turbed by the presence of a thermocouple, The
indicated temperature could have low values be-
cause of the thermocouple junction being raised
slightly above the surface of the capsule and
because of conduction of heat by the cooling air
stream from the very small amount of uninsulated
exposed wire. A review was. also made of the chart
temperature records of the experiments to detect
runs in which the temperature had risen to experi-
mentally undesirable magnitudes.

A bench-test apparatus used to mock up the
conditions of the MTR capsule tests is shown
schematicaily in Fig. 8.1, The electrically heated
test section which simulates the capsule consists
of a lin, section of an actual Inconel fuel capsule.
The annulus for cooling air corresponds to that
used in the reactor, and the pyrex portion serves
as a sight glass fo permit accurate temperature
measurements with an optical pyrometer at a point
180 deg from the thermocouple being tested,

Typical data from one of the 30 thermocouples
tested are shown for three different airflow rates
in Fig. 8,2, in which T is the unperturbed sQrfuce
temperature determined with the pyrometer at any
position along the length of the capsule. On each
of the three curves the temperature indicated by
the thermocouple 7, is shown at its contact po-
sition. For airflow rates of 189, 220, and 250 fps,
it can be seen that the thermocouple readings are
fow by 75, 88, and 105°F. The maximum difference
between T‘S and T[ for 30 thermocouples was 200°F

105
ANP QUARTERLY PROGRESS REPORT

UNGLASSIFIED
ORNL --LR-~DWG 515A

:- ™ \\ ‘ J,/Y
= ;\:“‘\\
\\\
\\ %y
N

 

 

 

 

- — ANNULUS
O BODY

 

 

 

 

 

 

 

 

 

TEST SECTION—--»—F-—- - THERMOCOUPLE
A N
I N
- SIGHT GLASS
TN
1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7

/ o 2
ELECTRODE -t /" 7

7

Fig. 8.1, Thermocouple Test Apparatus,

106

5500932
_DWG. 233734

    
    
  

 

     
   

1700 ’ e .i.____. e !,,, e [ —_— :
i \ CAPSULF CENTER |
| ! LINE ] ]
ENSOO 7 _ o ' T =1a85°F
N ) :
n
-
f;[_ 250 fps
% ! 2201 !
= . I e pPs |
W
k_ 1500 ;¢—§=146|°F
Ko : [
ENTRANGE ‘ v
AR // HBY fps
LT L / / ‘
e
! | / | ! :
a0 oAl ]
¢ e g Png
CAPSULE LENGTH (in)
ig. 8.2, emperature Measurements on
Fig. 8.2, Temperature M t MTR

Capsule Mock-up,

and the average difference at the medium airflow
rates used in the reactor was T100°F,

This error, together with several others of lesser
magnitude, led to the conclusion that the tempera-
ture-control system employed in the reactor experi-
ments tended to produce fuel-lnconel interface
temperatures averaging 1620°F and never higher
than 1750°F at the hottest point of the capsule
rather than the desired 1500°F., These findings
are in agreement with calculations made by H. F.
Poppendiek prior to the test experiments.

Temperature record charts of the 13 MTR cap-
sules irradiated so far in the MTR were examined.
In every instance of deep intergranular corrosion
and abnormally large grains, the temperature pattern
for the fuel-lnconel interface was found to be
either too high or too unsteady, For the cases in
which the capsule temperature-time profile was
within the limits originally set for the experiment,
depth of attack (2 to 4 mils) and grain growth were
not grossly different from those observed in bench
tests at 1500°F, The pattern of corrosion was,
however, still intergranular rather than that con-
sisting of subsurface voids as in the case in out-
of-pile tests.

An additional aid to interpretation of results from
the in-pile capsule tests was afforded by a study
of the effect of high temperatures on the grain
structure of Inconel. Samples of capsule material
were subjected to temperatures of 1800 to 2350°F
for various periods of time. Metallographic exami-
nations of these samples are now being made.
Preliminary results indicate that farge grains may
be found in Inconel which has been heated for
periods of time of less than 8 min duration at
temperatures of the order of 2100°F,

The Inconel capsules presently being irradiated
in the MTR contain both UF,- and UF 4-bearing
fuel mixtures. The recent mdlcahons in outsofe
pile tests that the use of UF, might result in less
corrosion suggested in-pile testing of UF -beurmg
mixtures. Temperatures of the capsules are being
carefully controlled so that the Inconel-fuel inter-
face temperature will remain approximately ]SOOOF
throughout each test.

LITR FLUORIDE-FUEL LOOP
W. E. Brundage M. T. Morgan

C. D. Baumann A. S. Olson
F. M, Blacksher W. W. Parkinson
C. Ellis 0. Sisman

R. M. Carroll

Solid State Division

The design of fluoride~fuel loops for use in the
LITR horizontal-beam facility was described in
previous reports.? The fabrication of components
for three of these loops was completed. The first
loop was inserted in the LITR but was removed
before start-up of the reactor because of a leak in
the fuel system. All fuel material was retained
within the loop apparatus. Examination to dezter-
mine the cause of the breakdown is in progress.

Development continued on in-pile pumps :for
loops to be operated in the horizontal~beam fa-
cilities of the LITR and MTR (cf., Sec. 2, ““Experi-
mental Reactor Engineering’’). Several models are
being constructed. A smaller uranium investment
and less dilution of fission products would be
attained by. using such pumps in place of the
present pumps which are operated at the reactor
face.

IN-PILE STRESS CORROSION AND CREEP

W. W, Davis J. C, Wilson
N. E. Hinkle J. C. Zukas
Solid State Division

Since intergranular corrosion of Inconel by molten
fused salts of the type NaF:ZrF -UF, has been

 

2W. E. Brundage et al., ANP Quar. Prog. Rep. Moar.
10, 1954, ORNL-1692, p 105, _

PERIOD ENDING JUNE 10, 1954

observed in bench tests in which a strain was
introduced into the Inconel, it is conceivable that
thermal stress may be associated with the inter-
granular type of attack seen in in-pile tests. To
investigate this possibility, an in-pile stress-cor-
rosion apparatys for the Inconel-fused salt systems
was designed. Several prototypes of this apparatus
are being bench tested.

Creep tests of Inconel in the ORNL Graphne
Reactor and in the LITR under conditions pertinent
to ANP reactor designs have been reported previ-
ously. Apparatus for tests in the higher flux of
the MTR is being built. Much of the mechanical
assembly is completed and the unit is almost ready
for wiring., It is planned to use Bourdon tube
extensometers rather than  thermal-expansion (bi-
metal) units since the former are in a more ad-
vanced stage of evolution. .

"REMOTE METALLOGRAPHY

A. E. Richt W. Parsley
E. Schwartz R. Ramsey
M. J. Feldman
Solid State Division

Solid, sandwich-type, UO,-type 347 stainless
steel fuel elements irradiated in the MTR: were
submitted to the Laboratory by Pratt & Whitney
Aircraft Division for metallographic examination,
These studies, which were at first intended to
yield information of interest only to the supercriti-
cal-water~cycle aircraft reactor, are also related
to certain power-package reactor designs because
of the similarity between the two types of fuel
elements both in materials and in required operating
conditions. Both reactors have operating pressures
of the order of 5000 psi, temperatures in the range
of 300 to 700°F, and high power generation rates.
Both fuel element prototypes are made as solid
sandwich-type elements clad with stainless steel
or Inconel.

A typical Pratt & Whitney fuel element test is
run with the element immersed in water in a sealed
capsule at one of a series of fixed temperatures
ranging from 100 to 700°F and with pressures of
3000 to 5000 psi. The power density of the fuel
runs was as high as several thousand watts per
cubic centimeter in the MTR., Results from the
five capsules examined so far provide information
about the effects of temperature and irradiation on
core density changes, bonding of core to cladding,
core hardness, and corrosion of the stainless steel

107
ANP QUARTERLY PROGRESS REPORT

cladding., No measurable core density changes
have been found and general corrosive attack on
the cladding is absent, Hardness measurements
taken on the core indicate a dependence between
particle size and hardness change. Cores with
particles less than 3 p in diameter show greater
hardness changes than do those in the 15- to 44-u

range. In general, it can be said that hardness

108

changes, expressedin DPH units, as great as ratios
of 2 to 3 occur for iradiation periods of several
hundreds of hours and for test temperatures as high
as 700°F. Cracking of the core in bending tests
{(1.5-in. radius bend) was found only in the less
than 3-p core somples. It is expected that this
work will be completed during the next several
months,
PERIOD ENDING JUNE 10, 1954

9. ANALYTICAL STUDIES OF REACTOR MATERIALS

C. D. Susano
Analytical Chemistry Division
J. M. Warde
Metallurgy Division
R. Baldock
Stable Isotope Research and Production Division

Research and development in analytical chemis-
try for the ANP Project was concentrated on the
problem of separating UF, from UF, in NaZrF -
base fuels and fuel solvents. Methods were in-
vestigated for the conversion of the fluorides to
chloride salts and the simultaneous extraction of
the chlorides into nonagqueous solutions. Boron
trichloride in a solution of dioxane appeared to be
promising for this purpose, since UCI, is soluble
in dioxane, whereas UCl, is essentially insoluble.
The separatioen is also feasible in the presence
of chromium chlorides. Chromium dichloride is
soluble in dioxane in contrast to CrCly, which is
insoluble. However, the conversion reaction was
found to proceed slowly.

A method was devised for the determination of
oxygen in NaZrF .-base fuels and fuel solvents in
which the oxides react with' anhydrous hydrogen
fluoride to produce water, quantitatively; the water
materially increases the conductivity of the hydro-
fluoric acid. It has been calculated that the
conductivity of anhydrous hydrogen fluoride is
doubled by the addition of 1 ppm of water, In the
study of the oxidation of trivalent uranium by
hexavalent uranium it was shown that the competi-
tive oxidation with hydrogen ion could not be
completely eliminated. Stability tests of trivalent
uranium in hydrochloric acid solutions were con-
ducted. Although evaluation of the data is not
complete, potentiometric measurements indicated
that the transition in color from purple to green,
which is considered to be evidence of oxidation,
is not accompanied by corresponding changes in
potential. A 1% aqueous solution of the tetra-
sodium salt of ethylenediaminetetraacetric ocid
effectively removed a film of calcium metasilicate
from Inconel. No evidence of corrosion was found
by chemical tests. The sulfur content of natural
gas was determined to be 0.5 mg/cu ft of gas.

Petrographic examination of NaF-ZrF -UF, fuels
reduced with metallic zirconium, with metallic
uranium, or with hydrogen has shown that the pre-
dominant phase in the reduction complexes ob-

tained is always a solid solution of reduced U**
in NagZroF .. In the RbF-UF, system, there
appears to be a compound, 3 RbF-UF3, which is
tan and isotropic, with an 7= 1,438, The NaF-UF,
system contains a compound of unknown compo-
sition which is lavender and uniaxially positive,
with refractive indices of approximately 1.552,
Examinations of melts in the NaF-CrF, system
were made, and four compounds were observed.

ANALYTICAL CHEMISTRY OF
REACTOR MATERIALS

J. C, White
Analytical Chemistry Division

Oxidation States of Chromium ond Uranium in
NquFs-Busie Fuels

A. S. Meyer, Jr, D. L. Manning
W. J. Ross
Analytical Chemistry Division

In extensive studies, covered in previous re-
ports,! no solvent has been discovered that will
selectively dissolve UF, or UF, from NaZrf .-
base fuels. Certain organic solvents preferentially

dissolved UCH, trom UCI,, but the equi!ibripm of

the reaction
UIV + Cr” @Ulll + Cr!ll

could conceivably be altered during the conversion
of the metallic fuel components by the method of
fusing the sample with NaAICl . Therefore, an
attempt is being made to dissolve the UCI, info
an organic solvent during the conversion of the
fuel to the chloride salts. '
Dioxane, which has been reported to dissolve
UCI, from U(:l3,l was selected as the most
promising solvent. |t may offer an additional ad-
vantage in that aliphatic ethers are reportgd2 fo

 

V). C. White et ol., ANP Quar. Prog. Rep. Mar. 10,
1954, ORNL-1692, p 107 ff,

2N. Y. Sidgwick, The Chemical Elements and Their
Compounds, VYol. Il, p 1024, Oxford University Press,
Oxford, 1950.

109
ANP QUARTERLY PROGRESS REPORT

dissolve CrCl, from CrCl, to yield stable solutions
which are void of reducing power. Thus, if the
chlorination of the sample could be carried out in
the presence of dioxane, UCi, and CrCl, would
be dissolved, while UCl, and CrCl, would remain
in the residue, It is hoped that the rate of reaction
between the solid chlorides is negligibly slow at
room temperature.

Preliminary tests were carried out by stirring
25-mg samples of UF,, UF ,, NaZrF, and NaZrF,-
UF, in solutions of aluminum chloride in dioxane
for periods of 4 to 24 hr, Visual observations and
qualitative tests showed no evidence of dissolution
of any of these samples. Similar negative results
were obtained when the tests were repeated with
solutions of aluminum chloride in ethyl ether,
tetrahydrofuran, and benzene, respectively.

Thermodynamic calculations indicated that BCI
is superior to AICI, as the reagent for the meta-
thetical chlorination of the fluorides of wranium,
chromium, and zirconium. In qualitative tests, 25-
mg samples of UF, and NaZrF .-UF , fuel were not
completely dissolved after 16 hr of stirring in a
solution of dioxane that had been saturated with
BCl,. When UF, was stirred with liquid BCl, at
0°C, there was no apparent dissolution of the
sample. Analysis of the residue for chloride indi-
cated that the reaction

4UF, + 3BCl, — 4UCI, + 3BF,

was only about 20% complete. Since the free-
energy change of this reaction is -110 kcal, it
appears that o slow reaction rate is responsible
for the incomplete dissolution of the samples.
Attempts were made to increase the rate of reaction
by treating the UF, and the fuel samples with
BBr, at reflux temperatures of 91°C. Quantitative
results of these experiments are not yet available,
but portions of the unreacted samples were visible
after refluxing for 3 hr, The reaction appeared to
be accelerated by addition of AICI; to the BBr,.

When 60 mg of UF, was stirred at room tempera-
ture for 6 hr with 60 ml of dioxane, which had been
saturated with BC|3, about 25% of the salt dis-
solved to form a green solution. The concentration
of uranium in solution did not increase significantly
when aluminum chloride was added and the stirring
was continued for 90 min. No tests of the solu-
bilities of CrF, and CrF, have yet been carried
out.

110

Since this method is the only approach which
appears to offer possibilities of the separation of
tetravalent uranium from a fuel containing another
element for which the oxidation potential is near
that of the U3 —> U4 couple, efforts will be
made to accelerate the reaction to rates suitable
for analytical applications, Because ethereal so-
lutions of BCI, are decomposed by heating, it will
probably be necessary to find a catalyst for the
chlorination reaction. Experiments are also being
carried out fo determine whether UCI, may be
sublimed by heating the fuels in an atmosphere

of BC|3.

Determination of Oxygen in NaZrF ;-Base Fuels

A. S. Meyer, Jr. J. M, Peele
Analytical Chemistry Division

Studies have been made to determine whether
microgram quantities of oxygen in NaZrF -base
fuels can be determined by measuring the water
generated by the reaction of the mefallic oxides
with anhydrous hydrofluoric acid. Calculotions of
the free energies for the reactions of the type

M,0, + 6HF (liquid) ~> 2MF, + 3H,0 (liquid)

have been carried out for the oxides which might
be present in NaZrF _-base fuels, These calcu-
lations indicate that the reaction is quantitative
for the following oxides: ZrOz, CrZO Fe203,
UQ,, NiO, and Na,0.

The conductivity of hydrogen fluoride is known

3'

to be very sensitive to traces of water. Literature
values?® for the equivalent conductivity of water
in hydrogen fluoride were used for calculations
which showed that the conductivity of anhydrous
hydroflueric acid is doubled by the addition of
1 ppm of water,

An apparatus is being constructed with which it
will be possible to measure the changes in conduc-
tivity of hydrofluoric acid that result from the
water produced by the fluorination reaction of
oxides in the fuel samples, The apparatus in-
cludes a drying still, a reoction vessel, and a
fluorothene conductivity cell. Provisions can be
made for carrying out the reaction with liquid
hydrotluoric acid at room temperature or with the
gaseous reagent at elevated temperatures.

 

3). H. Simons, Fluorine Chemistry, Vol. 1, p 240,
Academic Press, New York, 1950.
Oxidation of Trivalent Uranium by
Hexavalent Uranium

A. S. Meyer, Jr. D. L. Manning
Analytical Chemistry Division

Experiments were continued on the oxidmion:of
trivalent uranium by the urc:nyl ion UO
to the reaction _

203" + Uo,"™ + 4HT —3U%" 4+ 2H.0 .
In previous tests! the extent of the reaction was
determined by measuring polarographically the
concentration of the unreacted uranyl ion. This
measurement can also be made by titrating the
U4* formed with the ceric ion Ce?®, Several tests
were conducted in which known amounts of ucl,
were added to 100 ml of Q.01 N uranyl acetate
under varying conditions of acidity and tempera-
ture. After dissolution of the UCI,;, the concen-
tration of the U4 was determined by titrating with
0.1 N Ce**. The end point was obtained potentio-
metrically. The solutions were swept thoroughly
with helium and titrated under an inert-gas blanket,

Quantitative oxidation with the uranyl
not achieved in any of the tests.

occordin_g

ion was
indeed, when
solutions more acidic than 1 N were used as sol«
vent media, evolution of hydrogen was observed.
The reaction is favored, however, when the uranium
is added as the chloride instead of as the fluoride.
This is probably due to the greater dissolution
rate for UCI, in acid (several minutes) as compared
with the slow dissolution rate (several hours) for
UF, in this medium,

Efforts to eliminate the competitive reaction for
the oxidation of trivalent uranium in aqueous so-
{utions, _

3% ¢ HY —> U4t 4 UH,
have to date been unsuccessful. Future work will
involve investigation of possible
systems for the oxidation by uranyl ion.

nonaqueocus

Stability of Trivalent Uranium in Hydrochlorlc
Acid Solutions
A. S, Meyer, Jr. W. J. Ross

D. [.. Manning
Analytical Chemistry Division

Katz and Rubinowitch4 stated that the stability
of UCl, is greater in hydrochloric acid than in

 

4J. J. Katz and E. Rabinowitch, The Chemrsfry of
Uranium, p 458, McGraw-Hill, New York 1951,

PERIOD ENDING JUNE 10, 1954

““characteristic’ purple color

An investigation.of the sta-

water and reported a
for the trivalent ion.
bility of trivalent uranium in hydrochloric acid
media was made to ascertain the feasibility of
titrating U2, The potential of a solution of ucl,
in 1.0 N HCI at 0°C was f-ollowed with a plahnum
calomel electrode system. The initial potential
upon dissolution was +250 mv vs S.C.E., and the
potential decreased only 15 mv in 10 min. No
evidence of a purple color was observed ini this
experiment; the solution turned green upon dis-
solution. '

The experiments were repeated with concen-
trated hydrochloric acid at both 0 and -20°C., In
both cases, instantaneous formation of a .purple-
colored solution occurred, and the purple color
remained for 10 to 20 min before it changed to
green, The purple color remained longest in the
solution tested at the lower temperature., The
initial potential in all tests was of the order of
+150 mv, and the potential did not change appreci-
ably during the transition in color from purple to
green. The potential was essentially constant ot
+150 mv for 30 min. The constancy of the potential
of these solutions contradicts the usually accepted
hypothesis that the chonge in color from purple to
green represents oxidation to the U4" state. Some
quantitative data were obtained on the concen-
trations of tri- and tetravalent uranium in these
solutions by titration with ceric ion. In 0.1 N HCI
solutions, two distinct breaks in the potential
curve were noted, while only one break was ob-
served in concentrated acid solutions. The data
have not yet been thoroughly evaluated. The
anomaly of the change in color of the solutions
without an accompanying chcnge in potential will
be mveshgated

Removal of Film from Inconel Tubing

J. C, White
Analytical Chemistry Division

The tetrasodium salt of ethylenediaminetetra-
acetic acid was tested as a possible reagent for
removing ‘a film, probably calcium metasilicate,
which was found to be deposited upon the walls of
Inconel tubing following a wash with a detergent
Conklene. A 1% solution of the reagent in water,
heated to 75 to 80°C, effectively removed the film
rom a specimen of Inconel in a few minutes, The
metal had a bright finish after it was rinsed with
water. Chemical analysis of the reagent solution

11
ANP QUARTERLY PROGRESS REPORTY

after contact for 60 hr at 75°C with Incone!l showed
less than 1 ppm of chromium and nickel.

Determination of Sulfur in Natural Gas

G. Goldberg
Analytical Chemistry Division

A determination was made of the sulfur content
of the natural fuel gas received at the Y-12 Plant
because it is planned to use natural gas as a
source of heat in future corrosion tests. The
Referee® method was used in which the gas is
burned in an ammonia-rich atmosphere and the
condensable gases are collected in a tower filled
with glass balls and ammonium carbonate, The
reaction is

SO, + 2NH; + Hzo—%,(NH.fl)zSO3 .

The solution of ammonium sulfite is oxidized to
sulfate, and the suifur is determined as barium
sulfate., The sulfur content was found to be 0.5
mg/cu ft of gas, which is equivalent to 0.8 grain
per 100 cu ft. This concentration is essentially
the same as the sulfur concentration (1 grain of
sulfur per 100 cu ft) of the propane which is re-
ceived at the Y-12 Plant,

PETROGRAPHIC INVESTIGATIONS OF
FLUORIDE FUELS

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

With the shift in emphasis to reduced fuels, more
effort was devoted to the determination of the
optical properties of compounds containing UF,.
In NaF-ZrF -UF ; fuels reduced with metallic zir-
conium, with metallic uranium, or with hydrogen,
the predominant phase in the reduction complexes
obtained is always a solid solution of reduced U4*
in NaqzraF“.
from greenish-yellow, to greenish-tan, to a brilliant
yellow. The optic figure remains the same as that
for unreduced solid solution, but there is a definite
lowering of the refractive indices. Evidently the
amount of optic change in the crystal depends upon
the amount of reduction. In addition to the reduced
solid solution, minor amounts of UF,.2ZrF ,, which
is orange, uniaxially positive, and has refractive
indices of approximately 1.55, and an olive-drab
phase, which is pleochroic with refractive indices

The color of this complex varies

 

SL. M. Dennis and M. L. Nichols, Gas Analysis,
p 359, MacMillan, New York, 1929.

112

of about 1.54, are found. The composition of the
olive-drab phase is unknown. Sometimes a trace
of an isotropic yellow phase with n = 1.44 |s
present.,

In the RbF-UF, system, there appears to be a
compound, 3RbF.UF ., which is tan and isofropic,
with an n = 1,438, The NaF-UF; system contains
a compound which is lavender and uniaxially posi-
tive, with refractive indices of approximately 1.552.
The composition of this compound is not known.

Examinations of melts in the NaF-CrF, system
were made and four compounds were observed. The
compound Na,Crf, was previously reported as
isotropic and green, with n == 1.411. In composi-
tions with high CrF3 content, two phases were
present. The predominant phase was pleochroic,
with the fast direction yellow-green and the slow
direction blue-green, uniaxially positive, with 0 =
1.452 and E = 1.460, and polysynthetically twinned.
A second phase that was present in smaller amounts
was green and biaxially negative, with o small
optic angle and high dispersion. |t had a moderate
birefringence, and its refractive indices were near
1.52. A third phase was present in melts with
compositions near 50% NaF--50% CrF,. It was
yellow-green and uniaxially negative, with O =
1.444, and it had an estimated birefringence of
0.008.

The optical properties of the several compounds
that hove been synthesized were compiled and

published in an ORNL report.®

SUMMARY OF SERVICE ANALYSES

J. C. White W. F. Vaughan
C. R. Williams
Analytical Chemistry Division

Analyses of samples from corrosion tests of the
NaZrF ;-base fuel continued to be the major portion
of the work in the service laboratory, Samples of
fuel preparations containing mixtures of UF, and
UF, were received. In addition to the determi-
nations for the corrosion products iron, nickel, and
chromium, the concentrations of UF, and UF , were
determined, The method’ for hydrogen evolution

 

6T. N. McVay and G. D. White, The Optical Properties
of Some Inorganic Fluoride and Chloride Compounds,

ORNL-1712 (May 5, 1954).

’D. L. Manning, W. K. Miller, and R. Rowan, Jr.,
Methods of Determination of Uranium Trifluoride, ORNL -
1279 (Apr. 25, 1952).
has been used for the determination of UF,, and
following the determination of total uranium, the
content of UF4 is calculated from the difference

in concentration between total uranium ond UF3.
Samples from corrosion tests of prospective fufure

PERIOD ENDING JUNE 10, 1954

fuels which consist of fluorides of sodium, lithium,
and potassium or rubidium were also analyzed.
During this quarter the laboratory received 1163
samples and reported 1060 which involved 9357
determinations (Table 9.1). '

TABLE 9.1. SUMMARY OF SERVICE ANALYSES REPORTED

 

 

 

 

NUMBER OF SAMPLES NUMBER OF _DETERM!NATIONS
Reactor Chemistry 610 4435
Experimental Engineering : 370 4116
Fuel Production - 64 671
Miscelldneous 16 135
1060 9357

 

 

 

113
Part Il

SHIELDING RESEARCH
10. LID TANK FACILITY

C. L. Storrs
"GE-ANP
Jo M, Miller
Physics Division -
J. B, Dee
Pratt & Whitney Aircraft Division

G. T. Chapman D. K. Trubey

The Lid Tank Facility (LTF) has been used
primarily for studying special attenuation problems
which arise in aircraft shield design. One of
these studies which is of interest for crew-shield
design has been the penetration of slant-incident
neutrons through a hydrogenous shield. Inanother
study, comparison was obtained of secondary
(neutron-induced) gamma-ray produchon in lead cnd
in bismuth,

SLANT PENETRATIONS OF NEUTRONS
THROUGH WATER

G. T. Chapfian

Physics Division

The experiments for determining the slant pene-
tration of fission neutrons through water have been
continued at the LTF. In o previous experiment’
the neutron flux at the end of the duct was low
and it was possible to make measurements with
only small amounts of water between the detector
and the duct., For the present work, a new duct,
larger in diameter and shorter in length (Fig, 10.1),
was used, and the flux was increased to the
extent that measurements could be made 45 cm
from the end of the duct. The distribution of the
fast-neutron dose around the end of the duct is

plotted in Fig. 10.2,

The dose was integrated along planes at different
slant thicknesses (¢ ) of water and at angles (a)
of 0, 30, and 60 deg to the incident flux. The
results are given in Fig. 10,3, Owing to the large
size of the source, the distance to the integration
plane is not well defined at other than 0 deg, and
therefore the relative magnitudes of the curves
are in some doubt; however, the slopes are
probably significant.

The observed relaxation length for the 0- and
30-deg measurements was 7.2 cm. For the incident

angle of 60 deg, the relaxation length increased
from 7.8 to 12,5 cm as the slant thickness in-
creased. - This increase in relaxation length is
presumably due to the *‘short-circuiting’’ effect,
that is, the scattering of neutrons in a direction
which permits them to penetrate the shield along a
nearly normal path rather than along o slant path,

 

YC. L. Storrs et al., ANP Quar. Prog. Rep. Dec 10,
1953, ORNL 1649, p 120.

ORNL~LR-DWG 1045

¢ 6
D_UI\ST SOURCE
\
\
\ !
e . __-PLANE OF
f >\ e INTEGRATION
~ "»\-141
aQ° £\ I~ ]
\/f’ % %G T
; \
/“* a,-"'\ :
;
/
90-cm DUCT
7. 30-cm DIA -~ | 87.9cm

 

 

\
L W y |
|-—--"'28—in. DIA SQURCE ""

Duct Arrangement for Slab Penetration

 

 

Fig. 10,1,

Experiment.

117
ANP QUARTERLY PROGRESS REPORT

 

    
   
  

  
 
   
    

 

 

orn IR .,
135 | T T oo
. e 1T mMrep/hr )
@Ok ,i ——=ny P R — ' - —— i e !
i ; ;
: IO-‘ mrep/hr *
176 —— - \ . . i -
z i
5
=20 [
Ly
o
o
§ s
] 1 \ :
=
5 N
% 106 |- - ,1,” e e Y SN N S
2 ; i
w /'- N S mrspsar | \\ i
oo !-- - e b -\‘K_\\_ S N \
1G mrep/hr - ! \ \
| I - 4, . I
: ":‘10 "3 mrepsn- . \\ : \\
o TSN NN
‘ ‘~ NN N
90 ~— ,,,,,..,.,.,.‘,,,,,,,,.__,__,,A_u_,,,,‘,_.,.__\.‘_ L — L I
60 55 50 45 40 35 30, 5 20 15 10 5 Q

\
5
\
L

¥.DISTANCE FROM SQURCE CENTER LAvE (cm) ;
| " 7.3-cm DA X
Y A 90-~em DUCT

Fig. 10.2. Fast-Neutron Dose Distribution Be-
yond End of Duct,

 

 

 

 

 

 

DOSE UNITS x cos 4;

 

 

 

SLANT  THICKNESS {cm)

Fig. 10.3. Slant Penetration of Fast Neutrons in
Water.

SECONDARY GAMMA-RAY STUDY
D. K. Trubey
Physics Division
A study of secondary gamma-ray production in
materials considered for use as gomma-ray shields
has been undertaken ot the LTF., Measurements
have been made behind 6 in. of lead or bismuth
placed in borated water (1.1 wt % boron) at various

118

distances from the fission source plate (Fig. 10.4).
The 6-in. thicknesses of metal amounted to 145
g/sq cm for the bismuth and 161 g/sq cm for the
lead. The borated woter was contagined in a
36-in.-long aluminum tank placed adjacent to the

source, [he detector used primarily was o l-in.
anthracene scintillation crystal and o photo-
multiplier tube,

The gamma-dose measurements taken in the

plain water behind the borated-water tank in which
the lead or bismuth slabs were placed ot various
distances (d) from the source are shown in Figs,
10,5 and 10.6. It can be noted that the dose be-
comes less sensitive to slab position as & in-
creases. In both Figs. 10.5 and 10.6, curves
showing the gamma dose in plain water in the
Lid Tank and in plain water behind the boroted-
water tank have been added for comparison. The
slopes of these iwo curves differ because the
2,2-Mev hydrogen-capture gammas produced in the
plain water are attenuated more rapidly than the
harder-source gammas predominant in the water
behind the borated-water tank, In Fig. 10,5 the
effects of lead ploced 3.2 ond 8.4 cm from the
source in plain water are also shown,

A comparison of the dose measurements behind
lead and bismuth as a function of the position of
the metal, with the detector placed 120 cm from the
source, is shown in Fig, 10,7, Measurements ot
any other detector position would be similar. The
difference in the two curves for 4 = 47.6 cm is

ORNL-LR-DWG 933

 

 

 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘7 ,,,,,,,,, o
_ 1/ ALUMINUM TANK
METAL SLABS <% |
. S T TR ST e
§ A NN\ \ H,0 —

& i  BORATED Hy0 !
2 (14% B) |

., | — 2= 35 cm

- . -

Fig. 10.4. Arrangement for Measuring Secondary
Gamma-Ray Production in Materials Considered for
Use as Gamma-Ray Shields.
 

2 ‘ ORNL -flkDWG 1004

 

 

 

 

 

 

 

GAMMA-RAY DOSE {(mr/hr}

 

 

 

N 11% BORATED H,0, TANK AGAINST SOURCE |

 

 

 

 

 

 

90 100 110 120 130 140 150 180
z, DISTANCE FROM SOURCE TO DETECTOR (cm)

Fig. 10.5. Effect of Lead ofi Gamma Dose as o
Function of Distance in Waoter from Fission Source.

presumably due to the difference in the densities
of the two metals. The fact that the curves become
flat for d > 47 cm indicates that the slabs placed
at greater distances from the source were in such a
small neutron flux that the secondary gamma
production had become almost negligible, The
gamma-ray dose behind the slabs at d = 47.6 cmiis
therefore considered to be coused by source gamma
rays, including those originating in the water near
the source becouse of hydrogen capture of thermal
neutrons. The source dose can be subtracted
from the dose behind thé slabs at other positions
to determine the secondary gamma-ray dose at a
given distance from the source. The resulting
secondary doses for lead and bismuth are given in
Figs. 10.8 and 10,9, respectively, as a function of
distance from the front face of the slabs, since it
was considered that the origin of these gamma
rays was associated with the position of the slabs,

PERIOD ENDING JUNE 10, 1954

 

GAMMA-RAY DOSE (mr/hr)

 

 

 

 

 

 

 

 

 

 

90 100 MO 120 130 140 150 150
2, DISTANCE FROM SOURCE TO DETECTOR (cm)

Fig. 10,6. Effect of Bismuth on Gamma Dose as
a Function of Distance in Water from Fission
Source. ' '

A comparison of the doses from secondary gamma-
ray production in lead and in bismuth is presented
in Fig, 10.10, in which the dose at 100 cm from
the front face of the metal is plotted as a function
of the metal position. The experimental accuracy
does not justify concluding that the slopes are
different.,  Normalized thermal-neutron and fast-
neutron curves (measurements made in plain water)
have been included in Fig. 10.10 to show that the
secondary-gamma dose curve has a slope charac-
teristic of neutron curves. |

From this experiment little difference between
lead and bismuth secondary gamma-ray production
The data are to be extended in the
near future with the hope that it will be possible to
obtain a more complete understanding of fthe
phenomena, '

can be seen,

119
ANP QUARTERLY PROGRESS REPORT

 

 

 

 

 

 

ORNL—LR—COWG 930
10 L T -
. _ _ 1] o/ — ]
| b o]
I { i ! |
S e - i — e i ij —

 

 

 

 

 

 

§ ;6 in. OF Bi
i — - — — -

 

 

 

 

 

 

GaMMA - RAY DO3SE (mr/hr)

 

 

 

 

 

| \
FOUR 1'/2-in. Bi OR Pb SLABS IN 36-in. TANK
OF 1.1% BORATED H,O; TANK AGAINST SOURGE

2 - e

Y B _ e

 

 

 

w2l
) 10 20 20 40 50 60
d, DISTANGE FROM SOURGE TO FRONT FACE OF METAL (cm)

 

Fig. 10.7. Gomma-Ray Dose 120 cm from Source
as a Function of Position of Metal.,

 

DRNL!E—DWG 9321

    
   

 

 

 

 

 

 

 

 

2
W07 e T
: - T T T j ol
B e | L----—_- Lol
p ‘ - FOUR 1V2~in. Bi SLABS IN 36-in. TANK OF 11
[ | BORATED HZO; TANK AGAINST SQURLCE. 4‘
2 e ’ ‘ e i — — L
10 A
5|
=
=
£ oLl
w
w
o
&y
— -
o
o L
% 5:
g :
=
3
2 -
1o
10_2 . . i ‘ L i
70 a0 HO 130 150

< — d. DISTANCE FROM FRONT FACE OF METAL TO DETECTOR (ecm)

Fig. 10.9. Dose from Secondary Gomma-Rays
(Source Gamma-Roys Substracted) as o Function
of Distance from Bismuth Slabs,

120

  

     

ORNL-LR-DWG 932
2 ey
1o T LTI T B
5 - C e e e Lo - ]
FOUR 1Y%-~in. Pb SLLABS IN 3&-in. TANK N i
OF 1.1% BORATED H,0; TANK AGAINST SQURCE |
2 b -y §T e ' . e L —t

 

 

 

 

 

 

 

 

GAMMA-RAY DOSE (mr/hr)

 

 

 

 

 
 

| |
drr—e

e en

    
  

   

 

 

 

oL
70 80 90 100 110 120 130 140 150 160
z— 0, DISTANCE FROM FRONT FACE OF MZTAL TC DETECTCR {om)

Fig. 10.8, Dese from Secondary Gamma-Rays
(Source Gomma-Rays Substracted) as o Function
of Distance from Lead Slabs.

ORNL—-LR—DWG 923

 

 

   
   
 
  
  
  
  
  
 
 

 

10 —_— T
4 = T =
7~ FOUR i%2-n. Bi OR Pb SLABS IN 36-in —
5 TANK OF (4% BORATED H0; TANK

\\fi AGAINST SOURCE. -
\ i

e

 

 

 

 

T \””’7" ”7:7' T " . 77'7!' T - T

2 - \\ ————— — THERMAL -~ NEUTRON FLUX N -1
PLAIN H0, niy X 107 |

T ~—F— FAST-NEUTRON DOSE IN — 7777

e - PLAIN H0, mrep/hr X 107 -———

6 e RI NSUSSRRESEL [ S S —
|

i, AN ,+ i e ]

— - GAMMA DOSE BEHIND 4 ——

6 in. Pb, mr/hr

 

,,,77‘, B ) _—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[
107 |- } ~—GAMMA DOSE BEHIND -
& in B mesbr T
5
L — \
N
i
- ; i
10 1 ek
o 10 20 30 40 50 60

o, DISTANCE FROM SOURCE TQ FRONT FACE OF METAL {cm)

Fig, 10,10, Dose from Secondary Gamma«Rays
(Source Gamma-Rays Substracted) 100 cm from
Front Face of Meta!l as o Fuaction of Position of
Metal.
PERIOD ENDING JUNE 10, 1954

11, BULK SHIELDING FACILITY

. G. Cochran

E stabrook
. Flynn

. Haydon
Henry

ARE-O
xTvPET ™

F. C. Maienschein

H. E. Hunderford
E. B. Johnson
T. A. Love

R. W, Peelle

Physics Division

The Bulk Shielding Facilify (BSF) reactor has
been used for determining the power and power
distribution of loadings similar to those used in
the new Tower Shielding Facility (TSF) reactor.
The search for a better fast-neutron spectrometer
continues, but no substantial progress has been
made. Calculations of gamma-ray dose in a standard
divided-shielddesign have been extended to include
crew-shield penetration. '

GAMMA-RAY AIR-SCATTERING CALCULATIONS

F. C. Maienschein
Physics Division

Calculations of the dose received from air-
scattered gamma rays at the outside edge of the
crew compartment of an aircraft divided shield
were reported previously,! These calculations
have now been extended to determine the dose
inside the crew compartment.
was used for calculations of the gamma-ray pene-
tration through the crew shield, and in this case
the ANP-53 crew shield® was replaced by a simple
cylinder of about the same dimensions. |

A further calculation was made of the penetration
of the direct {unscattered) gamma radiation reach-
ing the crew position through the reor of the crew
shield. The buildup facters used were those
reported by Fano.?

Results of the air-scattering calculation indicate
that although the skew-beam flux at the outside of
the crew shield is extremely small compared with

 

1), L. Meem et al., ANP Quor. Prog. Rep. June 10,
1953, ORNL-1556, p ]09, F. Bly and F. C. Maienschein,
A “Calculation of the Gemma Rudiation Reaching the
ANP-53 Crew Shield, ORNL CF-53-5-117 (May 23, 1953).

2H. Goldstein, Reactor Handbook, Vol. 1, p 831, Tech-
nical Information Service, U, 5. Atomic Energy Com-
mlssmn, 1953.

Reporf of the Shielding Board for the Aircraft Nucleor
Propulsion Program, ANP-53 (Oct. 16, 1950).

41J. Eano, Nucleonics 11, 55 (1953).

The NDA method?

the radial-beam flux, the dose inside the crew com-
partment from the skew-beam flux is comparable to
the total radial-beam dose. This result was ex-
pected, since the skew-beam gomma rays, having
scattered through smaller angles on the averoge
than the radial-beam gomma rays, are of higher
energy and thus penetrate the crew shield more
readily.

The total dose inside the crew compartment from
air-scattered gamma rays, as calculated by this
method, is somewhat lower than that calculated by
the ANP-53 method? (Table 11.1); however, the
agreement is good if it is considered that for the
ANP-53 calculations it was assumed that all the
gamma rays were of 3-Mev energy and that all were
radially emitted from the reactor shield. The cal-
culations would be in poorer agreement if the
correction  {unknown) for difference in leakage
ratios between the ftwo reactors were applied.

For the direct beam, the large increase shown
by this calculation in comparison with the value
obtained by the ANP-53 calculation is coused
primarily by the gamma radiation that escapes
around the edges of the shadow shield. The
gamma-ray spectral medsurements at the BSF show
that a large number of low-energy (< 2-Mev) gamma
rays emerge from the reactor shield in such o
direction that they may strike the crew shield
without scattering. Unfortunately, the spectrometer
background at such positions was so large that
little value can be placed on the absolute magni-
tude of the dose. In general, for all results from
these calculations, the shapes of the curves ob-
tained are considerably more reliable thon the
absolute magnitudes.

The contributions to the fotal scattered dose of
various regions outside the reactor and crew shields
are shown in Figs. 11,1 and 11,2, The effect of
the shadow shield is shown in Fig. 11.2, which is
shaded proportionally to the dose contribution from
each region. Furthermore it is shown in that figure

121
ORNL-LR—DWG 974

0.0033
0959

06975 |
G.3257 | Lo.0z2a
.52 L0.0531

——

. . “

~SEZ INSERT L - —

T i T (WULTIPLY ALL VALUES 8Y 107)
UNOERLINED VALUES ARE SKEW BEAM DATA FOR ‘a AND £ AS SHOWN.
GEOMETRY APPLES TO RADIAL BEAM DATA ONLY.

VALUES ARE N
hr

 

180°
REACTOR CREW COMPARTMENT
lting from Air Scattering from Individual Cells — Radial and Skew Emission.

Fig. 11.1. Gamma-Ray Dose Inside Crew Shield Resu
123
TABLE 11,1, GAMMA-RAY FLUX AND DOSE AT THE CREW COMPARTMENT

 

 

FLUX OUTSIDE CREW DOSE INSIDE CREW

ME THOD COMPARTMENT, &/P COMPARTMENT, D
{(photons/sec/ watt) {e/hr/ watt)
Radial scattered 4.5 0.36 = 10—9
Skew scattered 0.016 0.19 x 10~7
Direct 15.0 130 % 10~
ANP-53 scattered* ' ' 1.2 x 10~
ANP.53 direct* | 1.2 x 10~
Skyshine** 114 x 10~°

 

 

 

*No correction made for reactor leakage.

**H, E. Hungerford, The Skyshine Experiments at the Bulk Shielding Facility, ORNL-1611 (to be issed).

e
ORNL~ LR~ DWG 1024

i oooos W 88;920 B 0 oo -10 "

 

7777 00001 TO [
LM*_W#J(DOOOO EZZ/A/ 0p00s &

 

a ’ B
REACTOR CREW COMPARTMENT

VALUES ARE N T (MULTIPLY ALL VALUES BY T

Fig. 11.2, Gamma-Ray Dose Inside Crew Shield from Air Scattering.

125
ANP QGUARTERLY PROGRESS REPORT

that air-scattering from regions far from the aircraft
makes an important contribution to the dose in the
crew compartment. In all cases, only the penetra-
tion through the side wall of the crew shield haos
been given, since the front- and rear-wall contri-
butions were negligible. Complete calculations
were carried out for the front wall for both radial
and skew emission in order to prove that these
contributions were actually small.

The dependence of the scattered dose (for all
angles) upon the prescatter energy is shown in
Fig. 11.3, and the dependence on the postscatter
energy in Fig. 11.4. It is interesting to note that
the average energy for the skew-emission dose is
higher than that for the radial-emission dose.

The scattered dose is presented as o function of
the source angle in Fig. 11.5 and as o function of

ORN[MG 2G4

 

 

 

 

)6

r
hr -watt
a

 

g~

 

 

 

SCATTERED GAMMA-RAY DOSE (
w

 

 

 

 

 

 

 

 

 

0 20 40 6.0 8.0 00 120
£, PRESGATTER GAMMA-RAY ENERGY [Mev)
Fig. 11.3, Scottered Gomma-Ray Dose as a

Function of Prescatter Energy.

126

the receiver angle in Fig. 11.6. The distribution
around the reactor shield (Fig. 11.5) clearly shows
the peak in the forward direction which is due to the
relatively favorable geometry for radiation starting
forward, After a drop of a factor of about 300 out
to 55 deg, the dose contribution rises rapidly of
the edge of the shadow shield and again falls off
because of geometry. It thus appears obvious that
the shadow shield should extend to larger angles.
It should, of course, be thinner with increasing
values of a. It also appears that the ANP-53 crew
shield was considercbly overshielded in the rear.
This was recognized at the time it was designed,

ORNL--LR—DWG BES

-
RADIAL EMISSION
D oeeme- GKEW BEAM |

 

 

 

hr

)

 

 

[

  

 

 

SCATTERED GAMMA-RAY DOSE (
3
&

 

 

o

1072 |

 

 

 

 

 

0~ L
0 1.0 2.0 3.0 4.0
£, POSTSCATTER GAMMA-RAY ENERGY (Mev)
Fig. 11.4. Scottered Gamma-Ray Dose as a

Function of Postscatter Energy.
wr

PERIOD ENDING JUNE 10, 1954

100° an

 

80° : OANL-LA-DWG 866

 
           
    
 

      
 
  

    
  
    

 

S0
-RADIAL EMISSION <
LT /,/.\ N

170°

 

 

SCATTERED GAMMA=RAY DOSE (M}
Fig. 11.5. Scattered Gamma-Ray Dose as a Function of Source Angle.
ORNL~LR-DWG 86T

o SKEW BEAM
* RADIAL EMISSION

  

\\\\‘:\ T ;.\\\‘\
VL
1
\ y Wy §
\ h
A

 

\

b
vy
VA

 

 

}
W
‘&\
~13 10——14 10—14 10—-13 10—-12 10—

R r .
SCATTERED GAMMA~RAY DOSE (hr‘m)

Fig. 11.6. Scottered Gamma-Ray Dose as a Function of Receiver Angle.

127
ANP QUARTERLY PROGRESS REPORT

but the excess shielding was supposed to com-
pensate somewhat for the presence of ducts in the
rear hemisphere of the reactor shield.

The distribution of the dose inside the crew
shield as a function of the angle around the crew-
comperiment midpoint (Fig. 11.6) shows a larger
contribution from scattered gamma rays from the
rear of the crew-shield side wall, as would be
expected becouse the crew compartment was taken
to be a long cylinder with walls of uniform thick-
ness. lhe variations observed in this case are
not nearly so large as those for the source angle.

The direct dose as a function of the energy of
the gamma rays emerging from the reactor shield is
shown in Fig. 11.7. The dependence of the dose
upon the energy is determined by the large angle
at which the major contribution to the direct dose
occurs, as shown in Fig. 11.8 in which the direct
dose as a function of the source angle a is pre-
sented. Figure 11.8 again shows the extreme im-
portance of the radiation escaping arcund the edges
of the shadow shield. The detailed calculations of
the gamma radiationreaching the crew compartment
will be published in o separate report,®

THERMAL-NEUTRON-FLUX PERTURBATION
BY GOLD FOILS IN WATER

E. B. Johnson
Physics Division

It is well known that when a detector that is not
infinitely thin is intfroduced into a neutron flux, the
flux is perturbed. The magnitude of the perturbation
depends upon the absorption cross section of the
detector and its physical dimensions and upon the
properties of the medium into which it is introduced.
[t is customary at the BSF to use the flux of the
ORNL Standard Graphile Pile® as the reference flux
for the calibration of foils and counters. However,
when g foil is used for thermal-neutron-flux measure-
ments in water, the flux perturbation caused by the
presence of the foil is different since the slowing-
down properties of water differ from those of graph-
ite. Klema and Ritchie’ arrived at a semiempirical

 

SE. C. Maienschein, F. Bly, ond T. A. Love, Sources
and Attenuation of Gamma Radiation from a Divided

Aircruft Shield, ORNL-1714 (to be published).

6¢. D. Klema, R. H. Ritchie, and G. McCamman,
Recalibration of the X-10 Standard Graphite Pile,
ORNL-1398 (Nov. 11, 1952).

128

correction for this effect for the 25-sq cm 5mil-
thick indium foils normally used for flux measure-
ments in water, The presence of these foils results
in a 22% depression of the flux from its unperturbed
value, Frequently it would have been desirable to
use gold foils for flux determinations in water, but
no value for the flux depression has been available,
Therefore an experiment has been completed ot the
BSF that was designed for measuring this effect,
Thermal-neutron-flux measurements were made
with both the indium and the gold (1-sq cm, 5mil-
thick) foils ot five positions along the north-south
center line of the reactor, The perturbation factor
of 1.22 was opplied, os usval, to the indium-foil
data, but no such correction was made on the gold-
foil data, The results are shown in Fig, 11.9. If it
is assumed that a probable error of +2% is associ-
ated with each measurement, there is no discernible
difference between the thermal-neutron-flux meas-
urements mode with indium foils using the perturba-
tion correction and those with gold foils using no
correction, While it is recognized that the

! . a T
presence of any obsorbing material must result in

a local depression of the flux, it seems that this
effect is within the probable errors of measurement
for small gold foils in a water medium, It should
be pointed out that such would not necessarily be
the cose if gold foils of larger dimensions were
used, These results substantiate those obtained
by P. M. Uthe® from a similar experiment performed
in the water tank at the thermal column of the ORNL
Graphite Reactor,

REACTOR POWER CALIBRATION TECHNIQUES?®

E. B. Johnson G. M. Estabrook
R. G, Cochran

Physics Division

Correlation of shielding data from the TSF with
those from the BSF requires that the power de-
terminaiions for the two reactors be made by com-
parable methods, Power determinations at the BSF
have been based on thermal-neutron-flux measure-
ments obtained by inserting foils between the

 

7E. D. Klema ond R. H. Ritchie, Preliminary Results
on the Determination of Thermal Neutron Flux in Water,

ORNL CF-51-4-103 (Apr. 24, 1951).
8p. M. Uthe, private communication.

&, B. Johnson et al., Povrer Calibration Techniques
for BSR, ORNL.-1723 (to be published).
PERIOD ENDING JUNE 10, 1954

ORNL~LR-DWG 1684

 

 

 

 

 

 

10_? L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

)

 

 

 

 

hr-watt -« Mev

 

 

 

 

 

e

 

 

 

GAMMA~RAY DOSE (

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1

4 6 8 10 12 14
£, SOURCE ENERGY (Mev)

 

Fig. 11.7. Direct Gamma-Roy Dose as o Function of Sour?:e Energy.

129
PROGRESS REPORT

ANP QUARTERLY

 

 

 

 

 

 

 

 
(ORNL-LR-DWG 417

ook

   

 

O INDIUM FOILS

 

 

X GOLD FOWLS

vy Swatt {neutrons semP -sec-watt)

THERMAL NEUTRON FLUX,

 

 

10 [_fi,.,_..“‘__.,.»ufi_*J__.,_.___.._"..:”. - e -
30 a0 50 50 70 80 90

DISTANCE ALONG REACTQR CENTER LINE FROM NORTH
FACE OF ACTIVE [-ATT[CE {cm)

Fig. 11.9. Comparison of Indivm- and Gold-Foil
Measurements in Water Surrounding BSF Reactor,’

plates of the fuel elements in the reactor lattice, ' °

However, the Tower Shielding fuel elements have
lead-shielded tops which make foil insertion im-
possible, and, in addition, the reactor control
chambers are located directly over the lattice,

It has been planned to make the flux measure-
ments in the TSF reactor by inserting cobalt wires
downthe center of the fuelelements. However, this
technique will yield only a total flux rather than

 

105 |, Meem and E. B. Johnson, Deferminotion of the
Fower of the Shield Testing Reactor — I. Neutron Flux
Measurements in the Water-Reflected Reoctor, ORMNL-
1027 (Aug. 13, 1951); E. B. Johnson and J. L, Meem,
Determination of the Power of the Bulk Shielding Re-
actor - {l. Neutron Flux Medasurements in Several
Beryllivm Oxide-Reflected Critical Assemblies, ORNL -
1438 (May 7, 1953).

PERIOD ENDING JUNE 10, 1954

the thermal flux needed for power-calibration caleu-
lations, Therefore, a method for obtaining reactor
power directly from total-flux measurements was
needed. ln the method used at the BSF for power
calibration, the total and epicadmium activities of
detectors placed throughout the reactor lattice are
obtained. The thermal-neutron flux obtained by this
methed is used to calculate the reactor power, since
the fission rote ond therefore the power level of an
enriched-fuel depend primarily on the
thermal-nevtron flux, . The relation between the
thermal-neutron flux as measured with foils and
the power produced as indicated by strategically
placed thermocouples has been very carefully
determined.'' It was found that the power of the

reactor

‘reactor can be expressed as

(1) = P = 4,264 x m‘“cgm ,

where P is the power in watts and G is the mass of
U233 in the volume over which 5”7 is the average
thermal flux. The constant contains the fission
cross section, energy release per fission, fast-
fission factor, and conversion factors.,

Various methods hove been investigated at the
BSF for simplifying ond adapting the currently used
power-calibration procedure for use with the TSF
reactor. The reactor power was calculated for each
of three loadings by using both total flux and ther-
mal flux as measured with small cobalt foils,
Loading 208 {Fig. 11.10) had a slab reflector of
beryllium oxide on all four sides, whereas loadings
22A (Fig. 11.11) and 24B (Fig. 11.12) were com-
pletely water reflected. The resulting dota indi-
cate that there is a constant ratio (within 1%) be-
tween the power calculated from thermal flux and
that calculated from total flux. This indicates that
if the constant in Eq. 1 is changed to 3.86 < 107 Y
the average flux in the equation can be considered
to be the total flux as measured with cobalt de-
tectors. Flux measurements made by using cobalt
wires have been obtained for the BSF reactor but
the calculations have not yet been completed.

 

115, L. Meem, L. B, Holland, and G. McCammon,
Determination of the Power of the Bulk Shielding Re-
actor — |ll.  Measurement of the Energy Releused per
Fission, ORNL-1537 (Feb, 15, 1954).

131
ANP GUARTERLY PROGRESS REPORT

CORNL-LR-ODWG {418

REACTOR GRID PLATE

>
am
=
)
wy
w
<
—
wl
=3
.
o
o
<I
0o
2
<1
—
w

SAFEYY RODS

7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\- CONTROL ROD

_/

PARTIAL FUEL ASSEMBLY

B8eO REFLECTOR

\

 

 

 

Fig. 11.10. BSF Reactor Loading 20B.

132
ACT !
/RE CTOR GRID PLATE

N\

PERIOD ENDING JUNE 10, 1954

CRNL-LR-DWG 1419

 

 

 

 

 

 

 

fij>W@%WU
DD
- )08)-)8)-,

 

 

\¥_///

 

OOOOOC
-

OOOOO

 

P
|

CONTROL ROD —

T
e
\

 

 

\ PARTIAL FUEL ASSEMBLY

Fig. 11.11. BSF Reactor Loading 22A,

HOOOHOG
| OOOOOO,

 

133
 

 

 

 

 

 

 

TN T
TN T TN LT
T T N Ty
)\1/\\!/

\'J\H/m
j))

=Ly

 

LT N N

Py

14 15
7w X
")lfi A A K

m TN \J
Q) N

)

A\
Woo>oo<
OO0

 

 

 

 

 

 

 

 

 

 

 
PERIOD ENDING JUNE 10, 1954

ks LEAKAGE-FLUX CHANGES DUE TO
ORNI~LR-DWG 1424

3»(10‘*r , ‘CONTROL-ROD SETTINGS
| J. D. Flynn R. G. Cochran

Physics Division

An experiment was performed ot the BSF to de-

termine the effect of the leakage flux of changes
in setting of the regulating rod and the safety rods
in order to obtain better correlation between suc-
cessive shielding measurements. At positions
close to the reactor foce, some effect would be
expected; however, at large distances from the
reactor, there should be no effect. The data (Fig.
11.13) indicate that there was no appreciable ef-
,,,,,,,,,,,,,,, B fect on the thermal flux ot distances of 40 to 80
’ M.
" In another experiment, the effect of changing the
quantity of U23% in o reactor loading was investi-
gated. A 110-g element was replaced with a 140-g
element on the side of the reactor opposite to the
side at which the measurement was made, and it
was found that the leakage flux was insensitive to
the quantity of uranium,

 

2

Nvyy / watt {neutrons/cm ~sec-watt)
I

 

o

 

3

 

 

 

THERMAL MEUTRON FLUX,
o

 

 

 

 

 

 

a0
DISTANCE FROM NORTH FACE OF REACTOR (cm)
Fig. 11.13. Comporison of Effects of Various

Control-Rod Settings on Leakage Flux of BSF
Reactor.

135

i AT A 08 A 8 A L A T T S S e s s T
ANP QUARTERLY PROGRESS REPORT

12. TOWER SHIELDING FACILITY

C. E. Clifford

T.V. Blosser
D. L, Gillilond!

.. B. Holland
J. L. Hull

F. N. Watson

Physics Division

Construction at the Tower Shielding Facility
(TSF) was completed May 1 (Fig, 12.1), and the
facility is now operating on a schedule of four
days per week. Critical experiments with the
reactor were started early in March, and installation
of the detector tank and experimental instrumenta-
tion was completely by the end of March. Since
that time, experimental operation of the facility
has been continued intermittently to permit final
adjustments and modifications of all equipment,

EXPERIMENTAL PROGRAM

Preliminary measurements have been taken for
the differential shielding experiments in a small
water tank (5 x 5 x 5 ft) which is known as the
detector tank, For these measurements the reactor
was in the 47-ton hemispherical water tank. The
relaxation length for the air-scattered neutron radi-
ation entering the sides and front of the detector
tank is 5.3 cm, which is to be compared to the
4,5-cm relaxation length obtained in the BSF
Skyshine experiments? and the 5-cm relaxation
fengthused in the 1953 Summer Shielding Session.3

Neutron- and gamma-dose measurements in the
reactor tank have been made with the 3-in. fission
counter, the 900-cc ion chamber, and the three-
chamber BF; counter. These measurements are
in qualitative agreement with those taken at the
BSF with a similar reactor loading.?

The first of a series of measurements for deter-
mining the background from ground-scattered
radiation has been completed with the reoctor and

 

10n loan from General Electric Company.

2H. E. Hungerford, The Skyshine Experiments at the
Bulk Shielding Facility, ORNL-1611 (to be issued).

3 Report of the 1953 Summer Shielding Session, ORNL-
1575 (to be issued).

4R. G. Cochron, J. L. Meem, T. E. Cole, and E. B.

Johnson, Reactivity Measurements with the Bulk Shield-
ing Reactor, ORNL-1682 (to be issued).

136

detector tanks 4 ft above the ground, The separa-
tion distances of the tanks were 35, 65, and 100 ft.
Since the power of the reactor is not absclutely
established, the dota from this experiment will
be reported first in terms of radiation intensity
relative to a nominal power, which is expected to
be within 30% of the octual power.

OPERATION OF THE REACTOR

The reactor was loaded to criticality on March
12. This loading consisted of 26 fuel elements,
the minimum number required for criticality, Since
that time, ancther loading (30 fuel elements ina 5
by 6 array) has been checked by a critical experi-
ment and is now being used for the preliminary
measuyrements, Both these loadings were run
earlier in the BSF reactor with TSF reactor fuel
elements, and the power distribution of the 5 by 6
element array was determined with BSF reactor
elements, The twe TSF reactor loadings are com-
parcble in shim-rod and regulating-red positions to

those used at the BSF,

During the experiments the reactor was run at
power levels of 1 to 10 watts at the 200-ft eleve
tion and at approximately 4 kw at just above the
poo! level, The instruments and controls performed
satisfactorily under both conditions.

Use has been made of the remote visual-indicator
system to observe operation and to position the
reactor and the crew compartment. The ease with
which it is possible to position the large loads by
remote control with this system has exceeded
expectations.

A system of slack-line limit switches has been
installed on the large hoists to prevent damage to
the hoist gear mechanism such as that experienced
during the run-in period. The operation of the
switches is satisfoctory, but, to odd assurance,
the cables will be kept taut by additional counter-
weights that will be odded to the floating sheave
blocks.
 

 

 

 

el

s

S

SRR
’%{\;}')“ s e

e
NI nY

e
o
o
Ry
b
S

] e 5
73:; 2

i

S

e
R

o0

SR
e

'.'&"A
3
3

S

Fig. 12.1. Completed Tower Shielding Facility.

i

2
B

v:~;;:§.‘:f:-‘a;@¢s_
S
S
i

T

et

G

T
Qgi::‘:fu; 7
o
T

J\" Y

 

PS6L ‘Ol INNr ONIANI AOIY3d
Part IV

APPENDIX
REPORT ND.

CF-54-3-65
CF-54-4-218
CF-54-5-51

CF.54-2-185

CF-54-4-6
CF-54-4-53

CF-54-4-221

CF-54-5-1
CF-54-5-196

CF-54-3-37
CF-54-3-194
CF-54-4-195
ORNL-1688

ORNL-1716

CF.54-4-47
CF-54-5-40

CF-54-5-47
CF-54-5-189
CF-54-6-6

CF.54-315

CF-54-3-193

13. LIST OF REPORTS ISSUED DURING THE QUARTER

TITLE OF REPORT

}. Aircraft Reactor Experiment

ARE Design Data Supplement
ARE Instrumentation List

Analysis of Critical Experiments

Il. Reflector-Moderated Reactor
Effects of Reactor Design Coenditions on Aircraft Gross
Weight
The Kinetics of the Circulating-Fuel Reactor

Reflector-Moderaféd Critical Assembly Experimental Pro-

gram

Radiation Damoge Elastomers, Lubricants, Fabrics, -and

Plastics for Use in Nuclear-Powered Aircraft
The Xenon Problem in the ART
Stresses in Heat-

Temperature Gradient and Thermal

Generating Bodjies
Iti. Experimental Engineering

50-Mw Design — Canned Be

Crosgs Sections

Discussions and Results of the Dielectric Heating Test .

Sodium Plumbing

from a  Molten Flueride 3Salt
in a Double-Tube

Turbulent Heat Transfer

Mixture to Sodium-Potassium Alloy

Meat Exchanger

IV;. Chemistry

Daia on the Ber-NuF System

Effects of Fission Products in Performance of a Reactor

Using Fluorides as Selvent for Fuel
Measurement of the Stability of Lithium Hydride
Analysis of Salt Mixtures

Fused Salt Compositions

¥. Metallurgy

Metallographic Examination of Forced Circulation Loop No.

2

High Flow Velocity and High Tempefufure Gradient Loops

»

E

- ®I ®O ¥ B I

orE» 00

AUTHOR(s)

B, Cottrell
G. Affel
B. Mills

- J.
M.

K.

Stumpf

Wilner

Ergen

. Scott
L.

Greenstreet

. Stumpf
. Wilner

J

M
L.
A

Meem

. Field

. Mills
. Mitls

. Southern

. Cottrell
. Mann

. Sglmon

. Orbar

R. F. Newton

E. Orban
F. F. Blankenship
C. J. Barton

G. M. Adamson

W. D. Manly

DATE ISSUED

3-2-54
4-7-54
5-12-54

5-21-54 .

5-5-54
4-8-54

4-15-54

5-3-54
5-21-54

3-9-54
3-8-54
4-27-54
8-14-53

To be issued

4-27-54
5.7.54

5-26-54
5-25-54
6-2-54
3-3-54

3-18-54

141
ANP QUARTERLY PROGRESS REPORT

REFPORT NO.

CF-54-4-162
CF-54-4-224
CF-54.5-88

CF-54-5-91
ORNL-1667

CF-54-4-205

CF-54-5-159

CF-54-5-160

ORNL-1701

ORNL-1702

CF-54-4-96
ORNL-1712

CF-54-5-41

CF-54-6-4

CF-54-5-201
CF-54-5-219
ORNL.-1682

142

TITLLE OF REPCRT
Status of Coelumbium, Beryllium, and Lithium

Report of Literature Survey of Beryllium

Heat Exchanger Fabrication

Free Energies of Formation of Oxides and Fluorides

Alkali-Meta! Nickel Oxides Containing Trivalent Nickel

Vl. Heat Tronsfer and Physical Properties

A Feasibility Study of Flow Visualization Using a Phos-
phorescent Particle Method

Effect of Oil Contomination on the Boiling Heat Transfer
Characteristics of Heat Exchangers and Solid Fuel-Plate

Reacters

Heat Caopacities of Composition No. 12, No. 44, and of
l(:st_.rl:‘5

Forced Convection Heat Transfer Between Parallel Plates
and in Annuli with Volume Heat Sources within the

Fluids

A Summary of Density Measurements on Molten Fluoride
Mixtures and a Correlation Useful for Predicting Den-

sities of Fluaride Mixtures of Known Composition

AUTHOR(s)

W, D, Manly

J. W. Woods

S.
T.

Vil. Anolytical Studies of Reactor Maoterials

Data on Aircraft Fuel Samples

The Optical
Chloride Compounds

Properties of Some Inerganic Fluoride and

VItl. Radiation Damage
Metallographic Analysis of Pratt and Whitney Capsules
1-4, 1-6, and 1-7

Release of Xenon from Fluoride Fuels: Proposal for an

Experimental Program

1X. Shielding

Secondaery Gamma Ray Study
Stant Penetration of Neutrons Through Water

Reactivity Measurements with the Bulk Shielding Reactor

T O ©Owr - woOm
T v U

- I ©=x
D OO9

. E. Hoffman

. Patriarca
M. Cisar

. Dyar
Borie, Jr.

. Smith

D. Paolmer
M. Winn

W. Rosenthal
L. Milles

. Powers

. Blalock

. Poppendiek

. Palmer

l. Cohen

N. Jones

R. Baldock

>

. N. McVay
. D. White

. J. Feldman

. E. Richt

M. T. Rekinson

A

et

. K. Trubey

. T. Chapman

. G. Cockran

al.

DATE ISSUED
4-6-54

4-13-54
5-17-54

5-5-54
2-26-54

4-30-54

5-19-54

5-20-54

5-11-54

5-14-54

4-14-54
5-5-54

5-5-54

6-2-54

5-26-54
5-28-54

To be issued
REPORT NO.

ORNL-1714

ORNL 1723

ORNL-1686
ORNL-1692

PERIOD ENDING JUNE 10, 1954

TITLE OF REPORT

Sources and Attenuation of Gamma Radiation from a Di-

vided Aircraft Shield

Power Calibration Techniques for BSR

X. Miscellaneous

Quantities and Reactions of Solid Surfaces

Aircraft Nuclear Propulsion Project Quarterly Progress

Report for Period Ending March 10, 1954

 

AUTHOR(s) DATE ISSUED

F. C. Maienschein To be issued

F. Bly

T. A. Love

E. B. Johaseon To be issuved
ef al.

E. P. Carter 4-7-54

A. W. Savolainen 4:15-54

143